A History of East Baltic through Language Contact

Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 A History of East Baltic through Language Contact Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 Leiden Studies in Indo-European Series editors Alexander Lubotsky Alwin Kloekhorst Tijmen Pronk volume 24 The titles published in this series are listed at brill.com/lsie Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 A History of East Baltic through Language Contact By Anthony Jakob leiden | boston Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 This is an open access title distributed under the terms of the cc by-nc-nd 4.0 license, which permits any non-commercial use, distribution, and reproduction in any medium, provided no alterations are made and the original author(s) and source are credited. Further information and the complete license text can be found at https://creativecommons.org/licenses/by‑nc‑nd/4.0/ The terms of the cc license apply only to the original material. The use of material from other sources (indicated by a reference) such as diagrams, illustrations, photos and text samples may require further permission from the respective copyright holder. This project has received funding from the European Research Council (erc) under the European Union’s Horizon 2020 research and innovation programme (grant agreement nº 716732). The Library of Congress Cataloging-in-Publication Data is available online at https://catalog.loc.gov lc record available at https://lccn.loc.gov/2023052014 Typeface for the Latin, Greek, and Cyrillic scripts: “Brill”. See and download: brill.com/brill‑typeface. issn 0926-5856 isbn 978-90-04-68646-5 (hardback) isbn 978-90-04-68647-2 (e-book) doi 10.1163/9789004686472 Copyright 2024 by Anthony Jakob. Published by Koninklijke Brill nv, Leiden, The Netherlands. Koninklijke Brill nv incorporates the imprints Brill, Brill Nijhoff, Brill Schöningh, Brill Fink, Brill mentis, Brill Wageningen Academic, Vandenhoeck & Ruprecht, Böhlau and V&R unipress. Koninklijke Brill nv reserves the right to protect this publication against unauthorized use. This book is printed on acid-free paper and produced in a sustainable manner. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 Contents Acknowledgements vii List of Tables ix Symbols and Abbreviations x Data Sources and Conventions xvi Introduction 1 part 1 Contacts with Known Languages 1 Baltic–Slavic Contacts 7 1.1 Early Slavic → Baltic Loans 7 1.2 Early Baltic → Slavic Loans? 29 2 Early Germanic → Baltic Loans 37 3 Baltic → Finnic Loans 45 3.1 Preliminaries 45 3.2 Baltic Loanwords with an IE Etymology 51 3.3 Analysis of Sound Substitutions 61 3.4 Loans from Proto-Finnic to Proto-Baltic? 84 3.5 Common Loans from Unknown Sources? 91 3.6 Analysis of Contact Relationship 113 4 Loanwords into Other Uralic Languages 4.1 Sámi 119 4.2 Mordvin 129 4.3 Mari 140 4.4 Permic 145 4.5 Conclusion 147 119 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 vi contents part 2 Contacts with Unknown Languages 5 Introduction 153 5.1 Research History 153 5.2 Methodological Considerations 159 5.3 Excursus: Illegal Root Structures 167 5.4 Preliminaries 171 6 Consonantism 173 6.1 ‘Nasalization’, *-VNT- ∞ *-VT6.2 Voicing Alternations 186 6.3 Sibilant Clusters 217 6.4 Other Irregularities 225 7 Vocalism 233 7.1 Initial Vowels 233 7.2 Alternations between Front and Back Vowels 244 7.3 Alternations between Low and High Vowels 257 7.4 Alternations between Monophthongs and Diphthongs 7.5 Length Alternations 268 7.6 IE *a 280 8 173 264 Analysis 283 8.1 Semantics 283 8.2 Stratification 302 Conclusion 308 Bibliography 311 Word Index 380 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 Acknowledgements The first seed for my dissertation, on which this book is based, was planted back in 2017 when I told Tijmen Pronk I’d be curious to study the whole subject of “substrate words” in Indo-European. This topic had interested me ever since I first flicked through Derksen’s etymological dictionaries. The idea that some of the words in the modern languages could have been borrowed from undocumented sources was something that seemed very appealing to me, especially in northern Europe where the historical record is so shallow. Tijmen put me in touch with Guus Kroonen, and a couple of years on, I was accepted onto Guus’ project (‘The Linguistic Roots of Europe’s Agricultural Transition’). It was amazing to have been able to conduct my research as part of a team, and the weekly group meetings, daily lunches and constant exchange of ideas (and idiocy) were a reliable source of inspiration and the perfect antidote to the mundane aspects of research. This team mentality led to three joint papers (Palmér et al. 2021; Kroonen et al. 2022; Thorsø et al. 2023), the workshop SubIndo-European Europe, which we hosted in August 2021, and a forthcoming volume based on the results of this workshop. Not only that, but the methodology expounded and applied in the second half of this work is as much a team effort as it is my own. I would like to sincerely thank Yvonne van Amerongen, Paulus van Sluis, Cid Swanenvleugel, Rasmus Thorsø, Andrew Wigman and honorary team member Axel Palmér for making these four years so special. My individual project took me in a number of unexpected directions. Sergejus Tarasovas has taught me the ways of philological pedantry, and shown me that there is light at the end of most rabbit holes. Countless insights in this work, both big and small, ultimately derive from our discussions (some more are to be published as Jakob forthc. b.). Alexander Savelyev has shown incredible patience in giving me a crash course in Turkic linguistics. Practically all discussions of Turkic languages in this monograph have greatly benefited from his input. Most significant, however, was my journey to becoming a Uralicist (which perhaps I could now call myself, having spoken at the Finno-Ugricist Conference in Vienna, August 2022) — something I’d never considered when starting out on this project. A significant catalyzer for me has been the work of Ante Aikio, whose accessible writing style was the gateway I needed, but I would also like to thank Petri Kallio, Sampsa Holopainen, Juho Pystynen, Abel Warries and Mikhail Zhivlov for giving me heaps of constructive criticism and ideas which have made this work considerably better, and especially Santeri Junttila, Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 viii acknowledgements not only for the same, but also for providing me with an enormous amount of literature I would not otherwise have been able to access. Dozens of other people have, at various moments, given me answers, made me question the answers I had, or made me realize I was asking the wrong questions. Without them, this book would not be half what it is, and I will mention just one dozen who happen to spring to mind: Lucien van Beek, Chams Bernard, Eugen Hill, Martin Kümmel, Giedrė Motuzaitė Matuzevičiūtė, Michaël Peyrot, Tijmen Pronk, Peter Schrijver, Zsolt Simon, David Stifter and Rémy Viredaz. The project of which this work forms a part would not have existed at all, of course, if it wasn’t for its initiator Guus Kroonen and Professor (now Emeritus) Sasha Lubotsky. They supervised me throughout the five years it took to write this monograph, and I know all too well that wasn’t the easiest of tasks. I am eternally grateful to them both for keeping their faith in me and helping me grow as a person. It has been an eventful few years, with the Covid-19 pandemic and war in Ukraine both massively impacting my personal life. The former — for the better, since it meant me and my wife could spend an unexpected and magical year together in Leipzig, the latter — decidedly for the worse, although for us it has been rather overshadowed by a serious struggle with my motherin-law’s health. I am incredibly thankful to Guus for being so understanding and allowing me to take the steps I needed. Finally, let me mention my parents, Dan and Barbara, my sister, Lara, and my adoptive parent, Natalya, who are always categorically supportive and who I can always rely on to be on my side, no matter what kind of pickle I’m in. Last and most of all, I am thankful to my wife Sasha. I know she has had to put up with a lot, including a lot of my absence, but she has brightened up every single day and given me a reason to carry on. My only hope now is that we finally get a chance to find out what “normal” means! Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 Tables 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 Words similar in Lithuanian and Vedic Sanskrit 3 Archaic features in Slavic loanwords 22 i-diphthongs in Early Proto-Finnic 67 Probable shared substrate words in Baltic and Finnic 112 Baltic loanwords in Sámi and Mordvin 148 Suggested criteria for identifying substrate borrowings 158 Comparison of deus and θεός 162 Possible examples of nasal alternations 185 Possible examples of *g⁽ʰ⁾ ∞ *k alternations 199 Possible examples of alternations involving dentals 206 Possible examples of labial alternations 212 Possible examples of an alternation *ž ∞ *š 217 Possible examples of sibilant metathesis 223 Possible examples of ‘prefixation’ 243 Possible examples of front ∞ back alternations 256 Possible examples of high ∞ low alternations 264 Possible examples of long ∞ short alternations 276 Distribution of borrowed animal names 286 Distribution of borrowed plant names 289 Distribution of borrowed agricultural terms 292 Chronology of cereal terms in Baltic 295 Distribution of borrowed apicultural terms 299 Distribution of borrowed terms for structures 301 Distribution of borrowed metallurgical terms 302 Alternations showing a geographical patterning 303 The *VND substrate 305 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 Symbols and Abbreviations For abbreviated titles, see the relevant subsection of the bibliography. Symbols * 1. (before lemma) reconstructed form; 2. (after lemma) unattested form of an attested lexeme ** counterfactual form # (in reconstructions) word boundary ∅ zero · introduces a gloss = ‘(historically) identical to’ ~ ‘belongs with (in unspecified capacity)’, ‘in variation with’ / 1. mutually exclusive alternative, 2. (in citations) demarcates co-authors : 1. paradigmatically alternates with; 2. correlates with (providing an analogical model) ∞ ‘irregularly alternates with’ < ‘developed from’, ‘derived from’ ≪ ‘analogically replaces’ (also used of semantic developments) ← ‘borrowed from’ / / phonological transcription [ ] 1. phonetic transcription; 2. alteration to quoted text or form ⟨ ⟩ orthographic transcription C consonant H laryngeal N nasal R resonant T dental stop V vowel General arch. bce c. ce archaic before Common Era century Common Era Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 xi symbols and abbreviations cf. comp. dial. diss. E ed., eds. e.g. et al. etc. fn. forthc. FS i.e. in prep. lit. loc. cit. N N.B. NE No. NW obs. op. cit. p., pp. p.c. poet. refs. S sp. s.v. SW trad. trans. usu. vel sim. viz. vol., vols. W confer, see compiler(s) dialectal dissertation east editor(s) exempli gratia, for example et alii, and others et cetera footnote forthcoming Festschrift id est, that is in preparation literature loco citato, at the cited location north nota bene, please note north-east number north-west obsolete opere citato, in the work cited page, pages personal communication poetic references south (unspecified) species sub verbo, under the corresponding entry south-west traditional translator(s) usually vel similia, or the like videlicet, namely, to wit volume(s) west Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 xii symbols and abbreviations Languages N.B. Abbreviations which only occur as second elements are given in [square brackets]. Alb. Albanian Arm. (Old) Armenian Aukšt. Aukštaitian Av. Avestan [B] Baltic Bel. Belarusian Bg. Bulgarian Bret. Breton Čak. Čakavian (dialect of Serbo-Croatian) Celt. Celtic Chuv. Chuvash CLuw. Cuneiform Luwian [Co.] Cornish CS Church Slavic ~ Bes. Besědy na evangelije ~ Ps. Sin. Psalterium Sinaiticum ~ Supr. Codex Suprasliensis Cz. Czech Du. Dutch E Estonian [E] English Eg. Egyptian F Finnish Fr. French [Fri.] Frisian Gaul. Gaulish Gm. Germanic Go. Gothic Gr. (Ancient) Greek H. Hesychius [HG] High German Hitt. Hittite HLuw. Hieroglyphic Luwian HLv. High Latvian Ic. Icelandic IE Indo-European Ingr. Ingrian Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 symbols and abbreviations [Ir.] Iran. It. K Kajk. Kash. Khot. Lat. [LG] Li. LSrb. Lt. [Luw.] Lv. M… Ma. Mac. Md. ME MHG ML MLG Mo… MoE MP Myc. Ngan. NP Nw. O… OCS xiii Irish Iranian Italic Karelian Kajkavian (dialect of Serbo-Croatian) Kashubian Khotanese Latin Low German Livonian Lower Sorbian Lithuanian Luwian Latvian Middle, Mediaeval [MBel., MCo., MDu., MGr., MIr., MR, MUk., MW see second element] Mari ~E Eastern (Meadow) Mari ~ NW Northwestern Mari ~W Western (Hill) Mari Macedonian Mordvin ~ E Erzya ~ M Moksha Middle English Middle High German Mediaeval Latin Middle Low German Modern [MoHG, MoGr., MoIr., MoLG, MoW see second element] Modern English (Manichaean) Middle Persian Mycenaean Nganasan New (= Classical) Persian Norwegian Old [OAv., OBret., OCo., OCz., ODu., OFr., OFri., OIr., OPl., OR, OSw., OTur., OW see second element] Old Church Slavic Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 xiv OE OHG ON ONovg. OS Osc. Oss. P… Parth. PIE Pl. Pr. PrG PrLt. PU R RCS Sá. SCr. SCS Skt. Sl. Slk. Sln. Sw. Tat. To. Tur. symbols and abbreviations Old English Old High German Old Norse Old Novgorodian (dialect of Old Russian) Old Saxon Oscan Ossetic ~ I Iron ~ D Digor Proto[PGm., PMa., PMd., PSá see second element] Parthian Proto-Indo-European Polish (Old) Prussian ~E Elbing Vocabulary ~ Gr. Grunau’s vocabulary ~ iii Third Catechism ~ TC Trace of Crete Prussian German Prussian Lithuanian Proto-Uralic Russian Russian Church Slavic Sámi (Saami) ~I Inari (Aanaar) Sámi ~ K Kildin Sámi ~L Lule Sámi ~ N North Sámi ~S South Sámi ~ Sk. Skolt Sámi Serbo-Croatian Serbian Church Slavic Sanskrit Slavic Slovak Slovene Swedish Tatar Tocharian Turkic Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 symbols and abbreviations Turk. Udm. Uk. Umbr. Ural. USrb. Võ. Vp. Vt. [W] YAv. Žem. xv Turkish Udmurt Ukrainian Umbrian Uralic Upper Sorbian Võro (South Estonian) Veps Votic Welsh Young Avestan Žemaitian Grammatical 1… (2…, 3…) abl. acc. act. adj. adv. coll. dat. du. elat. ess. f. gen. ill. inst. lat. m. n. obl. part. pl. pres. pret. prt. sg. first (second, third) person ablative accusative active adjective adverb collective dative dual elative essive feminine genitive illative instrumental lative masculine neuter oblique partitive plural present preterite participle singular Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 Data Sources and Conventions Reconstructions In general, following in part the practice adopted by OED³, I have avoided providing reconstructions unless this is necessary for the argument. The rationale is to maximize the emphasis given to attested data, and also reduce the need for me as an author to make a clear stand with regard to theoretical aspects of reconstruction where these are not strictly relevant to an argument. For instance, while one might object to my reconstruction of Lithuanian avìs ‘sheep’ as PIE *h₃eui-, it is unlikely anyone will object to its equation with Latin ovis ‘sheep’. I have made an exception in the case of evidence from Uralic languages, for which I have provided reconstructions quite systematically. This is partly a means to provide additional clarity for readers more familiar with IndoEuropean than with Uralic, but is also a reflection of my own process in dealing with these languages. In the following cases, my reconstruction differs from the established norm and/or requires certain clarification: East Baltic — Acute intonation is marked with the circumflex or caret ⟨ ˆ ⟩, following the use of this symbol as an indication of the broken tone in Latvian tonal orthography. Slavic — My reconstruction of the Slavic vowel system differs from the OCSbased one conventionally used (for instance, in the dictionaries of Berneker, Vasmer, Derksen, and ЭССЯ). The use of the standard reconstruction leads to anachronisms, such as an apparent claim that the Early North Russian source of e.g. Finnish dial. akkuna ‘window’ is more archaic than its Proto-Slavic ancestor. In general, I consider the Slavic vowel shift to be a Common Slavic, not ProtoSlavic development (cf. Olander 2015: 59–67), and therefore use a reconstruction with pre-vowel shift values.1 The reconstruction used in this work is as follows: 1 Differently from Olander, I do not operate with a Proto-Slavic predating the monophthongization of diphthongs. One reason for this is practicality: for instance, it is often impossible to decide whether Common Slavic i derives in any particular case from an earlier *ei or *ī. However, I also do not consider it likely that the monophthongization was a post-Proto-Slavic development: the absence of the second palatalization in North Russian (cf. also Holzer 2001: 39–40) does not necessarily imply that it branched off before the development *ai > *ē since there is no reason to exclude an intermediate stage */kæː/. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 xvii data sources and conventions This work ī ē i ȳ e u a Traditional ū i ā ě ь e y ъ o u a However, for ease of cross-referencing with works that operate with the standard notation, I have supplied the traditional reconstruction in brackets after each Proto-Slavic form. Finnic — In line with the “Leiden” tradition of not marking allophonic features, such as syllabic resonants, in reconstructions, I do not mark consonant gradation, as this is entirely predictable in Proto-Finnic (except for *s between unstressed vowels, where I have preserved the alternation with *h). Therefore, I reconstruct Finnish hammas (gen.sg. hampaan) ‘tooth’ as *hampas rather than *hambas, *hamβas or *hamp̆ as. Data sources In compiling this work, I have endeavoured to check all the forms cited in primary sources. I have generally avoided citing data which I was unable to independently verify, unless this is crucial to an argument. A selection of literature used to source the forms from the most important languages for this work is presented below. The orthography follows the cited sources unless otherwise indicated. References not provided in full can be found in the bibliography. All web links are valid as of 31 May 2023. Lithuanian Latvian Prussian Russian DLKŽ; LKŽ; Bendrinės lietuvių kalbos žodynas (ekalba.lt) Latviešu literārās valodas vārdnīca (tezaurs.lv); ME; LVPPV (tonal orthography follows ME; differences with LVPPV have been noted) PKEŽ and facsimile copies hosted at www.prusistika.flf.vu.lt; Trautmann 1910 Большой академический словарь русского языка [а-продел]; Толковый словарь русского языка (ed. Дмитрий Н. Ушаков); СРНГ; СДРЯ 11–14; СРЯ 11–17; СДРЯ. I have followed pre-revolutionary orthography in the use of the symbol ⟨ѣ⟩ (italics ⟨ѣ⟩), where this is etymologically relevant. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 xviii data sources and conventions Ukrainian Belarusian Polish Czech Словник української мови (sum.in.ua); ЕСУМ Тлумачальны слоўнік беларускай мовы (1977–1984); ЭСБМ Słownik języka polskiego (ed. Witold Doroszewski); SSP Příruční slovník jazyka českého (psjc.ujc.cas.cz); Gebauer; digitized resources at vokabular.ujc.cas.cz Slovak Slovník slovenského jazyka (slovnik.juls.savba.sk/?d=peciar); SSN Sorbian Schuster-Šewc; dolnoserbski.de; hornjoserbsce.de Slovene Pleteršnik (tonal orthography follows Pleteršnik; additional data from SSKJ² is given as in the source) Serbo-Croatian RJA; РСА; Skok (additional dialect data cited after Derksen 2008, 2015) Church Slavic Старославянский словарь (по рукописям x–xi веков); SJS; Miklosich 1865 Bulgarian Речник на български език (ibl.bas.bg/rbe/); БЕР Macedonian Дигитален речник на макдонскиот јазик (drmj.eu) Gothic Scandinavian English Dutch Low German High German Celtic Latin Albanian Greek Armenian Streitberg, Wilhelm. Gotisch-Griechisch-Deutsches Wörter buch (1910) and interlinear texts at wulfila.be/gothic/ Cleasby/Vigfusson 1874; Dictionary of Old Norse Prose and other digitized resources at onp.ku.dk; Blöndal 1989, SAOB, Norsk ordbok (no2014.uib.no, last accessed 31 May 2023) DOE; MED; OE. Middle English allophonic lengthening is not marked Philippa et al.; digitized resources at gtb.ivdnt.org Tiefenbach 2010; MndWb; Schiller/Lübben. MLG orthography generally follows the latter: allophonic lengthening is not marked; however, I have, after MndWb, distinguished ö from o AWb; Schützeichel 2004; MWb; Mittelhochdeutsches Handwörterbuch (ed. Matthias Lexer); DWb eDIL; GPC; Le dictionnaire diachronique du breton (ed. Martial Menard, devri.bzh) TLL; Lewis/Short; Walde/Hoffman; Ernout/Meillet Mann, Stuart E. An Historical Albanian-English Dictionary (1948); Demiraj 1997 LSJ; additional data from Frisk; Beekes 2010. Martirosyan 2009 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 data sources and conventions Finnish Karelian Veps Votic Estonian Livonian Western Sámi Eastern Sámi Mordvin Mari Permic Khanty Mansi xix SMS [a–mähistyä]; VKS [a–pitäytä]; SSA; SKES KKS Зайцева/Муллонен 1972; Зайцева 2010. Orthography follows the latter VKS EMS [a–puisklema]; VMS; Eesti keele seletav sõnaraamat (2009); other digitized resources at portaal.eki.ee; South Estonian: synaq.org Kettunen 1938; Viitso/Ernštreits 2012. Courland Livonian orthography follows the latter. Salaca Livonian: Winkler/Pajusalu 2009 Digitized resources at kaino.kotus.fi/algu/. Orthography after Bergsland & Magga Mattson (1993). Sydsamisk-norsk ordbok [South]; Korhonen (2007). Lulesamisk-svensk ordbok [Lule]; Sammallahti & Ocejohka (1989). Saamelais-suomalainen sanakirja [North]. Data from older sources is presented in updated orthography Digitized resources at kaino.kotus.fi/algu/. Orthography follows Sammallahti & Morottaja (1993). Inarinsaamelaissuomalainen sanakirja [Inari]; Sammallahti & Mosnikoff (1991). Suomi-koltansaame sanakirja [Skolt]. Kildin Sámi data is updated from Lehtiranta 2001 on the basis of Rießler 2022 MdWb; Серебренников et al. (eds.), Эрзянско-русский словарь (1993), Мокшанско-русский словарь (1998). Phonemic orthography simplified from MdWb, and updated to reflect the modern standard TschWb. Phonemic orthography after e.g. Aikio 2014 Digitized resources at dict.fu-lab.ru, Лыткин/Гуляев 1970 OstWb Artturi Kannisto (comp.), Wogulisches Wörterbuch (2014). Phonemic orthography follows secondary literature Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 Introduction The aim of this monograph is to place the East Baltic languages in their prehistoric linguistic context through the analysis of lexical borrowings. The work will be divided into two sections: in the first, I will critically assess the evidence for the established prehistoric contact relationships with Slavic (Chapter 1), Germanic (Chapter 2) and Finnic (Chapter 3) and examine proposals of contact with other Uralic languages (Chapter 4). The second half of the work will be devoted to the question of contacts with unknown languages, a complex and no doubt controversial subject, which has not yet had an extensive treatment. One of the aims is to establish applicable methodological principles for analysing this kind of material, and this half of the book can be seen as a practical demonstration and evaluation of these new methodological tools. The result will be a detailed catalogue of the contact relationships in which the East Baltic languages participated. In order to stratify these linguistic events, I will also attempt to incorporate evidence from other disciplines, specifically archaeology, archaeobotany, and genetics, to evaluate the context and nature of the individual contact situations. This will be particularly important in the analysis of contacts with unknown languages (Chapter 8), as we a priori have no other information about the other participants in these contact events. The focus of this work will be on East Baltic specifically. This is in itself unusual. Sabaliauskas (1990), for instance, stratifies the Lithuanian lexicon into the layers “Indo-European”, “Balto-Slavic”, “Baltic” and “Lithuanian”, without distinguishing a separate East Baltic layer. Discussions of vocabulary exclusive to the Baltic languages likewise often fail to demarcate East Baltic as a distinct unit (e.g. Zinkevičius 1984: 229–234 and Larsson 2018: 1687–1688 are only concerned with isoglosses involving Prussian). This reflects a wider tendency in the literature, where one can easily find grammars and handbooks on Baltic (such as Stang 1966; Endzelīns/Schmalstieg 1971; Dini 2014) and grammars and handbooks on individual East Baltic languages (e.g. Endzelīns 1923; Kazlauskas 1968; Zinkevičius 1980–1981; Forssman 2001), but very little discussion of the East Baltic languages together, and basically no systematic attempt at reconstructing a separate proto-language. There are, however, clear arguments for the separate study of East Baltic. Firstly, while the status of “Baltic” as a branch of Balto-Slavic has been disputed (Kortlandt 1977: 323; Derksen 1996: 1; Andersen 1996a: 63; Kallio 2008: 265; Kim 2018: 1974),1 the coherence of East Baltic as a subgroup appears to be univer1 Villanueva Svensson (2014: 164) mentions Иванов/Топоров (1958) and Harvey Mayer (e.g. © Anthony Jakob, 2024 | doi:10.1163/9789004686472_002 Anthony Jakob - 978-90-04-68647-2 This is an open access chapter distributed under the terms of the cc by-nc-nd 4.0 license. Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 2 introduction sally accepted (albeit often implicitly). This can be demonstrated by a small but robust set of innovations exclusive to East Baltic. Clear cases are the following:2 1. *ai and *ei merged into *ẹ̄ in certain environments (becoming further diphthongized to /ie/ in both standard languages), thus Lt. líepa, Lv. liẽpa ‘linden’ (= Pr. TC leipen acc.sg., R ли́ па) beside Lt. sniẽgas, Lv. snìegs ‘snow’ (= Pr. E snaygis, OCS снѣгъ) (Hirt 1892: 32–41; Stang 1966: 53–57; Hill 2016: 208).3 2. Probably related to this is the appearance of a prothetic v- in the word for ‘one’, cf. Lt. víenas, Lv. viêns (contrast Pr. iii ains) (cf. Fraenkel 1950a: 26–27; Petit 2010: 14). 3. A stem with -v- has been generalized in the paradigms of the 2sg. and reflexive pronouns, and corresponding possessive adjectives. Thus acc. sg. *ten (> Pr. iii tien, OCS тѧ) and dat.sg. *tebVi (> Pr. iii tebbei, OCS тебѣ) have been replaced by Lt. tavè, Lv. tevi and Lt. táu (< tãvi), Lv. tev, respectively (Petit 2010: 14; Hill 2016: 209–210). 4. Initial m- has been generalized throughout the paradigm of the 1pl. pronoun: cf. Lt. mū́sų, Lv. mũsu gen.sg. (against Pr. iii noūson, OCS насъ) (Forssman 2001: 44; Petit 2010: 14).4 Aside from this, a number of convincing isoglosses can be found between East and West Baltic,5 but also some seemingly non-trivial isoglosses between East 2 3 4 5 Mayer 1978) as sceptics of the Proto-Baltic theory, but the scepticism in both cases seems more directed at the Stammbaumtheorie in general and against Balto-Slavic unity in particular. Both use the term “Baltic” liberally in the traditional sense. Most lists only note differences between East and West Baltic without distinguishing innovations from archaisms (Endzelīns 1944: 17–21; Forssman 2001: 42–46), or include isoglosses with which Slavic also participates (Petit 2010: 12–17). Although the conditions of this merger are not fully resolved (see the discussions in Kuryłowicz 1956b; Stang 1966: 58–61; Mathiassen 1995; Petit 2003: 96–97), the high level of agreement between Lithuanian and Latvian shows that we cannot, at least, be dealing with a later areal development. I exclude: (a) the change *-tl- > -kl-, which is also shared by North Russian (Николаев 1989: 190–198; Зализняк 2004: 49), and is therefore to be considered an areal phenomenon which might have spread through an already diversified East Baltic; the development also seems to have taken place in the Prussian dialect of the Elbing Vocabulary (cf. in particular sasin-tinklo ‘snare’ where we cannot blame the graphical confusion between ⟨c⟩ and ⟨t⟩); (b) likewise, the loss of the neuter gender in nouns seems already to have been spreading to the dialect of the Prussian Third Catechism (cf. Endzelīns 1944: 84; Fraenkel 1950a: 28); cf. unds nom.sg. ‘water’ against Pr. E wundan. On alleged traces of the neuter in Finnic loanwords, see 3.3.3. See most recently Villanueva Svensson (2014) and Hill (2016), against which Kortlandt (2018). Here I would like to add another argument: the 1pl. and 2pl. pronouns, Lt. mẽs, jū̃s, gen. mū́sų, jū́sų correspond exactly (except for Innovation 4, above) to Pr. iii mes, ioūs, gen. noūson, ioūson. In OCS, we find мъı, въı, gen. насъ, васъ, where the oblique forms are old Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 3 introduction Baltic and Slavic (Villanueva Svensson 2014: 163; Kortlandt 2018: 176). For the purpose of this work, an agnostic stance can be considered acceptable, as the internal structure of the Balto-Slavic family does not have any bearing on the validity of East Baltic as a subbranch. While William Jones’ famous idea that Germanic was “blended with a very different idiom” can be seen as foreshadowing a whole subfield within Germanic studies (cf. Kroonen 2012: 240), the reputation of Baltic has developed quite differently. As Antoine Meillet (1913: 205) famously put it, a person who wishes to hear an echo of what Indo-European sounded like “va écouter les paysans lituaniens d’aujourd’hui” (despite Dini 2014: 45, fn. 21, I have verified this quotation to be genuine). This continues a legend present in non-specialist literature since the 19ᵗʰ century. Thus, the Encyclopædia Britannica (9ᵗʰ edition, 1882; cited per Klimas 1957) claimed that “whole Sanskrit phrases are well understood by the peasants of the banks of Niemen”, and one still often comes across claims that Lithuanian is “the oldest” (Encyclopedia of Linguistics, 2004, ed. Phillip Strazny, p. 119) or “most archaic Indo-European language still spoken” (as in the current online edition of the Encyclopædia Britannica, accessed 15 August 2023). It is true, of course, that Lithuanian is remarkably archaic in certain aspects. In terms of phonology, it probably can indeed lay claim to being the “most archaic”, and in nominal morphology its only serious competitor is Slavic (see the discussion in Erhart 1995). If we take the liberty of writing the Sanskrit sandhi variant -s (rather than usual -ḥ), then it is not difficult to assemble a collection of forms where Modern Lithuanian and Vedic Sanskrit appear almost identical (see Table 1, below). table 1 Words similar in Lithuanian and Vedic Sanskrit Lt. výras ‘man’ Skt. vīrás ‘man, hero’ Lt. sūnùs ‘son’ Skt. sū́nus ‘son’ Lt. ugnìs ‘fire’ Skt. agnís ‘fire’ Lt. šuõ, gen.sg. šuñs ‘dog’ Skt. śvā́, gen.sg. śúnas ‘dog’ (cf. Lat. nōs, vōs). To explain the Baltic oblique forms, it seems we have to assume a two-stage development: first, the strong stem *jûs spread throughout the 2pl. paradigm, yielding a new gen. *jûs-un; second, the vocalism of the 1pl. *nōsun was modified after the 2pl., resulting in a new stem *nûs-. These two non-trivial and consecutive developments seem to provide strong evidence of a common Baltic stage. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 4 introduction For context, compare the Modern Hindi bīr ‘hero’, āg ‘fire’, and sūnā ‘dog’,6 or the continuants of these words in other modern languages: Irish fear /fʲaɾ/ ‘man’, Icelandic sonur ‘son’, Slovene ógənj ‘fire’. The surface similarities in the above table are admittedly partly accidental, but mainly result from a phonological conservatism on the part of Lithuanian. This conservatism has no doubt led to the stereotype of Baltic as a ‘pure’ dialect which has had “little or no non-IE contact” (Nichols 1998: 254) and “has not mixed with any other Indo-European or non-Indo-European language” (Klimas 2002). Finnish in many respects holds a position similar to that of Lithuanian: many words in the modern language “appear almost bizarrely archaic” (Aikio 2022: 5), being identical to their reconstructed Proto-Uralic predecessors; thus e.g. muna ‘egg’ (< PU *muna), pesä ‘nest’ (< PU *pesä). At the same time, we know that the Finnic languages did not develop in isolation. In the Proto-Finnic lexicon, we can identify layers of loanwords from Slavic (cf. Kalima 1956; Kallio 2006), Proto-Norse (collected in LÄGLOS i–iii) and Baltic (see Chapter 3), while North Finnic also contains a significant lexical substrate from Sámi (Aikio 2009). Thus, a conservative phonology does not necessarily presuppose a conservative lexicon. I hope that this study will go some way towards dispelling the myth about the ‘purity’ of Baltic, and East Baltic in particular, in demonstrating that this branch, like any other, has a complex history and has been subject to numerous external influences. 6 According to Turner’s CDIAL. I cannot find the word for ‘dog’ in modern dictionaries, so it is perhaps obsolete, or at least dialectal (perhaps Turner’s source was John D. Bate, A Dictionary of the Hindee Language, 1875, p. 724). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 part 1 Contacts with Known Languages ∵ Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 chapter 1 Baltic–Slavic contacts 1.1 Early Slavic → Baltic Loans 1.1.1 Preliminaries The goal in this chapter is to establish the extent and nature of the earliest contacts between Baltic and East Slavic. While several studies have focused on the Slavic loanwords in Baltic (the pioneer articles being Brückner 1877; Endzelīns 1899; Būga 1911; Skardžius 1931), there have been few critical studies focusing on the earliest layer specifically: one often speaks of ‘early’ loanwords (e.g. Young 2009, Derksen 2020), but the actual material adduced encompasses ambiguous cases which could result in a skewed picture. The goal of this section is to identify the clearest and best quality data to substantiate the claim of early contacts. The aforementioned ambiguities are in part the result of “traditional” substitution strategies (Stang 1957: 52–55). For instance, the examination of Lithuanian proper nouns transcribed into Cyrillic in 13ᵗʰ and 14ᵗʰ century documents, led Būga (1911: 18) to conclude that Proto-Slavic length was still contrastive at this time. The idea that such length contrasts were maintained well into the literary period is hardly tenable; however, it is clear that the substitutions Slavic o → Lt./Lv. a and Slavic a → Lt. o, Lv. ā have continued into recent times. This must at least in part be based on “traditional” equivalences extrapolated from earlier loanword strata (i.e. “etymological nativization”; see Aikio 2006b: 18–23 for a discussion of the concept). Stang notes Lt. dial. notūrà ‘character, nature’ from Polish natura, a recent Latin loanword; and we can add here examples with Lt. o before a tautosyllabic resonant, which only became phonotactically possible in the last few centuries (see below), such as Lt. kortà ‘card’ (← Pl. karta; LEW 283), gvõltas ‘violence; uproar’ (← Pl. gwałt ← MoHG Gewalt; LEW 180). Levin’s (2003: 141–142; cf. Derksen 2020: 44) reconstruction of a ProtoLithuanian system with /aː/ and /ɒ/ seems to be more an attempt to force a phonetic explanation than something explicitly motivated by the data. True, a tendency for /ă/ to become rounded can be observed across the eastern edge of the Baltic territory, specifically in part of East Aukštaitian (but not on the Lithuanian–Belarusian border; cf. Zinkevičius 1966: 50–51 and Map No. 6), and as a conditioned change in Latgalian (Endzelīns 1923: 73–85). However, there is no reason to set this up as the most archaic system.1 1 On the substitution of Baltic *a with Finnic *a and *o, see p. 63. © Anthony Jakob, 2024 | doi:10.1163/9789004686472_003 Anthony Jakob - 978-90-04-68647-2 This is an open access chapter distributed under the terms of the cc by-nc-nd 4.0 license. Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 8 chapter 1 Moreover, as the Finnic loanwords from Slavic also clearly show the preservation of inherited length (cf. Kalima 1956: 33–42), indicating an early date, a layer of loanwords in Baltic from the same chronological period would hardly be surprising. Note the following examples showing a reflection of length in unstressed syllables which are shared by Baltic and Finnic (cf. Stang 1957: 53): Lt. pyrãgas, Lv. pìrãgs ‘pie’ (= F piiraa) Lt. sopãgas, Lv. zàbaks ‘boot’ (= K soappoa) HLv. žìvats² ‘animal’ (= F siivatta ‘livestock’) ← OR пирогъ (R пиро́г, gen.sg. пирога́ ) ← OR сапо́ гъ, gen.sg. сапога́ ← OR живо́ тъ, gen.sg. живота́ ‘life; animal’ (R dial. животы́ ‘livestock; property’) In Finnic, the smaller corpus of loanwords makes the suggestion of “traditional” substitution patterns less plausible, and the above examples must be accepted as early loanwords. However, this does not necessarily have any bearing on the age of these words within Baltic, and it cannot be ruled out that these too were borrowed at a later date following previously established nativization strategies. It is in fact highly difficult to identify the oldest layer of loanwords upon which the regular substitution patterns were originally based. The only unambiguous evidence of an early date of borrowing would be cases in which Baltic reflects phonemic contrasts subsequently lost in the history of East Slavic. I therefore limit myself to substitutions of this kind, and the resulting data will form the corpus for further analysis. 1.1.2 Reflection of yers Finding unambiguous examples of the reflection of the Slavic reduced vowels in the Baltic loans is more complicated than usually recognized. Note, for instance, the following examples involving sequences of the type *CŭRC: – Lt. kùrtas, Lv. kur̃ts ‘greyhound’ ← R, Uk. хорт ‘greyhound’ (= SCr. hȑt, Pl. chart); also → F hurtta, E hurt, Li. ūrta-pi’ņ ‘greyhound’ (Kalima 1952: 66) – Lt. obs. tùlkas, Lv. tul̃ks ‘interpreter’ ← OR *тълкъ, MR толк ‘interpreter’ (= OCS тлъкъ) – Lt. tur̃gus, Lv. tìrgus ‘market’2 ?← OR, ONovg. търгъ, u-stem, ‘market’ (= OCS тръгъ); also → F turku, E turg (gen.sg. turu) ‘market’ (Kalima 1952: 133) – Lt. pul̃kas, Lv. pùlks ‘crowd, troop’ ?← OR пълкъ ‘troop, regiment; crowd, throng’ (= OCS плъкъ)3 2 On the Latvian -i-, see fn. 16. 3 The rare F pulkka ‘regiment, troop’, quoted by Ahlqvist (1871: 209), shows a limited dialectal Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic–slavic contacts 9 While these are generally acknowledged as early loans, it should not be overlooked that similar substitutions are also attested in relatively late loanwords. Historically, in the absence of a phoneme /ŏ/ in either Latvian or Lithuanian, a foreign /oRC/ has been substituted as /uRC/. Note, for instance, Prussian Lithuanian kur̃bas, dial. kur̃vis ‘basket’ (← Prussian German korb, korw, Alminauskis 1934: 76; LEW 220) and Lt. dial. guñtas ‘roof shingle’ (← Pl. gont; LEW 176). A similar strategy is attested in Latvian loans from Estonian, e.g. puĩka ‘boy’ (← Võ. poig, E poeg ‘son’),4 kul̃da ‘ash pit’ (← Võ. kollõq, gen.sg. koldõ), cf. Thomsen (1890: 263–273). Most probably, the same applies for some dialectally isolated loans like Lv. dial. (Naukšēni) burtenis ‘empty beehive’ ← R dial. (Pskov, РЭС iv: 96) бо́ ртень ‘hive used to attract bees’ (ME i: 354); Lv. dial. (Aloja) turba ‘knapsack’ ← R dial. то́ рба (ME iv: 268); and High Latvian pulna ‘enough!’ ← R по́лно (cf. Būga 1925: 44).5 This observation raises suspicion with regard to the Aukšt. dial. bulvõnas ‘idol; dummy’, Lv. dial. (Vidzeme) bul̃vāns (ME i: 349), bulens (EH i: 251) ‘decoy bird’ ← Russian болва́ н ‘idol; stuffed animal, dummy’ (Даль² i: 111), which have consistently been described as early loanwords (e.g. Būga 1925: 40; LEW 33; Derksen 2020: 40). While a source with preserved *o is not probable in the Baltic area, it is conceivable that these were adopted from a dialect with dissimilative akanje typical of north-eastern Belarusian dialects (Wexler 1977: 79– 80).6 On the other hand, the distribution of the word within Baltic may not exactly favour such an interpretation; at least for Latvian, a certain role may here also have been played by Baltic German Bolwan, Bulwan ‘decoy bird’ (cf. Kiparsky 1936: 149). distribution and occurs alongside the allegedly younger F–K polkka, E polk ‘regiment’. The antiquity of this loanword seems doubtful to me (cf. Mikkola 1938: 53). Note F dial. hulkka beside holkka ‘crowd; troop’, considered to be of Germanic origin (cf. LÄGLOS i: 119–120). 4 Both formally and semantically, a more probable direct source seems to be Estonian Swedish poik ‘boy’ (Freudenthal/Vendell 1886: 165). 5 Lv. dial. tul̃pîties ‘to crowd’ ← толпи́ ться (Būga 1925: 43) may be based on the dial. 3sg.pres. то́лпится (cf. СРНГ xliv: 207), which must be the older form in view of the oxytone accentuation of толпа́ (Зализняк 2019: 208); cf. also Uk. то́ впитися (which must be analogical after 3sg.pres. то́ впиться). Despite ME (iv: 260), it does not seem likely that the Latvian word is cognate with Lt. dial. tùlpinti ‘make room for, economize’. The latter seems somehow to have been formed secondarily from til̃pti ‘fit, have enough room’ (compare Smoczyński 2018: 1484), although the details are unclear. Incidentally, the Latvian accentuation might also speak against an old loan (see 1.1.7). 6 Similarly, Endzelīns (1899: 298) refers to the pronunciation of unstressed *o as [u] in some Russian dialects (on this see ДАРЯ i: No. 2). This explanation seems possible for Lt. tulkocʒus (Brodowski 923) ‘pestle’, cf. Bel. таўка́ ч (Būga 1925: 763); perhaps also Lv. dial. (Endzelīns 1899: 299) grumada ‘assembly’, Lt. grùmada ‘crowd’ (Juška ii: 478), grummodas ‘Haufe Fliegen’ (Ruhig ii: 192) ?← Bel. dial. (*)гр[у]ма́ да, if not simply ← Polish gromada ‘flock, crowd’. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 10 chapter 1 The four cases cited at the start of this section could theoretically be dated to Proto-East-Baltic. In the case of Lt. tur̃gus, Lv. tìrgus, cognancy with Slavic is also an option (Endzelīns 1899: 299; REW iii: 123).7 In principle, the Slavic loan etymologies in all of these cases are encouraged primarily by circumstantial facts, for instance the existence of early parallel loans into Finnic from the same sources, although some phonetic details favour the loan etymologies.8 The Slavic word *pulka/u- (trad. *pъlkъ) ‘crowd; military regiment’ is a loan from Germanic (cf. OHG folc ‘people, crowd, troop’).9 It is possible that the Baltic words are parallel loans from a related Germanic source, possibly even West Germanic folk, rather than having been mediated through Slavic. In fact, several other suggested early Slavic loanwords in Baltic are ultimately of Germanic origin. Hirt (1898: 350) in this and other cases has assumed direct adoptions from Gothic (see also Chapter 2), while Būga (1922: 71) has preferred to assume a Slavic intermediary: – Lt. ãsilas ‘donkey’ ← OR осьлъ ‘donkey’ / ← Go. asilus ‘donkey’ – Lv. bruņas f.pl. ‘armour’ ← OR брънѧ ‘armour’ (also брънѣ f.pl., cf. СРЯ 11– 14 i: 321) / ← Go. brunjo ‘breastplate’10 – Lt. kãtilas, Lv. katls ‘kettle’ ← OR котьлъ ‘kettle’ / ← Go. katils* ‘kettle’ – Lt. stìklas, Lv. stikls ‘glass’ ← OR стькло ‘glass’ / ← Go. stikls ‘cup, chalice’11 7 8 9 10 11 Further cf. Alb. treg, dial. tregë ‘market’ (cf. Meyer 1891: 323–324; Jokl 1924: 88). If we reconstruct *trgʷ-, we might blame the non-acute accentuation in Slavic on the u-stem (Stang 1957: 79–82; Pronk-Tiethoff 2012: 242–244); the circumflex in Baltic might support a Slavic intermediary. Note that a reconstruction *trgʷʰ- would violate the IE root constraints. Bjorvand/Lindemann (2018: 1328–1330) assume the Baltic and Slavic words are loans from Germanic, cf. ON torg ‘market’, but it is doubtful there are any Norse loans in Proto-Slavic or in East Baltic (see Chapter 2). Significantly, their account fails to explain the Albanian data. For kùrtas, the correlation between Baltic k- and Sl. x- favours a Slavic → Baltic loan. For tùlkas, the acute accent in Baltic would be in disagreement with OIr. do-tluchethar ‘seek, demand’, Lat. loquor ‘talk’ (which rule out a root-internal laryngeal), which might favour a Slavic origin. See 1.1.7. As the substitution Germanic *o → Sl. *u (trad. *ъ) is unproblematic (Slavic had no *o), attempts to track down West Germanic forms with /u/ (cf. Pronk-Tiethoff 2012: 112–113) are unnecessary. For the Slavic word, Pronk-Tiethoff (2012: 122–123) considers it impossible to decide between a Gothic or West Germanic source. However, on the basis of OR пѣнѧзь ‘coin’ (СДРЯ 11–14 ix: 407) ← OS penning ‘penny’ (where we must assume an original *pę̄nę̄ʒe(trad. *pęnędzь) by dissimilation; cf. the close parallel in OCS мѣсѧць ‘moon’, dissimilated from *mę̄sę̄ce- (trad. *męsęcь) < IE *meh₁ns- ‘month’; Shevelov 1964: 320; Beekes 1982: 55), one would rather anticipate West Germanic *brunnjā (> OHG brunna, MHG brünne ‘chain mail, breastplate’) to be borrowed as Slavic *brǭnjā- (trad. *brǫnja) (> Old Russian **броунѧ). Therefore, a Gothic source seems preferable. Although the Slavic word seems to fit better semantically, it cannot be excluded that the Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic–slavic contacts 11 It is curious that all of the above words have attested equivalents in Prussian, viz. Pr. E asilis ‘donkey’, brunyos ‘armour’, catils ‘kettle’, sticlo ‘drinking glass’. This suggests yet another possible route these words could have taken: they may have entered East Baltic through Prussian. In fact, the realia seem to speak against ãsilas ‘donkey’ being an old word in Lithuanian. According to G. Piličiauskienė (p.c. April 2023), there is so far no evidence for donkeys in the Lithuanian zooarchaeological record, and no documentary evidence for them having been bred or traded. It is therefore quite probable that the word was taken from Prussian by Bretke for the purpose of translating the bible. On the other hand, it is not possible to rule out Bel. асёл, Pl. osioł ‘donkey’ as a source either, as this Slavic suffix appears as Lt. -ilas even in recent loanwords. Thus, South Aukštaitian bùsilas ‘stork’ (see Naktinienė et al. 1988: 48) ← Bel. бу́ сел (gen.sg. бу́ сла) cannot be an old borrowing in view of its short first-syllable vowel. Likewise, Lt. kazilaĩ ‘saw-horse’ can hardly be from OR ⁽*⁾козьлъı (pace Būga 1925: 41; Smoczyński 2018: 513), as this particular sense is a Polonism, and in turn a calque from German Bock (REW i: 590). In view of this, it is also possible that Lt. kùbilas ‘tub, barrel’ is a relatively recent loan from Polish kubeł ‘bucket’ rather than representing an early loan from the source of RCS къбьлъ*, OPl. (hapax, 15ᵗʰ c.) gbeł ‘bucket’ (cf. Būga 1925: 38– 39).12 As for the other words, no conclusions can be drawn as to their proximate source. Even if they were adopted through Prussian, it is unclear whether the Prussian words were themselves adopted directly from Gothic or via West Slavic. Given the multitude of possibilities, these words can hardly serve as evidence of direct early contact between Slavic and East Baltic. Incidentally, Levin (1974: 88) has suggested that certain other “general Baltic” Slavicisms spread along a trajectory from west to east; the following seem to be decent candidates: – Lt. kùmetis ‘serf, peasant’ in Suvalkia and Prussian Lithuania ?← Pr. E kumetis ‘gebuer’ ← Lechitic *kumeti- (trad. *kъmetь; > OPl. kmieć ‘serf, peasant’) – Lt. krìkštas ‘baptism’ ?← Pr. *kriksta-, cf. Pr. iii crixti lāiskas ‘tauffbüchlein’ ← Lechitic *krista- (trad. *krьstъ; > OPl. krzest ‘baptism’; SSP i: 257) 12 word also meant ‘glass (material)’ in Gothic; cf. the polysemy exhibited by e.g. English glass, and also Lt. stìklas, Pr. E sticlo in the sense ‘drinking glass’. For further discussion of the semantics, see Kiparsky (1934: 210–211). Both Slavic words are of West Germanic origin, cf. MHG kübel ‘bucket, tub; dry measure’ (Pronk-Tiethoff 2012: 126–127). Latvian kubls ‘tub, barrel’, like katls ‘kettle’ are taken as instances of syncope by Endzelīns (1923: 47), but could just as easily be relatively recent loans from the Slavic oblique stems in Pl. gen.sg. kotła, kubła; cf. similarly Lt. pãslas ‘ambassador’ ← Pl. paseł, gen.sg. pasła. The High Latvian kubyls cited by Būga (1925: 38) as evidence of an unsyncopated form is considered dubious by ME (ii: 297). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 12 chapter 1 For Lt. krìkštas, an ultimately West Slavic origin is attractive for semantic reasons. Continuants of Proto-Slavic *krista- (trad. *krьstъ) mean ‘baptism’ in traditionally Catholic areas (i.e. West Slavic and Slovene), while in Eastern Christianity they have the meaning ‘cross, crucifix’. Accordingly, a sense ‘baptism’ is unattested in Old Russian (СДРЯ 11–14 iv: 315–317), which practically rules out an East Slavic origin for the Lithuanian form, only attested in the sense ‘baptism’ (pace Skardžius 1931: 110; ALEW 607).13 Lt. krikščiónis, Lv. dial. (Zemgale) krišķāns ‘Christian’ may also have come via the same route (cf. Pr. iii crixtiānai* nom.pl.). This has both linguistic and non-linguistic implications. On the linguistic side, the -š- in Lt. krìkštas has sparked much discussion. Būga (1912: 3; 1925: 41) considered a possible German influence, while Endzelīns (1911: 60; 1937: 164) suggested contamination with Polish chrzcić /xš-/ ‘baptise’. However, the prevailing opinion is that the shift to -š- is due to the intrusive -k- (Endzelīns 1911: 35; Skardžius 1931: 110; Stang 1966: 14; Smoczyński 2018: 606). Whether this is chronologically plausible is uncertain; it is in any case worth noting that several words which show a pervasive -k- do not exhibit a subsequent shift to -š- in Lithuanian (e.g. Lt. áuksas ‘gold’ = Pr. E ausis; Lt. al̃ksnis, dial. alìksnis ‘alder’ = R ольха́ ; Lt. úoksas ‘tree hollow’ = Lat. ōs ‘mouth’). A Prussian transmission gives us a new possible explanation. In transitional Prussian–Lithuanian dialect areas, there would undoubtedly have been a significant level of bilingualism, providing the prerequisites for “etymological nativization” to take place. The regular correspondence between Lithuanian š and Prussian s may have been recognized by bilingual speakers, leading them to favour the seemingly counterintuitive substitution /s/ → /š/ over the phonetically more natural /s/ → /s/ (Mažiulis 1979: 147). A few other words suspect of being of Prussian origin may show a similar substitution strategy. At least the following can be cited: – Lt. (W Žem.; cf. LKA i: No. 82) bruñšė ‘roach’ ← Pr. E brunse ‘roach’ (PKEŽ i: 161) beside Žem. bruĩšė (< *brų̃ šė, cf. Trautmann 1910: 145) – Lt. (W Žem.) jū́šė ‘fish soup; slops, viscous liquid; mixture’ ← Pr. E iuse · juche 13 It is perhaps this semantic issue which encouraged Smoczyński (2018: 606) to interpret krìkštas as a back formation from the verb krìkštyti, 3pres. -ija ‘baptise’, which, in his opinion, is in turn derived from OR крьсти́ ти (Зализняк 2019: 365) ‘baptise’. However, the root stress in the verb rather points towards a denominal formation. Importantly, the Lt. form corresponds precisely to Pr. iii crixti-, pres.sg. crixtia ‘baptise’ (whose second -i- never shows a macron, implying the root here was also accented). As a result, the Lithuanian verb could be explained as a Prussianism, as well. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic–slavic contacts 13 – Lt. (W Žem.; PrLt.) kriáušė ‘pear’ ← Pr. E crausy ‘pear tree’, nom.pl. crausios ‘pears’. The dialectal limitation of this and the previous word speaks in favour of a Prussian borrowing (Būga 1915: 342).14 – Lt. šárvas, pl. šarvaĩ ‘armour; arms’ ← Pr. E sarwis ‘weapons’, which is no doubt ultimately from Go. sarwa n.pl. ‘weapons, armour’ (= OE searo, dat. sg. searwe ‘craft, wile’ and ‘weapon, armour’). Alternative etymologies (cf. LEW 965–966; PKEŽ iv: 65) derive from a need to explain the unexpected š-. – ⟨Ʒ́ ukmistras⟩ ‘Fischmeister’ (ClG i: 663)15 ← Pr. E suckis, iii suckans acc.pl. ‘fish’ From a non-linguistic perspective, this analysis removes one of the key arguments for an early informal Christianization from the east, which has been taken for granted since Būga (cf. Būga 1912: 11). Indeed, assuming a Prussian intermediary, one would hardly need to date such Christian terminology significantly earlier than Lithuania’s official adoption of Christianity in 1387. In Latvian, on the other hand, one does find a trace of early eastern Christianity: Lv. krusts, dial. krists ‘cross’, whose meaning points to an Old Russian source.16 Despite these ambiguities, the following ten cases seem to provide compelling evidence of early contacts between Slavic and East Baltic: – Lt. dial. bìrkavas, bìrkuva ‘a weight of 10 pūdai’, Lv. bir̃kavs ‘ship-pound’ ← ONovg. бьрковьскъ, a weight measure, frequent in Novgorod gramotas (cf. Зализняк 2004: 713 and Thörnqvist 1948: 29–32, where the ultimate connection with ML Birca, OSw. Biærkö is discussed) – Lv. krusts, dial. krists ‘cross’ ← OR крьстъ ‘cross’ (see above) 14 15 16 Note that kriáušė is now the standard word for ‘pear’, but the more widespread dialectal terms are the Slavic loanwords grūšià and dū̃lė. Although this is not the place to discuss the full material, I agree with Būga that the application of the RUKI law cannot be considered regular after *u and *i in East Baltic. The stressed form žùkmistras given by Brugmann (1897: 104), and found abundantly elsewhere (even in LKŽ) is apparently not attested (cf. Būga RR ii: 721). Mielcke (i: 341) and Nesselmann (1851: 552) have źukmistras without any stress mark (this form may ultimately derive from ClG). Kurschat (1883: 527) specifically notes that the word is “bei den Hafffischern ungebräuchlich”. Latvian -u- is perhaps to be explained as the result of a contamination with Lat. crux ‘crucifix’ (Endzelīns 1899: 301). Būga (1925: 42–44) rather blamed the -u- in Lv. rutks ‘radish’ (see below) and krusts on the Belarusian depalatalization of /r’/, but this is chronologically implausible; cf. Wexler (1977: 153). Lv. rutks may be explained by assuming a dissimilation *i–i > *u–i which would find a partial parallel in tìrgus ‘market’ < *turgus; however, it is usually assumed to have been influenced by ruds ‘red-brown’ (ME iii: 565). Note the other examples of a hesitation between -i- and -u- listed in Young (2009: 187). For an explanation involving Latgalian /y/, see Seržant (2006: 99–100). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 14 chapter 1 – ? Lt. Póvilas, Lv. Pãvils ‘Paul’ ← OR Павьлъ (Būga 1925: 44). However, see the discussion on p. 11 – Lt. pipìras ‘pepper’ ← OR пьпьрь, OPl. pierz, pieprz ‘pepper’17 – Lv. Pliskava ‘Pskov’ (whence German Pleskau) ← OR Пльсковъ (Būga 1925: 42) – Lt. ridìkas, dial. rudìkas, Lv. rutks ‘radish’ ← OR рьдькъı*, cf. редькови acc. pl. (13ᵗʰ c.), R, Uk. ре́дька ‘radish’, of West Germanic origin, cf. MLG reddik (1492 in Gaerde der Suntheit; MoLG Röddick) ‘radish’ (Pronk-Tiethoff 2012: 136–137) – HLv. dial. pluts ‘raft, barge’ (ME iii: 359) ← cf. R плот, OPl. ⟨plty⟩ nom.pl. (in Latin context; SSP vi: 167) ‘raft’. Probably of West Germanic origin; compare MDu. vlotscip ‘barge’, Du. vlot ‘raft’ (on which cf. Philippa et al. iv: 545–546; Kroonen 2013: 149)18 – Lt. obs. smirdas ‘peasant’ (cf. Būga 1925: 43), Lv. obs. smirds ‘poddany, subjectus’ (Elger 1683: 385; cf. ME iii: 966) ← OR смьрдъ ‘peasant’, cf. OPl. smard (SSP viii: 318) – Lt. šil̃kas (also plurale tantum: šilkaĩ ) ‘silk’ ← OR *шьлкъ, compare ONovg. шолкоу gen.sg. (Зализняк 2004: 541), R шёлк, Uk. шовк ‘silk’, most probably of Norse origin; cf. ON silki (REW iii: 387) – Lv. zizlis, also zizls ‘rod; spoke of a wheel’ (assimilated from *žizls?) ← OR жьзлъ ‘stick, staff’ Since Погодин (1903: 161–162), Lv. cìlvȩ̃ks, dial. cilȩ̄ks ‘person’ has been viewed as an early loanword from OR человѣкъ ‘person’ (cf. ME i: 382–383; Young 2009: 183; Derksen 2020: 40). There are two phonological issues with this derivation. First, Погодин’s Slavic preform *čilvēka- (trad. *čьlvěkъ) cannot be supported by any actual Slavic data. A slight improvement would be to start from a preLv. *cilavēkas with regular elision of the compounding vowel within Latvian (Endzelīns 1923: 187). However, even a form *чьловѣкъ is unattested. Although the evidence of OCS чловѣкъ is perhaps inconclusive (the word was only rarely spelled out), the complete lack of vocalization of prepositions in West 17 18 REW (ii: 341) has suggested the Slavic word may have been loaned through a Gothic *pipirs, and we might equally assume a borrowing through Gothic in East Baltic. Yet if the Slavic word is directly ← Lat. piper, as usually thought (cf. M. Matasović 2011: 118), then reconstructing such a Gothic term seems superfluous. A Slavic origin would be favoured by the non-initial stress of Lt. pipìras. By contrast, Lv. pipars ‘pepper’ is possibly borrowed from Swedish via Est. pipar (cf. ME iii: 221); compare Estonian Swedish pippar (Freudenthal/Vendell 1886: 163), OSw. pipar ‘pepper’ (EES s.v. pipar). A Germanic origin is rejected by Kiparsky (1934: 80), but without reference to the West Germanic data. Note that the similar MoHG Floß ‘raft’ goes back to *flōta- rather than *fluta-. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic–slavic contacts 15 Slavic would speak against Pl. człowiek deriving from a variant *čilavēka- (trad. *čьlověkъ; Havlová 1966: 80), and would favour an account starting from irregular syncope.19 At any rate, in East Slavic, the only usual form is of the type человѣкъ, cf. also Uk. чоловíк, Bel. чалаве́к ‘person’. As a result, a donor form with *i (trad. *ь) is questionable. Finally,20 Lt. grìkai, grìkiai, Lv. griķi ‘buckwheat’ are often interpreted as early loans from the (unattested) Russian *грьк- (Rūķe-Draviņa 1964: 125; Smoczyński 2018: 387; РЭС xii: 91) or directly from *грьча (> MR греча; Seržant 2008: 126). However, this appears to be excluded by the realia. Archaeobotanical evidence for buckwheat in the East Baltic region only certainly emerges in the 14ᵗʰ century, which coincides with the first documentary evidence (Sillasoo/Hiie 2007: 76; Grikpėdis/Motuzaitė Matuzevičiūtė 2020: 166). Moreover, the word is so far unattested in Old Novgorod-Pskov. Although an argument ex silentio, this fact is potentially significant, as terms for grains (пьшеница ‘wheat’, овьсе* ‘oats’, жито ‘barley’, ръжь ‘rye’) are abundantly present in the birchbark letters. The Russian word is not in fact found until the turn of the 16ᵗʰ century (gen.sg. гречки 1495, гречи 1498, гречихи 1500; СРЯ 11–17 iv: 132), where it emerges in Novgorodian trade books. In its German form, the word appears a century earlier in Riga (attested in 1383 as Kricken; Kiparsky 1936: 84; RūķeDraviņa 1964: 118), and crops up in the late 15ᵗʰ century in Prussian German Greck, Grick (PrWb ii: 513) and Old Polish grece, grice (1487/1488 in Latin context; SSP ii: 506). The forms with -e- are reminiscent of Middle German Grecken ‘Greeks’, greckisch ‘Greek’ (DWb ix: 256), opening up the possibility that the 19 20 Although seemingly a rather arbitrary suggestion, many of the Slavic reflexes of this word show irregular developments, which must result from frequency of use (to the lists in Berneker i: 140–141, and Havlová loc. cit. we can add R colloquial чек ‘person’). In addition, some Russian dialectal forms show an irregular raising, cf. Obojań чилэк (Шахматов 1915: 152), Rjazań цылье́къ (Даль³ iv: 1301), but so far, I have not identified any forms nearer to the Baltic territory. Other doubtful examples are the following: (1) Lv. dial. buca ‘barrel’ is more probably secondary for muca ‘barrel’ (as suggested by Mühlenbach in ME i: 344) under influence of R бо́чка ‘barrel’, rather than a direct loan from OR бъчи*, acc.sg. бъчьвь (СДРЯ 11–14 i: 331; on the Slavic form, cf. REW i: 113–114); (2) Lt. dial. cìrkva ‘church’ is a form apparently only recorded by Būga (1925: 42); in a genuinely old loan we would expect *cirkuva; (3) Lv. dukurs ‘polecat’ was probably adopted through E tuhkur (Kiparsky 1949: 65); (4) Lv. siruobs ‘notch at the end of a beam’ is hardly from OR *съроубъ (cf. R сруб ‘log frame’; ME iii: 848; Seržant 2006: 96); instead, the Lv. -i- may be epenthetic in the illegal cluster *sr-; (5) Lv. dial. timnica, timnice ‘dungeon’, timenîca ‘dark place; dungeon’ (compare OR тьмьница; Endzelīns 1899: 301; Būga 1925: 767) is perhaps rather built after tìmsa², dial. tima ‘darkness’ on the model of R arch. темни́ ца ‘dungeon’ rather than being directly borrowed from it. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 16 chapter 1 word was actually formed within German, and that R гре́чка as the ‘Greek grain’ is a calque coined in the context of Hanseatic trade. While this account remains quite uncertain, there is still no evidence that would allow us to backdate this word, or buckwheat cultivation, to the Old Russian period. 1.1.3 *TerT and *TarT As the East Slavic pleophony clearly predates the earliest texts, one has often hesitated to accept loanwords predating this development. With regard to the Finnic evidence, one has either assumed the existence of dialects which evaded pleophony (Mikkola 1894: 45–47; Kalima 1929: 165) or that a phonetically weak anaptyctic vowel was lost in the borrowing process (Setälä 1929: 34; Kalima 1956: 31–33; Kiparsky 1963: 83). The scepticism of earlier scholars is tied to a conception that any form lacking pleophony must necessarily belong to ProtoSlavic proper.21 However, this is certainly not the case; it has even been argued that the development spread through East Slavic as an areal feature (cf. Garde 1974: 112–115; Николаев 1988: 123–124; Крысько 1994: 18–19; Зализняк 2004: 39– 41). Bjørnflaten (2006: 66) has claimed that loanwords predating the East Slavic pleophony are only ascertained in Latvian. As far as Lithuanian is concerned, he refers to Zinkevičius (1987: 71), who argues that syncope cannot be excluded in any individual case. However, it is not clear why the same argument could not equally apply to Latvian: if we have Lv. pȩ̀lni ‘ashes’ (= Lt. pelenaĩ ) and ḕrglis ‘eagle’ (= Lt. erẽlis) (Endzelīns 1923: 47), then why not suggest Lv. kal̃ps ‘farmhand’ and HLv. kā̀rms² ‘building’ derive from an earlier *kalapas and *karamas, respectively? On the other hand, setting up hypothetical forms like these to explain away any relevant evidence (cf. Mikkola 1938: 25–26) would be circular. In the case of Lv. žer̃biņš ‘lot’ ← Bel. жэ́ рабя (Būga 1925: 37), the lack of lengthening before *rC proves that this sequence has arisen by syncope (Derksen 2020: 34, fn. 5; note also dial. žerebiņš, žeberis). Syncope is also quite imaginable in polysyllabic forms such as Lt. obs. čerpyčia ~ čerepyčia ‘roof tile’ ← Bel. чарапíца;22 Lt. karvõjus ‘wedding loaf’, dial. karavõjus ← Bel. карава́ й; 21 22 The position is exemplified by the statement of Колесов (1980: 69): “Все славянские языки изменили исходное сочетание типа *tort, но изменили по-разному. Следовательно, это изменение началось в праславянском языке” [“All Slavic languages modified the original sequences of the type *tort, but in different ways. It follows, therefore, that this change started in Proto-Slavic”]. The second statement, however, does not logically follow from the first. The first manuscript edition of Szyrwid has ćierpićia, which in the third edition is apparently corrected to ćierepićia. Both variants are also found in Bretke: cʒerpjcʒios nom.pl. and cʒerepijcʒes acc.pl. (see Skardžius 1931: 54–55; ALEW² s.v. čerepyčia). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic–slavic contacts 17 Lt. skavardà ‘frying pan’ ← Bel. скаварада́ (cf. Lv. dial. Vidriži skavardnīca, without lengthening of -a-). A case not explainable as a pre-pleophony borrowing is Lt. kalmaškà ‘light carriage’ ← Bel. калама́ жка (= Pl. obs. kolimaga ‘cart’; cf. Skardžius 1931: 99; Kiparsky 1948: 48; Zinkevičius 1966: 131–132). On the other hand, there are words in both Lithuanian and Latvian for which the assumption of syncope would seem ad hoc: – Lt. čérpė ‘clay pot; roof tile’ ← *čerpa- (trad. *čerpъ; > R черепо́ к, Bel. чэ́ рап ‘potsherd’, Pl. dial. trzop ‘clay pot’). – Lv. kal̃ps ‘servant, farmhand’ (= Lt. kálpas, if not from Latvian, cf. Derksen 2020: 34) ← *xalpa- (trad. *xolpъ; > OR холопъ ‘servant’) – HLv. kā̀rms², i.e. kùorms ‘building’ ← *xarma- (trad. *xormъ; > OR хоромъ ‘house, building’; СДРЯ 1387) – Lt. šálmas ‘helmet’ ← *šalma- (> OR шеломъ, шоломъ, OPl. in Latin context szłom ‘helmet’), earlier *šelma- (trad. *šelmъ). The last example must have post-dated the backing *CelC > *CalC, which Holzer (2001: 42) has described as a “North Slavic” areal change. Note that a similar backing does not seem to be reflected in Lt. šil̃kas ‘silk’, which I have accepted above as an East Slavic loanword. This might suggest a chronological or dialectal difference in the loan source. The word for ‘silk’ does appear to have had a backed variant already in at least part of Old Russian judging by Finnish sulkku, Veps šuuk ‘silk’ ← OR *шълкъ (cf. also R шёлк, Uk. шовк).23 Occasionally, it is difficult to decide between cognancy and borrowing. Lt. dial. (N) kar̃bas ‘basket’, Lv. kā̀rba ‘box; birchbark vessel, basket’ could be borrowed from OR (Novg.) коробъ ‘unit of measure’, R ко́ роб ‘bast or birchbark vessel’ (cf. Berneker i: 568; ME ii: 194; REW i: 629), but a regular cognate cannot be ruled out.24 A similar consideration applies to the derivative Lt. dial. 23 24 Mikkola (1894: 117) cited a Russian dialectal “шулк”, but it appears this is merely a hypothetical form based on Шахматов’s (1893: 296) claim of a sporadic shift o > [ọ ~ u] in Petrozavodsk Russian. This shift was later explained as a reflex of etymological *a (trad. *o) in accentually immobile words (see Л. Васильев 1929: 14). Since the vowel of R шёлк does not reflect an etymological *a, the Finnic u-vocalism should rather be taken as a direct reflection of *u (trad. *ъ; cf. Кулешов 2010: 349). Contra Berneker (cf. also LEW 220), it is not probable the Slavic word was borrowed from Germanic. First, the word is accentually mobile in Old Russian (Зализняк 2019: 527), which is atypical of Germanic loanwords (cf. Pronk-Tiethoff 2012: 242–244 with lit.). Secondly, from West Germanic *korb- (OHG korb, OS korf ‘basket, pannier’), I would anticipate Slavic *kurb- (trad. *kъrb-; see fn. 9); a Germanic loan may well underlie Cz. dial. (Machek 1968: 291) krb ‘dovecote’, SCr. kȑbulja ‘basket made of bark’ (РСА x: 449), as suggested by Berneker. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 18 chapter 1 (Szyrwid, S Aukšt.) kar̃bija, karbijà ‘kind of woven basket’ and its comparandum ONovg. коробьѧ ‘grain measure’ (Зализняк 2004: 749).25 1.1.4 Nasal Vowels The loss of nasal vowels appears to have happened very early in East Slavic; there is no trace of nasal vowels in any East Slavic dialect (cf. Shevelov 1979: 132).26 Although certain Norse loanwords seem at face value to have predated the loss of the nasal vowels,27 Шахматов (1915: 112–113) has suggested that this is illusory, and that syllable final nasals were simply omitted for phonotactical reasons. Indeed, omission plus compensatory lengthening better accounts for cases such as: OR Игорь (cf. MGr. Ἴγγωρ) ← OSw. Inguar (not *Ягорь); ижера ‘Ingrians’ ← Ingr. Inkeroin; ꙗкорь ‘anchor’ ← OSw. ankare (not *оукорь; cf. Thörnqvist 1948: 99).28,29 Similarly, for OR поудъ ‘a weight measure’ ← ON pund, one need not set up an intermediate form with *ǭ; instead, we could posit a direct substitution of *-unC- with *-ūC- (Brückner 1929: 142; Kiparsky 1934: 157); cf. similarly the names Асмоудъ, Веремоудъ (cf. ON Ạ́ smundr, Vermundr) mentioned in the chronicles (Thomsen 1877: 71–72; on these differently Николаев 2017: 28). An earlier nasal vowel is supposedly proven by Lt. pùndas ‘a weight measure’, which would be from Old Russian (Būga 1925: 28; Thörnqvist 1948: 75). The key argument here is that pùndas cannot be separated from the other 25 26 27 28 29 Pr. E tarbio (for *carbio) ‘meal box’ may, in turn, be analysed as a further Balto-Slavic cognate or as a loanword from Lechitic, cf. Pl. dial. krobia ‘large woven basket’ (PKEŽ ii: 117–118). Николаев (1995: 111) has claimed that a distinction between *ǭ and *ū (trad. ǫ and u) is preserved in some Carpathian Ukrainian dialects as /υu̯ / and /u̯ υ/, respectively, both of which supposedly had a number of allophones in free variation (idem: 107–108). I have been informed by S. Tarasovas that Николаев has rechecked the data using modern software, and (apparently) now rather considers these alleged reflexes to be phantoms. A key argument here is OR варѧгъ, MR ва́ рягъ (Зализняк 2019: 722) ‘Varangian’, cf. ON vǽringjar nom.pl., pointing to *vārę̄ga- (trad. *varęgъ). On the other hand, if we set up a source form *várjængja- (cf. Thomsen 1877: 121; Falk/Torp 1403), supported by MGr. βάραγγοι ‘Norsemen in service of the Greek emperor’, then we may also assume a direct substitution of -æng- with -яг-, without the need for an intermediary form with a nasal vowel. See below. Thörnqvist (1948: 105) ultimately settles on a reconstruction *ękorь, which is unlikely given the Norse a-. Note that Lt. iñkaras ‘anchor’ and Lv. ȩñkurs go back to Early MoHG enker ‘anchor’ (cf. ME i: 470) and do not support a Slavic nasal vowel, despite Berneker (i: 29). Perhaps this could account for OR грамота ‘letter, literacy, written document’ ← Greek γράμματα (compare fn. 10). Although geminates were lost in Middle Greek, this was not universal, and this geminate occurs in the usual environment where ‘spontaneous’ secondary gemination is also attested (Holton et al. 2020: 135–136). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic–slavic contacts 19 loaned weight measurement bìrkavas (cf. Derksen 2020: 37). On the other hand, Lithuanian also has a word pū́das (dial. also pū̃das), whose acute accentuation might suggest a relatively early loan (cf. 1.1.7). Since the variant pùndas may well have been adopted directly from, or have been influenced by, Baltic German pund (Alminauskis 1934: 106–107; Kiparsky 1948: 37; LEW 667), it can hardly be treated as a certain example. The only remaining plausible case is the ethnonym lénkas ‘Pole’,30 which has been derived from an early *lę̄xa- (trad. *lęxъ; > OR лѧхъ) (Brückner 1877: 103; Būga 1925: 33; LEW 356). Kiparsky (1948: 39; followed by REW ii: 84) has suggested the word may instead be taken from an unattested Lechitic source, yet there is no evidence this ethnonym was used as a self-designation by West Slavs.31 A unique archaism would not be too surprising, as names of ethnic groups often figure among the earliest borrowings,32 and the names of certain (presumably) Baltic tribes seem to have entered Slavic prior to the loss of nasal vowels: OR ꙗтвѧгъı acc.pl. (← *jātvingai), голѧдь (← *galind-); see p. 29. 1.1.5 Reflection of Slavic *ȳ One might anticipate that Slavic *ȳ (trad. y) would have been substituted with Baltic *ū, yet examples do not present themselves (Derksen 2020: 40). While the usual substitution for Slavic *ȳ (trad. y) in older East Baltic loanwords is long /ī/,33 in a number of cases, we also find -ui-, which is often interpreted as archaic (e.g. Būga 1911: 25; 1912: 10–11; Kiparsky 1948: 31; Seržant 2006: 97, fn. 7). However, it is remarkable that several of the examples involve a preceding labial: 30 31 32 33 Little faith can be given to the derivation of the ethnonym Unguras* ‘Hungarian’, recorded only in Daukša, from Slavic *ǭgra- (trad. *ǫgrъ, cf. Būga 1925: 24; not *ǭgura-, i.e. *ǫgъrъ: cf. Smoczyński 2018: 1561; Derksen 2020: 37). It is difficult to imagine what would motivate the early Lithuanians to borrow such an ethnonym, with there being no evidence of direct contact between Baltic and Hungarian-speaking groups, and with the latter having no particular folkloric significance in Lithuania. Instead, it seems obvious that this is a neologism created by Daukša on the basis of ML ungarus (Kiparsky 1948: 37), which would practically be proven by the fact that the Daukša himself uses the Polish loanword veñgras (viz. Wę́grų gen.pl.) in another passage. Neither is the more primary *lę̄d- (trad. *lęd-) reliably attested in West Slavic sources, although Hungarian lengyel (older lengyen) ‘Pole’, and the tribal name Λενζανηνοί mentioned in the 10ᵗʰ century De administrando imperio seem to imply a form *lę̄djān- (trad. *lędjan-; cf. REW ii: 84 for details and references). Thus OR литва ‘Lithuanians’ (see 1.2), *вьсь (cf. Шахматов 1916a: 19) ‘Vepsians’ ← *vepsä (REW i: 193), роусь ‘Rus’, probably ← Norse *rōþs- (REW ii: 551) and possibly Lv. krìevs ‘Russian’ (but see 1.1.6). Examples include Lt. dial. bagotỹrius ‘rich man’, kỹtras ‘sly’ (LEW 29, 261), tỹnas ‘fencepost’ (LKŽ; cf. on tuĩnas below), Lv. sìts² (= HLv. sèits) ‘full’ (Endzelīns 1899: 310; ME iii: 855). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 20 chapter 1 – – – – Lt. buĩlis ‘chervil, chives’, cf. MR быль (СРЯ 11–17 i: 364), Bel. быллё ‘weed’ Lt. muĩlas ‘soap’ ← Bel. мы́ ла ‘soap’ Lt. muĩtas, Lv. muĩta ‘toll’ ← MBel. мыто ‘toll’ Lt. smuĩkas, dial. smùikas ‘violin’ ← MBel. смыкъ (16ᵗʰ c.; ГСБМ xxxii: 3) ‘a kind of instrument’, cf. modern Bel. смык ‘bow (for a string instrument)’ – Lv. dial. vùikls² ‘deft, able’, cf. R dial. вы́ клый (Pskov, Novgorod; СРНГ v: 292) ‘experienced, able’ (ME iv: 676) In these cases, the diphthong need not demonstrate any particular antiquity, but may simply be a result of the strong velarization of labials before /ɨ/. A similar representation is found in Richard James’ 17ᵗʰ century English-Russian notebook (buic ‘a bùll’, muïla ‘sope’, but sït ‘satisfied’, yazïke ‘a tounge’; Ларин 1959: 23–24) and the Russian manual of Tönnies Fonne (buik ‘bulle’, muilo ‘seepe’ but iaszÿck ‘tunge’; Касьян 2012: 73–78; Hendriks 2014: 94). In Finnic, the substitution -ui- is also only attested after labials: Karelian vuitti ‘portion, share’ (← R выть); muila ‘soap’ (← R мы́ ло), cf. Kalima (1956: 41). Incidentally, this group need not be archaic in Finnic either; on the contrary, the Karelian buitto, puitto ‘as if’ must be recent (← R dial. бы́ дто, СГРС i: 242 ≪ бу́ дто; Kalima 1956: 65; РЭС v: 53). In Finnic, it is also remarkably difficult to find examples of *ū from Slavic *ȳ.34 In Prussian, however, /ui/ substitutes Lechitic *ȳ (trad. y) without exception: Pr. E waldwico ‘knight’ ← *valdȳkā (trad. voldyka), cf. OPl. włodyka ‘nobleman’; *suiristio (attested sutristio) ‘rennet’, cf. OPl. syrzisko ‘rennet’ (SSP viii: 155); wuysis ‘guard dog’ ← *vȳž- (trad. vyž-), cf. Pl. wyżeł ‘pointer’; Pr. iii zuit ‘genug’ ← *sȳta (trad. syto), cf. Pl. syty ‘full, satisfied’ (sceptically Levin 1974: 38–39). In East Baltic, there also remain a couple of examples of -ui- ← *ȳ without a preceding labial, which may indeed represent archaisms: – Lt. kùila ‘hernia’ ← OR *къıла, cf. MR ки́ ла ‘hernia, outgrowth’ (Зализняк 2019: 192)35 34 35 In unstressed position, we find *u after labials even in recent loanwords (F populi, E pobul ‘landless peasant; cottager’ ← R arch. бобы́ ль), so that cases like Veps kaput ‘hoof’ (← копы́ то) cannot serve as evidence. Even the *u in F dial. muula ‘lye’, Võ. mugõl, mukl ‘soap suds; lye’ (← OR dial. *мъıгло, cf. R мы́ ло ‘soap’; Ojansuu 1922: 139) is not necessarily probative, as it may as well be subsumed under the other examples of *ui, PF **muikla being phonotactically impossible. Kallio (2008a: 155) cites only Vt. dial. suura ‘home-made curd cheese’ (← сыр), but this is found alongside many other dialectal forms (e.g. Luditsa siira, Jõgõperä sõõra, Mati syyru) and singling out the variant with -uu- would seem like cherrypicking. Since Endzelīns (1899: 310; ME ii: 300), Lt. kuilỹs, Lv. kuĩlis, Pr. E ⟨tuylis⟩ */kuilis/ ‘boar’ have been derived from an Old Russian *къıль (thus also Trautmann 1910: 451; LEW 305; Derksen 2020: 41). However, such a source form cannot be set up (cf. Sabaliauskas 1968: Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic–slavic contacts 21 – Lv. obs. suits, suišs ‘excessive’ (whence probably Lt. suitus ‘abundant’ in Daukantas)36 ← OR съıтъ ‘full, satiated’, cf. R сыто будетъ ‘много, лишку, нескромны требованія твоя’ (Даль² iv: 386) – Lt. tuĩnas ‘branch in a wattle fence; fence post; picket fence’ ← OR тъıнъ ‘fence; defensive wall’37 These tie into the debate regarding the phonetic value of Slavic *ȳ (trad. y). Already since Miklosich (1878: 149–152), a minority viewpoint has been that *ȳ was pronounced as a diphthong /ui/, for which the evidence which has been adduced is so diverse that it seems difficult to dismiss out of hand (see Thomson 1927; Shevelov 1964: 377–379; Press 1986: 217–243; Касьян 2012: 84– 88).38 The theory that the development from *ū to /ɨ/ in Slavic went through an intermediate stage with a diphthong *ui (rather than through a direct delabialization as suggested by e.g. Kortlandt 1989a: 50) would explain the lack of loanword evidence for a stage with *ū, despite evidence for pre-vowel-shift values for all other vowels in the Slavic vowel system. According to Shevelov (1964: 378; and Levin 1974: 39), since -ui- still renders Belarusian -ы- (“which undoubtedly was and still is a monophthong”), the loanword evidence cannot be used. However, I see no reason to suspect that the above three loanwords should be late adoptions from Belarusian. As generally acknowledged (cf. Derksen 2020: 41), kùila cannot be a particularly young loanword, as it must have predated the change ъı > и after velars. Būga dates this to the 13ᵗʰ century in Belarus (1925: 52; cf. Колесов 1980: 155). Even though it might have taken place as late as the 14ᵗʰ century in Novgorod-Pskov (Зализняк 2004: 36 37 38 176; PKEŽ ii: 294). R dial. киля́ к, килу́ н ‘boar’ are clearly derived from ки́ ла ‘hernia’ (cf. горбу́ н ‘person with a hump’ < горб ‘hump’); compare the dialectal senses килу́ н ‘animal with a hernia; an animal (usu. piglet) with abnormal testicles; uncastrated boar’ (СРНГ xiii: 209). Perhaps, as Smoczyński (2018: 622) surmised, kuilỹs etc. was formed within Baltic from kùila ‘hernia’. In that case, HLv. kèiļs (= ķìlis², ME i: 388) ‘boar’ could have been similarly formed to ķìla ‘hernia, outgrowth’ (Bērzgale, EH i: 706), of the same origin. Alternatively, and perhaps more attractively, we could conceive of a relationship to Lt. kiaũlė ‘pig’ (see Sabaliauskas 1968: 175–177). The existence of the form suitis, attributed to Daukantas by Geitler (cf. LEW 937; also in Miežinis 1894: 232), is questioned by Būga (RR ii: 724). In the LKŽ, all of the data from Daukantas is listed under suitus. In view of PrLt. rùimas ‘space’ (cf. MLG rūm), šliùižė ‘sluice’ (cf. MLG sluse; Prussian German šlǖse; Alminauskis 1934: 129), šiùilė ‘school’ (cf. MHG schuole; Prussian German šoil; idem: 126), Prellwitz (1891: 35) has suggested that tuĩnas might be derived directly from MLG tūn ‘hedge, fence’, like Lv. dial. tûna, tũna ‘a fence of slanted planks’ (ME iv: 282). The circumflex accent seems to tip the balance in favour of a Slavic origin, however. The most striking indication perhaps remains the fact that this sound is rendered with the digraph ŭi in OCS: Glagolitic ⟨ⰟⰊ ~ ⰟⰋ⟩, Cyrillic ⟨ъı ~ ъи⟩. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 22 table 2 chapter 1 Archaic features in Slavic loanwords in Lithuanian and Latvian Reflection of yers *TerT groups Nasal vowels Slavic *ȳ -uo- for *ū Lt. Lv. 8 [+ 5] 2 [+ 2] 1 2 1 10 [+ 4] 2 [+ 1] 1 18+ 90–91), it is clear, in any case, that we are not dealing with a late Belarusian borrowing here. Based on this example, it would be reasonable to date the other two examples to an earlier date as well. In any case, there is nothing in these words that specifically favours a younger dating. 1.1.6 Reflection of Pre-Slavic *ọ̄ , *ẹ̄ In Latvian, Slavic *ū (trad. *u) is frequently substituted as /uo/, for instance: duõmât ‘think’, kàpuôsts ‘cabbage’, karuõgs ‘banner’, muõkas f.pl. ‘torment’, sùodît ‘punish, judge’, suõma ‘bag, satchel’ (← R ду́мать, капу́ ста, хору́ гвь, му́ ки, суди́ ть, сума́ , respectively). These have almost always been interpreted as archaic. Endzelīns (1899: 306) and Būga (1912: 14–15) suggested that they reflect a preserved Slavic diphthong *ou, but the communis opinio is now that they represent a monophthong *ọ̄ , predating the common Slavic raising to *ū (McKenzie 1919: 170; ME i: 533; Kiparsky 1948: 33–34; Young 2009: 178; Derksen 2020: 45). In Lithuanian, only one generally accepted example of this substitution is known: kuodẽlis ‘flax prepared for spinning’ ← R куде́ль. This state of affairs has led to the conclusion that the Latvian loanwords are generally more archaic than the Lithuanian ones (Bjørnflatten 2006: 67; Derksen 2020: 48–49). However, this is not consistent with the other lines of evidence for early loanwords, where the Latvian evidence does not significantly surpass the Lithuanian in any other category (see Table 2, above). It seems very unlikely that Latvian simply happened to borrow a large number of Slavic words containing *ū at an earlier date than Lithuanian. We therefore must agree with McKenzie (1919: 171) that treating the different reflexes as representing distinct chronological layers is unwarranted. One possible solution, hinted at by Derksen (2020: 43), is that Lithuanian and Latvian were in contact with distinct varieties of East Slavic. It is conceivAnthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic–slavic contacts 23 able that the reflex of Slavic *ū was pronounced closer to */ọ̄ / in the East Slavic dialect with which Latvian was in contact. A similar contrast in the source dialects probably underlies the difference in outcome of Slavic *ē (trad. *ě), which has yielded -ie- in Lithuanian but -ē- = */æː/ in Latvian, cf. Lt. dial. griẽkas but Lv. grȩ̀ks ‘sin’ (cf. Bjørnflaten 2006: 68; Derksen 2020: 39). In this respect, we can note sporadic instances of /’a/ from *ē recorded in modern north-western dialects, cf. NW dial. кяп ‘flail’, Arxangeľsk ря́ па ‘turnip’ (= R цѣп, рѣ́па; Николаев 1990: 60), ця́ лой ‘whole’, Vladimir медвя́ дь ‘bear’ (= цѣ́лый, медвѣ́дь; Галинская 1993: 39–40), which indicates that the difference was at least partly dialectal rather than diachronic (cf. Derksen 2020: 47).39 On the other hand, Seržant (2006) has offered a plausible alternative account for Latvian /uo/: he suggests that it is the result of a dialectal diffusion from east to west. Latgalian (High Latvian) has undergone a chain shift *ọ̄ > ū; *ū > ou ~ yu (Endzelīns 1923: 95–97). If this predated the influx of Slavic loanwords, one would expect East Slavic /ду́мать/ to have been adopted directly as High Latvian dùmuôt’. This would in turn be nativized in Low Latvian as duõmât, following the typical dialectal correspondences.40 Since it is generally assumed that the early loanwords in Latvian are of East Slavic origin, such a trajectory would not be surprising.41 Interdialectal borrowing might also explain the Žemaitian forms pọ̃ uks ‘down (of a bird)’, ọ̃ ustâ nom.pl. ‘moustache’, which ultimately derive from East Slavic пух ‘down’, усы́ ‘moustache’ (cf. Zinkevičius 1966: 79, fn. 27 and 84, fn. 30). These may be hypercorrections based on South Žemaitian dialects where *ọ̄ has merged into *ū (see the discussion in Būga 1912: 23–26). A similar explanation probably accounts for NW Žem. viẽšnė (= vẹ̃išnẹ) ‘cherry’, which Būga (1922: 177) has seen as an early loan from Slavic *vẹ̄šnjā (cf. R ви́ шня), but is more likely a hypercorrection based on South Žemaitian vĩ·šnẹ (cf. Aukšt. vyšnià). 39 40 41 Compare the substitution of this phoneme as *ǟ in the Finnic loanwords (Kalima 1956: 37–38), matching Latvian. In this respect, note that prùods² ‘pond’ (← пруд; ME iii: 400) seems to be an exclusively High Latvian word, so actually represents [prùds] (cf. http://vuordineica.lv/, s.v. dīķis). The spellings in ME and Ulmann (prohds ‘ein kleiner natürlicher Teich’; 1872: 212) are automatic transpositions of the dialectal form. From a High Latvian perspective, forms such as kìukûļi /kùkuļi²/ ‘corn cockle’ (← куколь) and pỳuka (cf. pùka², ME iii: 445) ‘fluff’ (← пух), which have undergone diphthongization of original *ū, might even be interpreted as more archaic (cf. Seržant 2006: 95). Similar evidence of interdialectal diffusion is shown by the Žemaitian diphthong /ie/, which occurs instead of the usual reflexes of *ẹ̄ in Aukštaitian loanwords, including words of ultimately Slavic origin, e.g. griẽks instead of *grẹ̃iks ‘sin’ (Būga 1912: 7–8; Zinkevičius 1966: 86; Derksen 2020: 40). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 24 chapter 1 While a number of examples supposedly reflecting a preserved Slavic *ẹ̄ (< *ei) have been presented, the only convincing one is the Latvian ethnonym krìevs ‘Russian’, which is apparently loaned from OR кривичи (Endzelīns 1899: 285–286, 304–305, ME ii: 284–285; Būga 1922: 177; Skach 2010: 137; Derksen 2020: 38).42 As this word is an ethnonym, a unique archaism is conceivable, but we may suspect that krìevs similarly results from interdialectal diffusion. This seems, at least, to be the best account for Lv. siẽrs ‘cheese’, an evident borrowing from East Slavic сыр ‘cheese’, but one whose vocalism has presented a problem (ME iii: 859). This can be resolved if we assume the word was adopted first as HLv. sìrs (Seržant 2006: 97) and then subsequently borrowed into Low Latvian. Such an explanation also works for the Lv. agent suffix -(e)niẽks (cf. Latgalian -inìks; Seržant 2006: 96–97), provided this is indeed loaned from East Slavic -ьникъ (for an analysis as cognate, see Derksen 1996: 185–186). In fact, the loanword evidence for a Slavic stage *ẹ̄ is very poor; almost all of the evidence traditionally adduced from Finnic is doubtful or demonstrably false. The Finnish and Karelian (Olonetsian) agent noun suffix -niekka (McKenzie 1918: 172; Mikkola 1938: 33–34; Kiparsky 1948: 31; Derksen 2020: 39) is problematic, as long vowels outside of initial syllables were not possible at the time of the earliest contacts with Slavic43 and the diphthong /ie/ is generally not permitted at all outside of initial syllables in most Finnish dialects. The vocalism must have something to do with the reanalysis as a compounding element (J. Pystynen p.c. June 2023), which is supported by its abundant attestation as a separate word in earlier Finnish (Vanhan kirjasuomen sanakirja, s.v. niekka). Aside from this, there is K viehkuri ‘gust of wind’ and miero ‘the (outside) world; township, village council’. That all of the examples are limited to Finnish and Karelian obviously speaks against this being a particularly archaic loanword stratum. However, Kallio (2006: 155) and Derksen (2020: 39) are both unconvinced by the suggestion of a Russian dialectal development i > e (Mikkola 1894: 57).44 While it is true that no regular dialectal change can be set up, the fact remains that both words are actually attested with /e/ in North Russian dialects. R dial. ве́хорь ‘strong gust of wind’ is relatively widespread (СРНГ iv: 208: Kem’, Petrozavodsk; СГРС 42 43 44 Other suggested examples such as Lt. obs. mieras, Lv. miêrs ‘peace’ are better interpreted as cognates with the Slavic forms (Derksen 2015: 316; ALEW 747). As witnessed by the reflection of yať in the loanwords K netäli, Võ. nätäľ -i, Li. nädīļ (< *nätäli) ‘week’ ← недѣ́ля ‘week’ and F veräjä, E värav (< *väräjä) ← верея́ ‘gatepost’ (< *verējā-, trad. *verěja; cf. OCS верѣꙗ ‘bar, bolt’). Skach’s (2010: 138) suggestion that these represent relics of a Russian dialect preserving *ẹ̄ seems completely gratuitous to me. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic–slavic contacts 25 ii: 88–89: Arxangeľsk), and most probably results from a contamination with вѣ́тер ‘wind’.45 Olonets (19ᵗʰ c.) мѣръ ‘community’ (Куликовскій 1898: 58; СРНГ xviii: 112) appears isolated, and I cannot explain it within Russian (cf. Мызников 2019: 496), but the Karelian data remains highly dubious grounds for assuming a layer of loans in Finnic with preserved *ẹ̄.46 On the other hand, Finnic does provide clear evidence of an earlier *ọ̄ (cf. Kalima 1956: 42; Kallio 2006: 155). The main disadvantage of the ‘dialectal diffusion’ scenario outlined above is that this situation in Finnic would have to be divorced from the superficially similar situation in Latvian. However, this is not necessarily a problem, as we are dealing with two distinct contact zones. In theory, it is possible that both scenarios are correct, and that an earlier layer of loanwords with *ọ̄ was bolstered by a later layer adopted with Latgalian ū. In this case, however, the Latvian evidence can only be used as indirect support of early contacts with Slavic. 1.1.7 Accentuation As Derksen (2020: 41–42) has observed, there appears to be a correlation between the intonation of the oldest Lithuanian loanwords and their accentuation in Old Russian. While circumflex is generalized in later loanwords, the circumflex examples from the oldest layer appear to correspond to Old Russian oxytones. As noted by Young (2009: 184–185), the same group appear to show a falling tone in Latvian: – Lt. pul̃kas 4 ‘crowd, troop’, Lv. pùlks — OR (14ᵗʰ c.) полкы̀ nom.pl., R полка́ gen.sg. (cf. Зализняк 1985: 134; 2019: 569; Pronk-Tiethoff 2012: 112; Николаев 2020: 290). – ? Lt. tuĩnas 4/2 ‘branch in a wattle fence, etc.’ — MR (16ᵗʰ c.) тыно́мъ inst.sg. (Зализняк 2019: 601);47 cf. Slk. dial. týň ‘branch in a fence’. 45 46 47 A parallel development is found in MUk. (17ᵗʰ c.) вѣхорь, which is actually attested in collocation with вѣтеръ (РЭС vii: 270). According to Kallio (2006: 155) the -hk- and -u- in viehkuri would favour an early borrowing. However, -hk- is also found in some very recent loanwords, e.g. F orehka ‘gingerbread, cookie’ < R орѣ́х ‘nut’, and is not probative. The -u- is probably due to the analogical introduction of the suffix -uri as in F tuhkuri ‘mink’, E tuhkur ‘polecat’ (← OR *дъхорь), F dial. pippuri ‘pepper’ (← Sw. dial. pipar), F ankkuri ‘anchor’ (← Sw. ankare) (Kiparsky 1949: 60). Note that -u- is also found in the younger variant vihuri. It should be admitted that modern Russian generally suggests accent paradigm (c); cf. early modern Russian тыновы́ й adj., and тыни́ ть, 3sg.pres. тыни́ т ‘to fence’ (Сл. Акад. vi [1794]: 344; also dial., cf. СРГК v: 543). The evidence for accent paradigm (b) supplied by Зализняк is very limited, but note also OCz. o-týniti ‘enclose, cover’, SCr. (Vuk) tíniti, 1sg.pres. tȋnīm ‘partition’ (RJA xviii: 333). Pronk-Tiethoff (2012: 90), apparently on the basis of SCr. dial. tȉn ‘partition wall’, analyses the word as having fixed initial stress, but does not take into account the evidence of the derived verb. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 26 chapter 1 – Lt. tur̃gus 2 (rare dial. turgùs 4) ‘market’, Lv. tìrgus — MR до торгу́ ‘to market’, торго́ выи adj., etc. (Stang 1957: 81; Зализняк 1985: 134, 2019: 534). On the other hand, many of the examples with a Lithuanian acute correspond to Old Russian barytones. Here we find a sustained tone in Latvian (Young 2009: 179–181): – Lt. bìrkavas ‘a weight measure’, Lv. bir̃kavs (ME i: 298; but [bìrkàu̯ s] in LVPPV 130) — R бе́рковец, cf. dial. (Pskov) бе́рковец ‘weight measure for flax’ (РЭС iii: 132–133) – Lv. kal̃ps, (Lt. kálpas 1) ‘servant’ — OR (Merilo) холωпъ, MR холо́ пъ (Зализняк 2019: 602); SCr. arch. hlȁp, gen.sg. hlȁpa (cf. Skok i: 671) – Lt. kùila 1 ‘hernia’ — MR ки́ ла, Uk. ки́ ла; cf. Cz. kýla, Slk. kyla, SCr. kȉla (Зализняк 1985: 132, 2019: 192; Derksen 2008: 265) – Lt. kùrtas 1 ‘greyhound’, Lv. kur̃ts (but LVPPV: kùrts) — R хорт, cf. dial. хо́ ртица ‘(female) greyhound’ (СРНГ li: 316); SCr. hȑt, Sln. hr̀t, gen.sg. hŕta (Derksen 2020: 41) – Lt. lénkas 1 (although LKŽ reports variants with 2, 3 and 4) ‘Pole’ — MR ля́ хи nom.pl. (cf. Зализняк 2019: 752) – Lt. Póvilas, Lv. Pãvils ‘Paul’ — OR Па́велъ (Young 2009: 180; Зализняк 2019: 842) – Lv. (Janševskis) suĩtums ‘Menge’ = dial. sùits², sùitâk² (cf. ME iii: 1116) — MR сы́ та nom.sg.f. = SCr. sȉta ‘satiated’ (Derksen 2008: 484; Зализняк 2019: 494) – Lt. šálmas 3 ‘helmet’ — Although synchronically oxytone in Middle Russian (cf. Зализняк 2019: 588; Николаев 2020: 313), the word must originally have been barytone; cf. SCr. šljȅm, Sln. šlẹ̀m, gen.sg. šlẹ́ma ‘helmet’ (PronkTiethoff 2012: 87) This correlation provides another argument in favour of the late origin of Lt. muĩlas 4 ‘soap’, which corresponds to the barytone MR мы́ ло, cf. Slk. mydlo, Sln. mílọ (Derksen 2008: 336), and of Lv. dial. bul̃vāns ‘decoy bird’, which Young (2009: 186) has noted as an exception to his accentological rules. Note that Young uses a much larger corpus of Latvian data, while I have limited myself to cases which unambiguously belong to the earliest period. However, the accentual rules seem generally to apply even within his larger data set. The same cannot be said of Lithuanian, where the circumflex clearly dominates in the remainder of the material (cf. Derksen 2020: 41). A remaining exception is Lt. šil̃kas 4 ‘silk’, the source of which appears to have been barytone (cf. Зализняк 2019: 568). I do not have an explanation for this form at this time. There is rather little evidence for borrowings of accentually mobile forms. The clearest examples, perhaps counter-intuitively, appear to have been adopAnthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic–slavic contacts 27 ted as acute in Lithuanian (note also Lt. bíesas ‘demon’, svíetas ‘world’, from accentually mobile бѣсъ, свѣтъ; Derksen 2020: 42): – Lt. obs. tùlkas 1 ‘interpreter’, Lv. tul̃ks — MR то́лка gen.sg., толковы́ и adj. (Зализняк 2019: 569); cf. R бе́з толку ‘in vain’ – Lt. čérpė 1 (but dial. also čer̃pė 2, čerpė ̃ 4) ‘clay pot; roof tile’ — MR че́репъ (Зализняк 2019: 602); cf. R черепа́ nom.pl. However, Lt. dial. kar̃bas 4 ‘basket’, Lv. kā̀rba, if it is loaned from Slavic rather than cognate, would represent an exception, cf. MR ко́ робъ, R dial. и́ з короба (Зализняк 2019: 527). All in all, the evidence is rather too scanty to draw any conclusions. Young (2009: 186, 187) in fact reaches the opposite conclusion (i.e. mobile nouns are borrowed in Latvian with falling tone) based on evidence which has not come into consideration here. The regular adoption of Slavic barytones with acute accentuation in East Baltic is, however, clear from my data set, and demonstrates that an inherited accentological contrast was still present in Slavic at the time of the earliest loanwords into Baltic. 1.1.8 Semantics, Dating and Context Much of the above evidence seems to indicate contacts in the context of trade. These include the following (the loans which are not as certain are given here in square brackets): – Words connected to the act of trade: ‘market’, ‘interpreter’ and perhaps ‘raft/barge’. Considering their general limitation to adverbial usage, it is possible that Lv. suiti ‘excessive’, suitāk ‘too much’; Pr. iii zuit ‘genug’ originated as trade jargon. Compare similarly Italian basta ‘enough!’, which has been borrowed widely, in many cases, presumably, through trade (cf. e.g. Snoj 2003: 33, s.v. bȃsta). – The names of vessels: ‘clay pot’, [‘tub’, ‘kettle’]; and weight measurements: Lt. bìrkavas, [pū́das]. – The names of specific trade items: ‘pepper’, ‘radish’, ‘silk’, probably ‘greyhound’, [‘glass’]. In view of the borrowed term bìrkavas, it would seem obvious to associate these trade relations with the Birka trade network (thus explicitly Būga 1913: 34–35); however, it is disturbing that there is no unambiguous evidence of direct Norse loanwords in East Baltic (see Chapter 2), and besides, the weight measure in question continued to be used after the collapse of Birka as a trade hub, being even recorded in the modern dialects of Pskov Region. At the same time, this term does place us in a rather narrow timeframe between the establishment of the eastern trade with Birka in the late 9ᵗʰ century (Ambrosiani 2005) and the loss of the reduced vowels in Novgorod Russian in the early 13ᵗʰ (Зализняк 2004: 60). Lt. čérpė ‘clay pot; roof tile’, if indeed transferred in a trade context, Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 28 chapter 1 would place the contacts firmly in a preliterary context, the 10ᵗʰ century at the latest. We can therefore assume that most of these loans were adopted between the 10ᵗʰ and 12ᵗʰ centuries ce. There are some words which suggest an unequal power dynamic between the two groups. First, there are the words for members of the lower echelons of society: ‘peasant’, ‘servant/farmhand’, [kùmetis ‘serf’], and words connected with the military: ‘helmet’, [‘armour’].48 Finally, there is the Latvian word for ‘cross’, which suggests that, like Finnic-speaking populations (Kiparsky 1952: 70–71; Kallio 2006: 156), Latvians were subject to early attempts at Christianization on the part of the Slavs. The only possible evidence of this in Lithuanian is the Christian name Póvilas ‘Paul’ (but see the discussion on p. 11). All of these loanwords are suggestive of Slavic cultural imposition, and therefore can be classed as typical ‘superstrate’ words (cf. Vennemann 2011: 240).49 Of at least 20 certain loanwords, six (just under a third) are found in both languages. in most cases a Proto-East-Baltic reconstruction can be provided. Nevertheless, this fact can be attributed to the small number of phonological changes which have taken place, and need not compel us to assume these loanwords were already present in Proto-East-Baltic. The idea of loanwords into separate languages is supported by the rather large number of ‘old loans’ limited to one of the two, while many of the shared loanwords represent trade terms which have been borrowed into numerous other languages. Contact with different East Slavic dialects seems to be implied, at least, by the establishment of different substitution strategies for the phoneme *ē (trad. *ě). In terms of the source dialect(s), the following can be said: – The borrowing of the weight measurement бе́рковец, frequent in NovgorodPskov sources and preserved in this area in the modern dialects, suggests contact with the dialect of Novgorod-Pskov, which is also supported by the early adoption of Latvian Pliskava ‘Pskov’. – In addition, the adoption of OR търгъ as a u-stem might also favour a Novgorod-Pskov source, as the morphological distinction between a- (trad. o-) and u-stems was much better preserved in this dialect than in the rest of Slavic (Николаев/Хелимский 1990; Зализняк 2004: 99–102, 112). 48 49 Here also belongs Lt. pul̃kas ‘regiment’, but I suspect that the dialectally better represented sense ‘crowd’ is original and the military sense may be due to more recent Slavic influence. Finally, a small number of loanwords are too vague to be categorized: ‘building’, ‘rod’, ‘fence post’. Since the original specific function of these borrowed terms cannot be determined, they may have been loaned in any number of contexts. Surprising is the word for ‘hernia’, as no other medical terms or terms for bodily defects appear in my corpus. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic–slavic contacts 29 As a result, the general picture is that the earliest contacts between speakers of Lithuanian and Latvian with East Slavic took the form of relatively incidental adstrate contacts with similar but distinct dialects of Old North Russian, primarily in the context of trade. At the same time, there is some limited evidence of Slavic cultural imposition, which suggests a degree of Slavic dominance in these contacts. 1.2 Early Baltic → Slavic Loans? There are very few loanwords which can be plausibly dated to the same period as the early Slavic → Baltic loans, and reference works on the subject (e.g. Лаучюте 1982, Аникин 2005) tend to focus on the much more extensive later layer of loanwords. Most of the plausible early loanwords which can be identified are ethnonyms. In the Novgorod First Chronicle, we find examples like OR лит[ъ]ва ‘Lithuanians’ (= R Литва́ , Pl. Litwa ‘Lithuania’; ← Lt. Lietuvà; cf. Lv. Lìetava, ME ii: 506); ꙗтвѧгъı acc.pl. ‘Jatvingians’, голѧдь ‘Galindians’ (cf. the region Galindia cited in Chronicon terrae Prussiae and the Γαλίνδοι in Ptolemy), корсь < *кърсь ‘Curonians’ (cf. Lt. Kur̃šas, Lv. Kur̂sa² ‘Curonia’); see the overviews in Fraenkel (1950a: 60–73) and Dini (2014: 290–312). A list of probable early loans has been given by Аникин (2014: 192), who divides them into several chronological layers. As ‘Proto-Slavic’ loanwords, he quotes R дёготь ‘birch tar’ and dial. перть ‘cottage’.50 As Baltic substrate words in early East Slavic, he cites дере́вня ‘village; (dial.) arable field’, dial. а́ лес ‘damp spot’, му́ма ‘bogey; (in children’s language) louse’ and пу́ сма ‘bundle’. Аникин evidently means to compare му́ма ‘bogey’ (Brjansk, СРНГ xviii: 344) with Lt. maũmas in the same sense, where *au > /u/ would suggest a very early date; but compare also Lt. dial. mū̃mas (LKŽ), which is evidently the origin of dial. му́ма ‘louse’ attested in Lithuania (Лаучюте 1982: 146). It seems far more probable that dial. му́ма is merely an arbitrary formation like Hungarian mumus ‘bogey’ (in children’s language; note also Lt. baũbas, bùbas in the same sense), and treating it as an exceptionally early loanword is unwarranted. Among the ‘early’ loanwords, Kiparsky (1973: 68–69) has mentioned па́ кля ‘(flax or hemp) tow’ (← Lt. pãkulos; cf. also Аникин 2005: 24) and ковш ‘ladle, 50 Аникин also cites R клѣть ‘storehouse’, with widespread Slavic cognates, as a Baltic loanword (cf. Lt. klėt́ is, Lv. klẽts). Here, he follows Eckert (1983: 86–87); however, Eckert’s main argument, namely that the Baltic word is derivable from the verbal root klóti ‘lay out’ is already refuted by Аникин himself (2005: 170), and there seems no other reason to prefer a Baltic source over an inherited cognate. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 30 chapter 1 water scoop’ (← káušas). There is no reason to consider the latter to have been adopted any earlier than the 14ᵗʰ century (cf. СРЯ 11–17 vii: 216); the same sound substitution is found in Vytolth 1386, Витовт 1396 ← Výtautas (Būga 1911: 36– 37), and remained usual throughout the 15ᵗʰ century (op. cit. passim). As for па́ кля, the substitution a → a clearly rules out an early date. The loss of the second syllable of the Russian word must be secondary; cf. dial. (N) па́ кула, Bel. па́ кулле, Pl. pakuły ‘tow’ (Лаучюте 1982: 18–19); compare the discussion of R dial. па́ ккула beside па́ кля ‘chaga (parasitic fungus)’ in Мызников (2019: 571–573). The following cases deserve a more detailed discussion: ‘bath-house’. ONovg. *пьрть ← Lt. pirtìs, Lv. pìrts ‘bath-house’ — The word is attested in the Sermon of Ilya of Novgorod (перьти dat.sg. in a 15ᵗʰ century copy, cf. СДРЯ 1772), where it probably referred to a kind of bath-house (Павловъ 1890: 19). This is supported by MR переть ‘bath-house’ (Pskov, 15ᵗʰ c.; Гальковскій 1913: 34; cf. Вахрос 1963: 157), пере́дка ‘hut’ (Pskov, 16ᵗʰ c.; СРЯ 11–17 xiv: 298) and the dial. (Novgorod, Karelia) derivative при́ переток ‘dressing room in a bath-house’ (Мызников 2019: 625; cf. СРЯ 11–17 xix: 245). The Baltic source has an impeccable internal etymology: it is a derivative of Lt. per̃ti, Lv. pḕrt ‘beat (e.g. with a besom); bathe’.51 Existing etymological discussions make the mistake of conflating the above forms with R dial. (Kem’) перть ‘Karelian cottage’ (Подвысоцкій 1885: 120). The latter, however, in view of its meaning and geographical isolation, is most certainly a recent loan from Karelian pertti ‘hut, cottage’ and not a direct continuation of the Old Russian form. Incidentally, the Novgorodian word has itself also been derived from Finnic (Būga RR ii: 516; Лаучюте 1982: 89; ALEW 899), yet given the meaning ‘bath-house’ in the earliest attestations, a Baltic origin is semantically more attractive. For a further discussion, see pp. 140–142. Despite the former’s narrow distribution, the Russian and Baltic words have often been interpreted as cognates (Vasmer 1909: 142; Trautmann 1923: 215; REW ii: 344–345; Nieminen 1953: 214–215; Derksen 2015: 358–359). The main argument for a native Slavic origin is the existence of R па́ перть ‘church porch’, which has a much broader distribution within East Slavic, and has an OCS cognate, папрьтъ (SJS iii: 14), in the same sense. However, that these contain the same root is not self-evident. Beside the semantic obstacle (Преображенский ‣ 51 The older meaning is ‘beat’ (cf. OCS пьрѣти сѧ ‘argue’, пьрꙗ ‘fight, dispute’). Lt. pirtìs is also attested as a verbal noun ‘bathing; flogging’, and it has been recorded as a root noun (Zinkevičius 1966: 265), which might make a direct connection with Skt. pr̥ t́ - ‘battle, strife’ possible. On the other hand, the attestation as a root noun seems to be limited to areas where root noun inflection became productive (cf. Zinkevičius 1966: 263). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic–slavic contacts 31 ii: 47), we can add that the variant папортъ (СРЯ 11–17 xiv: 148) might imply an original *purt- (trad. *pъrt-; with па́ перть due to yer assimilation? cf. Соболевскій 1910: 116–117). ? ‘arable field’. R дере́вня ‘village; (dial.) arable field’ ← Lt. dirvà ‘arable field’ (Schmid 1977: 51–53; Аникин 1998: 319, 2014: 192, РЭС xiii: 230) — Schmid argues in favour of a Baltic origin, noting that place-names containing the word are concentrated in the Upper Dniepr and north of Moscow, which in his opinion would be consistent with the area of Baltic influence. The loanword etymology implies the so-called “second pleophony”. Curiously, this development has been assumed in this word even by scholars who do not favour a Baltic origin (e.g. REW i: 341; Sławski SP v: 57–58; Derksen 2008: 136), where it appears only to have been motivated as a means to more directly equate the Baltic and Slavic forms. However, if we assume the words are cognates, the older reconstruction *derv- (Berneker i: 186), with a different ablaut grade, can hardly be ruled out (compare, with o-grade, ME tare ‘vetch seed; vetch’, MDu. tarwe ‘wheat’ < *tarwōn-). We can note that the second pleophony is usually a dialectal phenomenon that rarely has a pan-East-Slavic distribution. Николаев (2001: 88), besides дере́вня, cites only верёвка ‘rope’, but the latter could just as well be built analogically from вервь (in many places /ver’v’/) on the model of e.g. селёдка : сельдь ‘herring’. The main argument in favour of a loan is the word’s narrow distribution. On the other hand, a suitable Baltic source is unattested. Schmid (1977: 52) assumes an original syntagm *dirvinē žemē ‘arable land’ (cf. Lt. dir̃vinis ‘related to dirvà’) was subsequently substantivized in Russian. If the loan etymology is valid, it would be equally acceptable to start from an unattested nominal derivative *dirvinē ‘arable field’ already in Baltic. Despite the doubts of Аникин (РЭС xiii: 231), I find it at least possible that дере́вня ‘cleared land; arable land’ is the same word as R dial. (W) дере́вня ‘(pile of) logs’, Uk. arch. дере́вня ‘timber’ and is therefore derived from де́рево ‘tree’ (cf. Jēgers 1969: 79; Vaillant 1974: 608). A possible semantic path could be ‘felled trees’ → ‘area where trees are felled’ → ‘cleared land’. Needless to say, this remains hypothetical. ? ‘carrot’. R dial. борка́ н (Novgorod, Pskov, Smolensk, Kaluga, etc.) ← Lv. bur̃kãns ‘carrot’ (Karulis i: 155; РЭС ii: 222; Pronk/Pronk-Tiethoff 2018: 282) — Traditionally, the Latvian word has been derived from Slavic (Būga 1925: 48; REW i: 108); however, the limited distribution of the word within Russian implies the opposite directionality. The Russian vocalism would imply an early borrowing as OR *бърканъ (РЭС ii: 222), which, while theoretically acceptable, is rendered slightly awkward by the late attestation of the word within Russian (since 1564 apud СРЯ 11–17 i: 294; cf. Bentlin 2008: 247).52 ‣ ‣ 52 Аникин interprets dial. бурка́ н as a later Letticism, but the limitation of this form to Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 32 chapter 1 The picture is further complicated by Baltic German Burkane, Borkane, which could, phonologically speaking, just as well be the source of the Latvian and Russian words (Mikkola 1894: 91; РЭС ii: 223). Иллич-Свитыч (1960: 17) has argued that the Baltic German words are instead Balticisms, but does not address Kiparsky’s (1936: 201–202) argument that the stress — /burkánə/ — would speak against this. Kiparsky also argues against a Russian origin, stating that the form Purkahne (from 1577) is attested “lange vor Beginn der russischen Zeit”. Masing (1926: 80) connects the Baltic German forms to MLG brackannyen nom.pl. appearing among a list of edible roots in the Loccumer Historienbibel (15ᵗʰ c.).53 This is supplemented by Marzell (ii: 62–64) with some Scandinavian dialect forms that appear to be of Low German origin, cf. Early Modern Danish brekanne-rod 1550, barkena-roer 1738 ‘carrot’ (ODS s.v.), Sw. dial. (SW) barkanrot, barken-rot ‘carrot’ (with rod, rot, etc. ‘root’); cf. also Bentlin (2008: 248–249). Indeed, as Marzell states, it seems almost inconceivable that these forms are unrelated, yet the Low German a-vocalism is hardly reconcilable with Baltic German -u-. Most probably through Russian, the word has spread to F porkkana, Vt. borkkana ‘carrot’, Võro põrḱnas ‘carrot’ (Mikkola 1894: 91; Kalima 1956: 107; SKES iii: 604; Plöger 1973: 141; SSA ii: 375).54 It seems that Livonian borkõn must also be derived from East Slavic; at any rate, Kettunen (1938: 26) denies the possibility of a late Latvian loan. Beyond this, analysing the exact routes of borrowing is highly challenging, and the word can at best be characterized as a circum-Baltic term which has spread as a local Wanderwort. On Moksha puŕʿkä ‘carrot’, and for a discussion of the word’s ultimate origin, see pp. 229–231. ? ‘drying barn’. R ови́ н ‘drying barn’, Bel. авíн ‘granary’ ← Lt. javaĩ ‘cereals’ (Andersen 1996a: 154̄–155; Pronk/Pronk-Tiethoff 2018: 290) — The Slavic word is generally considered an inheritance from Proto-Indo-European (Berneker i: 455; REW ii: 249; ЭССЯ xiii: 187–188; Трубачев 1994: 7). This cannot be ruled out, but in view of the geographical limitation, a Baltic loan etymology looks attractive. Although no precise Baltic source is attested, one could certainly imagine a formation *javýnas, with the collective suffix -ýnas (Skardžius 1941: ‣ 53 54 Leningrad Region practically excludes such an interpretation. It is evidently the result of pretonic o > /u/ attested sporadically in the area (ДАРЯ i, No. 1); compare СРГК i: 97, where forms of the type /burkán/ are listed under the headword борка́ н. Masing cites the form as brackannige after Schiller/Lübben (i: 412), who set up this reading with a question mark. MndWb (i: 339) normalizes the form as brakannie, instead. Here, I have cited the actually attested spelling. The substitution of Russian pretonic *o as Võro. õ before tautosyllabic /r/ is paralleled at least by Estonian kõrts, Võro kõrtś ‘tavern’ < R корчма́ (Blokland 2005: 199–200). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic–slavic contacts 33 266–267; Andersen 1996a: 155; thus ‘a collection of grains’ ≫ ‘granary’). This would imply a development *jaw- > *( j)ew- > ov-. According to Andersen, we are rather dealing with a direct substitution *ja- → *a- due to the inadmissibility of an anlaut *jă- (trad. *jo-) in Proto-Slavic.55 In either case, we would probably be dealing with an early loan, although an exact dating is uncertain. ? ‘fish trap’. R dial. вя́ терь, вя́ тель (and variants, cf. СРНГ vi: 79–80; РЭС ix: 255–256), Uk. dial. в’я́тір, Pl. więcierz, Kash. wiącel ‘fyke net (kind of fish trap)’ ← Lt. vénteris ‘fyke net’ (Būga 1922: 298; REW i: 245; Аникин 2005: 111– 112) — This loan etymology is widely accepted. On the other hand, Brückner (1927: 620) has analysed the Lithuanian word as a loan from Polish, a position supported by Kiparsky (1948: 39, fn.).56 On phonological grounds, this is difficult to rule out (even though the distribution within Slavic is suggestive of a Balticism), especially since the word’s ultimate origin is uncertain.57 At a later date, the Baltic word was certainly borrowed into Russian ве́нтерь and Prussian German Wenter (on the latter, see Frischbier ii: 464, where an account of the realia is also given). ? ‘marshy spot’. R dial. (W) алёс ‘damp, marshy spot’; Bel. dial. алёс ‘alder forest in a swamp’; Pl. dial. (Lithuania) olesie ‘swamp in a forest’ (cf. Толстой 1969: 159; Черепанова 1973: 72; РЭС i: 158) ← Lt. álksna, Lv. dial. àlksna ‘alder thicket; marshy spot’ (Топоров/Трубачев 1962: 199; Аникин 2005: 85–86; Мызников 2019: 45) — The -k- in Baltic is intrusive, and the original form can be set up as *álisnā with syncope (Friedrich 1970: 70; Топоров ПЯ i: 53). A trisyllabic preform neatly accounts for the Lithuanian acute, and an unsyncopated variant is preserved in Szyrwid (SD1) alixnis, dial. (NE) alìksnis ‘alder’ (the reconstruction of two forms for Proto-Baltic as per Derksen 2015: 50–51 is unnecessary).58 ‣ ‣ 55 56 57 58 Primarily on the strength of R еба́ ть = Skt. yábhati ‘to copulate’ (1996a: 14, 155), Andersen assumes that original *je- did not develop to o- in East Slavic. However, the evidence is not quite clear-cut: at least Uk. оря́ бок ~ Lt. jerubė ̃ ‘hazel grouse’ would speak in favour of such a development (Andersen 1996a: 137–138 is misguided in disregarding the Baltic evidence for *j- in this word; see the discussion on p. 175). Note also that Andersen is forced to assume an ad hoc early loss of *j- in the words OR оже ‘if, that’ and оли ‘when, if’ (idem: 152–153; ~ Lt. jéi, Lv. dial. jà, Go. jabai ‘if’; Gr. ὅτε ‘when, as’; Dunkel 2014: 320–322). Contrast the more cautious wording in Kiparsky 1973: 69–70, 1975: 93–94. The derivation from Lt. vánta ‘besom’, Lv. dial. viẽtêt (EH ii: 798) ‘flog’ (LEW 1223–1224; ALEW 1405), assuming an original meaning ‘fish trap woven from twigs’ does not seem compelling. Note that the Kaišiadorys Museum encyclopaedia (accessed online at https://​ www.kaisiadoriumuziejus.lt/enciklopedija) specifically states that the distinguishing feature of a vénteris compared to other fish traps is the absence of a supporting frame. Schrijver’s assertion (1991: 42) that syncope “did not occur in Lithuanian” is simply false, cf. Zinkevičius 1966: 131–135. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 34 chapter 1 The Slavic words have been viewed as cognate in some recent sources (Andersen 1996a: 127; Derksen 2015: 50–51; ALEW 23), although none of these authors attempt to account for the -s- in the above words as opposed to *-x- in *alixā- (trad. *olьxa) ‘alder’.59 This phonological difference could favour a Baltic origin. In view of the distribution, the loan etymology looks highly attractive, although it is hampered by the absence of an attested donor form. † ‘birch tar’. R дёготь, Uk. дьо́готь, Pl. dziegieć, Cz. dehet ‘birch tar’ ← Lt. degùtas, Lv. dȩguts ‘birch tar’ (Mikkola 1894: 111; Zubatý 1894: 423, fn. 4; Būga 1922: 141; Kiparsky 1973: 68; Лаучюте 1982: 12) — The advantage of the loan etymology is that the Balto-Slavic verb *deg- ‘to burn’ is only found in the assimilated form *žeg- (< *geg-) in Slavic. However, this is hardly a decisive argument: if the formation were taken to be of Balto-Slavic age, the word’s semantic specialization would make fertile grounds for a preserved archaism. The main counter-evidence to a Baltic origin is the word’s existence in CzechSlovak (Trautmann 1923: 49; Brückner 1927: 109), yet this is somewhat circular given that other loanwords of this potential age are so few. Although the verb deg- is synchronically present in Lt. dègti, Lv. degt ‘burn’, the derivative *deguta- can hardly be recent. In Lithuanian, the suffix -utais rare (Skardžius 1941: 361; Ambrazas 2000: 103–104) and nowhere else is it deverbal.60 A comparable suffix is found only in Lt. ãšutas ‘horsehair’, where it looks old (= R осо́ т ‘sow thistle’, Pl. oset ‘thistle’), and Lt. ríešutas, Lv. riẽksts ‘nut’. Since the word clearly cannot be young in Baltic, it may as well be of BaltoSlavic age, and there is also no particular reason to consider it the source of the Slavic words. As an argument against a loan etymology, one can also point to the archaic-looking athematic OCz. dehet, gen.sg. dehte ‘turpentine tree’ (Gebauer i: 220; cf. Trautmann 1923: 49). † ‘bundle’ R dial. пу́ сма ‘bunch, bundle’ (Kursk, Voronež; СРНГ xxxiii: 142) ← Lv. puõsms ‘section, interval’ (Аникин 2005: 258) — Аникин claims that the older meaning of the Latvian word was ‘bundle’ (“связка”). However, this seems to derive from a misunderstanding of Karulis (1992 ii: 74), who merely suggests that the older meaning might have been “mezgls” in the sense ‘node on a plant ‣ ‣ 59 60 East Slavic *s would actually be the expected result of the progressive palatalization in this word, but in that case, we should expect Polish -sz-. On the other hand, the Polish variant is dialectally very limited and may be from East Slavic. It remains quite unclear (to me, at least) why the progressive palatalization did not occur in the word for ‘alder’ itself. There is a diminutive -ùtas, largely limited to Southern Lithuania (Ambrazas 1993: 56– 57). However, it is almost entirely restricted to velar-final stems and is therefore the result of distant dissimilation from *-uka- (Hasiuk 1970), which makes it unlikely that these represent an archaism (contra Ambrazas loc. cit.). The connection of these forms to the Lithuanian diminutive suffix -ùtis therefore remains unclear. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 35 baltic–slavic contacts stem’ (= “der Knoten beim Schilf oder Getreidehalm”, Seewald 1865: 68). In my opinion, this is not likely to be the oldest meaning, but is rather the result of a metonymical shift from ‘internode; interval’. An original meaning ‘interval’ is also supported by the Slavic cognate in R dial. па́ смо ‘length of yarn, lea’. All in all, Вершинин (iv: 384) is almost certainly correct in deriving the Russian words from Mordvin, cf. Md. E pusmo, M pusma ‘bunch, bundle’, meaning пу́ сма has nothing to do with the Latvian word. ∵ As can be seen from the above discussions, the main argument for analysing any word as a Baltic loan is its distribution. Only in the case of R /Bel. dial. алёс ‘marshy spot’ is a possible phonological argument available. In all the other cases, there is no phonological obstacle to treating the words as cognates. As a result of this and other ambiguities by way of loan sources, almost all of the examples must be considered uncertain.61 Nevertheless, I think that Old Novgorodian *пьрть ‘bath-house’, at least, is a highly probable loanword from Baltic. The distribution of this word would support the supposition made in 1.1.8 that the earliest contacts of the Balts 61 I have attempted to identify other words with a limited distribution which might be interpreted as Baltic loans, but these have mainly turned out to be problematic: (1) R dial. (N) ля́ га ‘swampy area; low, damp place’ has been equated with Lt. ⟨léngė⟩ ‘ein Wieschen zwischen zweyen Anbergen’ (Ruhig i: 76) (cf. REW ii: 65). However, this Lithuanian variant is only known from Ruhig, corresponding elsewhere to lénkė ‘swampy meadow; hollow’ (Smoczyński 2018: 688). The reliability of the Lithuanian form is therefore questionable. Compare, perhaps, a similar sporadic voicing in ⟨kengras⟩ ‘hager’ (Ruhig ii: 188) = keñkras (Kupiškis apud Būga in Juška iii: 76; cf. LKŽ s.v. kiñkras). For the Russian word, Николаев (1988: 135), offers an alternative etymology, comparing Vologda ля́ жа ‘damp, boggy place’, and deriving both from *lę̄djā- (trad. *lędja), with a suggested (albeit controversial) Novgorodian development *dj > g. Another account is given in Мызников (2019: 466). (2) R dial. (Vjatka) черв (Даль² iv: 607) ‘sickle’ has been considered cognate to Lt. kir̃vis, Lv. cìrvis ‘axe’ (Berneker i: 172; Trautmann 1923: 135; REW iii: 317; ЭССЯ iv: 171; Derksen 2015: 248). The word is known only from Даль, where it is cited alongside черп ‘sickle’. The latter has been recorded in other dialects (e.g. Perm, Беляева 1973: 689 and Arkhangelsk, Левичкин/Мызников 2014: 180), but черв does not seem to be. The form ⟨черва́къ⟩ · пила?, also cited here by Даль (followed by Zubatý 1894: 388, then Berneker and Vasmer) is not likely to belong here and is rather to be equated with dial. червя́ к ‘cross-cut saw’ (СРГС v: 274) which is probably a semantic extension of червя́ к ‘worm’. If we assume Даль’s черв was extrapolated from a phonetic [čerf], we might think of the facultative alternation /f ~ p/ reported in this dialect area (Сметанина/Иванова 2018: 208, cf. подчефре́ниться ~ подчепре́ниться ‘dress up’). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 36 chapter 1 and Slavs were with speakers of Old Novgorod-Pskov. Similarly, the dialectal distribution of R dial. борка́ н ‘carrot’ could favour this interpretation, but its analysis as a Baltic loanword is uncertain. The latter is clearly a trade item, while the term ‘bath-house’ rather seems to be linked to a certain cultural diffusion, and might theoretically be an indication of a Baltic substrate in Old Novgorod. The notion of a Baltic substrate reminds us of the evidence collected in a number of studies, chiefly by Топоров (Топоров/Трубачев 1962; Топоров 1972, 1988–1997), but going back to Būga (1923a) and Vasmer (1932), purporting to demonstrate a Baltic substrate in the hydronyms of the Upper Dnieper and Oka basins. The validity of this evidence has practically been taken for granted, and has remained absolutely central to discussions of the Baltic homeland (Zinkevičius 1984: 147–151; Gelumbeckaitė 2018: 1712; see also Grünthal 2012: 299–300 with lit.), soon also entering into archaeological discussions (Gimbutas 1963: 97; Rimantienė 1992: 137; Anthony 2007: 380). It is beyond the scope of this work to go through the evidence in any detail. However, Stang’s call for “tiefer gehende Sichtung und Diskussion” (1966: 2, fn.) seems to have largely remained unanswered, with later contributions rather looking to expand than critically assess the established material (cf. В. Васильев 2015 for a discussion of some of the issues).62 In any case, the alleged pervasiveness of a Baltic substrate in the hydronymy of this area contrasts starkly with the almost complete absence of evidence of early substratal loans on a lexical level.63 Finally, contrary to the claim of Аникин (2014: 192), there is no reason on the basis of this data to assume any loanwords from Baltic into Proto-Slavic. It is possible that such unidentified loanword layers do exist; however, as in the case of the word for ‘birch tar’, there are few if any phonetic criteria that would allow us to distinguish Baltic loanwords in Proto-Slavic from inherited cognates. At the current stage of research, it can be said that no entirely convincing cases exist. 62 63 Much of the evidence constitutes root etymologies, and these often permit alternative interpretations (see, for instance, the extended discussion of hydronyms of the type Велья in В. Васильев 2012: 545–550). The material is in need of a thorough critical review, and the results can certainly not be considered “hard facts” in the manner in which they are normally treated in the archaeological research. There is plenty of evidence for a later Baltic substrate in Belarus, Smolensk and the surrounding regions, as clearly visible in the material collected by Лаучюте (1982). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 chapter 2 Early Germanic → Baltic Loans There have been few focused studies attempting to isolate the earliest layer of Germanic loanwords in Baltic, and we must largely be content with the collections of Hirt (1898: 349–351) and Būga (1922: 64–65; Senn 1925: 46–53 and Alminauskis 1934: 19–22 to a great extent repeat Būga’s conclusions), as well as the later sceptical account of Otrębski (1966) and other comments found scattered in works of a more general character. This state of the field means that little has been done in terms of critically analysing and stratifying the material as a whole. The goal here is to present all of the (convincing) evidence for old Germanic loanwords in East Baltic, with a focus on the evidence for direct contacts with Gothic. Būga (1922: 65) divided his material into two groups, the first supposedly deriving from Gothic and the second from North Germanic: (a) – Lt. alùs, Lv. alus ‘beer’ ← Go. *aluþ (cf. OE ealu, gen.sg. ealoþ ‘ale’, ON ǫl ‘beer’) – Lv. klàips ‘bread’ ← Go. hlaifs ‘bread’ – Lt. midùs, dial. mìdus ‘mead’ ← Go. *midus (cf. OE medu, ON mjǫðr ‘mead’) – Lt. pẽkus ‘cattle, livestock’ ← Go. faihu ‘property, wealth’ – Lt. šárvas, pl. šarvaĩ ‘armour; arms’ ← Go. sarwa n.pl. ‘weapons, armour’ (b) – Lt. dial. gãtvė ‘cattle way’ (cf. Otrębski 1966: 63), Lv. gatve, gatva ‘path between fences, cattle way; street’ ← ON gata ‘passage, street’ – Lt. kviečiaĩ (acc. kviečiùs, dial. kvíečius), Lv. kvìeši ‘wheat’ ← ON hveiti ‘wheat’ – Lt. dial. kliẽpas ‘loaf’ ← ON hleifr ‘loaf’ Būga does not specify why he prefers to derive Lt. gãtvė from Norse. At first sight, Gothic gatwo ‘street’ appears phonologically more suitable (cf. ME i: 609; LEW 139). Senn (1925: 49), who follows Būga, observes that the word’s limitation to northwest Žemaitia and Curonia would favour a Nordic origin, but this is hardly decisive. Noting the Latvian variant gate ‘path between fences’, Zubatý (1892: 255) prefers to take the whole family from Low German (cf. ME i: 609, s.v. gate; Smoczyński 2018: 318–319), which seems possible; cf. Prussian German Gatt ‘opening; narrow passage’ (Frischbier i: 219). On the other hand, gãtvė is not easily analysed as an inner-Baltic derivative: the suffix -vė is rare and unpro© Anthony Jakob, 2024 | doi:10.1163/9789004686472_004 Anthony Jakob - 978-90-04-68647-2 This is an open access chapter distributed under the terms of the cc by-nc-nd 4.0 license. Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 38 chapter 2 ductive (Skardžius 1941: 379). Otrębski (1966: 63) takes Latvian gatuve as the original form, and assumes the more common forms with -tv- arose through syncope, but it is possible that gatuve is itself secondary (after the suffix -tuve; Endzelīns 1923: 280–282), and the coincidence with Go. gatwo ‘street’ is striking. For this variant at least, I believe an early Germanic origin should be preferred. Fraenkel (LEW 271, 326) follows Būga in deriving Lt. dial. kliẽpas and kviečiaĩ from North Germanic (see also ME ii: 356). This is motivated by the idea that Baltic *ẹ̄ must derive from *ei (Būga RR iii: 900–901; Senn 1925: 49–50; Alminauskis 1934: 21; on this, see also Endzelīns 1907). However, these authors’ denial that *ẹ̄ may derive from *ai lacked a solid basis, as Stang (1966: 53– 57) convincingly showed, and their formulation is now largely obsolete. As a result, there is no particular reason to posit a Norse origin for any of the loanwords.1 Senn (1925: 50), who accepts Būga’s reasoning, nevertheless presents a counter-argument: as Norse ⟨f⟩, outside of initial position, stood for the sound /ƀ/ (Noreen 1923: 40), we should expect ON hleifr to have given Lt. *kliẽbas. To explain the -p-, Senn is forced to assume an ad hoc contamination with Lt. kẽpalas ‘loaf’.2 Starting from Gothic, we can take the nom.sg. hlaifs or acc.sg. hlaif ‘bread’, with final devoicing, as the specific source. We may account for the vocalism of Lt. dial. kliẽpas ‘loaf’ and kviečiaĩ ‘wheat’ in two ways: (a) assume that in the Gothic dialect from which the word was taken, ai had monophthongized to */ę̄/, as probably in Wulfila’s dialect (cf. Wrede 1891: 165; Bennett 1949), and that this monophthong was adopted directly as East Baltic *ẹ̄; (b) assume that a preserved diphthong */ai/ was adopted directly as the Baltic diphthong *ai, which only later developed to *ẹ̄. Potentially relevant for resolving this matter are two etymologies presented by Vasmer (1922) supposedly pointing to a Gothic source: Lt. ýla, Lv. ĩlȩns, Pr. E ylo ‘awl’ ← Go. *ēla and the Latvian hapax glīsis (ME i: 627)3 ‘amber’ ← 1 Balaišis (1994) still maintains Būga’s view, but prefers to take the words in question from Gothic anyway. To do this, he is not only forced to assume a Gothic sound shift */ai/ > */ei/ in order to derive the relevant words from Gothic, but also an ad hoc change back from *ei > ai to explain Lv. klàips (idem: 11). 2 To be precise, Senn actually derives *kliẽbas from Slavic (cf. MBel. хлѣбъ) rather than Norse (the same explanation is taken up in Smoczyński 2018: 566). Berneker (i: 389) also derived the word from Slavic, but explained the -p- as the result of generalization from the nom-acc.sg. [xlěp]. However, this kind of phenomenon is unparalleled in the Slavic loanwords; see already Būga (1912: 31). Differently again (and implausibly), cf. Otrębski 1966: 53. 3 I do not have access to the Magazin der Lettisch-literärischen Gesellschaft 20/3, cited by ME, but see Kregždys (2012: 330, fn. 470; and also idem: 330–336 for an attempted etymology). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 early germanic → baltic loans 39 Go. *glēza-. This would imply a narrowing of *ē towards */ī/ which can be witnessed in occasional spellings of ⟨ei⟩ in place of etymological *ē and vice versa in the Gothic bible (suggesting they fell together in the language of later scribes; Marchand 1973: 51). Since this raising might well have been triggered by the monophthongization of *ai, we might take these etymologies as indirect support for option (a). However, it turns out that both of the reconstructed Gothic forms are problematic. While most sources seem to take the length of the first vowel in OHG ala ‘awl’ for granted (e.g. DWb² ii: 73; Kluge/Götze 7; EWAhd i: 135 s.v. âla), the evidence of the other Germanic languages points to short ă (see Kluge/ Seebold 22).4 Similarly, for ‘amber’, OE glær ‘amber’ (cf. DOE s.v. glǣ̆r) and MLG glar ‘resin’, glar(r)en* ‘smear with resin’ (Schiller/Lübben ii: 116), traditionally considered to contain a long vowel, are ambiguous and may just as well reflect short ă, as we do find in ON gler ‘glass’ and the Verner variant OE glæs (cf. ME glas) ‘glass’, OHG glas ‘glass, amber’ (see Meineke 1998: 141 with lit.). Evidence for a Germanic variant with -ē- seems to be limited to Pliny’s glaesum (note here the varia lectio ⟨glassū⟩). As a result, Vasmer’s Gothic reconstructions are based on very uncertain evidence. A similar assumption underlies Endzelīns’ (ME iv: 277) suggestion that Lt. tū́bas, Lv. tũba, Pr. E tubo ‘felt’ derive from an equivalent of ON þófi ‘felt’ in a Gothic variety in which *ō had become raised to *ū (cf. Marchand 1973: 52). Previously, Trautmann (1910: 451) and Būga (1922: 294–295) had taken the word directly from Norse, but this fails to explain the vocalism. An alternative account would be to assume the word entered East Baltic through Prussian. As In Lange’s dictionary (1773: 125), we find the following entry: “Glihſe (obſoletum) Börnſtein, Edelſtein[,] die Nordiſche Seefahrer nanten daher den Börnſtein Strandt, Glyswall.” It seems unclear whether the form *glīse was merely inferred by Lange on the basis of the cited toponym. 4 So, as universally agreed, ON alr ‘awl’. OE æl is given a short vowel in DOE s.v., which is supported by Middle English al (the ME form ēl cited by e.g. EWAhd loc. cit. is a figment, the examples with ⟨e⟩ being Kentish or West Mercian for *æ̆ ). Kluge/Seebold (loc. cit.) argue that High German might reflect *ă, too. In support of this, we can note that DWb² (ii: 73) cite a form allen dat.sg. from Peter von Ulm’s Cirurgia (c. 1430), which looks (at first sight at least) to point to a short vowel, and also the form ale ‘awl’ in the Elbing Vocabulary, where reflexes of MHG ā are regularly spelled ⟨o⟩ or ⟨oe⟩ (Braune 1876: 93–94; Trautmann 1910: xxv). A detailed study of the German dialectal evidence is obviously not possible here, but it is naturally far more straightforward to derive all the Germanic forms from a single ablaut grade than to set up a rare *ē/a ablaut alternation. As Kluge/Seebold points out, Skt. ā́rā- ‘awl’ may just as well reflect IE *Hol-, so there is no external evidence for Pokorny’s *ēlā (IEW 310). If we reject this variation, the only way to connect Lt. ýla (etc.), it seems, would be to assume a reduplicated *Hi-Hl-. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 40 chapter 2 Prussian had no *ō (cf. Smoczyński 2000: 66–70), it seems that a Gothic *þōb(or indeed ON þóf-) would most probably have been adopted here as *tūb- directly.5 Thus, we return to option (b), namely that the Gothic diphthong */ai/ was adopted directly as a diphthong in Baltic. If this is the case, there is no reason to separate Lt. dial. kliẽpas ‘loaf’ from Lv. klàips ‘bread’. Many doublets with and without monophthongization can be found, suggesting paradigmatic alternations may still have been present in Proto-East-Baltic: e.g. Lt. eĩti ~ Lv. iêt ‘to go’; Lt. žíedas, dial. žáidas ‘flower’ (cf. Chapter 3, fn. 81); Lt. saĩkas, Lv. sìeks ‘a dry measure’, etc. (Hirt 1892: 37–40; Kuryłowicz 1956b: 234; Petit 2003: 97). Another example of such an alternation is Pr. E caymis, Lt. káimas ‘village’ (and the derivative Lt. kaimýnas ‘neighbour’)6 beside Lt. kiẽmas, Lv. cìems ‘farmstead, village’. This word has often been understood as inherited and compared either with Lat. quiēs ‘rest, quiet’ (Uhlenbeck 1900: 69; Būga 1922: 70; LEW 251; Smoczyńsky 2018: 540) or Gr. κώμη ‘village’ (Zupitza 1896: 49; Trautmann 1910: 112). Both of these explanations encounter phonological issues, and far more attractive is the interpretation as a loanword from Go. haims ‘village’ (Hirt 1898: 347–348; Boisacq 1916: 544; ME i: 394; Derksen 1996: 215, 2015: 243–244; ALEW 565). This is favoured by the semantic closeness to the Gothic word; contrast the inherited cognate found in Lt. šeimà, Lv. sàime ‘family, household’. Several forms involving the phoneme /k/ have elsewhere been considered cognates, but the centum reflexes, as well as the close semantic and formal correspondence with Germanic, favour a loan origin:7 – Lt. kiẽmas, Lv. cìems ‘village, farmstead’ ← Go. haims ‘village’ — Contrast Lt. šeimà ‘family’ (and probably Skt. śéva- ‘dear’; see Chapter 3, fn. 16) – Lt. kviečiaĩ, Lv. kvìeši ‘wheat’ ← Go. ƕaiteis* ‘wheat’ — Contrast, if related, Skt. śvítna- ‘white’8 5 By contrast, the borrowing broakay ‘breeches’ ← MLG brōk or MHG bruoch (cf. Trautmann 1910: 314; PKEŽ i: 158) must have postdated the Pomesanian Prussian development *ā > /ō/. 6 Although the word belongs to the standard language, it is interesting that the LKŽ only cites Lt. káimas and káima from Žemaitia and Suvalkia. In view of this, we might suggest this is a borrowing from Prussian, which might potentially explain the acute accentuation, cf. Pr. iii kāimaluke ‘heimsucht’. On the other hand, the derivative Lt. kaimýnas ‘neighbour’ does not show this dialectal limitation, and is probably a genuinely East Baltic word. 7 Another possible example could be Lv. kàuns ‘disgrace, shame’, perhaps loaned from Go. hauns ‘lowly’ (Hirt 1898: 350; or rather from a corresponding noun, cf. MHG hōn ‘disgrace, shame’, MDu. hoon ‘humiliation’ < *hauna-), rather than cognate in view of the absence of correspondences elsewhere in IE (cf. Stang 1972: 27). 8 The exact correspondence with Germanic makes the assumption of a parallel formation based on Lv. kvitêt ‘flicker, glimmer’ (Otrębski 1966: 54; Sabaliauskas 1990: 41; ALEW 546–547; Pronk/Pronk-Tiethoff 2018: 302–303) entirely gratuitous. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 early germanic → baltic loans 41 – Lt. pẽkus ‘cattle, livestock’; Pr. i–iii pecku ‘cattle’ ← Go. faihu /fęxu/ ‘property, wealth’ — Contrast Skt. paśú- ‘cattle’9 – ? Lt. ker̃džius (secondary sker̃džius) ‘herdsman’ ← Go. hairdeis (nom.pl. hairdjos) ‘shepherd’ (Hirt 1898: 332) The word for ‘poppy’, Lv. maguône, Lt. aguonà is normally considered to have been loaned from Germanic (ME ii: 547; Sehwers 1936: 312; Sabaliauskas 1960a: 72; Smoczyński 2018: 6). As the short first-syllable vowel makes a late German origin improbable, Endzelīns (ME loc. cit.) suggests an Old Saxon mago* (attested in the compound magonhouut · papaver) as a proxy. However, given the absence of other evidence for Old Saxon loans in East Baltic, we might instead suggest a Gothic *mago. The main issue with this explanation is that the earliest evidence for the opium poppy in the East Baltic region dates to the Middle Ages (Grikpėdis/Motuzaitė Matezuvičiūtė 2020: 167), which is too late to be reasonably associated with Gothic contacts. As a result, no fully satisfactory explanation is available for this word, although it is certainly borrowed. See pp. 253–254 for further etymological discussion. By contrast, I see no reason to assume that Lt. alùs ‘beer’ (etc.) is a Germanic loanword (cf. Machek 1930; Kiparsky 1934: 78–79; Stang 1972: 13; Топоров ПЯ i: 80; Mallory/Adams 1997: 60), despite frequent claims to the contrary (Hirt 1898: 346; Būga 1922: 64; Kroonen 2013: 23; Derksen 2015: 53; ALEW 36; Smoczyński 2018: 22). Note that the word is not limited to northern Europe if Arm. awłi ‘strong fermented drink’ belongs here (Olsen 1999: 443; Martirosyan 2008: 155). Against a loan from Gothic speaks the absence of any trace of the stem-final dental, which ought to have been preserved there; cf. miliþ ‘honey’ (< *melit-). In an inherited context, the loss of the final -t can be accounted for by regular sound law (cf. e.g. Kortlandt 1989a: 44). According to Būga (1922: 66), all Gothic words entered East Baltic through Prussian, an opinion repeated by Senn (1925: 48: “weder Letten noch Kuren noch Litauer [hatten] an irgendeiner Stelle direkten Verkehr mit den Goten”; cf. also Senn 1943: 954). However, such conclusions must derive from the data, rather than from aprioristic assumptions. On phonological grounds, I have 9 Fraenkel (LEW 564–565) rejects this loan etymology due to the semantic distance. Yet since there does not appear to be any other word for ‘cattle’ attested in Gothic, it may well have been faihu; cf. the similar semantic range of the ON cognate fé ‘cattle; property, wealth’. Kortlandt (1978: 241) has attempted to explain the Baltic -k- as having spread from an oblique stem comparable to Skt. paśvás gen.sg. with his rule *ḱ > *k before u̯ + back vowel. The back-vowel criterion for this rule cannot be fulfilled, however, as only *-es can be reconstructed as an athematic genitive ending in Balto-Slavic. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 42 chapter 2 argued above that Lt. šárvas ‘armour’ (p. 13) and tū́bas ‘felt’ (pp. 39–40) may well have been borrowed through Prussian. I have also mentioned some words which ultimately derive from Germanic but which may have equally entered East Baltic indirectly, either through Slavic or Prussian. Due to the ambiguity, they cannot be used as evidence here: – Lv. bruņas ‘armour’ = Pr. E brunyos, OR брънѧ ‘armour’, Go. brunjo ‘breastplate’ – Lt. kãtilas, Lv. katls ‘kettle’ = Pr. E catils, OR котьлъ, Go. katils* ‘kettle’ – Lt. stìklas, Lv. stikls ‘glass’ = Pr. E sticlo, OR стькло ‘glass’, Go. stikls ‘cup, chalice’ A specifically Gothic source must be assumed at least for Lt. midùs ‘mead’, where -i- for expected *e can only reasonably be explained through the assumption of Gothic transmission (Hirt 1898: 346; Būga 1922: 65; unconvincing is Otrębski 1966: 55). Note also that German mete ‘mead’ is glossed alu in the Elbing Vocabulary, which does not exclude the possibility of Prussian mediation (there may have been dialectal differences which are not reflected in the attested evidence), but it certainly does not favour it. Similarly, the attested Prussian words for ‘wheat’ (Pr. E gaydis, G gaide, gayde) and ‘bread’ (Pr. E geytye, iii geits) differ from those attested in East Baltic and do not represent Gothic loans. Another alleged piece of evidence for direct contact with the Goths is Lt. gùdas ‘Belarusian; speaker of a different dialect’, Lv. guds ‘Belarusian raftsman; wandering merchant’ (ME i: 675), which has been taken from Go. guta* ‘Goth’ (Būga 1922: 67; LEW 174; Smoczyński 2018: 400; on the Gothic endonym, see Leumann 1986: 163–164 wih refs.), under the assumption that the word was used to refer to Slavs under Gothic rule. The medial -d- has been explained by assuming a pre-sound-shift loan from Germanic (Būga loc. cit.; Zinkevičius 1985: 73), which is hardly plausible, although the only other possibility is to assume an ad hoc contamination (cf. Karaliūnas 2004: 164).10 Karaliūnas (2004: 145–189) hypothesizes a native origin for Lt. gùdas. Noting the word’s pejorative value in folk literature, he suggests a derivation from a root *gud- ‘small, of poor quality’, which is set up on the basis of e.g. (į-)gùsti ‘get used to’, gùd-obelė ‘hawthorn; crab apple’ (obelìs ‘apple tree’), Pr. E gudde ‘bush’ (with which LEW 174 already suspected contamination). Despite the detailed treatment, I am not convinced that the evidence, mainly plant names, warrants 10 One would like to see the missing link in the gloss guti · krzyrzacy ‘Teutons’ in the Narev vocabulary (although the reliability of this vocabulary remains uncertain). On this word differently, see Karaliūnas 2004: 164–165. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 early germanic → baltic loans 43 the reconstruction of such a root.11 However, while the etymological connection with the name of the Goths is tantalizing,12 the -d- remains a significant stumbling block. Given the general assumption that the East Baltic loanwords from Gothic were mediated by West Baltic, it is remarkable that all of the certain Gothic loanwords in Prussian are shared with East Baltic. Two others have been suggested as unique to Prussian (Būga 1922: 66, Senn 1925: 47), but neither of these are certain: – ? Pr. E ilmis · bark ‘hay shelter’ ← Go. hilms ‘helmet’ (cf. ON hjalmr ‘helmet; hay shelter’; Lidén 1906; Trautmann 1910: 346; Тороров ПЯ iii: 42; PKEŽ ii: 24). Yet in view of *h- → ∅-, a Low German source seems more probable (Smoczyński 2000: 35–36; admittedly, a formally or semantically suitable source appears to be lacking).13 – ? Pr. E lapinis ‘spoon’ ← Go. *lapins (cf. OHG leffil, MDu. lepel ‘spoon’; Kluge 1907: 361; Trautmann 1910: 368; Endzelīns 1943: 202; Sabaliauskas 1990: 257). While this etymology still seems possible, it is now widely rejected in favour of a native etymology (Falk apud Топоров ПЯ v: 90; PKEŽ iii: 41–44). Collecting together the evidence for direct Gothic loans into East and West Baltic, we obtain a rather interesting picture (italicized words are those shared by Prussian; those in brackets were possibly obtained indirectly): – Agriculture: village, wheat, bread, ?poppy – Stockbreeding: cattle, cattle way, ?herdsman – Warfare: [armour (×2)] – Trade: [ felt], [kettle], [glass] – Other: mead 11 12 13 Lt. (į-)gùsti ‘get used to’ may well be backformed from gùd(r)inti ‘train, teach’, which belongs with gudrùs ‘smart, sly’ (on this word familiy, see also ALEW 439; Smoczynski 2018: 401). Combined with gùdė ‘whetstone’, one might imagine an original meaning ‘sharp’ (although this is by no means the only option). As to gùd-obelė, etc. I would rather assume the first element means ‘bush’, as in Prussian. The dial. gū́das ‘sad, gloomy’, in view of its acute root, must also be separated. Bearing in mind the various parallels adduced by Karaliūnas (2004: 162) whereby terms for other peoples have been generalized in the meaning ‘unchristened child’: R dial. лопь ‘Sámi; unchristened infant’ (cf. Мызников 2019: 450–451), Lv. krìevs ‘Russian’, krieviņš ‘unchristened child’ (ME ii: 284–285), it may be conceivable that the pre-Christian Balts, on the contrary, used the term *guda- as a pejorative designation for their ‘non-pagan’ (i.e. Christianized) neighbours (cf. the juxtaposition of the gudai and the pagan deity Perkū́nas in folklore; Karaliūnas 2004: 159–160). The vocalism of Pr. E kelmis ‘hat’ (Sabaliauskas 1990: 257) shows that it cannot be from Gothic hilms. I will not enter into a discussion of this word here. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 44 chapter 2 It is remarkable that the majority of the direct loanwords are connected in some way to stockbreeding and agriculture (note also p. 194 for a discussion of the possibility that Lt. rugiaĩ ‘rye’ is a Gothic loanword). Borrowings in this semantic field seem more likely to be indicative of an inward migration of Germanic speakers rather than incidental trade. A possible proxy for this migration could be found in the appearance of grave artefacts in the 5ᵗʰ century in Eastern Lithuania showing a remarkable similarity to those popular in the Carpathian Basin. Even though these artefacts may rather attest to trade routes (Bliujienė/Curta 2011), this does not rule out a small-scale migration. Importantly, the loanwords do not indicate the assimilation of an elite class. It is remarkable that none of the words associated with trade can be considered unambiguous direct loanwords. A second possible route for the incursion of Gothic-speaking populations could be a direct migration from the Lower Vistula region through Sembia and Žemaitia of “polyethnic warrior groups”, bringing with them new kinds of weapons as well as new burial customs (cf. Kurila 2021: 21). This migration could explain certain originally Gothic words shared between East and West Baltic. Note particularly Lt. šárvas ‘armour’, which I have argued was most probably borrowed through Prussian. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 chapter 3 Baltic → Finnic Loans 3.1 Preliminaries In this chapter, I have the following aims. The first task is to charaсterize the extent, chronology and nature of the contact relationship between Baltic and Finnic-speaking groups on the basis of the mutual loanwords. The second is to attempt to answer the question of whether some of the vocabulary shared between the two language families may in fact have originated in other, preBaltic languages spoken in the region before the arrival of the Balts and Finns. Before doing so, it is important to define the corpus of loanwords I will use for my analysis. In his 1890 magnum opus, Thomsen identified some 200 potential loans from Baltic into Finnic, of which he considered about 140 certain. To a large extent, Thomsen’s work has stood the test of time, and there are comparatively few really solid etymologies that have been proposed since. Despite a rather impressive amount of research into the subject of Baltic-Finnic loans, Petri Kallio (2008a: 265), 118 years later, still states that only “about 200” certain Baltic loanwords can be found in Proto-Finnic. This is more or less in line with Vaba (1990a), who labels 189 loan etymologies as certain. At the same time, Santeri Junttila’s dissertation (2016a) covers a corpus of almost 1000 etymologies proposed up until 2009, which implies nearly 7 new loan proposals every year since Thomsen. It would be beyond the scope of this work to discuss all the proposals, which would be a task of many years (Junttila in prep.). I have therefore limited myself to those which I have deemed reliable, starting with those of Thomsen (1890) and Kalima (1936). The material presented here is certainly incomplete, but hopefully sufficiently representative to allow for valid conclusions to be drawn. Insofar as the study of loan relationships cannot operate with strict sound laws in the Neogrammarian sense, the study of Baltic-Finnic loan relations has suffered from many of the same issues as long-range and pseudo-linguistic comparison. The first issue concerns semantics. To quote Robbeets (2004: 158): “The greater the semantic latitude permitted in external comparisons, the more likely it becomes that the apparent formal similarity is due to pure coincidence.” This criticism can of course apply even to comparisons within the Neogrammarian framework, but without the constraint of exceptionless sound © Anthony Jakob, 2024 | doi:10.1163/9789004686472_005 Anthony Jakob - 978-90-04-68647-2 This is an open access chapter distributed under the terms of the cc by-nc-nd 4.0 license. Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 46 chapter 3 laws,1 a low threshold for semantic similarity essentially leaves the scholar’s own imagination as the only limiting factor (cf. Rédei 2000). In the absence of any empirical approach to semantic shifts, such shifts ought to be approached with extreme caution. However, a cursive look at the state of the art in Baltic-Finnic loanword studies reveals that semantics have not been a primary consideration. Some representative cases of semantically questionable etymologies are given below. To avoid any risk of cherry-picking, I have limited myself to those assessed as “Relatively clear” in Junttila (2012):2 – F hiiva ‘yeast; sediment; froth (on beer)’, Vt. iiva ‘yeast, leaven’ (< *hīva); E dial. (Saaremaa) iive ‘froth’ ~ Lt. šývas ‘grey, whitish (usu. of horses)’ (Koivulehto apud Plöger 1982: 93; cf. Häkkinen 2004: 196; van Linde 2007: 35–37).3 – F huone, Võ. hoonõh (< *hōne̮h) ‘building; room’ ~ Lt. šónas, Lv. sãns ‘side’ (Koivulehto 1992b, supposedly in the sense ‘Nebenraum, Nebenhaus’; cf. the doubts in Häkkinen 2004: 221–222). – F kausta, E kaust, Li. kōsta (< *kausta) ‘side beam on a sledge’ ~ Lv. skàusts ‘withers; nape (of a person or animal)’ (Posti 1977: 264–265; SSA i: 333). – F ketara ‘sledge stanchion’, E kodar, Li. kõ’ddõrz ‘sledge stanchion; spoke (of a wheel)’ (< *ke̮tara) ~ Lt. keterà, sketerà ‘ridge, peak; crest (of the back)’ (Būga 1908: 72; Posti 1977: 265–266).4 – F sakara ‘point, protruding tip’, E sagar ‘(wooden) hinge’ (< *sakara) ~ Lt. stãgaras ‘(dry) stalk, branch’ (Kalima 1936: 203 with “?”; SSA iii: 144). – F sampi (< *sampi); E samb, gen.sg. samma (Setälä 1902: 149–150; not in VMS — a Finnish loan?) ‘sturgeon’ ~ Lt. stambùs ‘big, beefy; coarse-grained’ (Liukkonen 1999: 124).5 1 Santeri Junttila (p.c. April 2023) has argued that substitution rules can be treated similarly to sound laws, and I would indeed recommend a strict approach. However, in practice, multiple substitutions for a single sound can and have been assumed. Even if we attribute such variation to different chronological stages or dialectal differences in the source language, this is rarely independently verifiable, with the result being that ‘substitution laws’ are more flexible than traditional sound laws. 2 The choice of this article is merely dictated by convenience, and I do not mean to single out Junttila as a particular offender in this domain. Rather, the lax approach to semantics exhibited by this article is characteristic of the subfield in general. Note that Junttila (in prep.) now rejects Koivulehto’s Baltic etymology for *hīva and doubts the one for *hōne̮h. 3 This suggestion “rescues” Thomsen’s (1890: 218) unsuccessful comparison with Lt. sývas (often pl. sývai) ‘sap’. The mention of šỹvas ‘Hausbier’, a hapax recorded in a daina (šyvū acc.sg. ‘kvass’, Niemi/Sabaliauskas apud LEW 996), is hardly sufficient to make the comparison “relatively clear”. Note that “šývis ‘mould’”, cited by van Linde (2007: 35), is the result of his misunderstanding of German Schimmel ‘grey horse’ (cf. Nesselmann 1851: 520). 4 The further comparison with Md. E kodorks ‘twining plant stem’, M kodərks ‘vegetable tops’ (cf. SSA i: 351; Grünthal 2012: 317) is also semantically implausible. 5 Liukkonen points out that the word has in Finnish also been applied to other large fish, and Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic → finnic loans 47 – F uksi, E uks, Li. ukš (< *uksi) ‘door’ ~ Lt. úoksas ‘hollow; cavity’ (Koivulehto 1993a: 34; SSA iii: 369 with “?”).6 The theoretical possibility of a semantic shift should not be considered sufficient grounds for an etymology to be accepted, as we always have to reckon with the possibility of chance resemblance. In the case of *hīva, for instance, even if it is accepted that ‘yeast’ could plausibly be referred to as ‘grey matter’ (van Linde 2007: 35–37), the etymology cannot be considered by any means certain as there is no trace of a sense ‘yeast’ in Baltic, nor of a sense ‘grey’ in Finnic, and thus the proposal that a semantic shift occurred depends itself on the correctness of the borrowing proposal. Semantic parallels must be specific and trivial: Posti (1977, see above) simultaneously presented two etymologies connecting Baltic words for body parts with Finnic words for parts of the sledge. But if Finnic *jalkas ‘runner’ is derived from *jalka ‘leg’, then surely the side beam would be the ‘arm’ and by no means the ‘withers’ or ‘neck’. Even if some of the above etymologies are actually correct, the sheer semantic distance means that they can never be regarded as “relatively clear”. The possibility of chance resemblance between Baltic and Finnic words may also be increased by the simpler phonotactics of Finnic as compared to Baltic. Just focusing on word-initial position, Finnic *k- can equally stand for Baltic *k-, *g- or *sk- which results in a potentially significant increase in the ‘hit rate’ when searching for Baltic donors. Finnic *r- would regularly substitute 11 phonotactically acceptable Proto-Baltic anlauts (*r-, *sr-, *pr-, *br-, *spr-, *tr-, *dr-, *str-, *kr-, *gr-, *skr-). As a brief illustration of the possibility of chance resemblance, I searched through the LKŽ for potential Baltic comparanda for Finnic words with cognates I was able to verify in Samoyed (some 140 items). These Finnic words can obviously not be classified as Baltic loanwords, although a few fairly good matches can be found. One such example has in fact been treated as a possible Baltic loanword in certain sources: F lampi ‘pond’ was hesitantly compared attested in the general meanings ‘big fish; fish god’, but these are clearly metaphorical extensions of ‘sturgeon’. 6 Koivulehto defends his comparison of the Finnic word for ‘door’ with the Baltic word for ‘tree hollow’ by referring to Lat. ōstium ‘door’, an indirect cognate of the Baltic word. But the Latin word represents a different formation, such a meaning is unknown in Baltic, and a development ‘tree hollow’ ≫ ‘door’ is itself almost inconceivable. Not to mention that, providing the comparison with Latin is correct, the Baltic -k- must be secondary (LEW 1165–1166). Its recent nature would apparently be supported by the rare form úosvauti (Jablonskis apud LKŽ) ‘search for a tree hollow (of scout bees)’, and the absence of the RUKI law. The root meaning is ‘mouth’ (IEW 821, without the Baltic word). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 48 chapter 3 with Lt. dial. klampà ‘swamp, muddy spot’ by Kilian (1986: 494), and independently by Liukkonen (1999: 78–79). Junttila (2012: 282; 2016a: 103) categorizes this example as “dubious” rather than “erroneous”. On the other hand, the word may equally be regularly cognate with Ngan. ľüŋhə ‘boggy spot’ (UEW 235; Aikio 2014c: 86). Since only one of the two etymologies may be correct, this a clear illustration of the risk of false positives in loanword research. Allowing a rather small level of semantic flexibility, we can make a number of additional false comparisons, for example: – PF *kaiho ‘grief, loss; yearning’ (= Ngan. koče ‘illness’, Aikio 2014b: 3–5) ~ Lt. gaižùs, Lv. dial. gàizs² (EH i: 379) ‘bitter, acidic’ — cf. the derived verb Lt. giẽžti ‘feel an unpleasant sensation (in the throat)’, also ‘long for, request insistently’ – PF *kaiva- ‘dig’ (cf. Ngan. kajbu ‘shovel’) ~ Lt. nu-káivinti ‘wear out (the soil); exhaust’ – F dial. kumpu- ‘well up’7 (= Ngan. koŋhu ‘wave’) ~ Lt. gum̃ bas, Lv. dial. gum̃ ba ‘bump, bulge, excrescence’ – PF *lanci ‘damp lowland’ (= Ngan. li̮ntə ‘plain, valley’) ~ Lt. sklandùs ‘slippery, smooth, flat’ — The semantic connection would be ‘flat’ : ‘flat land’ – PF *mene- ‘go’ (= Ngan. mi̮nsi̮) ~ Lt. mìnti (pres. mẽna), Lv. mĩt ‘trample’ — the Lt. word is also attested in the meaning ‘go, tread’ – PF *nüci ‘scythe handle’ (= Ngan. ńir ‘axe handle’) ~ Lt. dial. (Juška) gniutìs, Lv. dial. gņuta ‘thin plank used to attach straw to a roof’ In drawing up a corpus of etymologies upon which further conclusions can be based, only the clearest cases should be used. To this end, I have excluded all etymologies which involve speculative or non-trivial semantic shifts. That is not to say that I deem these etymologies impossible, but simply that it would be misguided to base any further conclusions on them. Their acceptance should rather be informed by the analysis of the clearer cases. However, even etymologies which show perfect semantics cannot necessarily be regarded as certain loanwords from Baltic to Finnic. As an example, Lt. tóšis, Lv. tãss and F tuohi, E toht, Li. tū’oigõz (< *tōhi) all mean ‘birchbark’; however, since neither the Baltic nor the Finnic word has a clear etymology,8 it 7 The Finnic word shows an irregular vowel, but I wonder if it could be explained by an (irregular) assimilation *o–u > *u–u, parallel to the recently proposed sound law *e–ü > *ü–ü (Aikio 2021: 171). Note that a generally high frequency of stems of the shape *u–u in Finnic was already observed by E. Itkonen (1948: 133). 8 See Smoczyński (2018: 1498). Lv. dial. (Kurzeme) tâst² ‘hew’ probably shows secondary length after the preterite (ME iv: 151), cf. the derived dial. (SW) tastît ‘hew’, and the similar phenomenon in Lv. tèst, dial. têst (ME iv: 175–176) ‘hew, chop, adze’, where we indeed find a Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic → finnic loans 49 is difficult to make any conclusion with regard to directionality (Kalima 1936: 171; LEW 1107; cf. Bednarczuk 1976: 54). Although Junttila has argued that there are no Finnic loanwords in Proto-Baltic (see 3.4), he concludes that a few cases could represent “parallel borrowings from a shared source, perhaps a lost substrate language” (2015: 31).9 In view of this possibility, only when the Baltic source word has a solid IndoEuropean pedigree can a specifically Baltic → Finnic directionality be proven. As a result, I think it worthwhile to limit my corpus of loanwords to those which have an Indo-European etymology. Ideally, this means regular, unambiguous cognates beyond the neighbouring Slavic and Germanic, although I have also included examples which require some additional (more or less trivial) assumptions with regard to word formation and semantics. In a small number of cases, an etymology has been excluded because the derivation from Germanic is equally plausible: – F olut, Vt. õlud, Li. vȯ’l (< *olut) ‘beer’ ~ Lt. alùs, Lv. alus ‘beer’; Pr. E alu ‘mead’; Sln. (dated) ǫ̑ ł ‘beer’ = ON ǫl ‘beer’, OE ealu, gen.sg. ealoþ ‘ale’ (LÄGLOS ii: 310; Junttila 2012: 273)10 – F rastas, E rästas, Li. rastā (< *rasta(s); see p. 103) ‘thrush’ ~ Lt. strãzdas, Lv. strazds ‘thrush, starling’ = Ic. þröstur ‘thrush’ (Qvigstad 1893: 259; LÄGLOS iii: 130–131)11 – F terva, E tõrv, Li. tȭra ‘tar’ ~ Lt. dervà ‘tarry log; tar, resin, pitch’, Lv. dar̂va ‘tar, pitch’ = ON tjara, OE teoru ‘tar’ (LÄGLOS iii: 289–290)12 9 10 11 12 vacillation between 1sg.pres. tèšu and dial. tešu. These are to be equated with Lt. tašýti ‘carve’, which cannot be connected with tóšis on formal grounds. In this connection, he cites e.g. F kinnas, E kinnas, Li. kīndaz ‘mitten, glove’ (?→ South Sámi gamhtse, 18ᵗʰ c. ⟨kamtes⟩ ‘leather glove’) ~ Lv. cìmds ‘glove’ (cf. Thomsen 1890: 187; Kalima 1936: 118; Posti 1953: 36–37; SSA i: 336). However, there are several generally accepted loanwords which could easily have been mentioned in the same context. Take, for instance, F vuota ‘hide, pelt’ ~ Lt. óda, Lv. âda ‘skin; hide, leather’ (Thomsen 1890: 205; Kalima 1936: 183). While the direction of borrowing has apparently never been doubted, and Koivulehto (2000: 104) has even explicitly ruled out a substrate word, the Baltic word remains unetymologized (cf. LEW 515–516; Smoczyński 2018: 883; ALEW 826). Due to the ambiguity in the analysis, these and similar cases have been excluded from the dataset. While *o ← *a might favour a Baltic source, the final *t seems rather to favour a Germanic one. Against deriving the Balto-Slavic words from Germanic, see p. 41. LÄGLOS favour a Baltic origin due to the final *-as in Finnic, but the adoption of Germanic *-us as Finnic *-as does not seem impossible (see Koivulehto 1981: 193). For further discussion of the Indo-European background, see pp. 203–205. Sá. N darvi, Sk. tâ´rvv ‘tar’ (< *te̮rvē) is hardly a loan from Finnic (Aikio 2006b: 32). Instead, it may be a Norse loan. For the vowel substitution, compare the example Sá. N gavja ‘(fine) dust’, Sk. kõbjj ‘dust, dandruff’ (< *ke̮pje̮) ← Germanic *heuja- (cf. Ic. hý ‘down, fluff, dust’; Aikio 2006a: 24). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 50 chapter 3 – F vaaja, E vai, Li. vaigā (< *vakja) ‘wedge’ = Sá. S vuevjie, L vuojvve ‘clothing insert’ (< *vuovjē) ~ Lt. vãgis ‘peg, wedge’, Lv. vadzis ‘wall hook, wedge’ = Nw. dial. vegg, OHG weggi ‘wedge’ (LÄGLOS iii: 344) – F äes, E äke, Li. ä’ggõz (< *äkes) ‘harrow’ ~ Lt. akėč́ ios, Lv. ecêšas; Pr. E aketes pl. ‘harrow’ = OE egeþe, OHG egida ‘harrow’ (Koivulehto 1971: 591; LÄGLOS iii: 429; Junttila 2012: 273) The same goes for F kaima, E kaim ‘namesake; relative, companion’, Li. kāima ‘neighbour’ (= Sá. N guoibmi, Sk. kuei´mm ‘companion’), for which, rather confusingly, a Germanic loan etymology is almost never suggested. Semantically, the Finnic word is no closer to Lt. kiẽmas, Lv. cìems ‘farmstead, village’ than it is to Go. haims ‘village’, ON heimr ‘world, realm’ (in place names ‘village’); the existence of a derivative Lt. kaimýnas, Pr. iii kaimīnan acc.sg. ‘neighbour’, Lv. kaĩmiņš (different suffix!) ‘neighbour; resident of the same village’ (cf. Thomsen 1890: 177; Kalima 1936: 105) is hardly of any relevance, since such a derivative cannot have given the Finnic words directly. Perhaps closest to the attested Finnic sense comes the compound MDu. oom ‘uncle’, OE ēam, OHG ōheim ‘maternal uncle’ (< *awa-haima-).13 I have also omitted words limited to Livonian, such as the following: – Li. kil ~ kiļ ‘black woodpecker’ and palāndõks ‘pigeon’ both predate developments specific to Latvian (viz. the palatalization in Lv. dzil̂na ‘woodpecker’ and loss of nasal in baluôdis ‘pigeon’). However, these loanwords need not date to Proto-Baltic, either. Endzelīns (1914b: 102) associates these cases with the so-called ‘Curonianisms’ in Latvian dialects, which is not implausible. – Li. kǭla ‘sandbank’ < *kalla < *kalva (cf. Lv. kal̃va, Lt. kalvà ‘hill; sandbank’) has undergone a number of phonological developments within Livonian, but since these are specific Livonian changes, the word likewise need not be dated to Proto-Finnic. In general, I have erred on the side of caution, and taken the liberty of leaving out etymologies which seem problematic to me for any reason. No explicit attempt at exhaustivity has been made in this survey, but the following hopefully covers the most unambiguous evidence. 13 A slightly different case is F ranta, E rand, Li. rānda (< *ranta) ‘coast, shore’, which has been etymologized both as a loan from Baltic (Lt. krañtas, kránta ‘shore; precipice’; Būga 1908: 30; Терентьев 1990: 30; Liukkonen 1999: 117–119) and from an unrelated Germanic source, cf. ON strǫnd (< *strandō-) ‘coast, shore’. Both etymologies are formally and semantically unproblematic; it does not seem possible to choose between them (LÄGLOS iii: 127; Junttila 2012: 282). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 51 baltic → finnic loans 3.2 Baltic Loanwords with an IE Etymology In total, I have identified 70 certain loanwords which have a strong IndoEuropean etymology. I present these below, organized by semantic category. Where no reference is given, the comparisons are already present in Thomsen 1890. On the Baltic side, I have prioritized East Baltic attestations, but have cited Prussian equivalents wherever these are available. For economy of presentation, I have usually presented the IE etymology by citing a single cognate from a non-contiguous branch. These etymologies are generally well-established and can be found in standard reference works. Any additional discussions have been confined to the footnotes. 3.2.1 Kinship – F dial. ativo, atima ‘visiting relatives; married woman visiting her parents’, Vp. adiv, -on ‘guest; marriageable woman’ (< *ate̮iva / *ate̮ivo)14 ~ Lt. ateĩvis, dial. atéiva ‘foreigner, newcomer’ (to at-eĩti ‘come, arrive’, cf. Skt. éti ‘go’) (Kalima 1939–1940: 211–214)15 – F heimo, Võ. hõim, Li. aim ‘tribe, kin’ (?< *he̮imo; see 3.3.1.4) ~ Lt. šeimà, Lv. sàime ‘family, household’; Pr. iii seimīns ‘gesinde’ (= OIr. cóim ‘dear’)16 – F morsian, K moršien, E mõrsja, dial. mõrsija ‘bride, newlywed’ (< *morcijan, obl. *morcijame̮-) ~ Lt. martì, acc.sg. mar̃čią ‘son’s wife; bride’, Lv. mā̀rša ‘brother’s wife’; Pr. iii mārtin acc.sg. ‘bride’ (cf. Lat. marītus ‘married (man)’, Gr. μεῖραξ ‘girl’ < *mer-ih₂-)17 14 15 16 17 To explain the absence of the assibilation *ti > *ci in the word ativo ‘visiting relatives’, Kalima (1939–1940: 212) has posited an Early Proto-Finnic *atei̯vo. Despite Koivulehto (1972: 628), the development here must be distinguished from pre-Proto-Finnic *ej > *ij, which, at least in non-initial syllables, clearly predated the assibilation of dentals, cf. *vecitä (= F vesiä) ‘water’ part.pl. (< *vete-j-tä) (Kallio 2012: 35). Instead (also in view of vowel harmony), we must reconstruct the Proto-Finnic diphthong -e̮i- (which emerged due to the reduction of *-aj- under certain conditions, see Kallio 2012: 32–34). The diagnostic (South Finnic) forms for this reconstruction are unfortunately unattested. Forms in -eĩvis are limited to Lithuanian, but as the suffix *-vīs is rare and unproductive (Skardžius 1948: 379), they may represent an archaism. Stang (1972: 28) can see “keine einleuchtende Verbindung” with the Irish word, but it is semantically very close to the Baltic ones. The eDIL (s.v. cáem) glosses the Irish word as ‘dear, precious, beloved; belonging to the family’. In Middle Irish, the word is often used substantively in the sense ‘relation, comrade’. In addition, a semantic parallel can be found between Skt. śéva- ‘dear’ beside OE hīwan ‘household, family’ (< *ḱei(H)-uo-), which is most probably also from the same root. This is, of course, merely a root equation, but since the word martì is one of only two Lithuanian nouns with a nominative in -ì (the other being patì ‘wife’), it seems very likely it is inherited. For Latin -a-, see Vine (2011: 265–266). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 52 chapter 3 – F dial. nepaa, arch. nevat (< *ne̮pat) ‘nephew, niece’; E nõbu ‘cousin’ (< *ne̮poi)18 ~ Lt. obs. nepuotis (gen.sg. -ies) ‘nephew, niece’ (= Skt. nápāt‘grandson’) – E sõsar, Li. sõzār, ?Võ. sysaŕ ‘sister’ (< *se̮sar) ~ Lt. sesuõ (obl. sẽser-) ‘sister’ (= Skt. svásar-)19 – F tytär, E tütar, Li. tidār ‘daughter’ (< *tüttär) ~ Lt. duktė ̃ (obl. dùkter-); Pr. iii duckti ‘daughter’ (= Skt. duhitár-) 3.2.2 Body Parts – F hammas, E hammas, Li. āmbaz (< *hampas) ‘tooth’ ~ Lv. zùobs ‘tooth’, Lt. žam̃ bas ‘edge, hem; (dial.) blade’ (= Skt. jámbha- ‘tooth, jaw’) – F napa, E naba, Li. nabā (< *napa) ‘navel’ ~ Lv. naba ‘navel’; Pr. E nabis ‘navel’ (= YAv. nāfa-)20 3.2.3 Adjectives – F ahdas, Vt. ahaz, Li. ǭ’dõz (< *ahtas) ‘narrow, cramped’ ~ Lt. añkštas ‘narrow’ (*aNž- + *-sta-; = Skt. aṃhú- ‘narrow’, cf. LEW 11)21 – Ingr. haljas ‘verdant’ (cf. F haljakka ‘pale, faded’), E haljas, Li. ǭļaz ‘green’ (< *haljas) ~ Lt. žãlias, Lv. zaļš ‘green’; Pr. E saligan */zaljan/ (~ Skt. hári- ‘fallow, yellow-green’) – F keltainen, E kollane, Võ. kõllanõ (< *ke̮ltaine̮n) ‘yellow’ ~ Lt. geltónas, Lv. dial. dzȩltãns ‘yellow’, cf. Pr. E gelatynan, probably for */geltajnan/ (~ YAv. zairita‘pale yellow’)22 18 19 20 21 22 In Estonian, we are dealing with an affective derivation by clipping, compare F tyttö, K tyttö, tytöi ‘girl’ ≪ *tyttär. In North Finnic, we find F sisar, Vp. sizar, with an irregular -i-. Kallio (2018: 225, fn. 6) also takes Võro sysaŕ from *sisar, and considers two independent loans to have taken place. Indeed, the change *i > y between two sibilants is paralleled by Võ. sysalik ‘lizard’ (< *sisalikko), and sys ~ sis ‘then’ (cf. E siis). Võro [y] elsewhere primarily occurs as an allophone of /õ/ before n. I still wonder whether our word could irregularly reflect *se̮sar after all. The assumption of two independent loans is not very economical, and does not help explain the North Finnic *i. The word for ‘navel’ is more likely to be loaned from Baltic than from Germanic. In Germanic, the meaning ‘navel’ is usual for the suffixed *nablan- (> ON nafli, etc.), while the more basic *nabō (> ON nǫf ) means ‘nave (of a wheel)’ (Kalima 1936: 141). The comparison has been considered uncertain (e.g. Kalima 1936: 86; SKES i: 4) because the verbal root seen in F ahta- ‘cram, stuff’ appears to be inherited; however, the semantic development of the latter is most probably the result of secondary convergence with the Baltic loanword (Koivulehto 1998: 244; Aikio in prep. 51). Furthermore, Võ. atma, 3sg. ata ‘cram’ implies PF *akta- rather than *ahta-. Traditionally, one has compared the noun F kelta ‘the colour yellow’, Vt. dial. kõlta ‘egg Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic → finnic loans 53 – Vp. kurdeh, Vt. dial. kurrõ (gen.sg. kurtõõ), E dial. (Mulgi) kurre (< *kurte̮h),23 elsewhere kurt -i ‘deaf’ ~ Lt. kur̃čias, dial. kurtùs, Žem. kur̃tas ‘deaf’ (cf. Lv. kur̃ls, kùrls, dial. kur̃ns, YAv. karəna- ‘deaf’) – F taaja, dial. tavea, K toakie (< *takja) ‘dense, frequent’ ~ Lt. tánkus, adv. tánkiai ‘dense, frequent’ (= Parthian tng ‘narrow, tight’) (Liukkonen 1999: 140–142)24 – F tyhjä, E tühi, Li. tijā (< *tühjä) ‘empty’ ~ Lt. tùščias, Lv. tukšs ‘empty’ (= Skt. tucchyá-) 3.2.4 Nature – F dial. (W) hako ‘conifer branch; needle’, E hagu, dial. haga ‘fine lopped branches; branch, stick’, Li. dial. (W) a’g ‘conifer needle’ (< *hako ~ *haka) ~ Lt. šakà, Lv. dial. (ME iii: 642) saka ‘branch’ (= Skt. śā́khā-) (Būga 1908: 30; Ojansuu 1921: 6)25 23 24 25 yolk’ with Lt. geltà ‘jaundice; (dial.) the colour yellow’ (Thomsen 1890: 172; Kalima 1936: 115). However, F kelta may easily be a recent back formation based on pairs such as F puna ‘the colour red’ ~ punainen ‘red’, and even Lt. dial. gel̃tas, Lv. dial. (ME i: 543) dzȩlts ‘yellow’ are not attested in early sources. I prefer to take *ke̮ltaine̮n ‘yellow’ directly from Baltic *geltâna- (or even *geltaina-? cf. Lv. dial. dzȩltains, ME i: 542) with adaptation to the Finnic adjectival suffix *-inEn (cf. Liukkonen 1999: 39). An apparently old derivative is Lv. dzȩl̂ta ‘ground cedar’, which has also been borrowed into Finnic (see below). If we assume analogical generalization of the weak grade, K (Olonets) kuuris ‘deaf’ also appears to be regular from *kurte̮s (cf. e.g. kieral ‘at once’ < *kertalla; Kalima 1924: 166–167). In this case, F dial. (SE) kuurne and K kuurnis may be borrowings from Olonets Karelian with hypercorrect -rn- (cf. in particular Olonets kuuru ~ F kuurna ‘chute’; see Kalima 1924: 164–166). Given the complementary distribution of *kurte̮s and *kurne̮s, an analogical origin of the latter seems more promising than the assumption of a second, independent loan from Baltic *kur̂nas (Kalima 1936: 124; Junttila 2019: 42). F dial. tavea replaces tavia (< *taɣja) under the influence of the adjectival suffix -ea (cf. dial. lavea, older lavia < *lakja ‘broad’; T. Itkonen 1982: 123). The surprising reflex in Karelian is paralleled by North Karelian voakie ‘peg’ (< *vakja) and roakie ‘limb’ (< *rakja) (idem: 124–125). The acute in Lt. tánkus remains unexplained, as admitted by ALEW (1072) and Smoczyński (2018: 1446), but the IE etymology seems difficult to reject. Thomsen (1890: 244) compared a different Baltic word, Lt. žãgas ‘hayrick, heap’, Lv. obs. schaggas f.pl. ‘Laub, feine belaubte Ruthen, dergleichen sie zu Badequäste brauchen’ (Lange 1773: 272). Junttila (2017: 139) has defended this etymology, assuming the senses in both Finnic and Baltic developed from an earlier ‘trunk’. In support of this, he adduces the Latvian hapax zȩga ‘body’ (in Rucava apud ME iv: 702). The Finnic senses are more easily derived from ‘branch’, however (only the eastern F, K hako ‘rotten or submerged fallen tree’, Vp. hago ‘fallen tree; snag’ would be derivable from ‘trunk’); and an original sense ‘trunk’ would also be hypothetical in Baltic. Junttila (in prep.) instead assumes that the original meaning in Baltic may have been ‘branch’, but this still presupposes one additional hypothesis in comparison to the etymology suggested here. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 54 chapter 3 – F halla, E hall, Li. ǭla (< *halla) ‘frost, hoarfrost’ ~ Lt. šalnà, Lv. sal̂na (LVPPV: sàlna) ‘frost, hoarfrost’ (~ Lt. šálti = Oss. ID sæl- ‘freeze’)26 – F helle (obl. helte-) ‘hot weather’, K dial. (Olonets) helleh adj. ‘sweltering’ (< *helteh) ~ Lt. šil̃tis ‘heat’, cf. Lt. šil̃tas, Lv. sìlts ‘hot’ (= MW clyd ‘sheltered, warm’) (Ojansuu 1921: 7; Kalima 1936: 100) – F kelta-lieko ‘ground cedar’, Vp. dial. küud ‘smoke tree’, Vt. kõlta ‘clubmoss?; a plant used to die (eggs) yellow’ (cf. VKS), E kold ‘clubmoss’ ~ Lv. dzȩl̂ta ‘ground cedar’. Plants named for their use as dyes (cf. Lv. dzȩltãns ‘yellow’) – F metsä, E mets; Li. mõtsā, Võ. mõts (?< *me̮cca) ‘forest, wood’ ~ Lt. mẽdžias, Lv. mežs ‘forest’; Pr. E median ‘wood’ (= R межа́ , SCr. mèđa ‘border between fields; boundary strip’; to Skt. mádhya- ‘in the middle’)27 – F routa, Vt. rõuta (< *routa) ‘frozen ground’ ~ Lt. grúodas ‘frozen ground’ (= Lat. grandō ‘hail’, cf. Rasmussen 1999: 152–153) – F takiainen, dial. takkiainen, E takjas, dial. takijas (< *takkijas) ‘burdock’ ~ Lt. dagỹs, Lv. dadzis ‘burdock, thistle’ to Lt. dègti, Lv. degt ‘burn’ (= Skt. dáhati; cf. Lt. dagùs ‘prickly, bitter, harsh’, LEW 85–86) – F taula, E tael, Li. da’ggõl (< *takla) ‘tinder (fungus)’ ~ Lv. dagla, daglis ‘tinder (fungus)’, whose primary use is as fuel; to Lt. dègti, Lv. degt ‘burn’ (= Skt. dáhati) 3.2.5 Wild Animals – F ankerias, E angerjas, Li. aņgõrz (< *anke̮rjas) ‘eel’ ~ Lt. ungurỹs; Pr. E angurgis */angurjəs/ ‘eel’ (~ Lat. anguīlla)28 26 27 28 Thomsen (1890: 220) compares F halli, E hall ‘grey; grey animal’, Li. aļ ‘grey seal’ directly with Lt. šal̃nis ‘grey cattle’. However, these are more probably both productively formed from the respective words for ‘hoarfrost’ (cf. Kalima 1936: 95). Kalima (1936: 11) considers the comparison phonologically difficult, although the main reason he doubts the loan etymology is that he suspects the Finnic word to be cognate with Taz Selkup mači̮ ‘forest, tundra’. However, the Selkup affricate derives from ProtoSamoyed *-j- (cf. Janhunen 1977: 85) and the word shows no regular correspondence with the Finnic word. It is rather cognate with Md. EM moda ‘earth, soil’ and Finnish muta ‘mud’ (< PU *muďa, Aikio 2002: 22–23). Note that the Selkup word is no longer mentioned in SKES (ii: 343), who nevertheless follow Kalima and consider the Baltic etymology uncertain. The initial u- could be explained as an East Lithuanian dialecticism, in which case the word must have been borrowed from there into the other dialects (Derksen 2015: 479). Vowel assimilation (cf. Būga RR ii: 509; Otrębski 1955: 26; LEW 1163) seems less probable. An alternative account is that the u-vocalism originated in the zero-grade (*h₂ngʷʰ-; cf. Smoczyński 2018: 1561), implying an old ablaut variant. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic → finnic loans 55 – F herhiläinen ‘hornet’, E herilane ‘wasp’ (< *herhiläinen) ~ Lt. šìršė, Lv. sir̂senis ‘hornet’; note Lt. dial. šir̃šilas, Pr. E sirsilis (= Lat. crābrō)29 – F hirvi ‘elk’, E hirv, Li. īra ‘deer, roe-deer’ (< *hirvi, obl. *hirve-) ~ not in East Baltic, cf. Pr. E sirwis ‘roe-deer’ (= MW carw ‘deer’)30 – F karva ‘animal hair; coat colour’, E karv ‘hair, fur; bristle’ ~ Lt. gauraĩ ‘fur; bodily hair’, ?Lv. gauri ‘pubic hair’ (Ulmann 1872: 73 “Scheint nicht sehr bek[annt]”) (= OIr. gúaire ‘animal hair, bristle’) – F kiiliäinen, Võ. kiińläne ‘botfly’ (< *kīlijäinen ~ *kīliläinen); E kiil, -i ‘dragonfly; botfly’ (< *kīli) ~ Lt. gylỹs ‘gadfly’, dial. ‘sting’ (~ gélti ‘to sting’ = Arm. kełem ‘torment, afflict’) (Mikkola 1906: 78) – F käärme ‘snake’, Li. kīermõz ‘woodworm’ (< *kärmes / *kärmeh), E dial. (Saaremaa) kärm, kärv -i ‘snake’ (< *kärmi) ~ Lt. dial. kirmìs ‘worm; snake’, Lv. cìrmenis, dial. (Kurzeme) cirmis ‘maggot’ (= Skt. kr̥ mí- ‘worm’) – F vaapsainen, Vp. bapshaine, Vt. vaapsia, E vapsik (?< *vapsas) ‘wasp’ ~ Lt. vapsvà, dial. vãpsas, Lv. dial. vapsene; Pr. E wobse ‘wasp’ (= Pahlavi wpc /wabz/)31 3.2.6 Animal Husbandry – F dial. ehkonen (dial. hehvo; standard hieho), E dial. õhv, Li. õ’v (< *e̮hva) ‘heifer’ ~ Lt. obs. ašva, ešva ‘mare’; perhaps Lv. ⟨ôssa⟩ ‘mare’ (Elger 1683: 133; cf. Karulis 1992 i: 468) (= Lat. equa)32 29 30 31 32 Lt. dial. šir̃šilas is extremely rare; the l-suffix may have been added secondarily within Finnic, cf. in particular *mehiläinen ‘bee’ (Nieminen 1934: 32–35; Kalima 1936: 100). Lt. šìrvas ‘grey, dapple-grey’ could be cognate if it originally meant ‘roe-coloured’, but more likely it represents a contamination of Lt. šìrmas ‘grey, dapple-grey’ (= Lv. sir̃ms) and šývas ‘grey, whitish (usu. of horses)’ (= Pr. E sywan, SCr. sȋv ‘grey’). Particularly note that the acute accentuation would be in conflict with MW carw (< *ḱr-uo-; cf. Zair 2012: 94–95). On Lt. kárvė ‘cow’ etc., see Chapter 6, fn. 109. F vaapsainen, K dial. vuapsahane, Vp. bapshaine reflect a derivative *vapsahainen (cf. F muurahainen ‘ant’). Possibly, these are built on a more basic *vapsas continued by Vt. vaapsaz (absent from VKS; cited after SKES 1580) ‘wasp’. E vaps-ik ‘hornet’, in any case, is the result of suffix substitution (Nieminen 1934: 35). The long vowel attested in these forms is the result of the sporadic but frequent secondary lengthening before *-Cs- (T. Itkonen 1987: 195–196). Li. vaps, nom.pl. vapsūd ‘wasp’ (< *vapso or *vapsoi, Nieminen 1934: 33–34), may well be a later loan from Latvian. The secondary nature of F h- is supported by the presence of h-less variants well outside of the area of Estonian influence (Junttila in prep.). Baltic *ašvā was already moribund at the time of its earliest attestations, being replaced with kumẽlė in Lithuanian and ķève in Latvian: Both versions occur side by side in Szyrwid: “kumełe[,] aſzwa”. Likewise, Elger has “kiêwa, D ôssa”. In Bretke, eſchwų gen.pl. only occurs as a marginal gloss to kumelių (see ALEW 60), and in Ruhig, the word is semantically specified: ‘eine Stutte großer Art’ (Ruhig i: 8). For a discussion of the semantics, see 3.6.1. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 56 chapter 3 – F hanhi, E hani, Võ. haah́ (< *hanhi, obl. hanhe̮-) ‘goose’ ~ Lt. žąsìs, Lv. zùoss ‘goose’; Pr. E sansy ‘goose’ (= Skt. haṃsá-) – F oinas, E oinas, Võ. oinas, dial. oonas (< *oinas) ‘ram’ ~ Lt. ãvinas, Lv. àuns ‘ram’; Pr. E awins ‘ram’ (cf. Skt. ávi- ‘sheep’) – F paimen, Vt. dial. paimõõ, Li. paint (< *paime̮n) ‘shepherd’ ~ Lt. piemuõ, obl. píemen- ‘shepherd’ (= Gr. ποιμήν) – F villa, E vill, Li. vīla (< *villa) ‘wool’ ~ Lt. vìlna, Lv. vil̃na ‘wool’ (= Skt. ū́rṇā-) – F vuohi, Vt. voho (< *vōhi, obl. *vōhe̮-) ‘goat’ ~ Lt. ožỹs, Lv. âzis ‘he-goat’; Pr. E wosee ‘she-goat’ (= Skt. ajá-)33 – F vuona, Vt. võdna, Võ. vu̬ u̬n, Li. ūoņõz (< *vōtna) ‘lamb’ ~ not attested in Baltic; cf. OCS агнѧ ‘lamb’ (= Lat. agnus, Gr. ἀμνός)34 On Võ. pahr ‘boar’, see p. 73. 3.2.7 Agriculture – F herne, E hernes, Li. jērnaz (< *hernes / *herneh) ‘pea’ ~ Lt. žìrnis, Lv. zir̃nis ‘pea’ (= Lat. grānum ‘grain’) – F pelu, usu. pl. pelut, Vp. pelu (?< *pe̮lut nom.pl.)35 ‘straw chaff’ ~ Lt. arch., dial. pẽlūs, Lv. pȩlus pl.; cf. Pr. E pelwo ‘chaff’ (= Skt. palā́va- ‘chaff’, Lat. pulvis ‘dust’ < *pel-ou-) ́ en- ‘lin– F siemen, E seeme, Li. sīemt (< *sēmen) ‘seed’ ~ Lt. sėmuõ, obl. sėm seed, seed; (obs.) sowing’; Pr. E semen ‘seed’ (= Lat. sēmen) – F vannas, Vt. vadnaz ‘ploughshare’, E dial. (W) vannas ‘plough beam’ (< *vatnas) ~ unattested in East Baltic; cf. Pr. E wagnis ‘coulter’ (= Gr. (H.) ὀφνίς ‘ploughshare’) (Paasonen 1909b)36 33 34 35 36 For the Votic form, cf. toho < *tōhi, obl. *tōhe̮- ‘birchbark’. The substitution *gn → *tn (see also *vatnas, below) appears to suggest that the cluster *kn was not yet licenced at the time of borrowing. According to Paasonen (1909b: 17), *kn had developed into *nn already in Early Proto-Finnic, cf. F ynnä, Li. īņõ, (Salaca) ǖnis ‘together’ (< *ük(t)-nä ess.sg. from *ükci ‘one’), Võ. nännüt ‘see act.prt.’ (< *näk-nüt). Kallio (2008b: 313–314), however, argues that *kn was preserved in ‘Core Finnic’, but examples like Estonian näinud < *näk-nüt might show restored *k. If Paasonen is correct, the word *sakna ‘sauna’ must have post-dated these loanwords (Kallio 2008b: 315; although Kallio’s preGermanic etymology can hardly be consistent with this chronology). The word is generally plurale tantum in Finnish. I cannot establish whether the situation is similar for Veps, as the form is absent from Зайцева/Муллонен 1972. I can trace it back as far as a Central Veps “peлu” cited in SKES 516. LÄGLOS (iii: 368–369) leave open the possibility of a Germanic origin. The Norse data (cf. ON hapax vangsna obl.sg., Nw. dial. vangsne, (17ᵗʰ c.) vagnsne) seem to point towards a proto-form *wagnVsnan-. According to Kroonen (2013: 565), the umlaut in Swiss wägese ‘ploughshare’ could favour a reconstruction *wagnisan- (cf. the forms in Schw. Id. xv: 770–774, where folk etymology is instead suspected). Since Pr. E ⟨wagnis⟩ can stand for */wagnas/, a Baltic source does not raise any phonological issues, while the neuter s- Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic → finnic loans 57 3.2.8 Technology – F ansa ‘trap, noose’, E aas, Li. ǭz ‘noose, loop; handle’ (< *ansa) ~ Lt. ąsà, Lv. ùosa ‘handle, eyelet’; Pr. E ansis ‘(pot)hook’ (= Lat. ānsa ‘handle’)37 – F kirves, E kirves, Li. kīraz (< *kirves) ‘axe’ ~ Lt. kir̃vis, Lv. cìrvis ‘axe’ (~ Gr. κείρω ‘crop, shave’; cf. Trautmann 1923: 135)38 – F mäntä, E mänd (< *mäntä) ‘stirring stick’ ~ Lt. meñtė ‘mixing stick; trowel, paddle’ (= Skt. mánthā- ‘churning stick’) – F niisi, usu. pl. niided, E niied, Li. nīdõd (< *nīci, pl. *nītet) ‘heddle’ ~ Lt. nýtys, Lv. nĩtis f.pl. ‘heddle’, cf. Lt. nýtis ‘warp thread’ (~ Lv. dial. nĩt, nìt² ‘thread (a needle)’; OIr. sníid ‘twist, bind’) – F dial. pahla ‘(fishing) rod’, E dial. (W) pahl ‘spit, skewer’ (< *pahla) ~ Lt. baslỹs ‘fencepost, stake’ from Lt. bèsti ‘drive in, stick in’, Lv. best ‘dig, bury’ (~ Lat. fodiō ‘dig, pierce, thrust’) (Kalima 1928: 102–103)39 – F ratas, E ratas, Li. (Salaca) rat (< *rattas) ‘wheel, cartwheel’ ~ Lt. rãtas, Lv. rats ‘wheel’ (= Skt. rátha- ‘chariot’, Lat. rota ‘wheel’) – F rattaat, E arch. rattad, Li. rattõd (< *rattahe̮t pl.) ‘cart’ ~ Lt. rãtai, Lv. rati pl. ‘cart’ (= Skt. rátha- ‘chariot’; see above) – F siula, K šikla (< *sikla) ‘side net in a seine’ ~ Lt. tiñklas, Lv. tìkls ‘net’, cf. Pr. E sasin-tinklo ‘snare’ (~ Gr. τείνω ‘stretch, pull tight’) (Koivulehto 1979a: 267– 269)40 – F silta, E sild, Li. sīlda (< *silta) ‘bridge’ ~ Lt. tìltas, Lv. til̃ts ‘bridge’ (to Lt. dial. tìlės ‘bottom of a boat; planks (as paving)’; for the semantics, cf. OR мостъ ‘bridge; pavement, floor’; СДРЯ 11–14 v: 25–26)41 37 38 39 40 41 stem *wagnas, obl. *wagnis- reconstructed for Germanic by Karsten (1915: 84–85) remains purely hypothetical. For a similar reason, it is difficult to derive the Prussian word from Germanic (pace Smoczyński 2000: 132–133). A Germanic origin (Sammallahti 1998: 123), cf. ON ǽs ‘eyelet (in a shoe)’ (< *ansjō-) is formally less straightforward. On R dial. черв, which hardly belongs here, see Chapter 1, fn. 61. It seems natural to compare Lt. kir̃sti, Lv. cìrst ‘chop, cut’ (= Skt. kr̥ ntáti ‘cut off’). However, the dental would not be lost in the formation *kirt-u̯ ia̯ -, which means the root must be identified as *ker-. The rare deverbal suffix *-vīs appears to form agent nouns (Leskien 1891: 348; Skardžius 1948: 379). On -hl- < *-sl-, see Aikio 2015a: 44. Krevinian ⟨pahlis⟩ ‘stake’ cited by VKS, on the other hand, like Livonian pǭ’lõz ‘stake’, is from Latvian pàlis ‘stake’, with the orthographic sequence -Vh- simply standing for *V̄ as in Krevinian ⟨pählin⟩ ‘head, leader’ (= Vt. päälin), ⟨śohla⟩ ‘salt’ (= Vt. soola). Within Balto-Slavic, the same root is continued in Lv. tît, 1pres. tinu ‘wrap, wind’ (IEW 1065–1066; hence probably the intonation tîkls given by LVPPV), and R dial. тенёта, Slk. arch. tenatá n.pl. ‘net snare’ (whose suffix can be compared with that of R решето́ , Slk. rešeto ‘sieve’; Vaillant 1974: 697); see ALEW 1280. Further, OR тьло ‘ground, bottom’, Pr. E talus ‘floor’, OE þel ‘plank (of wood); plate (of Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 58 chapter 3 – F tempaa-, E tõmba-, Li. tȭmbõ- (< *te̮mpaita-) ‘pull, tug’ ~ Lt. tem̃ pti (3sg. -ia), Lv. tìept (1sg. -ju, -u) ‘stretch, tighten’ (= NP tāb- ‘twist, spin’)42 – F torvi, E dial. tõri (< *torvi, obl. *torve̮-) ‘horn (for blowing)’ ~ Lv. tàure ‘hunting horn’, Lt. taurė ̃ ‘chalice, drinking horn’, cf. Lt. tauras ‘aurochs’ (= Gr. ταῦρος ‘bull’) – F tuulas (< *tūlas) ‘night fishing; fishing spear’; F tuulaalla, Vp. dial. tuľhuuda (< *tūlahe̮la-) ‘spear-fish by torchlight’ ~ Lv. dũlis ‘torch for night fishing’, also ‘torch to fumigate beehives’, Lt. dū́lis ‘fog; smoke to fumigate beehives’ (cf. Hitt. tuhhae-ᶻⁱ ‘produce smoke’) 3.2.9 Other – F jo ‘already’, E jo, dial. ju ‘already, indeed’, Li. jõ, ju (< *jo) ‘already’ ~ Lt. jaũ, Lv. jàu ‘already’; Pr. iii iau ‘je’ (= OCS оу, ю ~ Go. ju; further Gr. (Hom.) αἰεί ‘always’; cf. Dunkel 2014: 352–353)43 – F liika, E liig (< *līka) ‘surplus, extra; odd (number)’ ~ Lt. liẽkas, Lv. lìeks ‘surplus; odd (number)’ (= Gr. λοιπός ‘left over’) – F reuna, Vp. röun ‘edge’ (< *re̮una), ?Võ. rõõnõq ‘strip (of fabric)’ (?< *re̮une̮k)44 ~ Lt. briaunà ‘brim, edge’ (~ ON brún ‘eyebrow; brow (of a hill); strip of cloth’, Skt. bhrū́- ‘eyebrow’) (Būga 1908: 42)45 42 43 44 45 metal)’, in compounds ‘floor’. Probably, all of these words are related to Lat. tellūs ‘ground, earth; the Earth’ (< *telH-nu- with regular laryngeal loss? cf. van Beek 2011: 162–165) and OIr. talam ‘earth, ground’, although the formations are all different. Root cognates have been suggested in several branches, but most of these are semantically uncompelling (cf. IEW 1064–1065). The long -ā- in Iranian is unexpected (for a suggestion, see Cheung 2007: 389), but the etymological equation seems in principle attractive. Here probably, if reliably attested, Ic. obs. þömb ‘bowstring’ (which need not originally be from ‘gut’; Ic. þömb ‘belly’ is perhaps to be separated as a Reimbildung to vömb ‘belly, rumen’). The Baltic etymology, suggested by Thomsen (1890: 174), is considered by LÄGLOS (i: 140) to be “lautlich problematisch”, and a Germanic etymology is preferred (thus also SSA i: 238; Häkkinen 2004: 278). However, a Germanic origin is unattractive, as not only is the word unattested in Norse, the substitution Germanic *u → Finnic *o lacks convincing parallels. The Baltic etymology, on the other hand, does not pose any phonological issues (see 3.3.1.6). The Võro form may belong here if it originated in the eastern dialects showing õu : õõ gradation (cf. dial. lõõnõq ‘south’, gen. lõunõ). The word is indeed recorded primarily in this dialect area, although VMS reports a couple of stray attestations from further west. If true, the Võro gen. ryynõ would have to be analogical. Semantically, we can compare the sense ‘strip of fabric’ in Norse. For this polysemy, compare also Lt. dial. brunìs ~ brùnė ‘eyebrow; dull edge’ and further comparanda in LEW 57. The attractive analysis of Pronk (2015: 333) would see the forms with *-n- as continuants of an original singulative *h₃bʰru-n-, while Lt. bruvìs, Skt. bhrū́‘eyebrow’, etc. would reflect a fossilized dual *h₃bʰru-h₁. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic → finnic loans 59 – F suola, E sool, Li. sūol (< PF *sōla) ‘salt’ ~ Lv. sā̀ls (gen.sg. sā̀ls), dial. sā̀lis² (gen.sg. sā̀ļa; EH ii: 470) ‘salt’ (= Gr. ἅλς) (Būga 1924b: 104)46 – F vuoro ‘turn, shift’, E voor -u ‘turn, time’ (< *vōro) ~ Lt. vorà ‘line, row’, dial. ‘turn, shift’ (from Lt. vérti, Lv. ver̃t ‘pierce; thread, string together’) (Koivulehto apud Häkkinen 2004: 1514)47 3.2.10 Inner-Baltic Etymologies There is a small group of words which can be analysed as derivatives within Baltic, even though their ultimate origin is unknown. In these cases, the direction of borrowing can nevertheless be considered certain: – E õis, dial. heis ‘flower’ (< *häici, obl. häite-), whence F heiti- (< *häiti-), E õitse-, Võ. häitse- (< *häiticce-) ‘to bloom’ ~ Lt. žíedas (dial. žáidas), Lv. ziêds ‘flower’; cf. Pr. TC zaidiantẽ acc.sg. ‘blossoming’ (~ Lt. žydėt́ i ‘to bloom’) (Топоров/Трубачев 1962: 247; Mägiste 1970)48 – F haara, E dial. haar (gen.sg. haara), Võ. haro, Li. a’r, nom.pl. a’rūd (< *hara ~ *haro) ‘branch, fork’ ~ Lv. zars (?→ Lt. dial. žãras) ‘branch; prong’ also ‘ray of light’49 (to Lt. žėrėt́ i ‘glow, sparkle’) – F härkä ‘ox, bull’, E härg, Li. ǟrga ‘ox’ ~ Lt. žìrgas, Lv. zir̂gs; Pr. E sirgis ‘horse, steed’ (~ Lt. (ap-)žer̃gti ‘sit astride’) – F härmä, E härm, Li. ǟrma (< *härmä) ‘hoarfrost’; F harmaa, Vt. harmaa (< *harmaka), Võ. haŕm, -i (< *harmi) ‘grey’ ~ Lt. šarmà, Lv. sar̂ma ‘hoarfrost’; Lt. šìrmas, Lv. sir̃ms ‘grey, dapple-grey’50 – F luuta, E luud, Li. lūdõ (< *lūta) ‘broom’ ~ Lt. šlúota, Lv. sluôta ‘broom’ (to Lt. šlúoti ‘sweep’)51 46 47 48 49 50 51 On the Mordvin and Permic words for ‘salt’, see pp. 137–138. For the development from ‘row’ to ‘turn’, compare R о́чередь ‘line, row; turn’. The Baltic verb is related to OCS (Supr.) проврѣти* ‘thrust through’, Bg. вра ‘shove, thrust’ and is generally considered to belong with Lt. at-vérti, Cz. otevříti, Lat. aperō ‘open’ (e.g. LIV 227– 228). The etymological comparison with OHG kīnan* ‘sprout, come forth’, OE cīþ ‘sprout, shoot’ (Walde/Pokorny i: 544; LIV 161–162; Kroonen 2013: 287) is not certain (cf. ALEW 1506), as the Baltic *d is unexplained. Assuming an earlier present-tense formant (cf. Smoczyński 2018: 1735) remains ad hoc. Cf. zarus zaruodama ‘casting rays (of the sun)’ in folk songs, ME iv: 691–692; compare English beam or Lv. stars ‘ray of light; (dial.) branch’; see ME loc. cit. Baltic *šar̂mā- ‘hoarfrost’ is a derivative of *šir̂ma- ‘grey’. The semantics can be illustrated by several parallels: (1) Lt. šer̃kšnas ‘hoarfrost, rime; grey (of animals)’ (= Sln. srẹ̑n ‘hoarfrost’ ~ CS срѣнъ ‘greyish-white’ < *ḱersno-, cf. ME iii; 722), (2) F halli, E hall ‘grey; grey animal’, Li. aļ ‘grey seal’ (< PF *halla ‘hoarfrost’, see fn. 26), and not least (3) ME hore-frost ‘hoarfrost’, cf. OE hār ‘grey, hoary; grey-haired’. See also Liukkonen (1999: 38). According to Kortlandt (1995), šlúo- regularly reflects *ḱleh₃u-, and the Baltic words are to Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 60 chapter 3 – F puuro ‘porridge’, E puder ‘porridge; mash’ (< *putro) ~ Lv. putra ‘porridge, mash; a kind of soup’, Lt. putrà ‘gruel, skilly; flour soup’. Compare Lt. dial. (E Aukšt.) pùtelis ‘thickened soup of oatmeal and milk’52 – F rako, E pragu (< *rako) ‘crack, crevice; gap’ ~ Lt. spragà, Lv. spraga ‘gap (usu. in a fence), crack’ (cf. Lt. sprógti, Lv. sprâgt ‘burst, crack’)53 – F rouhi-, Vp. rouhi- ‘grind (coarsely), crush’, E rõhu- ‘press down, oppress’ (< *rouhi-) ~ Lv. kràusêt ‘crush’, Lt. dial. kraušýti ‘barge, shove’ (cf. Lt. krùšti ‘pound in a mortar’) (Kalima 1936: 156)54 – F seiväs, E teivas, Li. tāibaz, Võ. saivas (?< *ste̮ipas; see 3.3.1.4) ‘post, stake’ ~ Lt. stíebas ‘stalk, trunk, pillar’, ?Lv. stìebrs ‘stalk, reed, rush’ (~ R сте́бель ‘stalk’, SCr. stáblo ‘tree; trunk’)55 – F tapa, E dial. taba, tava (< *tapa) ‘custom, habit’ ~ Lt. dial. dabà, Lv. daba ‘way, custom’ (~ OCS по-добати, Go. ga-daban ‘befit, be suitable’)56 – F tuura, E tuur (< *tūra) ‘ice chisel’ ~ Lv. dùre ‘fist’, dial. ‘ice chisel’ (from the verb Lt. dùrti, Lv. dur̃t ‘stab, poke, prick’)57 52 53 54 55 56 57 be compared with Lat. cluere, cloāre ‘purify’. However, the implied phonological development in Baltic is doubtful (cf. Villanueva Svensson 2015). Furthermore, both Latin variants are hapaxes used to explain a divine name, and therefore raise suspicions (cf. Rix 1999: 519). Since the soup is generally thickened with flour, probably from Lt. pùsti ‘swell up’ (LEW 681–682), like pùtos ‘foam, froth’. Other motivations are possible; compare the secondary sense ‘eat sloppily’, or even “Nuog putros tik pilvas išsipūtė” (Aukštadvaris, LKŽ) ‘All I got from that putrà was a bloated stomach’. Note that Mikkola (1896a: 121) was unconvinced by this derivation, and derived the Baltic words instead from Finnic. The metatony is awkward, so the association with the verbal root may be secondary. If the initial s- is due to lexical convergence, the rare dial. pragà beside Lt. próga ‘opportunity’, dial. ‘forest clearing’ (cf. progas ‘Lücke’, ClG i: 1219) seem to suggest an analysis *pra-gā- (Smoczyński 1998: 255–256); compare Lt. próperša ‘thawed patch of ice; break in the clouds’ beside praparšas (Szyrwid; see Chapter 6, fn. 82). In that case, the root is perhaps that of Lv. gãju ‘went’. Finnish louhi- ‘chip away (stone), quarry’, for which Thomsen (1890: 194–195) has suggested another Baltic etymology, most likely represents a secondary alternant of rouhi- (for similar cases, see Nikkilä 1999: 130–134), perhaps under the influence of lohjeta (lohke-) ‘chip, break (intr.)’. The Baltic acute is unexplained. Note the similarly obscure acute in Lt. stámbas ‘stem, stalk’ (= Skt. stambha- ‘pillar’). The Baltic word must have developed from a verbal base meaning ‘be suitable’. Arm. darbin ‘blacksmith’ is rather to be derived from Urartian (Yakubovich 2009: 267–270), which also makes the appurtenance of Lat. faber ‘smith; artisan’ less certain (Pronk 2019a: 152; but differently see Simon 2022: 71). Although this meaning is limited to a small area in northern Latvia, so it cannot be entirely excluded that this sense arose under Finnic influence (cf. Thomsen 1890: 169), the derivation from the cited verbal base is semantically satisfactory; compare the parallel derivative in Žem. durà ‘ice chisel’ (LEW 113). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic → finnic loans 61 – Vp. värpita-, dial. värbita- ‘spin, rotate (a spindle)’, Li. vērbikšõ, dial. vǟrbõkš ‘spin (thread)’ (?< *värpi- ~ *värpe-) ~ Lt. ver̃pti (pres. ver̃pia) ‘spin (thread)’, Lv. vḕrpt ‘spin; (refl.) wind round’ (Thomsen 1890: 240; Posti 1946: 386)58 – F vielä, E veel, Võ. vi̬il̬ (< *vēlä)59 ‘still, yet’ ~ Lt. vėl̃ , dial. vėl̃ e, vėl̃ ei ‘again, still’, Lv. vêl ‘still, yet’ (?< a fossilized adverbial derivative of Lt. vėlùs, Lv. vȩ̂ls ‘late’ with a development ‘lately’ > ‘recently’ > ‘still’, cf. Būga 1923–1924: 95–96) 3.3 Analysis of Sound Substitutions 3.3.1 Vocalism 3.3.1.1 *ǣ → *ē; *ā → *ō While it seems natural to interpret the first-syllable vowel in PF *sēmen as an ́ en-), this is exact equivalent of Baltic *ē (thus e.g. Kalima 1936: 68; cf. Lt. sėm rather a notational fallacy. The Proto-Baltic precursor of Lt. ė was almost certainly a low vowel */æː/. It remains low (in part) to this day in standard Latvian; likewise, the low vowel realizations /aː/ and /æː/ are still present in certain East Aukštaitian dialects (cf. Bacevičiūtė et al. 2004: 124–125), and one can still find the spelling ⟨a⟩ for Proto-Baltic *ā in early Lithuanian texts deriving from Prussian Lithuania, such as the Wolfenbütteler Postille, the Mažvydas Catechism and sporadically elsewhere (Palionis 1995: 46). As shown by Lehtinen (1967: 150–151) and Aikio (2012b: 232), Finnic *ē has also developed from an earlier low vowel *ǟ (e.g. PF *kēle- ‘tongue’ < pre-PF *kǟlə < PU *kälə). This raising must have predated the emergence of secondary *ǟ resulting from contraction over PU *x and *ŋ (e.g. PF *pǟ ‘head’ < PU *päŋə, UEW 365), which was no longer subject to raising.60 Therefore, we should 58 59 60 According to Junttila/Holopainen (2022: 112–113), the connection to weaving is an East Baltic innovation, but the original semantics, even within Balto-Slavic, are difficult to establish (cf. LIV 691 s.v. *u̯ erp- fn. 1); therefore, it is unclear whether Pr. iii powiērpt ‘leave, forsake’, OR вьрпати* (attested вьрпеши 2sg.pres.) ‘tear, rob’, CS на-врапити ‘invadere’ (Miklosich 1865: 399) even belong here. At least from a semantic point of view, it is tempting to compare OE warp, OHG waraf ‘warp (in a loom)’ (Persson 1912: 497–499; Trautmann 1923: 353), which could be connected by assuming Kluge’s law. Compare the OED’s definition of warp: “The threads which are extended lengthwise in the loom, usually twisted harder than the weft or woof” (emphasis mine). Livonian vēl, ve’l represents an independent loan from Latvian (cf. Suhonen 1973: 237). This chain of developments can be attractively analysed as a push shift. The fact that the Baltic loans underwent this raising in Finnic suggests, by extension, that they predated the loss of intervocalic *ŋ and *x (or at least the vowel contraction). The fact that these phonemes are not represented in the Baltic loanwords is not surprising, as no corresponding phonemes are present in Baltic. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 62 chapter 3 rather state that Baltic */æː/ was borrowed as pre-PF *ǟ, and the subsequent raising in Finnic and Lithuanian must be considered parallel, unrelated developments. A confirmation of this chronology is provided by the substitution of PB *ā as PF *ō in e.g. PF *vōhi ← PB *âžē ‘goat’. Although in fact entirely analogous to the case of */æː/ → *ē, only this substitution has sparked any significant debate. One has either operated with a Proto-Baltic reconstruction *ō (Kalima 1936: 66–67; cf. Mikkola 1930: 443, fn.) or assumed that the Finnic people came into contact with a Baltic dialect in which *ā had become rounded, be it Curonian (Nieminen 1934: 59), High Latvian (Endzelīns 1932: 255), or “North Baltic” (Kallio 2008a: 272). However, all of these speculations are rendered unnecessary by the insight that Finnic *ō has itself developed from an earlier *ā (Pystynen 2018: 72–75). The following additional pieces of support can be presented for this chronology: a. Baltic *ā-stems are overwhelmingly adopted as Finnic *a-stems, cf. *halla ‘frost’, *villa ‘wool’. Likewise, Baltic *-ē was adopted as Finnic *-ä in *mäntä ‘whisk’ (← *mentē). b. The substitution *ō → *ū, *ou is naturally accounted for if Proto-Finnic lacked a phoneme *ō at the time of borrowing (see pp. 68–69). At the same time, a couple of arguments can be put forward in favour of a Baltic rounded vowel: a. Baltic *ā-stems are occasionally adopted as Finnic *o-stems, as in F heimo ‘tribe, kin’ ← Baltic *šeimā- (on which see 3.3.3) b. F vohla, dial. vohli ‘kid’, if loaned from Baltic (cf. Lt. ožẽlis ‘kid’), implies an underlying *vōhl-, with shortening of the vowel before a consonant cluster (Koivulehto 2000: 104). The latter case can be explained easily provided the syncope of the medial *-e̮is a late development (Kallio 2007: 241): Baltic *âžel- would be adopted into Finnic as *āše̮la, which subsequently developed to *vōhe̮la (with automatic *vbefore *ō-) and finally to *vohla by syncope. This incidentally nicely accounts for the absence of the development *wo- > *o- (Posti 1953: 72; Aikio 2014b: 10) in Finnic. The Proto-Finnic status of the word vohla remains doubtful, however, as the word is limited to the dialects of Western Finland, which seems suggestive of a local innovation.61 61 The oft-quoted Estonian vohl is found only in the Kuusalu coastal dialect in the far north (according to VMS), and is probably a loan from Finnish. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic → finnic loans 63 3.3.1.2 *e → *ä ~ *e; *a → *a ~ *o In some cases, short *e is substituted as *ä, cf. *mäntä ‘whisk’, *värpi- ‘to spin’ and perhaps the second syllable of *tüttär ‘daughter’ (< *dukter-). This is the expected substitution, as Proto-Baltic *e was probably an open vowel */æ/, as it remains to this day in the modern languages. In my opinion, here also belongs *kärmes ‘snake’, which is most probably derived from a full-grade variant *kermstill attested in Lv. cḕrme ‘roundworm’, Lt. kermenaĩ ‘bee larvae’ (Thomsen 1890: 98; Liukkonen 1999: 54).62 Besides this, we find examples of the substitution *e̮: *se̮sar ‘sister’, *ne̮pat ‘nephew’, *ke̮lta ‘ground cedar’, *ke̮ltaine̮n ‘yellow’, *pe̮lut ‘chaff’, *te̮mpaita- ‘pull, tug’, *e̮hka ‘heifer’. Kallio (2008a: 270) has argued that the conflicting reflexes might be explained if Baltic *e was phonetically *[ε], standing somewhere in between Finnic *e and *ä. An interesting fact, however, is that all of the examples involve a Baltic back vowel in the second syllable, which suggests *e̮ could have been a sort of compromise between the backvocalic stem-vowel and front vowel of the initial syllable.63 The two substitution strategies for Baltic *e are mirrored by the similar situation with regard to Baltic *a: here the usual substitution is Finnic *a (of which there are many examples), beside which examples of *o can also be identified. Here, a similar solution could be proposed by suggesting that Baltic *a was in fact *[ɔ] (Kalima 1936: 64–65; Steinitz 1964: 338). Among the loanwords with a clear Indo-European background, four certain examples show an *o: *oinas ‘ram’, *torvi ‘horn’, *morcijan ‘bride’ and *rouhi- ‘crush’. These examples would be consistent with Nieminen’s theory (1957: 199–201) that Baltic *o reflects *a where a front vowel follows in the next syllable. However, this theory encounters counter-evidence (e.g. *hanhi ‘goose’), and is typologically questionable (Steinitz 1964: 336). A typologically more apt observation is that Baltic *a in all three examples is found adjacent to a labial, viz. *avinas, *taurē, *martjan, *krauš-. It is therefore possible that the substitution with Finnic *o was a reflection of an allophonic rounding in a labial environment within the Baltic donor dialect. Nevertheless, we could only talk of a tendency here, as no rounding is found in *hampas ‘tooth’, *karva ‘(animal) hair’ or *vapsas ‘wasp’. All in all, the evidence is rather too limited to convincingly identify conditioning factors. 62 63 An ablauting *kerm- : *kirm- in Proto-Balto-Slavic might be required to account for the unexpected reflex *ir as opposed to *ur after a labiovelar, cf. MW pryf ‘worm, maggot, fly’ (see Kortlandt 1978: 240; Matasović 2004: 350). This is an argument in favour of interpreting *me̮cca ‘forest’ as archaic. See below. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 64 chapter 3 The substitution observed in PF *härmä ← *šârmā- ‘hoarfrost’ is quite unclear.64 The expected back vocalism seems to be found in F harmaa, Võ. haŕm ‘grey’, which Liukkonen (1999: 38) has plausibly analysed as inner-Finnic derivatives of a noun *harma; however, no regular derivational process can account for the shift to front vocalism within Finnic. Although similar cases of secondary vocalism are sporadically observed in Finnic (Saukkonen 1962; Nilsson 1996: 186), these normally concern words of an expressive character. See also 3.5.3. 3.3.1.3 *i → *i ~ *e; *u → *u ~ *ü Baltic *i is normally reflected as Finnic *i in loanwords (*villa ‘wool’, *silta ‘bridge’, etc.). In some cases, however, we find Finnic *e, instead, viz. *herneh ‘pea’ (← Baltic *žir̂nīs), *herhiläinen ‘hornet’ (← *šir̂š-) and *helteh ‘hot weather’ (← *šilta-). This hesitation could simply be attributed to a more centralized pronunciation of /i/, as is found in modern Lithuanian (Pakerys 2003: 24–25). On the other hand, conditioning factors may be identified: Kalima (1936: 70) has attributed the lowering to the influence of a following resonant. Yet, as Ritter (1998) has pointed out, it is hardly a coincidence that all the examples feature an initial *š or *ž in Baltic.65 This might be phonetically understood if we suggest that Baltic *š and *ž were realized as retroflex consonants, as usually assumed for Proto-Uralic *š. Retroflexion tends to be disfavoured in the environment of front high vowels, and may be accompanied by concomitant vowel lowering (Hamann 2003: 94, 99–100). However, the substitution *ä in *härkä ‘ox; bull’ (← *žir̂ga-) might suggest that we are dealing with a genuine sound change in a Baltic dialect,66 which would provide evidence that the source language was not a direct 64 65 66 Since Thomsen (1890: 221), one has generally referred to a Latvian sȩr̂ma (cf. ME iii: 819; EH ii: 478) to support the reconstruction of a Baltic source form *šer̂mâ. However, this Latvian form is probably the result of a secondary dialectal development (Endzelīns 1923: 36–37) and cannot be projected back to Proto-Baltic. Ritter in fact operates with a rule the *i is lowered after both *k- and *š-. However, the examples with *k- are unconvincing. For *kärmeh ‘snake’, I posit an original e-grade; see above. In view of its distribution, F kelles, kelle, K dial. (N) kelleš ‘split log; large round chip; thick slice, chunk’ is more likely to be loaned from Sámi (cf. Sá. N galda ‘block of wood; tree stump’, Sk. kõldd ‘block; wooden lure’) than the opposite (contra Kalima 1936: 115). This is supported by the fact that the substitution of Finnic e → Sá. *e̮ is practically unparalleled (Aikio 2006b: 32), while the opposite (i.e. Sá. *e̮ → Finnic e) is known to have occurred (see Aikio 2009: 77). If true, then the association with ‘something split’ would have arisen secondarily within Finnic, and the connection to Lt. skìltis ‘clove; slice; piece cut off’ looks more tenuous. This is far preferable to seeing the source in the deverbal noun žargà ‘spread legs’ Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic → finnic loans 65 ancestor of the attested East Baltic languages. The fact that we find *ä–ä as opposed to *e̮–a in this word (see above) possibly implies that lowered *i was still phonologically distinct from *e. As a substitution for Baltic *u, we find both *u (in *kurte̮h ‘deaf’ and *putro ‘porridge’) and *ü (in *tüttär ‘daughter’ and *tühjä ‘empty’). This is noted by Thomsen (1890: 100) and Kalima (1936: 71), but not commented upon. Again, one might attribute this vacillation to a ‘laxer’ pronunciation of Baltic short vowels (cf. Pakerys 2003: 24–25), but Koivulehto (1971) has compared the frontvocalic forms to doublets such as F rastas ~ dial. rästäs ‘thrush’. In light of this, Kallio (2008a: 269) writes “[t]here are no reasons to think that the substitution *u → *ü had anything to do with the actual pronunc[ia]tion of Proto-BaltoSlavic *u”. In my view, it is anachronistic to use post-Proto-Finnic vacillations such as that in the word for ‘thrush’ to explain phenomena in Early Proto-Finnic (see 3.5.3). In the word for ‘daughter’, the front vocalism can be explained in the context of Finnic vowel harmony: we may assume that the choice of front vocalism was triggered by the second syllable of Baltic *dukter-. Such an explanation does not really work for ‘empty’, however; although it is phonetically possible that the second-syllable vowel in Baltic *tuštja was allophonically fronted in the neighbourhood of *j, this is an ad hoc assumption, especially in view of the back-vocalic *haljas ‘green’ and *anke̮rjas ‘eel’. It therefore seems that the explanation should be at least partly phonetic, although multiple factors may be at play. The length of Baltic *ī and *ū is reflected in the Finnic loans, cf. *tūlas ‘spear for night-time fishing’, *kīli- ‘gadfly’, *nīci ‘heddle’. 3.3.1.4 ?*ẹ̄ → *e̮i ~ *ī; *ō → *ou ~ *ū A very interesting case as regards vocalism is PF *he̮imo ~ *haimo ‘tribe, kin’.67 Here one finds reflexes of a diphthong *e̮i throughout all of Finnic except in Livonian and South Estonian, where we instead find *ai (cf. Li. aim, Leivu aim). The following words show a similar pattern, showing *ei in “Core Finnic”, and *ai elsewhere (Kallio 2014: 159): – F heinä, E hein ~ Li. āina, Võ. hain ‘hay’ (~ Lt. šiẽnas) – K dial. leinä, E lein ‘grief, sorrow’ ~ Seto lainalinõ ‘sorrowful’ – F leipä, E leib ~ Li. (Salaca) laib* (Winkler/Pajusalu 2009: 107), Leivu laib ‘bread’ (~ Lt. dial. kliẽpas, ON hleifr) 67 (Liukkonen 1999: 55–56, taken over by Junttila in prep.), which lacks the required semantic specialization. For a more detailed account of this problem, see now Jakob forthc. d. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 66 chapter 3 – F reikä ‘hole’, E dial. (insular) reig ‘wound’ ~ Võ. raig ‘scab’ – F reisi, E reis ~ Võ. arch. raiź ‘thigh’ (~ Lt. ríetas) – F seinä, E sein ~ Li. sāina, Võ. sain ‘wall’ (~ Lt. síena) – F seiväs, E teivas ~ Li. tāibaz, Võ. saivas ‘post, stake’ (~ Lt. stíebas) – F seiso-, E seisa- ~ Võ. saisa- ‘stand’ It is remarkable that the majority of these examples have Baltic comparanda.68 Only the last is inherited from Proto-Uralic, where the cognates (e.g. Sá. N čuožžut, Eastern Mansi tuńś- ‘stand’) point to PU *saŋća-.69 Therefore, it is normally assumed that the e-vocalism shown by ‘core Finnic’ is an innovation, and that Livonian and Võro preserve an archaism (Thomsen 1890: 101– 102; Koivulehto 1979b: 140). Since inherited *aj is normally preserved as such throughout Finnic,70 one must then speak of a “sporadic development” (Kallio 2014: 159). In view of the systematic distribution of the reflexes, it is likewise unattractive to assume multiple layers of independent loanwords (e.g. Uotila 1983: 7–8; Viitso 1998: 12). The lack of any clear conditioning factors (cf. Kallio 2018: 258– 259),71 instead rather suggests that we should reconstruct two diphthongs for Proto-Finnic, which I will provisionally notate *ai (> F ai, Võ. ai) and *?i (> F ei, Võ. ai). Of course, the diphthong *?i must somehow be part of the Proto-Finnic phonemic system, and so the number of options is limited. If we consider the 68 69 70 71 Both *reikä ~ *raika and *leinä ~ *laina have been derived from Baltic, too. Liukkonen (1973: 17–25; cf. Sammallahti 1998: 127; SSA iii: 60) compares the former with Lt. riẽkti ‘slice (e.g. bread); plough for the first time’, riekė ̃ ‘slice’, but this is nothing more than a (semantically weak) root etymology. Nirvi (1964: 153–154) has derived the latter from Lt. klíenas, Lv. kliêns², kliẽns ‘thin, lean’, but this again requires unsubstantiated assumptions with regard to semantics (van Linde 2001: 291–293). See Kallio (2007: 231–232; 2012: 35–36). Pystynen (2014a) rejects this reconstruction and prefers *sańća- (thus also Sammallahti 1988: 549); however, *ńć does not normally develop into *js in Finnic, instead simply becoming *s, cf. *osa ‘part, share’ (< *ońća), *kusi ‘urine’ (< *kuńćə). At the same time, *-ŋć- (> -ŋś-) > *-js- would be a typologically similar development to *-ŋs- > *-ws- found in F jousi ‘bow’ (< *joŋsə). I wonder whether such a PU reconstruction could also explain the difference between Khanty *ᴧāńć- ~ *ᴧī̮ńć- (Vakh li̮ńť-, Kazym ᴧɔńś-) ‘put, set’ (< *saŋća-) and *kus- (Vakh-Vasjugan kŏs-, Kazym χŏs-) ‘urinate’ (< *kuńćə-). On this differently, see now Pystynen apud Живлов (2023: 144). For example, F aivot, Vp. aivod, Võ. aivõq ‘brain’ (< *ajŋə, UEW 5); F kaiva-, E kaeva-, Li. kōva- (< *kauva- < *kaiva-; Kallio 2016: 55) ‘dig’ (< *kajwa-, UEW 116–117); F aita, E aed, Võ. aid ‘fence’ (< *ajta, Aikio 2014b: 1–2). Kallio’s own solution seems to be to assume a residual Baltic ‘substratal tendency’ to confuse *ei and *ai, but this is clearly anachronistic, not to mention that it is precisely Võro and Livonian, which have been subject to the most persistent Baltic substrate influence, that have preserved the supposedly more archaic form. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic → finnic loans table 3 (*ī) *äi 67 i-diphthongs in Early Proto-Finnic *ui *oi *ai based on kallio 2018 i-diphthongs reconstructed for Early Proto-Finnic by Kallio (2018), the result is rather interesting (see Table 3, above). Three possible diphthongs appear to be missing: *ei, *e̮i and *üi. It is attractive to assume that one of these corresponds to our diphthong *?i. Our choice is narrowed down the fact that Estonian and Votic show a partial back-vocalic inflection for this group of words, cf. E leibu, Vt. leipoi part.pl. < *-oita < *-a- j-ta (Kallio 2014: 159). As this can hardly be analogical, it is a compelling argument in favour of original back vocalism, but not necessarily in favour of an original *ai. I would therefore like to suggest the Proto-Finnic reconstruction *e̮i.72 In this case, we can assume a regular fronting *e̮ > *e in “Core Finnic” conditioned by the following palatal resonant, triggering an automatic shift to front harmony (i.e. *le̮jpa > *leipä). If we generalize this sound law to any tautosyllabic palatal, we could also explain the fronting of *me̮cca ‘forest’ (Li. mõtsā, Võ. mõts) to *meccä (F metsä, E mets) in “Core Finnic” (Santeri Junttila p.c. March 2022), as *c must have still remained a palatal consonant in Early Proto-Finnic. At the same time, we can assume a regular lowering *e̮i > *ai in Livonian and South Estonian (cf. Viitso 1978: 95–97).73 72 73 The relevance of forms like Sá. N suoidni ‘hay’ < *šajna, need not be overstated, as it is possible that the Sámi loans were adopted independently (see 4.1, particularly p. 124 onwards). The change *a > *e̮ in the word for ‘stand’ could be explained as a raising due to the influence of the following palatal cluster *saŋća- > *saŋ́ śa- > *se̮jsa- (cf. Ravila 1935: 32, fn. 1; Viitso 1978: 97). Although ad hoc, attributing a unique change to a unique environment is better than assuming a sporadic change with no conditioning factors. For more discussion of these points, see Jakob forthc. d. The diphthong *e̮i seems only to be found in evidently late words like *le̮ikka- ‘to cut’ (cf. Kallio 2018: 260; even here we find E dial. leika-), *pe̮ippoi (> F peippo ‘finch’, Vt. põippõ ‘chick’). An exceptional case is F leivo, Võ. lõiv ‘lark’ (< *le̮ivo), normally taken as a Germanic loan (LÄGLOS ii: 190–191), although Schrijver (1997: 309) considers the possibility of a parallel substrate borrowing. It is possible that the preservation of *e̮i is due to the stem-vowel *o, cf. Võ. hõim (but Leivu aim ← Livonian?; cf. Pajusalu, Krikmann & Winkler 2009: 293–294; Jakob forthc. d.). This would not only explain põippõ, but also the Votic back-vocalic forms sõiso- (in NE Estonian dial. also sõisa-) ‘to stand’, õimo ‘kin, relatives’. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 68 chapter 3 According to Kallio (2018: 262) *ei has arisen secondarily in later ProtoFinnic due to contraction, cf. F seimi, Võ. seiḿ ‘manger’ < *sewi-mi.74 As a similar contraction took place in the loanword PF *oinas ‘ram’ (← Baltic *awinas), there seems to be no chronological objection to the assumption that the Baltic loans predated the emergence of Late Proto-Finnic *ei. The absence of *ei in Early Proto-Finnic may be explained by the development of earlier *ej into *ij (*pīmä ‘milk’ < *pejmä, of Iranian origin,75 cf. Av. paēman- ‘mother’s milk’, Holopainen 2019: 178–180). If the above account is correct, the Baltic loans must have postdated this change. There are two Baltic loans, however, which are argued to have predated the change *ej > *ij (cf. Toivonen 1917: 27–28; Kallio 2008a: 273): – F liika, E liig ‘surplus, extra’ ~ Lt. liẽkas, Lv. lìeks ‘surplus’76 – F tiine, Vp. tineh, Li. (Salaca) tīn ~ Žem. dienì nom.sg.f. ‘pregnant (of animals)’ (Lõo 1911: 86; Kalima 1936: 169 with “?”). If we assume a specifically East Baltic source for the Finnic loans, then the most natural solution is to assume that *ī in these cases is a direct substitution of East Baltic *ẹ̄. Phonologically, such a substitution would not be unexpected given the absence of long *ē in Early Proto-Finnic (see p. 61). In this case, one might imagine a chronological difference, with *e̮i representing an earlier, still diphthongal, pronunciation of Baltic *ẹ̄ (Liukkonen 1973). However, this is not strictly necessary, as the substitution *ē → *e̮i in the absence of a corresponding long monophthong is also quite conceivable. The same substitution is found for Swedish /ē/ in recent loanwords in Finnish, where inherited *ē has developed into /ie/, e.g. F kreivi ‘count’ ← Sw. greve /greːvɛ/ (Thomsen 1870: 56–57; Būga 1908: 23–24).77 This analysis seems to be confirmed by the substitutions of Baltic *ō, which is not diphthongal in origin. Here, we also find two Finnic equivalents: *ū (in 74 75 76 77 In that case, Võro hõim would then need not have been borrowed from North Estonian (cf. Kallio 2021: 125). Apparent examples of *ei often show irregularities. For ‘manger’, some languages show reflexes of *soimi instead (> F dial. soimi, E sõim, dial. soime). For the verb *peittä- ‘cover, hide’, containing the causative suffix *-ttA-, South Estonian pi̬it̬ ä- appears rather to suggest *peettä-. The substitution PU *e for Iranian *a is more or less regular in the position adjacent to a palatal, cf. *sejtə ‘bridge’ (← *saitu-, cf. YAv. haētu- ‘dam’), *rećmä ‘rope’ (← *raćman-, cf. further Chapter 4, fn. 6). Note that, morphologically, Baltic *lẹ̄kas ‘surplus’ more probably reflects an earlier *laikas (= Gr. λοιπός ‘left over’). This might explain the substitution of Baltic *ē as *e̮i in some Livonian loanwords from pre-Latvian, provided these postdated the Livonian raising *ē̮ > *ī̮ (> Courland ȭ, Salaca ǖ; Kallio 2016: 49); compare Li. kȭidaz ‘weaver’s reed’ (← Lv. šķìets, Lt. skiẽtas), lȭiga ‘surplus’ (← Lv. lìeks, Lt. liẽkas). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic → finnic loans 69 *lūta ‘broom’) and *ou (in *routa ‘frozen earth’), which can only be understood as two alternate substitution strategies for a foreign phoneme *ō; compare similarly F housut ‘trousers’ (← Sw. arch. hosor ‘leggings’), ruusu ‘rose’ (← Sw. ros); Thomsen 1870: 51.78 The parallelism between *ẹ̄ → *?i and *ō → *ou is, incidentally, another argument in favour of interpreting *?i as *e̮i. Against the direct substitution *ẹ̄ → *ī, however, speaks Finnic *tīneh ‘pregnant (of animals)’. As already noted by Kalima (1936: 169), the absence of the change *ti > *ci in Finnic can only be understood if *ī in this word is of secondary origin. Therefore, one has assumed an earlier *tejnəš (Koivulehto 1972: 627–628; Liukkonen 1999: 144; Aikio 2014c: 90–91). The assumption that the change *ej > *ij postdated the assibilation *ti > *ci is potentially problematic. The following facts would speak against this chronology: – The Baltic loans *routa and *torvi evidently postdated the symetrical change *ow > *uw (cf. Pystynen 2018: 53).79 At the same time, the Baltic loans predated the assibilation (see 3.3.2). – If Finnic *?i should be interpreted as *e̮i, the example F reisi, Võ. arch. raiź (< ?*re̮ici) would show assibilation but lack the raising to *ī. In my opinion, there is only one possible Uralic reconstruction which could safely account for Finnic *tīneh, namely *tüjnəš. The change *üj > *ij would run parallel to the established change *iw > *üw witnessed in F syvä, Võ. süvä ‘deep’ < PU *tiwä (Aikio 2015b: 9).80 Importantly, the aforementioned example immediately confirms the suggested chronology, as the initial sibilant in F syvä implies an intermediate stage *civä, whereby the change *iw > *üw must have post-dated the assibilation of *t. As PU *e and *ü merged in Mari, this new 78 79 80 For *routa, Junttila (2016b: 226) prefers an original *graudā, reconstructed on the basis of R гру́ да ‘mass, heap’, Pl. gruda, Sln. grúda ‘clod (of earth)’. However, this is hardly necessary from a phonetic point of view, and using an actually attested Baltic form as a source is of course preferable; note that the Slavic words are probably unrelated to Baltic *grôdas (see Villanueva Svensson 2015: 315). Compare PF *ūtin ‘mosquito curtain’ < PU *owdəm(ə) (= Eastern Mansi åml; Komi (Permjak) e̮n, pointing to *o(-ə), see Aikio in prep. 81–82); PF *tūli ‘wind’ < PU *towlə (cf. Ma. W tul ‘storm’, Komi te̮v ‘wind’; see Aikio 2012b: 243); PF *kūsi ‘spruce’ < PU *kowsə (cf. Komi-Permjak ke̮z, Northern Mansi χɔwt, North Sámi guossa ‘spruce’; Collinder 1960: 407; Живлов 2014: 139). On the principle of symmetry in sound changes, see now Jakob forthc. d. The change also has a potential parallel in Vp. silöi, E siil, Li. tsīl ̦ ‘hedgehog’ (< PF *sīli < ?*ćüjələ), cf. the cognates Ma. W šülə and Hungarian sün, older szül ‘hedgehog’, which suggest a rounded first-syllable vowel (Aikio in prep. 127). The reconstruction of PF *kǖ ‘adder’ is also too uncertain for it to constitute a counter-example (< *küü ?< *küjü ?< PU *kejəw, cf. Md. E dial. kijov ‘snake’; see Pystynen 2017). For more discussion, and on the other possible exceptions, see Jakob forthc. d. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 70 chapter 3 reconstruction still allows for an equation with Ma. E tüž, dial. tü.üž ‘pregnant (of animals)’ (E. Itkonen 1953: 183; Aikio 2014c: 90–91), but seems to speak against the comparison with Baltic. Note that the existence of Baltic loanwords in Mari is itself questionable (see 4.3). In conclusion, despite the opposite conclusion of Kallio (2008a: 273), the Finnic reflexes of Baltic *ẹ̄ and *ō can be explained with the assumption of a monophthongal pronunciation in Baltic,81 and therefore an East Baltic origin of the loanwords (for a discussion, see 3.3.4). This, incidentally, can be seen as an argument for a Baltic origin of PF *he̮ina ‘hay’ and *se̮ina ‘wall’, even though the corresponding Baltic words do not have reliable Indo-European cognates beyond Slavic (cf. OCS сѣно ‘grass, hay’;82 стѣна ‘wall’; on the latter, see also pp. 219–220). 3.3.1.5 *eu → *e̮u In view of the etymology PF *re̮una ~ Lt. briaunà ‘edge’, several scholars (Būga 1908: 42; Kalima 1936: 75) have argued that the Baltic loanwords predated the change *eu > *jau.83 While it is often assumed that this is a common BalticSlavic change (Kortlandt 1989a: 48; Matasović 2008: 105), it does not appear to have been shared by Prussian (Levin 1974: 5, fn. 4; Derksen 2010: 38),84 and 81 82 83 84 That we find a diphthong in *paime̮n ‘shepherd’ (~ Lt. piemuõ) need not be an issue. It is known that Baltic *ẹ̄ was a conditioned development. If, for instance, it only arose under stress (cf. Hirt 1892: 37–40; Kortlandt 1977: 323), the word for ‘shepherd’ would have ́ exhibited an alternation, viz. nom.sg. *pệmōn, gen.sg. *pâimenés. It is possible that the allomorph *pâi- was generalized in the dialect which donated the form to Finnic (note that the Finnic form must in any case be from an oblique form, see further 3.3.3.3). In the case of *häici ‘flower’, a form with *âi is actually attested in Lt. dial. žáidas ‘flower’. A remaining question is how this conclusion can be reconciled with the evidence that the Gothic loans in East Baltic predated the monophthongization (see pp. 38–40). Guus Kroonen (p.c. August 2022) suggests Du. heen ‘upright sedge, Carex stricta’ as a possible cognate, noting that the plant is used as animal fodder. The other example, F leuka, E lõug, Li. lȭga ‘chin’ ~ Lt. liaukà ‘gland’ is highly doubtful for semantic reasons (cf. Nieminen 1945: 45; Junttila 2016b: 222–223, whose alternative does not fare much better). The Elbing Vocabulary consistently shows ⟨eu⟩. Note that the glide in Pr. E piuclan ‘sickle’ was not adopted analogically from the full-grade (Arumaa 1964: 87), but is instead probably from inherited *-i̯- (see Hackstein 1992). The Third Catechism offers very little evidence: for *jau clearly speaks iaukint ‘üben’ (= Lt. jaukìnti ‘tame, train’, OCS оучити ‘teach’). On the other hand, the pret. driāudai ‘furen (sie) an’ beside imp.pl. draudieiti seems to show a similar pattern of ‘breaking’ under stress otherwise observed only in e-diphthongs (cf. tiēnstwei ‘reytzen’ beside imp.pl. tenseiti */tenséiti/ ‘reitzet’, etwiērpt ‘vergeben’ beside imp.sg. etwerpeis */etwerpéis/ ‘verlasse’, cf. Kortlandt 1998: 124), and would imply an earlier *driēud- : *dreud-V́ -. This interpretation would be supported by Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic → finnic loans 71 the idea that the diphthong *eu may have been preserved in Baltic until quite recently remains plausible (Kallio 2008a: 274). Thus, Nieminen’s chronological concerns (1945: 53–55) can be disregarded. A valid criticism of Nieminen (1945: 43–45) concerns the Baltic reconstruction: it is indeed true that the general hesitation between /Cr/ and /Crʲ/ in Lithuanian dialects (Zinkevičius 1966: 153–156) makes the Lithuanian anlaut non-probative. On the other hand, Lv. braũna ‘shed skin, scale; husk’ (cf. ME i: 327), which would support Nieminen’s reconstruction *braunā-, is semantically remote and better kept separate (LEW 57; ALEW 147). Therefore, I see no particular reason to doubt the Baltic etymology for Finnic *re̮una. From a phonological perspective, a Baltic loan predating the change *eu > *jau seems preferable to the direct substitution *rʲau- → *re̮u- (Kulonen 1988). 3.3.1.6 Non-initial syllables In Early Proto-Finnic, the vowel contrasts in non-initial syllables were very limited (cf. Kallio 2008b: 269). In fact, it seems possible that only the archiphonemes *A and *E existed at the time of the Baltic loans, and that later *i and *u can be interpreted instead as *Ej and *Ew (Kallio 2012: 31–32). This explains the adoption of both *i and *u as *E, cf. *kärmes (← *kermis), *anke̮rjas (< *angurjas; cf. Kallio apud Junttila 2015a: 19). The phoneme *o in non-initial syllables may have synchronically still been *aw. This would explain *jo ‘already’ (< *jaw ← Baltic *jau), which as a prosodically unstressed particle may show developments typical of unstressed syllables (similarly E dial. ju, showing the development *-o > -u). Baltic *ō, outside of initial syllables, seems to have been substituted as *a, as in *ne̮pat ‘nephew’ and possibly *se̮sar ‘sister’ (see p. 82).85 3.3.2 Consonantism Compared to the Slavic loans (see Kalima 1956), the Baltic loans predated several Early Proto-Finnic developments affecting the consonants: namely *š > *h (e.g. *haljas ‘green’ ← Baltic *žaljas), *ti > *ci (e.g. *silta ‘bridge’ ← Baltic *til̂ta-) and *tj > *cc (in *me̮cca ‘forest’ ← Baltic *medja-),86 as well as the metathesis 85 86 pievſſen acc.sg. ‘pine’ in the Trace of Crete (Lemeškin 2014: 142; in our interpretation: */piēusen/), provided this is correctly read (differently see Kaukienė/Jakulytė 2015: 46– 47). If this is the case, iaukint must be understood as an East Baltic loanword. There are, however, a couple of examples which show non-initial *i and *u in Baltic loanwords, namely *pe̮lu(t) ‘straw chaff’ (adopted as *pe̮le̮w ~ *pe̮lə̑w?), and *oinas ‘ram’ (adopted as *owe̮jnas ~ *owə̑jnas?). Koivulehto 1986 (cf. also 1979a: 290, fn.) discusses a couple of convincing parallels among the Germanic loanwords: F otsa ‘forehead’, E ots ‘end, front; forehead’, Li. vȱntsa ‘forehead’ (← *anþja-, cf. ON enni ‘forehead’) and probably F maltsa, E malts ‘orache’, Li. mõltsõz Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 72 chapter 3 *-wR- > *-Rw- (e.g. *torvi ‘horn’ ← Baltic *taurē) and the development *-ln- > *-ll- (e.g. *villa ‘wool’ ← Baltic *vil̂nâ). A few aspects deserve a more detailed discussion: 3.3.2.1 *š → *h; *s → *s It seems that Finnic speakers were able to reliably distinguish the two Baltic sibilants, consistently substituting Baltic *š with *š. The only apparent exception to this rule is PF *hanhi ‘goose’, but this can be explained as the result of a rather trivial assimilation. In fact, P. Kallio points out to me (p.c. March 2023) that there are no old words with the combination *h–s in Proto-Finnic, so that the assimilation might even be treated as regular. The final *h (> -∅) in F herne ‘pea’, käärme ‘snake’ can be of analogical origin. Due to the change *s > *h between unstressed vowels, stems in *-es and *-eh are indistinguishable in most oblique cases, e.g. gen.sg. *-ehen. Such a vacillation is also known in inherited words, e.g. F kaarne ‘raven’ beside Vp. dial. karni̮š ‘crow’, Li. kārnaz ‘raven, crow’ (= Sá. N gáranas, Komi ki̮rni̮š ‘raven’ ?< PU *karnəš ~ *kärnəš). The exact dating of the change *š > *h is difficult, but the sibilant pronunciation must have been preserved until after the arrival of Finnic speakers in Fennoscandia. A layer of older Germanic loans show the substitution *s → *h, e.g. F ahjo ‘furnace, forge’ ← *asjō, cf. Sw. ässja, OHG essa ‘furnace, forge’ (LÄGLOS i: 5–6); keihäs ‘spear’ ← *gaizas, cf. ON geirr; cf. also Koivulehto 1984: 193–195. Furthermore, the earliest Sámi loans from Finnic still show *š for Finnic *š (cf. Sá. N vašši ‘hatred’ ~ F viha ‘hatred’ (< PU *wiša; Aikio 2006a: 41)). Juho Pystynen (2016) has presented an argument which could show that this sound law even post-dated Proto-Finnic. The word haah́ ‘goose’ in some peripheral South Estonian dialects (Seto, Lutsi, Kraasna) apparently shows a development *Vn > *V̄ before *š, which would seem to parallel the common South Finnic change *Vn > V̄ before *s (cf. E maasikas, Võ. maaśk ‘strawberry’ < *mansikka; Kallio 2014: 162). This might suggest that South Estonian originally preserved a sibilant *š longer than the rest of Finnic, and the change to *h only diffused into this dialect area at a later date. This remains highly tentative, however, especially since the equally peripheral Leivu and Kraasna vahn ‘old’ (= F vanha < PU *wanša, on which see 4.4) would seem to contradict this sound law.87 87 ‘goosefoot, orache’ (vocalism after mõltsi ‘green’? Kettunen 1938: 222) (← *maldjō-, cf. OSw. mäld; Ritter apud LÄGLOS ii: 248), although admittedly the latter is of obscure origin (cf. Kroonen 2013: 351). Note also Sá. N fihčču, K võhč ‘(seal’s) flipper’ (< *fiččō ← Norse *fitjo < Germanic *fetjō-, cf. ON fit ‘webbed foot; flipper’). In view of this, it is perhaps preferable to opt for Pystynen’s alternative account that South Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic → finnic loans 73 Nevertheless, this observation might explain the occurrence of a couple of exclusively South Estonian loanwords which have undergone the sound change *š > *h, in particular Võ. pahr, (Hargla) parh ‘boar’ ~ Lt. par̃šas ‘piglet, castrated boar’ (= Lat. porcus ‘pig’) (Kalima 1936: 145) and perhaps Võ. eherüs, Mulgi eerus ‘trout’ ~ Lt. ešerỹs ‘perch’ (Ojansuu 1921: 5–6). Still, it remains possible that these indeed belong to the earliest layer of loanwords, and were merely lost after South Estonian split off from the rest of Finnic. 3.3.2.2 Initial *c- and *stKallio (2007: 235, 241–242; 2014: 157) has argued in favour of reconstructing a phoneme *c for Proto-Finnic on the basis of the South Estonian evidence. While there indeed do appear to be some compelling examples of South Estonian -dś (sporadically) reflecting Proto-Finnic *-ci (< *-tə), the status of this phoneme in initial position is less certain. Kallio’s only example is the word for ‘pig’: F sika, E siga ~ Võ. tsiga, yet this word’s etymology is uncertain, the traditional comparison with Mordvin *tuvə (> E tuvo, M tuva) ‘pig’ (UEW 796) being phonologically irregular (Aikio 2015a: 46). Therefore, it cannot be proven that the South Estonian ts- goes back to an earlier *c-.88 Moreover, in all other cases where an initial *c- would be expected on etymological grounds, we find s- in South Estonian: – F sinä, E sina = Võ. sina ‘you (sg.)’ (cf. Kallio 2007: 242) < *tinä (UEW 539) – F syvä = Võ. süvä ‘deep’ < *tiwä (UEW 525–526) — According to VMS, the word is practically limited to South Estonian and adjacent Tartu dialects, while in North Estonian, it is only found as a relic in the western periphery. – F sitkeä, E sitke, Võ. sikkõ ‘tough, durable’ ~ Sá. N dađgat ‘firm (of body parts)’ (Sammallahti 1999: 74–75) — As the Võro term shows regular *tk > kk, a loan from North Estonian is improbable. – F silta, E sild, Võ. sild ‘bridge’ ← Baltic *til̂ta — -Attested throughout all of South Estonian, including the language islands. In the Yhteissuomalainen sanasto (YSuS) online database,89 Kallio has adduced several other examples of *c- based on correspondences between initial *tsin Võro and affricates in Karelian and Veps; however, the data encompasses at least seven distinct correspondence patterns between the three languages. Furthermore, the majority of the words are clearly onomatopoeic (e.g. K dial. 88 89 Estonian originally preserved a form *hansi (← Baltic *žans-) which first developed to *hāsi and only then was assimilated to *hāhi. For more arguments for a late dating of Finnic *š → *h, see Pystynen (2023: 355–356). It is possible that Võ. ts- is due to the secondary influence of Latvian cũka ‘pig’. Hosted at https://sanat.csc.fi/wiki/Luokka:Yhteissuomalainen_sanasto. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 74 chapter 3 čivissä ‘rustle’, Võ. tsibisemä ‘whisper’; Vp. čiraita ‘sizzle’, Võ. tsirisemä ‘buzz’) or belong to semantic fields where expressivization could be anticipated (K čirkku ‘small bird’, Võ. tsirk ‘bird’; Olonets čongie, Võ. tsunǵma ‘root about’). Kallio (2007: 242) has himself acknowledged that the Karelian and Veps data are largely irrelevant. It is clear that the above data provide no evidence for a contrast between *sand *c- in initial position in reconstructible Proto-Finnic. Given that it remains possible that *c- > *s- took place in initial position earlier than it did in intervocalic position, I prefer to reconstruct the word for ‘bridge’ as *silta (and not *cilta) for Proto-Finnic. ∵ The Finnic cognates of F seiväs ‘post, stick’ are interesting in two respects. Not only do they show reflexes of the unclear diphthong *?i (probably = *e̮i, see 3.3.1.4), but also an unclear alternation between initial t- and s-. This correspondence has some clear parallels. Compare the following: – F seiväs, Vt. seiväz, Võ. saivas ~ E teivas, Li. tāibaz ‘post, stake’ (~ Lt. stíebas) – F siipi, Vt. siipi, Võ. siib ~ E tiib, Li. tībõz ‘wing’ – F seipi ~ E teib ‘dace’, Li. teib ‘ide’90 (~ Lv. obs. stiepats ‘dace’; see p. 97) – ? F saparo ‘short tail’ ~ Li. tabār ‘tail’ (?~ Lv. dial. stebere ‘tail’)91 The agreement between the words for ‘wing’ and ‘post, stick’ is striking: in both cases, the distribution between s- and t- is almost identical,92 yet not geographically contiguous. If the reason was ‘unstable’ substitution strategies (Kalima 1936: 160) or independent loans (Kallio 2018: 258–259), we should expect a more or less random distribution. Since North Estonian and Livonian do not con90 91 92 Nirvi (1961: 152) and Heikkilä (2013: 583) have adduced E dial. taivikas to support a ProtoFinnic reconstruction with *?i. This form apparently derives from Wiedemann’s dictionary (non vidi; cf. Nuutinen 1987b: 109) where it occurs alongside numerous other variants (among which teivikas and täivikas). P. Kallio (p.c. March 2023) informs me that the form first appears in the second (posthumous) edition of his dictionary and is perhaps the result of a mere printing error; note also that Ariste (1975: 471–472) leaves out the variant with -a-. In any case, none of these forms are likely to be South Estonian, as VMS only records teib and variants across the north and on the islands. Although the variant in -äi- must be somehow secondary, the Proto-Finnic vocalism is quite possibly to be reconstructed as *äi, anyway (cf. Kallio 2018: 261). But note F sapa, E saba(!) ‘tail’ (< *sapa). While saivas is purely South Estonian, siib has apparently spread into neighbouring Central Estonian dialects (see VMS), and is also attested in northeast coastal Estonian, which must probably be attributed to influence from Votic and/or Finnish. Nevertheless, I think it is possible that the distributions of the words were originally identical. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic → finnic loans 75 stitute a subgroup of Finnic languages, the fact that the same two languages happened to ‘reborrow’ the words in question is a remarkable coincidence. Nuutinen (1987c: 61) and Vaba (1997: 177) have suggested that the loanwords were adopted into an already dialectally diverse Proto-Finnic, but the fact that the vocalic reflexes of *e̮i (see 3.3.1.4) straddle the two groups makes this very awkward. Heikkilä (2013: 586–587) has argued that the above evidence would prove that a cluster *st- was licensed in Early Proto-Finnic. While a phonological solution would be welcome, the assumption that a phonotactical restriction against initial consonant clusters could have been relaxed in Early Proto-Finnic before being reinforced again later on, though not impossible, is certainly uncomfortable, especially given that there are no examples of the alleged Finnic *stamong the Germanic loanwords. Nevertheless, in lieu of an alternative solution, I have used the notation *ste̮ipas in this chapter.93 3.3.2.3 Syllable structure At the time of the contacts with Baltic, Finnic still seems to have had a fairly strict maximum syllable structure *CVC. The avoidance of heavy clusters can be observed in *ahtas ‘narrow’ (← *aNštas), *takja ‘dense’ (← *tan̂ kjV-) and *sikla ‘side net in a seine’ (← *tinkla-), which show the regular loss of a nasal before two consonants (cf. PF *kanci ‘cover’ ~ *kat-ta- ‘to cover’; Posti 1953: 56–59).94 In the case of *morcijan ‘bride’, which reconstruction seems to be confirmed by North Karelian moršien (contrast hoaśśa ‘hayrack’ < *hāsja ?← Sw. hässja, LÄGLOS i: 62), an epenthetic vowel appears to have broken up the heavy cluster *-rtj-. In PF *tühjä ‘empty’, a similar cluster *-štj- was resolved to *-šj-. The single example of *vōtna ‘lamb’ is problematic, because it appears that CRtype clusters could not occur after long vowels even in relatively recent loanwords. Although a convincing explanation is lacking, it is potentially relevant that the essive form *vōt-na ‘year ess.sg.’ (> F vuonna ‘in the year’), where the 93 94 Considering that I do not reconstruct an initial *c- in Late Proto-Finnic, one might consider that this is what underlies the correspondence *t- ~ *s-. However, this is chronologically problematic, as in my model *c- (or rather *ć-) would still have been present in Early Proto-Finnic, at the time of the contacts with Baltic. Furthermore, a palatal affricate *ćwould be a phonetically unlikely substitution for a foreign *st- (I thank Santeri Junttila for pointing these issues out to me). Note that Posti does not adduce this Baltic evidence and considers the possibility of a very early dating for this change. An early dating is not excluded by the Baltic evidence, as the sound change may have been productive over a long period. Aikio (2022: 11) even reconstructs this rule for Proto-Uralic. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 76 chapter 3 long vowel might have been restored early on due to analogical pressure from other case forms, would have been formally identical to the word for ‘lamb’. Phonotactic constraints also explain the rarity of geminates, which are typically found as substitutions for voiceless stops in Germanic and Slavic loanwords. In the material collected in 3.2, only two contain a geminate: *rattas ‘wheel’ and *tüttär ‘daughter’, yet only a handful of others (e.g. *ate̮ivo ‘visiting relatives’, *hako ‘branch’ and *ne̮pat ‘nephew’) could have theoretically tolerated a geminate in Early Proto-Finnic. Steinitz (1964: 337) has proposed that the examples with geminates represent a younger layer, while Junttila’s (2017) explanation is that geminates were restricted to disyllabic stems. On the basis of such limited data,95 it is difficult to draw any firm conclusions. A substitution *t → Finnic *tt seems necessary to account for PF *tüttär ‘daughter’ ← Baltic *duktēr. As Proto-Finnic possessed a cluster *-kt- (> Võ. -ttand -ht- elsewhere; Posti 1953: 38–43; Sammallahti 1977: 133; Kallio 2014: 156), it is unclear why we do not find reflexes of *-kt- in this word. Posti (1953: 45) has suggested that the substitution strategy was conditioned by the position of the stress in Baltic,96 but Kallio (2007: 237) sticks to the view that *tüttär shows an “exceptional” development from earlier *tüktäri. Since the evidence does not permit the reconstruction of *-kt- at any stage in Proto-Finnic, it seems necessary to assume that Baltic *dukter- was perceived as */tü(k)ttär/ by Finnic speakers, and realized as *tüttär when subjected to Early Proto-Finnic phonotactics.97 A number of loan etymologies have been proposed in the literature which show a geminate after a heavy syllable, such as the following examples in Thomsen (1890: 74), and Kalima (1936: 53): – F laukki (dial. laukas, laukko; K dial. laukka), Li. laik (< *laukki) ‘blaze; blazefaced animal’ ?← Lt. laũkas, Lv. làuks ‘blaze-faced’ – F pirtti ‘cabin’, Vp. perť ‘house, cottage’ ?← Lt. pirtìs, Lv. pìrts ‘bath-house’ – F dial. kääppä, Vt. tšääppä, E kääbas ‘burial mound’ ?← Lt. kãpas, Lv. kaps ‘grave, burial mound’ 95 96 97 Junttila, of course, uses a larger corpus, but besides two new proposals, the only other example with a geminate he classed as certain (cf. 2017a: 142) is *vakka ‘wooden container’ (on this, as well as *hakkaita-, see pp. 100–102). The “probable” etymology F–K huttu ‘flour porridge’ ← Lt. šùsti (3pret. šùto) ‘stew, steam, sweat’ is a mere root etymology, as a word of appropriate meaning is not attested in Baltic. He also (2017a: 141–142) proposes to compare F obs. (18ᵗʰ century hapax?) lappa ‘thin plate’, Vp. lapak ‘flat, shallow’ with Lt. lãpas ‘leaf’; on F kukka and variants, I refer to his discussion (idem: 134–137). Note in this context that *-kt- > *-tt- is apparently regular in Finnic after unstressed syllables; cf. F sädettä ‘ray, beam part.sg.’ (< *sädek + *-tA). On the Sámi words for ‘daughter’, see Chapter 4, fn. 22. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic → finnic loans 77 While these etymologies are semantically strong, Nikkilä (1982: 254) emphasizes the rarity of such a syllable structure among the Baltic loan material, an observation which has recently found support in an extended treatment by Junttila (2019), who has argued that in fact no such words can be counted among the early Baltic loans. Since Nieminen (1953), pirtti has generally been regarded as a Slavic loan (although see pp. 140–142 for a detailed discussion). As for laukki (etc.), Junttila (2019: 61–62) argues that the relevant forms should be seen as inner-Finnic derivatives of a more primary *lauka, cf. F dial. laukama ‘bare patch (of land, fur)’. With regard to kääppä, Junttila (2017: 133; 2019: 55) has stated that the vowel results from a secondary lengthening. This cannot be accepted: ‘sporadic’ secondary lengthening is only observed under specific conditions (cf. T. Itkonen 1987), and cannot simply be invoked as a license. If Finnic *kǟppä goes back to Early Proto-Finnic, it would have to reflect a trisyllabic preform *käŋəppä or *käxəppä with contraction of the vowel sequence, as in *kāri ‘curve; rib of a boat’ < *ke̮ŋərə (UEW 126; Aikio 2015a: 58).98 The similarity with the Baltic word is therefore probably coincidental. 3.3.3 Declinations 3.3.3.1 Reflection of Baltic *-s The nominative ending *-as of the Baltic a-stems is abundantly reflected in Finnic, e.g. *hampas ‘tooth’, *oinas ‘ram’, *rattas ‘wheel’, *ste̮ipas ‘post’ and perhaps *vapsas ‘wasp’. This is also the case for the adjectives *ahtas ‘narrow’ and *haljas ‘green’, which are evidently based on Baltic masculine nominative singular forms. There are, however, several words which show no trace of *-s. These fall into the following categories: – Words with suffix replacement: *ke̮ltaine̮n ‘yellow’ (fn. 22), *herhiläinen ‘hornet’ (cf. *mehiläinen ‘bee’, *kimalainen ‘bumblebee’, Nieminen 1934: 32–35), E vapsik ‘hornet’, and the i-stems E kurt ‘deaf’, dial. kärv ‘snake’ (cf. Nieminen 1944: 249) – The adjectives *tühjä ‘empty’, *līka ‘surplus’, which could equally be based on feminine or, more probably, neuter (~ predicative) forms. Behind *takja ‘dense’ perhaps lies a acc.sg.f. *tan̂ kjan, or an adverbial form *tan̂ kjai.99 98 99 Whatever reconstruction we use, it is about time we abandon the comparison with Mansi (South) kε̮p, (East) käp ‘small hill’ (still repeated in SSA i: 484; van Linde 2007: 84; Junttila 2017: 133); Mansi *ä implies Proto-Uralic *i, *e or *ü in the initial syllable. Futhermore, this word can hardly be separated from Mansi (South) kε̮mp, (East) kämp in the same sense, thus suggesting a Proto-Mansi *kämp. In fact, the paradigm käp, obl. kämp- is still recorded for the Middle Lozva dialect by Munkácsi/Kálmán (1986: 190). Although the u-stem adjectives in Bretke appear to have been largely unspecified for Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 78 chapter 3 – Nouns: *he̮ina ‘hay’, *hirvi ‘elk’, *häici ‘flower’, *härkä ‘ox’, *me̮cca ‘forest’, *pahla ‘rod, spit’, *sikla ‘side net in a seine’, *silta ‘bridge’, *sōla ‘salt’. It has often been suggested (Thomsen 1890: 112; Būga 1924b: 104, fn. 4; Nieminen 1944: 243–248; Иллич-Свитыч 1963: 42–44; Kortlandt 1993: 47) that the last group of words reflect Balto-Slavic neuters. There is indeed independent support for a neuter in two cases: Baltic *šẹ̄nas ‘hay’ is cognate with the neuter OCS сѣно ‘grass, hay’ and *tinklas ‘net’ contains the neuter instrument suffix *-klas found in Pr. E -clan (piuclan ‘sickle’) and Slavic (e.g. Czech) -dlo. In addition, several scholars (Иллич-Свитыч 1963: 78; Kortlandt 1993: 47; Derksen 2015: 466) have compared *silta ‘bridge’ with the neuter Skt. tīrthá- ‘ford; descent to the water’. However, this is a false comparison, as the Sanskrit word rather belongs with tárati ‘pass, cross; overcome’ < *terh₂- (= Hitt. tarah-ᶻⁱ ‘overcome’); see EWA i: 650; ALEW 1277. It is difficult to evaluate this evidence. First and foremost, there is no independent evidence that the East Baltic nominal neuter ending was originally *-a, matching Slavic, and not *-an, matching Prussian.100 Furthermore, in some cases, the absence of the *-s in Finnic is the only evidence adduced in favour of an original neuter, which runs the risk of circularity. For instance, of the unsuffixed cognates to Lt. par̃šas ‘piglet, castrated boar’, only OHG farah (nom.pl. farhir) is neuter, where we might consider analogy after lamb ‘lamb’, kalb ‘calf’, while OE fearh (pl. fearas) ‘young pig’ and Lat. porcus are masculine. Given this ambiguity, it can hardly be stated (with Иллич-Свитыч 1963: 48) that Võro pahr would prove an original neuter for Baltic.101 Besides, an original masculine gender is secured for East Baltic *žir̂gas ‘male horse’ not only in view of the meaning, but also by Pr. E sirgis.102 100 101 102 gender, even here there is some level of syncretism with the jā̆-stems (Specht 1932: 276– 279), and due to the overall transfer of original u-stems to ja-stems in Latvian, this tendency is probably to be dated to Proto-East-Baltic, at least. On the adverbial suffix *-jai applied to old u-stems in Latvian, see Endzelīns (1923: 461–462). In Lithuanian, the originally pronominal ending -a occurs in predicative adjectives, but this does not imply that it was present in nouns, as it is logical that the ending would have first spread to adjectives; cf. the secondary spread in Pr. iii sta wissa ‘das alles’ (cf. also zuit ‘genug’, with expected apocope; PKEŽ iv: 273). In Prussian, there is also some evidence for this ending in participles in predicate function, e.g. Pr. iii isrankīt postāt ‘erlöset werden’ (Endzelīns 1944: 199). A more extreme case is the mention of the isolated and surely secondary Veps dial. kouvaz (beside usual kauh) ‘ladle, scoop’ as evidence of a vacillation between neuter *kaûša and masculine *kaûšas within Baltic (Иллич-Свитыч 1963: 82; cf. Derksen 2015: 234). More counter-evidence could be retrieved from other widely accepted loan etymologies in *-a, which have not been mentioned here due to the lack of an Indo-European etymology. The proposed Baltic sources of Finnic *ätälä ‘aftermath’ (see 3.5.3) and *kataka ‘juniper’ Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic → finnic loans 79 Moreover, there is an alternative available. Already Thomsen (1890: 112), followed by Kalima (1936: 78), has suggested that Finnic forms lacking *-s may have been abstracted from other Baltic case forms. The most obvious option that comes to mind is the genitive singular *-ā, although one has generally taken the accusative *-an as the most probable basis. Иллич-Свитыч (1963: 42) denies the latter possibility, stating that final *-n would not have been lost in Finnic by sound law, but he does not consider the possibility of analogy. A Finnic form such as *he̮inan (← *šẹ̄nan acc.sg.) could easily have been apprehended as a genitive-accusative singular form, on which basis a new nominative such as *he̮ina could have been backformed. An accusative source form must be assumed at least for PF *morcijan ← Baltic *martjan acc.sg., where the final *-n has not undergone reanalysis as an oblique form in Finnic, but has instead assimilated into the *me-stems, the only common category with nominatives in *-An (e.g. *sütän, *sütäme- ‘heart’). Finnic *hirvi ‘elk’ is not easily explained on the basis of a Baltic a-stem, and may instead reflect a Baltic feminine *širvē as in Lt. vìlkė ‘she-wolf’ to vil̃kas ‘wolf’ (Nieminen 1940: 378); similarly, Finnic *vōhi ‘goat’ may well be from a feminine *âžē as attested in Pr. E wosee (Thomsen 1890: 205). The form *häici ‘flower’ perhaps likewise presupposes a different formation (such as *žaîdē instead of *žaîdas), but this cannot be supported by any Baltic-internal data. The Finnic reflexes *-jas (*anke̮rjas ‘eel’, *haljas ‘green’) and *-es (*hernes ‘pea’, *kirves ‘axe’) seem to echo the dichotomy between the East Baltic nominative *-īs (e.g. *žirnīs > Lt. žìrnis, Lv. zir̃nis ‘pea’) and *-jas (e.g. *žaljas > Lt. žãlias, Lv. zaļš ‘green’; cf. Thomsen 1890: 114–117; Kalima 1936: 79–80). While Lt. ungurỹs ‘eel’ belongs to the former category, Pr. E angurgis might presuppose the existence of an earlier *angurjas.103 Admittedly, the word for ‘pea’ may have arisen from an earlier i-stem (cf. Nieminen 1957: 206; Skardžius 1941: 53), although 103 (see pp. 84–85) have masculine cognates in Prussian, viz. Pr. E attolis, kadegis, and *vakja ‘wedge’ (see p. 50) is masculine in Germanic (cf. OHG weggi). Иллич-Свитыч (1963: 128– 129) assumes an original neuter for *vaha ‘wax’ due to the Germanic evidence (which is not regularly cognate, see pp. 217–218), but the evidence he adduces for accent paradigm (b) — which he would predict in Slavic in the case of an original neuter — is marginal; almost all the evidence points to accent paradigm (c) and therefore an original masculine (cf. Зализняк 1985: 137). The grapheme ⟨g⟩ in the Elbing Vocabulary, in its function as representing a glide, only occurs after stem-final resonants in cases where East Baltic shows nom.sg. *-jas; compare wargien ‘copper’, kragis (read *kargis) ‘army’, saligan ‘green’. Contrast Pr. E tuylis ‘boar’ (~ Lt. kuilỹs), singuris ‘goldfinch’ (?~ Lv. žĩguris ‘sparrow’). Lithuanian ungurỹᵴ may be the result of a general preferrance for *-īs in polysyllabic words (cf. Lt. kumelỹs ‘colt’ beside Lv. kumeļš ‘colt; male horse’). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 80 chapter 3 no trace of this is found in the Baltic languages. A probable original i-stem is *kärmes ‘snake’. It appears that the distribution between *-jas and *-īs is essentially related to syllable weight. While the variant *-īs is clearly the productive type and occurs with after all syllable structures, almost all nouns with a nominative *-jas which can be reconstructed back to Proto-East-Baltic have a light first syllable. The distribution thus corresponds more or less exactly to that of Gothic -jis and -eis (Sievers’ law; see Sommer 1914: 242 and passim):104 – *karjas ‘war, army’ (> Lt. obs. kãrias, Lv. kaŗš) = Pr. E *kargis, Go. harjis ‘army’ – *keljas ‘way’ (> Lt. kẽlias, Lv. ceļš) – *kraujas = */kravjas/ ‘blood’ (> Lt. kraũjas) = Pr. iii krawia ‘blood’, Skt. kravyá‘bloody’ – *medjas ‘forest’ (> Lt. mẽdžias, Lv. mežs) = Pr. E median ‘wood’, Go. midjis, Skt. mádhya- ‘middle’ – *teljas ‘calf’ (> Lt. tẽlias, Lv. teļš) – *svetjas ‘guest’ (> Lt. svẽčias, Lv. svešs) – *varjas ‘copper’ (> Lt. obs. vãrias, Lv. vaŗš) = Pr. wargien In view of this distribution, it is likely that *-jas and *-īs both reflect the same proto-form (i.e. *-ios).105 Usually, one has assumed that *-īs went through an intermediate stage *-ijas (Sommer 1914: 227), and proof of this has been seen in Estonian dial. takijas ‘burdock’ ← Baltic *dagīs (Stang 1966: 190; Zinkevičius 1980: 217–218; Kortlandt 1977: 324; 2018: 182). However, the analysis of the Finnic form is somewhat problematic. First of all, it appears that Finnic *-ja- and *-ija- were likewise in complementary distribution, whereby the disyllabic reflex was automatic after a heavy syllable (Ritter 1977; see *morcijan ‘bride’, discussed above). The preserved -kin Estonian takjas, dial. takijas ‘burdock’ shows we are dealing with an original 104 105 Lt. -ias has become somewhat productive in adjectives (particularly after dentals, apparently to avoid consonant alternations such as t : č, occurring in u-stems?). Nevertheless, a similar tendency can be observed here as well; note Lt. šlãpias, Lv. slapjš ‘wet’, Lt. žãlias, Lv. zaļš ‘green’, Lt. naũjas, Lv. naujš ‘new’ (= Go. niujis). Forms such as mẽdis ‘tree’ beside Lt. dial. mẽdžias ‘forest’ and Žem. svetỹs ‘guest’ beside svẽčias must result from analogy (Būga RR ii: 509). Taking *-īs from *-iHo- per Hill 2016: 214 is unnecessary, and moreover, the contraction *-ijV- > *-jV- in the oblique cases would be irregular (compare uncontracted Lt. eldijà ‘dugout canoe’ which corresponds exactly to OCS ладии ‘boat’). It seems more likely that a satisfactory solution can be found starting from an model based on syllable weight. At first sight, the correspondence between Lv. âzis, beside gen.sg. âža with Gothic hairdeis (< *-ijas) beside nom.pl. hairdjos is remarkable. For Baltic, we may suggest that the development *-j- > *-ij- only occurred after a heavy syllable and before a short vowel, with later generalization of *-j- in the oblique cases. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 81 baltic → finnic loans geminate, which can also be seen in F dial. takkiainen, Vt. dial. takkiaz, takkiain, K takkis-heinä ‘burdock’. The disyllabic suffix in Finnic *takkijas can thus be viewed as a symptom of the weight of the initial syllable, and does not directly have any bearing on the reconstruction of the Baltic word. The question remains, however, as to why the Finnic word has a geminate in the first place. This would certainly not be expected on the basis of Baltic *-g-, which could only usually be substituted by a single *-k-. The usual explanation (Thomsen 1890: 231; SSA iii: 258) is that the Finnic word has been influenced by the verb *takkista- ‘stick, hinder’; compare OE clīfe, OHG klība ‘burdock’ (< clīfan, klīban ‘adhere, stick’). In this case, the introduction of the stem *takkinto a borrowed form *takjas would have automatically resulted in *takk-ijas due to the aforementioned phonotactic rules. This would allow us to assume Baltic *dagīs is in fact secondary for *dagjas (with the expected suffix variant after a light syllable). On the other hand, the assumption of contamination is never exactly compelling. A possible alternative solution presents itself if we indeed start from *dagijas, namely that the introduction of a geminate in Finnic was necessitated by the inadmissability of the sequence *-ija- after a light syllable. However, we must admit that other examples of Baltic *-īs do not show any evidence of an earlier *-ijas (Sommer 1914: 228; Kalima 1936: 79–80); cf. *hernes ‘pea’ and *kirves ‘axe’ noted at the start of this section. 3.3.3.2 Vocalic stems As touched on above, the Baltic feminine ā-stems were generally adopted in Finnic as a-stems, cf. *ansa ‘loop’, *halla ‘frost’, *karva ‘(animal) hair’, *re̮una ‘edge’, *tapa ‘way, custom’, *villa ‘wool’, *lūta ‘broom’, possibly *takla ‘tinder’. On the other hand, there are a few examples which appear as o-stems, cf. *hako (beside *haka) ‘branch’, *he̮imo ‘tribe’, *putro ‘porridge’, *vōro ‘turn’. As for *he̮imo, it has been suggested its stem vowel represents an innerFinnic development. Since the Finnic a- and o-stems coincide in the oblique plural, Nieminen (1934: 19) has suggested that an analogical shift to an o-stem might have been encouraged by the frequent plural use of the word in the sense ‘relative’. He supports this with some alleged traces of the original a-stem in Karelian dial. heima-kunda ‘tribe’ and Võ. dial. hõimanõ ‘relative’ (absent from VMS). A similar account seems to be required to explain ativo ‘visiting relative’ beside the a-stem atima;106 cf. the collocation olla ativoissa ~ atimoissa ‘visit relatives’, lit. ‘to be in guests’ (SMS), which is ambiguous between an a- and 106 With sporadic dialectal *-v- > -m-, cf. Nikkilä 1999: 14–17. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 82 chapter 3 o-stem. A similar explanation could perhaps account for the co-ocurrence of haga and hagu ‘branch, stick’ in Estonian (but cf. Junttila in prep. s.v. *hako). For *he̮imo, another account may be to assume a Baltic accusative form *šeiman as a source, with the otherwise attested substitution *a → *o (note also the labial environment; see p. 63). Differently on the o-stems, see Holopainen, Kuokkala & Junttila (2022: 126–130). The Baltic ē-stems appear to have been adopted either as *A-stems (*mäntä ‘whisk’; perhaps *tūra ‘ice chisel’, cf. Lv. dùre), or as *E-stems (cf. *torvi ‘hunting horn’ ← Baltic *taurē). The latter substitution may also be accounted for by assuming a loan based on the acc.sg. *-en. The category of feminine jā-stems with a nom.sg. *-ī is now only represented by two common Lithuanian nouns, namely patì ‘wife’ and martì ‘son’s wife; bride’. This group was originally a larger, however. In the Elbing Vocabulary, twenty-five words are attested with a nominative in -i (Levin 1973; 1974: 48–49). The only other loanword for which such a nominative in attested is *žansī ‘goose’ (= Pr. E sansy), which form might directly account for Finnic *hanhi. Note that the Finnic i-stems apparently did not exist at the time of the Baltic loans, so that an e-stem would be the closest match (Junttila 2015: 18– 19). However, the word for ‘goose’ shows good evidence for an earlier consonant stem (Nieminen 1957: 200–201; Zinkevičius 1966: 266), and in either scenario, it is difficult to rule out an i-stem accusative *žansin as the basis for borrowing. 3.3.3.3 Consonant stems F obs. nevat ‘nephew, niece’ (whence Sá. N neahpát, S neapede ‘sister’s son or daughter’) apparently belonged to the same inflectional class as F kevät ‘spring’, gen.sg. kevään. It must have been loaned from Baltic *nepōt-s (most probably on the basis of acc.sg. *nepōti-n → gen.sg. *nepate̮-n; see below). For *se̮sar ‘sister’, a bolder solution is required. We could start from an earlier oblique form *sesari-n (compare secondary Lt. sẽserį, after dùkterį ‘daughter’?), matching Skt. svásāraṃ acc.sg. On the other hand, it would also be possible to start from a Baltic nominative singular *sesōr. The loss of final resonants has often been dated very early (Schmalstieg 1983: 152–154; Jasanoff 2002: 34–35), but the Slavic evidence suggests a fairly recent loss (Kortlandt 1979b: 264, 1983; Pronk 2018: 301), and there is no clear argument as to why it should be early in Baltic, either. Note that some forms such as pirmuonìs ‘forebear’ (in Daukša a consonant stem, cf. pirmůnés gen.sg.), schirſchonis ‘hornet’ (Bretke; see ALEW²), and others, look to be built on nominatives in *-ōn, suggesting the loss of final resonants in fact occurred not long before the historical period.107 107 Note, however, that the southern Žemaitian forms entered in the LKŽ under šuõn, vanduõn, piemuõn show a secondary development (Zinkevičius 1966: 196–197). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic → finnic loans 83 The words *sēmen ‘seed’ and *paime̮n ‘shepherd’ are nevertheless better derived from an oblique stem in *-en-. As with the examples described in the previous section, a plausible source may be the accusative *sēmenin, which would be adopted as Early Proto-Finnic *sǟmenen, on which basis a new nominative singular *sǟmeni > *sēmen could have been backformed. In the case of *tüttär ‘daughter’, both a nominative *duktēr and an accusative *dukterin could come into question. 3.3.4 The dialectal origin of the Baltic loans It has been noted that the Baltic loans in Finnic seem in certain aspects more akin to West Baltic than East Baltic (cf. Nieminen 1957: 188; Vaba 1998: 182–184; Kallio 2008a: 275), and it has been suggested the loans were adopted from some other unknown Balto-Slavic dialect (Junttila 2016b), or at least partly adopted from Proto-Balto-Slavic itself (Kallio 1998: 212, 2008a: 265; Koivulehto 1999: 9– 11). The evidence of a particular connection to Prussian is not strong. I have accepted two etymologies where the source form is only found in Prussian: *hirvi ‘elk; deer’ (~ Pr. E sirwis) and *vatnas ‘ploughshare’ (~ Pr. E wagnis); however, these do not represent West Baltic innovations, and may well once have existed in East Baltic, too. Based on the inflection, PF *vōhi ‘goat’ also stands somewhat closer to Prussian (~ Pr. E wosee), but the Prussian form represents an archaism (in East Baltic we find innovative forms: Lt. ožkà < *âž-(i)kā-; Lv. kaza ← R коза́ ), and it cannot be excluded that suitable forms were previously present in East Baltic (cf. Endzelīns 1933: 80–81). A similar argument can be put forward with regard to Pr. E angurgis ~ Lt. ungurỹs ‘eel’. On the other hand, there are some forms which betray innovations that are limited to East Baltic: – The form *ahtas ‘narrow’ reflects an innovative form with the adjective suffix *-stas. This suffix has been somewhat productive in East Baltic (Skardžius 1941: 324–325), but not elsewhere in Balto-Slavic,108 and Slavic continues a more archaic u-stem *ǭzu-ka- (trad. *ǫzъkъ) > OCS ѫзъкъ ‘narrow, tight’ (= Skt. aṃhú-). 108 In Prussian it is found in one form, iii angstainai ‘in the morning’ (cf. Lt. ankstì ‘early’). However, according to Petit (2005), this word is derived from the verb attested in Lt. dial. ant-stóti ‘to begin’, cf. Lt. apstùs ‘abundant’ to ap-stóti ‘surround’, atstùs ‘distant’ to at-stóti ‘(obs.) move away’. The suffix may therefore not be akin to that of Lt. añkštas ‘narrow’. The form iii auckstimiskan ‘Obrigkeit’ is an error: all 8 other attestations show au(c)kt- (PKEŽ i: 113). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 84 chapter 3 – Both *kärmes ‘snake’ and *hernes ‘pea’ reflect semantic shifts which are only attested in East Baltic, cf. the more archaic meanings of Pr. E girmis · made ‘worm, maggot’, and Pr. E syrne, OCS зрьно ‘grain’ (= Lat. grānum). – If my analysis of *līka ‘surplus’ and *he̮ina ‘hay’ as showing direct substitutions for Baltic *ẹ̄ can be upheld, this would be a strong argument in favour of a specifically East Baltic origin for the Finnic loanwords. I therefore consider the most likely source of the Finnic loanwords to be an East Baltic dialect. It still remains probable that the source of the Finnic loanwords was not a direct ancestor of the extant Baltic languages. One possible argument for this is the evidence for a dialectal lowering *i > *e after *š, *ž and before *R (see pp. 64–65). A further indication is the lack of any evidence for early Finnic loanwords in the attested East Baltic languages, as will be argued in the following section. 3.4 Loans from Proto-Finnic to Proto-Baltic? As with loans from Baltic to Finnic, the only reverse loans which could be considered certain are those with regular Uralic cognates, particularly in nonadjacent branches. In most cases where a Finnic to Baltic loan has been suggested, it has been done so on this basis, although for the most part the suggested comparanda predate our modern understanding of Uralic sound changes, and cannot be upheld. An exemplary case is the word for ‘juniper’, attested in Lt. kadagỹs, Lv. dial. kadags, kadȩgs (ME ii: 131), Pr. E kadegis and F kataja, E kadakas. Setälä (1909) connected the Finnic words with a plethora of Uralic material, which led Kalima (1936, cf. p. 12) to exclude the word from his treatment of the Baltic loanwords. The idea that the Baltic word should be derived from Finnic became quite pervasive in the literature, at least among Uralicists (SKES 170; RūķeDraviņa 1955: 404–409; Kiparsky 1959b: 424; Bednarczuk 1976: 48; UEW 165; cf. SSA i: 326–327). Already Collinder (1955: 79) noted that the Finnic vocalism was problematic, and was sceptical towards the etymology; however, UEW still accepted a link with the Sámi and Mansi material. In reality, there are clear phonological problems with all of the Uralic comparanda (see also van Linde 2001: 288–290). Here I present the data along with the possible PU reconstructions:109 109 I omit Sámi *ke̮sŋe̮s (> S gasnges, N gaskkas) ‘juniper’, already considered doubtful by Setälä, and Mari E lume-ɣož, W lə̑me-kož ‘juniper’ (TschWb 352) in which the second element is simply kož ‘spruce’ (UEW 165), cognate with Finnish kuusi ‘spruce’. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic → finnic loans 85 – – – – Finnic *kataka ‘juniper’ < *kaTaka / *ke̮Taka (where *T = *t, *d, *ď or *č) Sámi N goahcci, Sk. kuä´cˈcev ‘conifer needle’ < *koččawa / *kaččəwa Komi dial. kač-pomeľ ‘juniper’ < ?*käččV / *käčkV Mansi (East) köäsp ‘juniper’ ~ (West) kǟšäp < *käČ(k)V- ~ *ke/iČ(k)V- (where *Č = *ć or *č; the Mansi forms do not regularly correspond to each other) The words indeed bear a certain similarity, but they cannot be related by sound law. Only the Komi and Eastern Mansi forms could theoretically be cognate, but since the word is irregular within Mansi, and the development *ä > Komi a (cf. Aikio 2021: 167–168) is somewhat dubious, this is most probably due to chance. In any case, the Finnic word cannot be related to any of the others. Mikkola (1930: 442) presented another argument in support of the word being native to Finnic, namely the suffix *-aka. This suffix is present in other tree names, e.g. F pihlaja, E pihlakas ‘rowan tree’ (< *pićlä ~ *pećlä, UEW 376), F dial. petäjä, Li. piedāg ‘pine’ (< *pečä, UEW 727). According to Mikkola, the suffix stands quite alone in Baltic. However, he overlooked an important example. Lt. mẽdžiaga, which now means ‘material’, is preserved in older texts and Belarusian language islands in the sense ‘tree; wood’. The original form is probably *medaga, cf. Lv. dial. mȩdaga ‘timber’, while mẽdžiaga shows the influence of the root word mẽdis (gen.sg. mẽdžio) ‘tree; wood’. Since, in each case, the suffix has clearly been added within Finnic (being absent from the other Uralic comparanda),110 one may ask whether this ‘tree suffix’ *-aka was actually imported from Baltic (or from somewhere else). Another word for which a Finnic → Baltic loan is often assumed is Lt. šẽškas, Lv. sȩsks ‘polecat’ ~ K dial. (Olonets) hiähky, Vp. hähk ‘mink’ (Wichmann 1911: 253; Kalima 1936: 102–103; Kiparsky 1949: 46–47, cf. Kiparsky 1972; Mägiste 1959: 171; ALEW 1179). This was the only loanword of this type positively assessed by Junttila (2015a: 27), who stated “the sound correspondences between the Uralic words are flawless”. However, this is clearly not the case.111 Mari E šaške, W šäškə ‘mink; otter’ reconstructs to PMa. *šäškə, while PMa. *ä has no regular origin and is not usually found in inherited words (E. Itkonen 1953: 203–207; for a more detailed discussion of the Mari word, see p. 143). The Samoyed comparanda, Tym Selkup 110 111 In the case of *pihlaka, the unsuffixed form is widely preserved: Vp. pihľ (gen.sg. -än), Vt. (Цветков) pihl-puu, E dial. (insular) pihl, Võ. pihl. Junttila still defended the Baltic origin with the argument “there are no less than three possible Baltic derivational explanations for Lith. šeškas”. This would rather speak for the opposite: if scholars cannot agree on the origin of the Baltic word, then probably none of the proposals are fully satisfactory. This is indeed the case: all proposals mentioned present semantic and phonological issues, cf. ALEW loc. cit. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 86 chapter 3 tȫt, Kamas ća’n (= ⟨ťśa͕’ǹ̥ ⟩, Donner/Joki 76) ‘otter’, are justifiably rejected by Aikio (2015a: 45): Selkup *ȫ could perhaps reflect Proto-Samoyed *oj, judging by Taz Selkup tȫti̮- ‘vomit’ (< *tojtə̑ apud Janhunen 1977: 164–165), but cannot be squared with the other Uralic data, nor can such a reconstruction even account for the Kamas form. The addition of Mator tit ‘otter’ to this cognate set by Helimski (1997a: 362) only complicates matters, as the development *ä > Mator -i- is only supported by dubious examples (cf. idem: 99). The invalidity of these cognates was later also admitted by Junttila (in prep. s.v. hähkä). Bednarczuk (1976) has suggested that a whole host of other comparisons represent loans from Uralic into Baltic. As Junttila (2015a) has already written an extensive article criticizing Bednarczuk’s views, and it seems that his conclusions can generally be upheld,112 I will limit myself to the examples which have plausible cognates in other West Uralic branches: (a) ‘lake’. F järvi, E järv ‘lake’ ~ Lt. jáura ‘boggy soil which cracks and dries out in the summer’ (LKŽ); cf. Sá. N jávri, Sk. jäu´rr (< *jāvrē); Md. E eŕke, M äŕʿkä (< *ärkə; *-kə is a diminutive suffix); Ma. E jer, W jär ‘lake’ (< *jer, cf. Aikio 2014b: 135–137) — This loan was first suggested by Būga (1908: 95; 1922: 238–241), although it was not until its independent discovery by Nuutinen (1989) that it received widespread acceptance among Uralicists (Sammallahti 1998: 249; van Linde 2007: 45–46; Junttila 2012: 281; Aikio 2012a: 107). Most reference works (SKES 132; UEW 633; SSA i: 259) have considered järvi to be a native Uralic word. Indeed, a reconstruction *jäwrä (e.g. Sammallahti 1998: 249) can account for most of the data. The metathesis *wr > *rv in Finnic is regular (cf. Koivulehto 1979a: 279).113 The loss of the initial glide in Mordvin is paralleled by Md. E ej, M äj (< *jäŋə) ‘ice’ and E ezńe, M äźńä ‘joint’ (< *jäsən), cf. Bartens (1999: 46).114 The loss of *w in Mordvin is probably paralleled by Md. M dial. (Penza) śeńi ‘a kind of fish, ?ide’ (< *sewnə ~ *säwnə, UEW 437– 438), while the same development can potentially be posited for Mari, cf. tić ‘full’ (< *täwdə). 112 113 114 I would like to point out that the claim that “a Finnic two-syllable a-stem cannot be dated [to] PU if it has a long vowel in the first syllable” (2015: 20) is accurate only for pre-ProtoFinnic, but not for reconstructible Proto-Finnic, in which long vowels can occur in such an environment if they result from contraction, as in e.g. F pyörä ‘wheel’ < *pi/eŋärä (cf. also Plöger 1982). Prior to this, the standard reconstruction was *järwä, but the assumed metathesis in Sámi would be ad hoc. Bartens claims that the initial glide in Moksha dial. (Penza) jäŕʿkä ‘lake’, jäj ‘ice’, (etc.) shows the preservation of *j, but it is rather a secondary prothetic glide as proven by its appearance in words with no etymological *j: cf. Md. M dial. jäľ ‘hem’ (< PU *älä), jäľďä ‘mare’ (cf. Sá. N áldu ‘reindeer cow’). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic → finnic loans 87 The only irregularity is the stem vowel: while Sámi and probably Mari point to *ä–ä, Finnic unequivocally suggests *ä–ə (Aikio 2015a: 41).115 Despite this irregularity, Ante Aikio (in a discussion forum) has recently suggested the revival of Wichmann’s (1902: 165) old comparison with Samoyed *jörä ‘deep’ (> Tundra Nenets joŕa, Taz Selkup kori̮, Alatalo 2004: 327; cf. Janhunen 1977: 47; reconstruction given per A. Aikio). If this comparison is correct, then the word can certainly not be a Baltic loanword in Uralic, although some details admittedly need to be worked out. The question now is whether a Finnic → Baltic loan can be proposed (cf. Senn 1943: 953; Bednarczuk 1976: 48). In my opinion, we must probably answer here in the negative, primarily for semantic reasons. In East Lithuania, whence the majority of the attestations in the LKŽ derive, jáura clearly refers specifically to a kind of boggy, infertile soil that dries out and hardens in the summer. The meaning seems to have broadened to ‘bog’ in Žemaitia, but nowhere does the word refer to a water body. Therefore, a Finnic origin is semantically unattractive. I would also question whether this word really can be compared with Lt. jū́ra, Lv. jũra, Pr. E luriay */jūrjai/ ‘sea’ (as in Trautmann 1923: 335, etc.). From a semantic perspective, ON aurr ‘mud, mire’ seems a closer match.116 Lt. jū́ra ‘sea’, while corresponding with the Uralic forms semantically, cannot be compared formally; moreover, if it is related to Arm. ǰowr ‘water’ (Meillet 1920: 251–252; Olsen 1999: 787),117 this would effectively exclude a Finnic origin. (b) ‘leather (strap)’. F hihna, E dial. ihn, Li. (Kettunen) nī’n̦ ‘leather strap or belt’ ~ Lt. šikšnà, Lv. siksna ‘untanned leather; leather strap or belt’; cf. Sá. S sesnie ‘untanned hide left to moult’, L sassne ‘tanned reindeer leather’ (< *se̮snē);118 Md. E kšna, M šna (< *(šə)šna) ‘worked leather; leather strap’; Ma. 115 116 117 118 The expected Finnic *jarvi would appear to be found in Vt. jarvi and Li. jǭra; however, Salaca Livonian järu seems to prove a Proto-Livonian *järru < *järvi (Grünthal 2012: 313; Kallio 2016: 46); compare likewise Salaca jämde, but Courland Livonian ja’mdõ ‘thick’ (< *jämetä). Also, Votic jarvi (dial. järvi, cf. VKS: 306) must be recent in view of Krevinian järvi (Kettunen 1930: 125–126, cf. the 17ᵗʰ century toponym Järfwenkylä). True, aurr and jū́ra are often combined under a single etymology (e.g. IEW 78–81), which would appear to be supported by OE ēar ‘sea’. However, it still remains difficult to explain the initial glide in the Baltic form (see the following footnote). The outcome of initial *i̯- in Armenian remains controversial (see Martirosyan 2008: 706– 707 with lit.; Olsen/Thorsø 2022: 203–204), but this etymology seems quite compelling to me. It is preferable to the comparison of the Baltic term with Skt. vā́r ‘water’, Lat. ūrīna ‘urine’ (e.g. Derksen 2015: 215), as this leaves the Baltic *j- unaccounted for; an analogical *j- from the full-grade, postdating *eu > *jau- is hardly possible for Prussian at least, since the latter development does not appear to have occurred there (see fn. 84). The Eastern Sámi languages (Sá. i šišne, Sk. še´šnn) reflect an irregular form *šišnē. Accord- Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 88 chapter 3 E šüštö, W šəštə (< *šü̆štə) ‘tanned leather (used for harnesses); harness, strap of such leather’ (Thomsen 1890: 223; Kalima 1936: 101). Aikio (2009: 151) stresses that the correspondences within Uralic are irregular (cf. already UEW 786). He therefore assumes that the Mordvin and Mari words represent independent loans from Baltic. Grünthal (2012: 318) agrees, stating that the expected Mordvin cognate of Finnish hihna would be *šokšna. However, the loss of pre-Mordvin high-vowels in unstressed syllables is a welldescribed phenomenon (Надькин 1988: 7); cf. similarly E dial. kšta-, M dial. šta‘wash’ ~ F huuhto- ‘rinse, wash’ (< *šušta-?) and Md. E šta (dial. kšta), M dial. šta ~ Ma. W šəštə ‘wax’.119 The initial kš- (< *č-) in Erzya appears to be regular (cf. even Md. E dial. gžniva ‘stubble’ ← R dial. жни́ во, Juho Pystynen p.c. October 2021). The Mari form, on the other hand, is probably indeed irregular, as the expected reflex of Proto-Uralic vowel combination *i(-a) in Mari is *ŭ (Aikio 2014a: 156). As with the word for ‘lake’, the irregularities here are quite modest. Due to the existence of apparent cognates in West Uralic, the direction of loaning has occasionally been questioned (Mikkola 1930: 440–441; Mägiste 1959: 171; Bednarczuk 1976: 53; cf. Karulis 1992 ii: 180). Indeed, the Baltic word does not have an acceptable etymology (ALEW 1183; cf. Holopainen 2019: 249, fn. 43),120 so that a loan from Finnic to Baltic would seem more probable than the opposite. On the other hand, the irregular Mari form, nonUralic phonotactics (medial CR-cluster), and the occurrence of the phoneme *š make it unlikely we are dealing with a genuinely inherited word in Uralic (cf. J. Häkkinen 2009: 47; Aikio 2015a: 44–46). It therefore cannot be excluded that the word was adopted into Baltic and the West Uralic languages from some other source (Junttila 2015a: 31). 119 120 ing to Aikio (2009: 151), these are later loans from Finnic. While this is probably true, note that Aikio has later characterized West *s ~ East *š as a common feature of palaeoLaplandic words (2012a: 85); compare Sá. N siekkis ~ K ši´ŋŋg ‘dewclaw’, N sáhppasat ~ K šaahpreš ‘small intestine’. Holopainen (ibid.; cf. also Pystynen 2020a: 83) reconstructs *śišta for this word, but it seems only the Mordvin form might be able to reflect such a preform: we would expect Mari *šŭštə and Komi *śeš(t) instead of the attested śiś ‘candle’ (cf. ež ‘surface’ < PU *iša ‘skin’). If the Komi *-i- shows a special development (or is unrelated), then Udm. śuś ‘wax, honeycomb’ and Mari *šĭštə could perhaps reflect PU *ćeštV vel sim. A promising suggestion has been made in van Sluis et al. (2023: 226) who compare the Baltic words with MW cen, Bret. kenn ‘skin, hide; scales’, providing a Proto-Celtic reconstruction *kisnā-. This Celtic form is traditionally compared instead to ON hinna ‘membrane’ (LEIA C–55 with lit.; IEW 929; Kroonen 2013: 226), which still, however, cannot be ruled out on formal grounds. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 89 baltic → finnic loans (с) ‘alder’. F leppä, E lepp, Li. liepā (< *leppä) ‘alder’ ~ Lt. líepa, Lv. liẽpa; Pr. TC leipen acc.sg. ‘lime tree’; cf. Sá. N leaibi, S liejpie ‘(grey) alder’ (< *leajpē; Sk. leä´pˈp, K lie´hp ← Finnic); Md. E ľepe, M ľepä (< *lepə) ‘alder’ (Sammallahti 1977: 139) — To my knowledge, a Finnic → Baltic loan has never been suggested, although the Uralic words have traditionally been treated as cognates (e.g. UEW 689). Sammallahti’s Baltic → Uralic loan etymology was accepted by Koivulehto (1992a: 173–174) and Aikio (2012a: 74) although it has often been qualified as uncertain (Suhonen 1988: 611; SSA ii: 64–65; Häkkinen 2004: 595; van Linde 2007: 107–109). The Uralic words do not show regular sound correspondences, as has long been recognized (E. Itkonen 1946: 306 attributes the irregularities to “dem allgemein bei den Baumnamen zu beobachtenden lautlichen Schwanken”). The Mordvin form has been explained as a loanword from Finnic (Sammallahti 1977: 139; Aikio 2012a: 108). This would explain the irregular vocalism, but the existence of Finnic loanwords in Mordvin requires further substantiation. On the surface, the Mordvin forms imply *lippä or *lüppä, while Sámi suggests *leipä. Sammallahti assumes that Finnic and Sámi borrowed the word from Baltic independently, and that Finnic *leppä “was better suited to the sound system”. This is rather a strange claim, since we know that Baltic *ẹ̄ is regularly substituted by Finnic *e̮i or *ī in loanwords, as discussed in 3.3.1.4, cf. *he̮ina ‘hay’ ← Baltic *šẹ̄na-, while the substitution PF *e ← PB *ẹ̄ is completely unparalleled. Another issue with assuming independent Baltic loans is the semantics. The Uralic words all refer to the ‘alder’, while in Baltic, the word means ‘lime tree’. As noted by van Linde (2007: 109), these trees are not very similar to each other, so if a semantic shift can be assumed at all, it would be difficult to imagine it occurring twice. Grünthal’s (2012: 321) proposal to assume a third independent borrowing into Mordvin exacerbates the issue. Aside from equivalents in Slavic (R ли́ па, Slk. lipa, SCr. lȉpa ‘lime tree’), the Baltic word has no other Indo-European cognates. The traditional etymology comparing Lt. lìpti ‘to stick’ (Trautmann 1923: 155; REW ii: 44; Smoczyński 2018: 697) fails to explain the acute attested throughout Balto-Slavic (Kortlandt 1985: 121; Derksen 2008: 279; ALEW 669). Thus, if there is any relationship between the West Uralic words for ‘alder’ and the Balto-Slavic word for ‘lime tree’, then it would have to be indirect. This seems a fairly decent candidate for a shared substrate word (cf. Matasović forthc.), although in view of the difference in meaning, the possibility remains that the similarity is coincidental. ∵ Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 90 chapter 3 The only Lithuanian words for which a Finnic origin can be said to have gained general acceptance are late loanwords mediated through Latvian (cf. Thomsen 1890: 68–71). The most widely accepted example is Lt. laĩvas, Lv. laĩva ‘boat’ ← Finnic *laiva (> F laiva, E laev, Li. lǭja) (Mikkola 1930: 443; Kalima 1936: 129; LEW 335; Smoczyński 2018: 660). As noted by Junttila (2015a: 24), the accentual relationship between Lithuanian and Latvian implies a post-Proto-Baltic diffusion. The direction of borrowing was confirmed by the discovery of a convincing Germanic etymology, cf. ON poet. fley ‘ship’ (< *flauja-, Koivulehto 1970; LÄGLOS ii: 159–160; SSA ii: 39). In a similar semantic field, note Lt. bùrė, Lv. buŗa ‘sail’, which Kalima (1936: 148) considered to be most probably from Finnic *purjeh (> F purje, E puri, Li. pūŗaz; cf. also Mikkola 1930; Bednarczuk 1976: 47; SSA ii: 435). In an extended treatment, Nieminen (1955) has argued that the Lithuanian word was borrowed from Latvian, and that the word is indeed a Finnic loan (cf. LEW 65; Smoczyński 2018: 165). Incidentally, Koivulehto (1970: 182, fn. 27) has suggested a Germanic origin here, too (← Norse *buri- > ON byrr ‘sailing wind, favourable wind’). The same route was taken by Lt. dial. aĩrė, áirė (Būga 1924a: 24; LKŽ kartoteka) ‘oar’ ← Lv. aĩris, dial. aĩre ‘oar’, ultimately from Germanic *airō-, cf. ON ár, OE ār ‘oar’, for which Endzelīns (ME i: 13; Zeps 1962: 100), probably correctly, assumes a Finnic intermediary (F airo, E aer ‘oar’). Another plausible case is Lt. dial. asiaĩ ‘rough horsetail’, Lv. aši (secondary ašķi, cf. ME i: 146–147) ‘horsetail, Equisetum’, which might be analysed as a loan from Finnic *hosja (> F hosia, E osi, Li. vȯžā) ‘(rough) horsetail, Equisetum’; cf. Thomsen (1890: 253). If this etymology is correct, however, it would have to postdate the change *š > *h and therefore cannot be interpreted as contemporaneous with the Proto-Finnic loans from Baltic (Junttila 2015a: 25).121 These words are of little interest for our purposes. As the above discussion has shown, there are no cases in which Finnic can be conclusively shown to have been the donor language into Proto-East-Baltic, even if this cannot always be excluded. It seems quite possible that there are no Finnic loans in ProtoEast-Baltic at all, despite the relatively large number of loans in the opposite direction. While this could imply something about the power balance within 121 Frankel (LEW 797, followed by Smoczyński 2018: 1185) has suggested the same for Žem. skárda ‘sheet metal’. In his opinion, this was borrowed through Latvian skãrds from E kard, Li. kārda (< *karta) ‘sheet metal’ (cf. Endzelīns 1924: 120–121). However, a pan-Baltic distribution is implied by an attestation from Pelesa (Belarus) provided in the LKŽ. In North Finnic, the word appears to be limited to Ingrian, where it might have been borrowed from Votic (cf. Thomsen 1890: 138, fn. 1), and it is possible that this word diffused through South Finnic fairly late. This word must be considered in the context of evidence for metal production in the Baltic region (Būga 1923: 3). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 91 baltic → finnic loans the contact relationship, the most probable analysis appears to be that the Baltic language which donated loanwords to Finnic is not the direct ancestor of any extant Baltic language. This could be attractively interpreted in the context of a Baltic speech community being absorbed by a Finnic one (in other words, a Baltic substrate in Finnic; cf. Kalima 1936: 190), a hypothesis which is also supported by other lexical data (Kallio forthc.; see 3.6). 3.5 Common loans from unknown sources?122 The theory that certain words within Finnic derive from an unknown ‘autochthonous’ or substratal language is chiefly associated with the Estonian linguist Paul Ariste (1962, 1971),123 whose views on the subject seem to be regarded as synonymous with the theory itself (cf. Kendla/Viikberg 2015). Essentially, Ariste observed that words of unknown etymology tended to cluster in certain semantic fields, particularly geographical terminology, “somatic words” (1962: 17) and fish names (1971: 10–11, 1975). As the only criterion for identifying substrate words was the absence of an etymology, it is not surprising that the theory failed to achieve widespread acceptance (Saarikivi 2004: 188): the clustering of etymologically obscure words in particular semantic fields may be a statistical argument in favour of a linguistic substrate (cf. Aikio 2004, Saarikivi 2004), but the suggestion becomes circular when applied on the level of an individual lexeme. Thus, when the Finnic cognates of saari ‘island’ are reduced to a reconstructed Proto-Finnic *sāri, what we are left with is a single, isolated data point — a single witness. In the absence of comparative data, we may speculate that the word is of foreign origin, but this cannot be substantiated with any positive evidence. A proposal built on the absence of an etymology alone is naturally very vulnerable. For instance, Ariste suggested that Estonian aed ‘fence’ was a substrate word (1962: 17), but this has since turned out to have an impeccable cognate in Khanty (Aikio 2014b: 1–2), and there are competing etymologies for many other suggested substrate words, some of which are now widely accepted (Kendla/Viikberg 2015: 143–147; Kallio forthc.). 122 123 This sub-chapter will be published, in a slightly modified form, as Jakob forthc. c. I have unfortunately been unable to access Ariste’s monographic treatment (Keelekontaktid: eesti keele kontakte teiste keeltega. Tallinn: Valgus, 1981), although judging by the discussions in Kendla/Viikberg 2015, it appears most of the relevant material was already discussed in his earlier articles. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 92 chapter 3 Rather little evidence from Finnic has been adduced beyond the lists presented by Ariste. For instance, K. Häkkinen (2004) discusses the possibility of a “proto-European” origin where Ariste had previously suggested it (thus s.v. helmi ‘pearl’, liha ‘meat’, saari ‘island’) but does not expand the corpus, even though many words are otherwise labelled as lacking an etymology and could, at least as far as the semantics are concerned, be decent candidates (hiki ‘sweat’, mahla ‘sap’, tavi ‘teal’, etc.). Support for a substrate loan origin has been furnished in specific cases by internal evidence, such as unusual phonotactics or morphology (J. Häkkinen 2009: 37–38; Aikio 2012a: 84; Живлов 2015), but even suggestions of this kind may be vulnerable and run the risk of circularity. For instance, both J. Häkkinen and Живлов (op. cit.) cite the internal cluster -mm- as evidence of non-Uralic origin, yet Aikio (in prep. 12) has argued in favour of such a cluster in native vocabulary. Furthermore, even if a word’s phonotactics would indeed rule out an inherited origin, we can still not in principle exclude that the word’s source will be later identified in an attested language. Since Ariste, some attempts have been made to elevate the West Uralic substrate theory both on a general theoretical level (Напольских 1990, 1997; Wiik 1992; Helimski 2001), and with reference to new linguistic data (Aikio 2004; 2012a: 80–88; Saarikivi 2004), but it is only in the last decade that we have seen a real surge of interest in the area (cf. Живлов 2015; Kendla/Viikberg 2015; Aikio 2015a: 45–47; Soosaar 2021). These studies show an increased focus on phonological and phonotactic criteria for identifying substrate words. Aikio also identifies cases (and later patterns) of irregular correspondence between Sámi varieties (2004: 14–16; 2012a: 85). This is important, as it allows us to move beyond the “single witness” problem, allowing multiple proto-forms to be treated as independent comparanda in support of a substratal origin. What can be remarked upon is that the results achieved in this area by Uralicists seem to have been largely independent of those achieved by IndoEuropeanists (on which see 5.1). Of the cited authors, only Soosaar draws on any Indo-European evidence previously mentioned in this connection, noting the suggestion that F leivo, E lõoke and OE lāwerce ‘lark’ may be parallel loans from an unknown language (Schrijver 1997: 309).124 Otherwise, Indo-European evid124 Напольских (1990: 129; 1997: 200, fn. 5) does refer to some literature from the first half of the 20ᵗʰ century, namely Feist’s theory of a lexical substrate in Germanic and Pokorny’s theory of a non-IE substrate in Celtic. Kallio (1997: 126–128) can be considered responsible for bringing the American school of thought to the attention of Uralicists (Aikio 2004, Saarikivi 2004), although as discussed in 5.1, this particular branch of research was rather light on specific data. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic → finnic loans 93 ence has rarely figured in the discussion of possible palaeo-Baltic borrowings in Finnic. Besides a brief comment by Junttila (2015a: 31) that certain lexical isoglosses between Finnic and Baltic may represent “parallel borrowings from a shared source, perhaps a lost substrate language”, the potential relevance of the Baltic evidence to this debate has not been recognized. Combining Baltic and Finnic evidence could be a further way to resolve the “single witness” problem, and allow us to substantiate proposals of substrate origin based on positive comparative evidence. However, the Baltic evidence can only be considered an independent witness of a shared substrate word where a direct loan relationship with Finnic can be ruled out. Where a Finnic word can, on phonological grounds, be treated as a Baltic loanword, it cannot constitute independent evidence, and while the possibility that the word was loaned into Finnic and Baltic from a third unattested source remains a theoretical possibility, it cannot be substantiated (compare, for instance, the examples on p. 49, fn. 9). Thus, in order to find reliable evidence for a shared substrate in Finnic and Baltic, which I will refer to here as the “palaeo-Baltic” substrate, we will need to identify words which are clearly related but which cannot be considered direct borrowings from one attested language to the other, thus presupposing the involvement of some third source. In this section, I will try to identify cases in which the Baltic and Finnic evidence complement each other and support the supposition of a palaeo-Baltic lexical layer in both language families. After a case study on fish names, I will attempt to identify phonological criteria which might allow us to distinguish substrate lexemes, and finally present a couple of good candidates. 3.5.1 Fish names Aside from an old inherited term for ‘fish’ (Lt. žuvìs, Lv. zivs = Gr. ἰχθῡ́ς, Arm. jukn ‘fish’), very little of the fishing-related vocabulary in Baltic can be traced even as far as Proto-Balto-Slavic. A common term for ‘eel’ can be reconstructed on the basis of Lt. ungurỹs (→ Finnic *anke̮rjas), Pr. E angurgis and — with divergent suffixal vocalism — R у́ горь, Cz. úhoř, Sln. ugọ́ r ‘eel’.125 Beyond this, just a couple of common Balto-Slavic terms can be cited, each having an obscure ultimate origin.126 This situation can be explained in at least two ways. On the one 125 126 Based on the inherited word for ‘snake’: Lt. angìs ‘adder’, Lv. uôdze ‘viper’, Pl. wąż, Lat. anguis ‘snake’, etc. (LEW 1163). The best example is Lt. šãmas, Lv. sams ~ R сом, Pl. sum, SCr. sȍm ‘wels catfish’; beside this, we find Lt. lýnas, Lv. lĩnis ~ Pr. E linis, R линь (gen.sg. линя́ ), Cz. lín, Sln. lȋnj ‘tench’ (note the mismatch in intonation!). See Pronk (2022: 270). Note my discussions of the words for Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 94 chapter 3 hand, we might simply assume that early Balto-Slavic speakers did not engage much in fishing and did not distinguish many kinds of fish. Alternatively, and more probably, we can assume that an originally richer fishing terminology has largely been replaced, likely the result of changing subsistence practices and language contact. Unlike Baltic, Finnic has inherited a relatively rich range of fishing terms from its linguistic parent. As well as the generic noun F kala, E kala, Li. kalā ‘fish’, there are inherited words for specific fish species (F särki ‘roach’, säyne ‘ide’, E dial. tötkes ‘tench’), and vocabulary related to fish (e.g. F kute- ‘spawn (of a fish)’, suomus ‘scale’) and fishing (F pato ‘fishing weir; dam’); cf. Aikio 2022: 24. It therefore cannot be stated that Finnic has undergone massive lexical replacement in this semantic field to the same extent as Baltic. Nevertheless, loanwords in this semantic domain would be unsurprising: the Baltic Sea represents a particular ecosystem featuring species that would not have been familiar to speakers of Indo-European or Uralic languages before they reached the Baltic coast, such as the whitefish, Baltic herring, Atlantic salmon, and sea mammals like whales and seals. The Latvian ethnologist Pēteris Šmits (see P. Schmidt 1930: 87) already noted that a substantial number of fish names in the region were of unclear origin, which he associated with an ancient autochthonous fishing population. This idea was repeated in Benita Laumane’s monograph on Latvian fish names (1973: 14; cf. Лаумане/Непокупный 1968: 76; Ariste 1975: 468), and the same semantic field has been the focus of a number of devoted studies (Герд 1970, 1981; Ariste 1975; Sausverde 1996). Although most of the fish names mentioned by Šmits are also present in Latvian, the majority of these are transparent loanwords from southern Finnic. On the other hand, a couple of the fish names he cites do have a wider distribution. I will treat these here in more detail: (a) ‘whitefish’. F siika, E siig, Li. sīgõz (< *sīka) ‘whitefish’ ~ Lt. sỹkas; ON síkr (attested in kennings), whence Nw./Sw. sik ‘whitefish’ — Already before Šmits, the word for ‘whitefish’ had been labelled as a possible loanword from ‘an aboriginal people’ by Būga (RR ii: 561). The word also featured among Ariste’s lists of substrate words (1971: 11; 1975: 470–471), and was treated as such in a separate article by Герд (1981: 52).127 The question is whether there is any positive evidence that the word was adopted from a palaeo-Baltic source. 127 ‘ruffe’ (p. 275), ‘salmon’ (pp. 258–259) and ‘sturgeon’ (pp. 218–219, 236–237), which show irregular correspondences between Baltic and Slavic. Janne Saarikivi has made the same suggestion at the 13ᵗʰ Finno-Ugricist Conference in Vienna, August 2022. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic → finnic loans 95 Several sources have treated ON síkr as an inherited cognate of R сиг and Lv. sĩga ‘whitefish’ (Falk/Torp 965; REW ii: 621; de Vries 1962: 475), implying Finnic *sīka was borrowed from Norse. However, the dialectal distribution of the word within Russian clearly favours its interpretation as a Finnic loanword (Kalima 1919: 217; Thörnqvist 1948: 247–248; Герд 1981: 52) and the Latvian word is also generally explained from Finnic, which indeed seems likely (Thomsen 1890: 279; ME iii: 851). Thomsen (loc. cit.) has considered Lithuanian sỹkas a loan from Baltic German Siek ‘whitefish’ (with voiceless /s-/; cf. Kiparsky 1936: 181–182), which is itself probably from Estonian siig (Anderson 1938: 148), and SKES (p. 1013) would even take the Norse word from Finnic, which LÄGLOS (iii: 231) acknowledge as a possibility. As a result, depending on our analysis, all of the evidence can be explained as ultimately deriving from Finnic, or from Norse. In other words, we return to the “single witness” problem, and no positive data can be presented in favour of the substrate hypothesis. In this particular case, the Baltic evidence is furthermore most probably irrelevant to the word’s ultimate origin. Although the word remains without a convincing etymology, that fact alone is insufficient to substantiate a hypothesis of palaeo-Baltic origin. (b) ‘herring’. F silakka, E dial. (rare) silakas ‘Baltic herring; salted herring’ ~ Pr. E sylecke, Lt. sil̃kė, Lv. dial. (?) silce (cited for Rēzekne, see ME iii: 840) ‘herring’ — E silk (gen.sg. silgu) ‘(salted) herring’ and Li. siļk (nom.pl. sīlkõd) ‘herring’ are usually quoted here, but due to the awkward syncope128 and mismatch in stem vowel, a direct equation with F silakka seems phonologically problematic. Most probably, Li. siļk is loaned from Lv. sil̃ķe, ̧ which is itself from Lithuanian (ME iii: 840), but E silk is not well accounted for. In view of the trisyllabic Pr. E sylecke, it is attractive to assume that Lt. sil̃kė has arisen through syncope from *silekē or *silikē (Būga 1916: 143).129 Trautmann (1910: 426) has assumed svarabhakti here, but there is simply no other evidence for such a phenomenon in Prussian.130 This fact also rules out Brückner’s (1877: 131) preform *sildkē and derivation from Slavic.131 Other etymologists 128 129 130 131 Contrast E harakas, dial. arak, Võ. harak, Li. arāgõz (= F harakka) ‘magpie’. Alternatively, we could directly compare Estonian silk and assume a variant *silk-, which may further support the non-IE etymology (see below). Trautmann cites J. Schmidt (1875: 209), but accepts neither of Schmidt’s supposed parallels (gelatynan and salowis, cf. Trautmann 1910: 336, 417). The fact that svarabhakti is reported by Becker (1904: 262–263) to be frequent in Pervalkas (as also in South Kurzeme dialects, Endzelīns 1923: 106; Becker is the source of the ‘Curonian’ form ⟨ſīlĕke⟩ cited by Trautmann) has little bearing on our understanding of a Prussian dialect some 600 years and a hundred miles removed from it. The preform is itself anachronistic, as the R diminutive селёдка must derive from a virtual Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 96 chapter 3 have derived the Baltic words from ON síld, OSw. sīldh ‘herring’ (e.g. Solmsen 1904: 585; Smith 1910: 141; Falk/Torp 966), but this requires an entirely unparalleled and phonetically unexpected development *ld > lk (Smoczyński 2018: 1168). In sum, all existing loan etymologies require unjustified assumptions, and cannot be upheld. Šmits (P. Schmidt 1930: 87) also noted the similarity of the Baltic and Finnic words with ON síld, and assumed they were independently borrowed from a substrate source (cf. similarly Преображенский ii: 274; Būga RR ii: 561; Герд 1980; Кузьменко 2013: 514–515, fn. 4). As lengthening is not regular before *ld (Noreen 1894: 320–322), the long -í- either implies a disyllabic preform *silidor *siled- (cf. Falk/Torp 966; Kroonen 2013: 436) or a metathesis from *sīþlō(Smith 1910: 141; Noreen 1923: 172).132 In favour of the former clearly speak the early loanwords into Sámi (N sallit, L sallet ‘herring’ < *se̮lētē) and Slavic (R сельдь, Pl. śledź ‘herring’ < *silidi-, trad. *sьlьdь).133 The disagreement between Baltic *sile/ik- and Norse *sile/id- would certainly favour the interpretation of these words as parallel loans from an unknown source. The irregularity is reminiscent of that between ON hnot, OHG nuz (< *knud-) ‘nut’ and Lat. nux (?< *knuk-) ‘nut’ discussed by Kroonen (2012: 248) and van Sluis (forthc.). One possible explanation for such a phenomenon could be a word-final neutralization of stops in the source language, such as we find in North Sámi (cf. mádjit, gen.sg. mádjiga ‘beaver’). However, this is merely a typological parallel. Other possible explanations can no doubt be suggested, and as we have no criteria to decide between them, we may limit ourselves to the observation that the correspondence is irregular. Likewise, the Finnic words are not easy to explain as loans from Baltic, primarily because of their back vocalism. Already Mikkola (1903: 28) compared the Finnic and Baltic words, but stated that the direction of loaning is unclear. Since Posti (1962), however, the Finnic words have generally been derived from Middle Swedish *sill-laka (cf. sill-lake 1700) ‘herring brine’ (cf. SSA iii: 180; 132 133 *silid-ikā- or *-ukā- (trad. *sьlьdь/ъka), which should have turned up in Baltic as *silidukē, or the like. Mažiulis (PKEŽ iv: 107) starts with a Baltic preform *sildikē, but in that case, the loss of *d is completely unmotivated. Note the parallels in Ic. bíldur (since 17ᵗʰ c.) ‘lancet, device for bloodletting’ = OHG bīhal ‘axe’ < *bīþla- (cf. EWAhd ii: 36–37 with lit.), and ON sáld ‘sieve’ < *sēþla-, cf. OCS сѣти* ‘sift’ (see Kroonen 2017: 105, fn. 1 and 108, fn. 8). The connection with Du. zeelt ‘tench’, which would support this reconstruction, is uncertain on semantic grounds. For the Slavic reconstruction, cf. Mikkola (1903: 28), Būga (1916: 143), Thörnqvist (1948: 78). I fail to understand the alternative reconstruction *sildi(trad. *sьldь), favoured by REW (ii: 606–607), which ought to have yielded R **солдь, Pl. **słudź(?). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 97 baltic → finnic loans LÄGLOS iii: 237). There are serious problems with this explanation, the most important being the single *-l- in Finnic. If even Sw. sill has been loaned into Finnish with a geminate (cf. F silli), it is difficult to conceive of how sill-lake, where the geminate is further reinforced by a morpheme boundary, could turn up with a singleton /l/. There is no reason to suspect an original geminate would have been shortened in Swedish or Finnic (pace Posti 1962: 285).134 Thus, we are only left with a rescue solution such as the assumption of a contamination with F salakka, E dial. salak ‘bleak (type of fish)’, itself of unclear origin (SSA loc. cit.). We are faced, therefore, with three similar preforms — Baltic *sile/ik-, Norse *sili/ed- and Finnic *silakka — whose relationship cannot adequately be accounted for either by cognancy or by borrowing. I would therefore argue that this is a good candidate for parallel borrowing from a palaeo-Baltic source language. ∵ Some additional terms relating to fishing are shared between Baltic and Finnic and lack a plausible Indo-European etymology. At least the following can be cited: – F seipi, E teib ‘dace’, Li. teib ‘ide’ (?< *stäipi, -e-; see p. 74) ~ Lv. obs. stiepats ‘chub’, i.e. Steepats ‘Alantsbleyer’ (Lange 1773: 325; ME iii: 1079) (Nuutinen 1987b)135 — The Baltic stem *stẹ̄p- has no apparent further etymology (no attempt is made in ME iv: 1079; Laumane 1973: 79 speculates on a connection with Lv. dial. stipt ‘to become rigid’). – F toe, Vt. tõgõ, Li. tǭgõz (< *toke̮h) ‘fishing weir’ ~ Lt. takišỹs, Lv. tacis ‘fishing weir’; Pr. E takes ‘(mill) weir’ (Thomsen 1890: 226)136 — Some connection with Lt. tekėt́ i ‘to flow’ is often assumed (Miklosich 1886: 348; LEW 1052; PKEŽ iv: 181), but the formation has remained problematic (cf. the spec134 135 136 The Swedish compound does not appear to have ever been very frequent, and was probably never fully conventionalized, while the occasional spelling with -ll- in older Finnish sources could be due to Swedish sill. Secondly, the semantics are possible, but awkward, as a two-stage metonymical shift must be assumed from ‘herring brine’ (unattested in Finnic) to ‘salted herring’ (unattested in Swedish), followed, in several languages, by a further generalization to ‘Baltic herring’. However, see Posti (1962: 286) for a possible parallel. Nuutinen (op. cit. 109–110) points out that the suffix -ats has had some productivity in fish names, e.g. dial. šķaunats (ME iv: 22) ‘carp’. The Latvian word is much more easily explained from *tacsis < *tacisīs with syncope than, as often suggested, through reanalysis of a nom.sg. *taciss. Prussian takes must, however, be taken for */takiss/ (= Lt. dial. tãkišas); compare Pr. E crays, */kraiss/ ‘hay’ (= Pr. G kraise ‘hay’, cf, craysewisse ‘a grain tax’, on which see Chapter 7, fn. 13). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 98 chapter 3 ulative analysis as *tak-kiš- with the root of Lt. kìšti ‘to stuff’ in Smoczyński 2018: 1441). There is nothing in these comparisons that would rule out a transmission of the word through Baltic into Finnic,137 meaning that we have no positive evidence for a palaeo-Baltic origin, although there might potentially be some statistical significance if numerous shared fishing terms turn out to be of unclear origin. We may tentatively add the word for ‘salmon’ to this list (Laumane 1973 apud Ariste 1975: 468), whose semantics would make a loanword very probable a priori: – F lohi, E lõhe, lõhi (< *lohi, -e̮-); Sá. N luossa, Sk. luõss (< *luose̮) ‘salmon’ ~ Lt. lašišà, Lv. lasis ‘salmon’ (Thomsen 1890: 194) The Baltic word has further comparanda in Pr. E lalasso */lasasā/, R лосо́ сь, Pl. łosoś, and ON lax, OHG lahs ‘salmon’, which cannot strictly be combined under a shared proto-form. As I suspect that Lv. lasis and Lt. lãšis have resulted from syncope from an earlier *lašišīs, a potential irregularity in the Finnic transmission could be the absence of any reflection of the second *š (the existence of a Proto-Baltic form with syncope is questionable; see the detailed discussion on pp. 258–259). However, this evidence remains rather tenuous and open to interpretation. 3.5.2 Finnic short vowel vs. Baltic long vowel Even if the word for ‘herring’ seems to be a reasonable candidate for a palaeoBaltic substrate word, it would be nice to find some patterns that would help to identify such parallel borrowings in Finnic and Baltic, for example correspondence patterns which do not occur in direct loanwords. In this context, I would like to examine the Baltic vowels *ē and *ā. The usual substitutions we find for Baltic *ē (= *ǣ) and *ā in words with a clear Indo-European pedigree are Finnic *ē and *ō (see 3.3.1). On the other hand, several examples of short *a as a substitution of Baltic long *ā were collected by Koivulehto (1990: 152, 2000: 105–106 and passim; cf. also Kallio 2008a: 207). In his opinion, these loanwords must belong to an earlier layer predating the rounding of Proto-Baltic *ā, a development he assumes to explain the supposedly later substitution with Finnic *ō. However, it has now been shown that Finnic *ō developed from an earlier *ā, and so the innovation took place on the Finnic side (Lehtinen 1967: 150–151; Aikio 2012: 232). As noted by Pystynen (2018: 72–75), this points to the opposite conclu- 137 While in the most certain Baltic loanwords, *o ← *a is only found in the neighbourhood of a labial (cf. p. 63), the data is insufficient to rule out a chance correlation. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 99 baltic → finnic loans sion, namely that the loanwords showing short *a must be later, postdating the raising of Early Proto-Finnic *ā (> *ō) but predating the emergence of a new phoneme *ā: (a) Early (b) Late (c) Post-Baltic īǖ ū īǖ ū ē ō īǖ ū ē ō ǟ ā ǟ ē PB *ǣ → PF ǟ PB *ǣ → PF ä While Pystynen’s account does indeed explain the facts, it seems unattractive to view the raising of original *ǟ and emergence of a new *ǟ as unrelated phenomena. The two developments seem to be interpretable as a push shift caused by the loss of intervocalic *ŋ and *x. The resulting contractions (e.g. *kaŋərə > *ka.ərə > *kāri ‘curve; rib of a boat’) can be seen as having motivated the raising of the earlier low vowels (cf. footnote 60). In this context, it is unnecessary to assume that Proto-Finnic went through a stage in which *ā was absent, as in system (b). If we examine the examples which supposedly show short reflexes of Baltic *ē and *ā, it is notable that none of them have a completely evident IndoEuropean etymology. In five cases, the Baltic word lacks any plausible comparanda entirely: 1. E vähk, gen.sg. vähi, Li. vē’jõz (< *vähi)138 ~ Lt. vėžỹs, Lv. vêzis ‘crayfish’ (Thomsen 1890: 241) — The Baltic word has no clear etymology (cf. LEW 1235–1236; ALEW 1419).139 2. Li. vägāli ‘burbot’ ~ Lt. vėgėlė ̃ 3ᵃ, Lv. vêdzele ‘burbot’ (Thomsen 1890: 77) — Although limited to Livonian, the assumption of a later loan from Žemaitian (Thomsen 1890: 141–142) or Curonian (Endzelin 1914: 102; 138 139 Estonian -k is a secondary excrescent consonant (cf. mähk ‘sapwood’ < *mäihä). The i-stem may indicate a very recent origin (Junttila 2015a: 181), but it could also be secondary (cf. 3.3.3 on E kurt, dial. kärv). The Livonian form appears on paper to suggest something like *vähjes, which could suggest an originally different inflectional type. Salaca Livonian väji* ‘crayfish’ may rather represent a loan from Leivu väi (cf. Pajusalu, Krikmann & Winkler 2009: 293) or Estonian vähi (P. Kallio p.c. February 2022). The connection with NP gazīdan ‘bite, sting’ is formally impossible (Cheung 2007: 117–118) and that with Skt. vāhaka- ‘a kind of insect’ very uncertain (KEWA iii: 198). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 100 3. 4. 5. 140 141 142 143 144 chapter 3 Nieminen 1957: 199) does not help to explain the short first-syllable vowel. The Baltic word has no clear etymology (cf. LEW 1212; ALEW 1392). F apila, dial. apelias (?< *ape̮ila ~ *ape̮lja)140 ~ Lv. âbuõls, cf. Pr. E wobilis ‘clover’ (Thomsen 1890: 156; Kalima 1936: 94 with “?”) — The Baltic forms cannot be separated from Lv. dâbuõls, Lt. dóbilas ‘clover’, with an unclear initial d-. It is generally assumed that the d- was lost due to contamination with the word for ‘apple’ (Lv. âbuõls; Pr. E woble) and/or influence of Lv. ãmulis ‘mistletoe’ (cf. dial. amuols ‘mistletoe; clover, wood sorrel; daisy’; ME i: 235; LEW 99; ALEW 26–27). While Lv. (d)âbuõls does indeed appear to have been influenced by the word for ‘apple’,141 the similarity of Baltic *dâbila- and *âbōla- seems hardly sufficient motivation for the former to have lost its initial stop, which is a typologically unusual development.142 If the word is not of Indo-European origin, the *d- ∞ *∅- alternation might be attributed to the source language(s). A potential parallel is found in the plant name ME doder, MHG toter, totter ‘dodder’ beside Lt. jùdros, Lv. idra, dial. judras (ME ii: 115), Võ. judõr, (Hargla) jutr, Li. ju’ddõr ‘false flax, Camelina’.143 E hakkama ‘begin; grasp’, ?Li. akkõ ‘grasp, catch’ (cf. Junttila 2017a: 131) ~ Lv. sâkt ‘begin’, Lt. šókti ‘jump, spring (into action)’ also dial. ‘start suddenly (esp. of weather phenomena)’ (Vaba 1992: 222; Holopainen/Junttila 2022: 97) — The original meaning is probably ‘jump’: cf. ME sterten ‘jump, spring (up, forth); come suddenly into a state or condition’ > modern start (16th century) ‘begin’. The connection with Gr. κηκίς ‘ooze, viscous liquid (of blood, pitch, fat, etc.)’ (LIV 319; ALEW 1213) is semantically unconvincing. F varhainen, dial. varas, E varane, Võ. varahinõ, Li. varāz, va’rri ‘early’ (< *varas, *varahine̮n);144 Sá. N vuoras ‘old; old man’, Sk. vuõrâs ‘old man; VKS cites Vt. apila only from the botanical notes of Gustav Vilbaste. Perhaps this is a Finnish loan. Note that e.g. Standard Latvian âbuõliņš ‘clover’ synchronically appears as if it is a diminutive of âbuõls ‘apple’. Koivulehto (2000: 107) suggests that the d-forms could instead be secondary, but since he does not provide any explanation for the d-, this cannot be considered a fully-formed hypothesis. The relationship between the Baltic and Võro/Livonian words is unclear (LEW 196). A loanword is conceivable in either direction (cf. Sommer 1914: 197), as well as in a relatively recent timeframe (Junttila 2012: 273). Liukkonen (1999: 152) suggests a semantic shift ‘old’ ≫ ‘long ago’ ≫ ‘early’, citing as a parallel Hungarian rég ‘long ago’ and its derivative régi ‘old’ (but this shows the opposite development). Another possibility could be to start from the sense ‘fully grown, ripe’ (cf. Kildin Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic → finnic loans 101 grown up (e.g. of a reindeer calf)’ (< *vuore̮s) ~ Lt. obs. voras, Pr. iii urs, acc.pl. urans */ūrá-/ ‘old’ (Liukkonen 1999: 151–152) — The Baltic word is isolated. No cognates are suggested by LEW (1274), PKEŽ (iv: 211) or Smoczyński (2018: 1695).145 Even though the latter example has an equivalent in Sámi, the second syllable vowels do not correspond, meaning that no common proto-form can be set up (cf. Liukkonen loc. cit.).146 Likewise, the other examples have comparanda in Slavic, but in two of three cases, the comparison is phonologically irregular, suggesting the words in question postdate Proto-Balto-Slavic: 6. F lapio, dial. lapia, E labidas, Li. lä’bḑi (< *lapita) ‘spade’ ~ Lt. lópeta, Pr. E lopto ‘shovel, spade’ (Thomsen 1890: 197 with “?”) — The Baltic forms are clearly related to OCS лопата ‘(winnowing) shovel’, but the correspondence is irregular.147 Note that if we start from Baltic *lâpetâ, the Finnic second syllable vowel *i is also unexpected, especially if we consider that the suffix *-etA is frequent in Finnic, while *-itA is otherwise unknown (cf. Koivulehto 2000: 110–111).148,149 7. F lava ‘platform, deck’, E lava ‘(sleeping) platform; bench (in a sauna)’, Li. lovā ‘bed; bench (in a sauna)’ (< *lava); Sá. N luovvi, Sk. lue´vv ‘raised platform (for storing meat)’ (< *luovē) ~ Lt. lóva, Lv. lâva ‘bunk (for sleeping); 145 146 147 148 149 Sámi vūras ‘large (of fish)’), with a subsequent development to ‘timely’ as in SCr. dòspijeti ‘ripen, mature; be on time’, and finally to ‘early’. As the word is only attested in older lexical sources, the circumflex given by Trautmann (1910: 127), Fraenkel (LEW 1274), and other authors, does not appear to have any basis (cf. Būga RR ii: 720). The word is essentially limited to Prussian Lithuanian, and may be a Prussianism (cf. Smoczyński 1983: 171, fn. 15), but the derivative vorùšis ‘frail person’ reported from Linkmenys implies a broader distribution. The form ùrupė (rather *ū́rupė?, cf. the river name Ū̃rupiai in Luokė) (= vórupė) ‘old river bed’, cited by Juška (apud LKŽ), is, contra Smoczyński (2018: 695), hardly reliable evidence for ablaut. Could it be a Sembian Prussianism with regular */ūr-/ < *wār-? It is unlikely that Sámi shows suffix replacement. On the contrary, we would expect retention of the suffix *-ēs to have been encouraged by the more usual synonym *poarēs ‘old’ (> Sá. N boaris, Sk. puä´res). There is no indication that the Baltic word represents a derivative with lengthened grade (Fraenkel 339–340; Smoczyński 2018: 724), and the comparison with Lt. lãpas ‘leaf’ is better abandoned. On the other hand, it is possible that a variant with *-i- existed in Baltic, as in Lt. dial. vedigà ‘adze’ (LKA i: 87), mẽdiga ‘material’ (for vedegà, mẽdžiaga), and this might underly Prussian lopto, cf. Pr. E wedigo ‘Carpenter’s axe’, Lv. dial. vȩdga ‘ice chisel’. Koivulehto (2000: 114) also discusses F lapa ‘shoulder blade’, but this is rather an inherited word and cognate with Inari Sámi lyepi and Eastern Mansi lūp ‘shoulder blade’ (Aikio 2015b: 13). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 102 chapter 3 bench in a sauna’ (Wiklund 1896: 45–46; Kalima 1936: 131) — The Baltic word is cognate with R dial. ла́ ва ‘bench; platform for washing clothes’, Cz. dial. lava ‘bench (along a wall)’, but has no further etymology.150 8. F vakka, E vakk ‘oval container made of bark; dry measure’, Li. vakā ‘woven basket; dry measure’ ~ Lt. vókas ‘(eye)lid; woven grain basket’, Lv. vâks ‘lid’, dim. vâcele ‘woven grain basket; dry measure’ (Koivulehto 2000: 114–115) — The Baltic forms must be connected to R вѣ́ко ‘eyelid’, dial. (СРНГ iv: 101) ‘lid of a basket or wooden vessel; grain basket’, Cz. víko ‘lid’; however, the vocalic relationship between the Slavic and Baltic words is irregular.151,152 As a result, we are faced with a situation where all of the Baltic loanwords whose Indo-European background is certain show long reflexes of Baltic *ā and *ǣ in Finnic, which is actually what we should expect in the case of direct loanwords, while all the plausible examples in which Finnic shows short vowels lack an Indo-European etymology, being at best common Balto-Slavic. In this context, we may venture the conclusion that the two different substitution patterns do not represent different chronological layers, as was previously assumed, but rather betray a distinction between direct and indirect contacts. A possible explanation for this could be that a substrate language underlying Baltic had undergone a sound change (such as open-syllable lengthening) which resulted in phonetically long vowels, while a related substrate underlying Finnic retained short reflexes.153 150 151 152 153 Fraenkel (LEW 387) suggests a derivation from the root of Lt. liáutis ‘cease’ (note this verb in the sense ‘abgeschnittet, verstümmelt werden’ appears to be unattested); however, the semantic connection between this verb and ‘raised platform or deck’ is by no means trivial. Furthermore, one would anticipate the palatal onset of the verb to be preserved in such a derivation, as in paliovà ‘break’ < pa-liáuti. The derivational chain set up by Smoczyński (2018: 726), involving an unattested verbal form *lóvyti, involves too many hypothetical stages to be taken seriously. From an o-grade *uoh₁k-, I would anticipate Lt. *úoka-; cf. the discussions in РЭС vi: 196 and Derksen 2015: 509. A potential parallel is the word for ‘turnip’, Lt. rópė ~ R рѣ́па, which is, however, almost certainly of non-IE origin; see p. 237 for a discussion. Md. E vakan ‘vessel, bowl’, as already noted by Paasonen 1896: 36, is hardly from R dial. (СРНГ iv: 9) вага́ н ‘wooden trough’. Contra van Linde’s (2007: 177) claim that *-k- is a usual substitution for foreign *-g-, this substitution actually seems to be highly exceptional. The only generally comparable example listed in Paasonen (1903: 17) is Moksha dial. avkə̑s ‘August’. The Erzya word could instead be seen as cognate with the Finnic word, with a suffix as in Md. kućkan ‘eagle’ < PU *kočka. For more length alterations, compare the examples collected in 7.5.1. A similar example could perhaps be F leppä, E lepp, Li. liepā ‘alder’ ~ Lt. líepa, Lv. liẽpa ‘lime tree’. On this word in detail, see p. 89. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic → finnic loans 103 It must be acknowledged, however, that this theory is to a large extent built on a theoretical postulate (“Proto-Finnic always had a phoneme *ā”) and can be viewed as a potential house of cards. Should a convincing Indo-European etymology be discovered for any one of the Baltic source words, we would be forced to accept a Baltic → Finnic loanword, and with it, the possibility of a substitution *ā → *a. In that case, we would be compelled to accept an alternative solution, such as Pystynen’s chronological one, and we might as well apply that explanation to all of the examples. Thus, although the theory potentially carries more weight than Ariste’s in that it identifies a linguistic pattern in the data, its vulnerability is only exacerbated, as it depends not only on a single word lacking an etymology but on a whole set of words lacking one. 3.5.3 Irregular front vocalism Koivulehto (1971) collects some material which would show Finnic front vowels as substitutes for back vowels in loanwords, but does not concern himself with any explanation of this phenomenon. I will not address the Germanic evidence, which is beyond the scope of my study. As for the Baltic evidence, *tüttär ‘daughter’ and *tühjä ‘empty’ are open to interpretation (see p. 65). Two other frequently cited examples (e.g. Kalima 1936: 66; Koivulehto 1971: 577; Nuutinen 1989: 498) show front and back variants within Finnic: – F rastas, Vt. dial. rassa ‘thrush’, E dial. raastas, Li. rastā ‘starling’ ~ F dial. rästäs, Vt. dial. räsäz, E rästas, Võ. rästäs ‘thrush’154 – F ankerias, E angerjas, Li. aņgõrz ~ K (Olonets) ängeriäs, Vt. (Kukuzzi) ängeriä, E dial. (Vaivara) änger( jas) ‘eel’ In both cases, the front-vocalic form appears to be secondary. This is shown by the lack of clear dialectal patterning: E dial. änger( jas) is rare and marginal, while rästas is attested throughout Estonia (see VMS s.v.). In North Finnic, the fronted variants are in principle infrequent. It is clearly anachronistic to blame these dialectal effects on a borrowing event many centuries prior.155 The transfer of back-vocalic words to front harmony is a typical expressivization mechanism in Finnic (cf. Saukkonen 1962; Nikkilä 2002: 132; Vaba 2011: 749), and both words show other signs of expressivization, e.g. introduction of the 154 155 Compare similarly the bird name F varpunen, E varblane ~ E dial. (E) värb, värblane, Vt. värpo ‘sparrow’, of Slavic origin. Uotila (1986: 213) and Vaba (2011: 749) suggest that the words in question were originally disharmonic, with this discrepancy only being resolved in the individual languages, but it is hardly believable that the violation of vowel harmony was permitted in Early ProtoFinnic only to be reinforced again in Late Proto-Finnic (compare Pystynen 2018: 70–72). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 104 chapter 3 primarily non-native phoneme /č/ in Karelian račoi ‘thrush’ or irregular suffix substitution in F dial. angerva (SMS s.v. ankerias). In this context, we can examine the following case: – Vt. ätälä,156 E ädal (secondary hädal) (→ F dial. ätälä, Ojansuu 1916: 202), Võ. ätäl ‘aftermath’ ~ Lt. atólas, Lv. atãls, Pr. E attolis ‘aftermath’ (Thomsen 1890: 159) — For various etymological analyses, none of which are convincing, cf. ME i: 149; Witcak 2001; Kabašinskaitė/Klingenschmitt 2004: 89–95. See also p. 232. The consistent front vocalism shown in Finnic is difficult to explain starting from the attested Baltic forms. While Li. (Kettunen) a’ddõl ‘aftermath’ does indeed suggest a variant with back vocalism, according to Kettunen (1938: 2), the word should be reconstructed *ate̮la rather than *atala,157 therefore neither representing a back-vocalic equivalent to the Estonian forms, nor being straightforwardly derivable from Latvian (see also Gāters 1953: 155, who offers an unconvincing solution). As a result, this example is not directly comparable with those cited above, where equivalent front and back variants were attested dialectally. Furthermore, there are no other indications of ‘expressivization’ in this word. To resolve this problem, we might suppose that the irregularity is the result of an indirect loanword relationship. There is otherwise possible evidence that the Baltic word was borrowed from a non-IE source in its irregular comparandum in Slavic (see p. 232), although it cannot be entirely excluded that the front vocalism in Finnic is merely secondary, as in the word for ‘thrush’.158 We can also note the word for ‘sleigh’: F reki, E regi, Li. re’ggõz, whose evocalism is unexpected on the basis of Lt. rãgės, Lv. ragus, ragavas pl. ‘sleigh, sledge’. The traditional etymology for Baltic connects these to Lt. rãgas, Lv. rags ‘horn’, based on the “horn-like” shape of the sledge’s runners (thus ME iii: 465, LEW 685; ALEW 964; Smoczyński 2018: 1105). Needless to say, this is merely guesswork, and does not account for the Finnic evidence (cf. Kalima 1936: 66). A Proto-Baltic variant *regē can hardly be posited on the strength 156 157 158 Ojansuu (cf. SSA iii: 499; Junttila 2012: 272) assumes the Votic word was adopted from Estonian, but apparently only because he takes the latter as a late Latvian loan, which is hardly necessitated by the data. Compare Li. vie’ddõl ‘liquid’ (< *vetelä) as against madāl ‘low’ (< *matala). In a footnote, Vaba (loc. cit.) notes a form ⟨addal⟩ from Hupel’s dictionary, but this must be a printing error: the German-Estonian part of the dictionary has ⟨aͤddal⟩ (Hupel 1818 ii: 417). In North Finnic, there is yet another suspiciously similar word: F odelma, Ingrian oelma ‘aftermath’ (< *ote̮lma). The derivation from F ota ‘spear, thorn’ (SSA ii: 258) does not seem particularly convincing. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic → finnic loans 105 of Lv. dial. rȩgavas (which is probably due to a secondary dialectal development, cf. Endzelīns 1923: 36–37) and toponymic evidence (contra Nieminen 1957: 202).159 On balance, while the vocalism of the examples adduced here is indeed problematic and has not yet found a satisfactory solution, the evidence is rather limited. While the proposal of parallel loanwords from a palaeo-Baltic source might provide a possible explanation, it is uncertain whether there are sufficient examples for such a proposal to be justified. We may conclude that the search for phonological criteria to distinguish direct and indirect loanwords between Baltic and Finnic has yielded only modest results. In the following, I will tackle the question from a slightly different perspective, and treat two case studies in detail. 3.5.4 The word for ‘thousand’ First, we will examine the word for ‘thousand’, which is generally accepted to be a Baltic loanword in Finnic (Thomsen 1890: 232–233; Kalima 1936: 170–171; SSA iii: 318). The data are as follows: – F tuhat (obl. tuhante-), E tuhat, Li. tū’ontõ (< *tuhat, obl. *tuhante̮-) ‘thousand’ ~ Lt. tū́kstantis, Lv. tũkstuôtis ‘thousand’ Despite the consensus, it has always been clear that the East Baltic forms do not represent a suitable source as attested. As a result, one has instead operated with a hypothetical Baltic source such as *tūšamti- (Būga 1908: 138; Nieminen 1957: 190; Lühr 1993: 124; Liukkonen 1999: 15),160 a reconstruction based primar159 160 A slightly different issue is posed by F rieska, E rõõsk, Li. rȭskõ ‘fresh, unleavened’, which is compared to Lt. prėś kas in the same sense. These forms can be reconstructed to ProtoFinnic as *rē̮ska, yet such a form would presuppose an Early Proto-Finnic *rǟska, in violation of vowel harmony (Pystynen 2018: 71–72; a similar issue faces Vaba’s derivation of E lõõts from Lv. plẽšas ‘bellows’, on which see Holopainen/Junttila 2022: 64). There are two possible solutions. First, the back vocalism could be secondary, an unusual development which, however, does have a parallel in F mela, E mõla ‘paddle’ (< *melä, cf. Sá. N mealli, Md. M miľä ‘oar’; Kallio 2014: 161). The alternative solution is to assume a younger loan, which would also be supported by the young syllable structure *CV̄ CC- (cf. Junttila 2019: 36). However, none of the other loan evidence can support the existence of Baltic loanwords in Late Proto-Finnic. It is perhaps of relevance that the Baltic word has an irregular cognate in OHG frisc ‘fresh’ (see p. 271), although since the issue with this loanword is mainly chronological, it is uncertain whether the unexpected Finnic vocalism can be explained away by positing a loanword from an unknown source. Kalima (1936: 57, 86–87) sees a parallel for the substitution *kst → *š in F dial. ahingas (?← Estonian, Junttila 2016b: 226), E ahing, Li. a’ņgõz (< *ahinka ~ *ahinkas) ‘fishing spear’ ~ Lt. ãkstinas, HLv. obs. (Bezzenberger 1882: 275) akstyns ‘thorn, goad’ (Thomsen 1890: 157). However, this comparison is best abandoned, as the Finnic stem-final velar is also unexplained (cf. Liukkonen 1999: 15). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 106 chapter 3 ily on the Finnic form, and unsupported by the comparative data. If the alleged Baltic *š reflects IE *ḱ, then it remains to be explained why no trace of this phoneme is found in Go. þusundi, ON þúsund ‘thousand’.161 If we assume *š reflects IE *s with RUKI law, then it remains to be explained why we do not find a RUKI reflex in Slavic (cf. OCS тъıсѫщи ‘thousand’). Moreover, in both scenarios, the actually attested East Baltic data is unexplained. The only way to reconcile the Germanic and Slavic evidence is to reconstruct a medial cluster *-ts-: the *t would be lost in Germanic, and would block the effects of RUKI law in Slavic. From this starting point, there is no room for a Baltic form with *-š-. Instead, the Baltic evidence can only be accounted for by assuming an irregular metathesis to -st-. As a result, Pijnenberg (1989: 104–105) has reconstructed an underlying *tuHt-h₁s-nt-ih₂- (in his notation *tūt-sn̥ t-ī) ‘eine große Quantität bildend’. However, the root *tuHt- (a supposed extension of the root of Skt. tavás- ‘strong, powerful’) lacks external parallels, meaning his semantic reconstruction is ad hoc, and moreover, the Baltic metathesis remains irregular (see also Lühr 1993: 118).162 In view of the problems in reconstructing a common proto-form, Stang (1966: 282; 1972: 49) has suspected that the word for ‘thousand’ is in fact of nonIE origin. Indeed, as discussed in 6.3.2, there are possible parallels for an irregular alternation between *st and *ts, which might be an indication of parallel borrowing.163 If the Indo-European word cannot be analysed as inherited, we 161 162 163 A reconstruction of the type *tuHs-(d)ḱmt- (Bugge 1888: 327; Leumann 1942: 126–128; Kroonen 2013: 554) has usually been suggested based on a notion (in my view misguided) that this word contains the Indo-European word for ‘hundred’. The development of *-s(d)ḱ- to *-s- in Germanic is implausible (Hirt 1896: 343; Pijnenberg 1989: 101; Gorbachov 2006: 8) and not supported by any other evidence. The *m is usually reconstructed on the basis of Pr. iii tūsimtons acc.pl. ‘thousand’, but this, like ON þús-hund ‘thousand’, is more probably a folk-etymological distortion after the word for ‘hundred’ (cf. Lt. šim̃ tas; Hirt 1896: 345–347; Vaillant 1958: 647). As Hirt pointed out, the word-internal *-sḱ- should have given Germanic *-sk- by sound law, so any sequence -sh- must necessarily be of secondary origin. Indeed, we would expect an old *-m- to have been preserved in East Baltic (Stang 1966: 100). A somewhat similar irregularity is seen in the word for ‘wax’, F vaha, E vaha, Li. vǭ’ ‘wax’, which cannot be regularly derived from Baltic *vaškas (> Lt. vãškas, Lv. vasks ‘wax’). Here, as in the examples discussed below, one has assumed the generalization of a weak consonant grade (Thomsen 1890: 76; Kalima 1936: 171). Since the irregular cognate in OHG wahs, ON vax ‘wax’ can be seen as an indication that the Baltic word is of non-Indo-European origin (see pp. 217–218), one may suggest the same for the irregular Finnic form. It must be admitted, however, that the Finnic word could be of Germanic origin, after all: the substitutions Germanic *h → Finnic *k and *s → (*š >) *h are known from other early loanwords, (e.g. *kaltas ‘bank, shore’ ← Gm. *halþaz, cf. ON hallr ‘slope, hill’; PF *kana ‘chicken’ ← Gm. *hanan-, cf. ON hani ‘rooster’; see LÄGLOS ii: 20, 35. On *s → *š, see p. 72), while Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic → finnic loans 107 may suggest that the Finnic word likewise represents a borrowing from a nonIndo-European source, rather than a direct loan from an Indo-European one. Note that there are a couple of other cases where Finnic shows *h as against Baltic *s, neither of which have a watertight IE etymology: – F laiha, E lahja, Li. lajā (< *laiha) ‘thin, lean’ ~ Lt. líesas, Lv. liẽss ‘thin, lean’ — Regarded as etymologically obscure (ALEW 670; Smoczyński 2018: 698– 699).164 – F lahto, Vp. dial. lahk, -on (< *lahto) ‘bird trap’ ~ Lt. slãstai pl., Lv. slasts, usu. slazds ‘trap, snare’ — Etymology uncertain (LEW 827; Smoczyński 2018: 1219). The former has also been explained as showing a reflex of Baltic *š due to RUKI law (Kallio 2008a: 267).165 While it appears likely that the RUKI law must have applied after *u and *i at some point in pre-Proto-Baltic, the more typical attested reflex is -s- (cf. Endzelīns 1911: 29–60; Stang 1966: 99). As the exact chronology of these developments is difficult to establish, it cannot be ruled out that Finnic reflects an earlier Baltic *laiša-. In this context, we might favour the comparison of the Baltic word with OE lǣs, OS lēs adv. ‘less’ (Kroonen 2013: 324) and further with OHG līso ‘mild, soft’, ?Gr. λιαρός ‘mild, warm’ (Osthoff 1910: 325–326; Heidermanns 1993: 370), which I think cannot be ruled out. However, the *h in Finnic *lahto can hardly be blamed on the RUKI law,166 and the irregularity in this word might be compared with that found in the word for ‘thousand’, and assumed to be an indication of shared substrate origin. Still, given that the substitution *s → *h is well known from Germanic loanwords (Koivulehto 1984: 193–195), an alternative way out might be to suggest that the word for ‘thousand’ is of Germanic origin, a solution which has almost never 164 165 166 the development *kš > *h is regular in Finnic (Posti 1953: 7–9), cf. F mehiläinen ‘bee’ (< PU *mekšə, UEW 271). Thus Germanic *wahsa- → pre-PF *wakša > *vaha can be considered quite plausible (contra LÄGLOS iii: 350 with further lit.). Lt. láibas (?→ Lv. dial. laĩbs) ‘thin, lean’ cannot be linked by any known derivational process (contra LEW 329–330; Derksen 2015: 268–269). The traditional explanation has been to assume the generalization of a weak consonant grade (Kalima 1936: 58–59; Posti 1953: 61–62), but such a theory applied to Proto-Finnic is in principle problematic, as the phonologization of consonant gradation postdated ProtoFinnic (see Viitso 1981; Nahkola 1995). Not only that, but *sC-type clusters did not undergo gradation in Proto-Finnic at all (cf. Posti 1953: 9), meaning that such an explanation is excluded for lahto. In any case, *s was only ever weakened to *h between unstressed vowels. Nieminen (1934: 28) has in fact suggested that the RUKI law may be responsible in the case of *lahto by positing a donor form *slakštā- or *slagždā- (cf. Lv. obs. slagzds; ME iii: 912) with an intrusive velar. The dating of a dialectal by-form in Latvian to Proto-Baltic does, however, feel anachronistic. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 108 chapter 3 been suggested. True, the Baltic suffixal syllable *-ant- does come closer to the Finnic data than Germanic *-und- (but see Koivulehto 1981: 193). An obstacle to both the Germanic and Baltic etymologies could be the short *u in Finnic. This has not usually been viewed as a problem, or even remarked upon.167 Such short reflexes have been attributed to the fact that long vowels were originally only possible in e-stems (Koivulehto 1981: 193). However, if such a phonotactic limitation did once exist, there is plenty of evidence the Baltic loanwords postdated it (cf. Plöger 1982: 93). Compare the following etymologies: – F tuulaalla, Vp. dial. tuľhuuda (< *tūlahe̮la-) ‘spear-fish by torchlight’ ~ Lv. dũlis ‘torch for night fishing’ – F tuura, E tuur (< *tūra) ‘ice chisel’ ~ Lv. dial. dùre ‘ice chisel’ – F luuta, E luud, Li. lūdõ (< *lūta) ‘broom’ ~ Lt. šlúota, Lv. sluôta ‘broom’ The substitution of Baltic *ō as Finnic *ū in the last example can only be understood if this loanword predated the raising of early Proto-Finnic *ā to *ō, demonstrating that this must belong to a chronologically earlier period (see above). We might suggest that Finnic *tuhat belongs to an even earlier layer, but this feels ad hoc without other supporting evidence. Aside from ‘thousand’, there is one more possible example of the substitution of *ū as *u among the Baltic loanwords: – F kulo ‘wildfire; last year’s grass’, E kulu, Li. ku’l ‘last year’s grass’ ~ Lv. kũla ‘last year’s grass; old hair of an animal’, Lt. dial. kū́lymas ‘last year’s grass’ Here again, the Baltic source word is of uncertain origin,168 and the direction of loaning has often been declared uncertain (Thomsen 1890: 190; Kalima 1936: 121–122; SKES ii: 234–235). Therefore, there is no solid evidence that would support the substitution *ū → *u among the Baltic loanwords, but even if such a substitution is accepted, we are still left with the awkward Finnic *h. Next, we have to address the words for ‘thousand’ in Mordvin and Mari. While the vocalism in Md. E ťožań, M ťožäń ‘thousand’ seems to match that of Finnic, Mordvin *ť- normally only occurs in words of affective or obscure 167 168 Thomsen (1890: 99) simply remarks that both long and short *u are substituted as short *u in Finnic, while Kalima (1936: 71) passes over the short reflexes in silence (similarly Kallio 2008a: 272). Nieminen (1957: 190) writes dryly: “Das ū der ersten Silbe wurde bei der Entlehnung durch ŭ ersetzt”. The Lithuanian word looks deverbal, which suggests a comparison with West Aukštaitian iš-kūlýti ‘dry up, deteriorate’, yet the latter itself looks be denominal (cf. 3pres. -ija). Nieminen (1934: 26) connects Lv. kàlst (1sg.pret. kàltu) ‘dry out, wither’, but the vocalism and intonation are prohibitive. The further connection with Gr. (Hom.) κήλεος* ‘burning (of fire)’, καίω ‘kindle, set on fire’ (Walde/Pokorny i: 376; ALEW 617; Smoczyński 2018: 625) is formally possible but not compelling. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic → finnic loans 109 origin (Bartens 1999: 46). In view of Mari *tü̆žem (> E tüžem, W təžem) ‘thousand’, we might assume that Mordvin ’o results from a ‘breaking’ from *ü (cf. E. Itkonen 1946: 300–301; Mägiste 1959: 174–175; Keresztes 1986: 170). At first sight, a close parallel would appear to be found in Md. E dial. śokś ‘autumn’ < *sükćə, but the initial palatal in the latter is evidently due to a secondary assimilation from the more usual form sokś, and cannot be associated with the palatal in the word for ‘thousand’. At any rate, Mordvin -ń might be derivable from an earlier *-m, which is strongly supported by the form ťožəm, gen. ťožməń recorded by Paasonen (MdWb 2411–2412) for the Erzya village of Seńkino.169 The result is that the Volgaic forms could possibly go back to a common proto-form *tüžäm(ə), but cannot be compared directly with the Finnic forms. Since a derivation directly from Baltic involves a similar issue with regard to the medial *š and an additional issue by way of the final *-m,170 these forms can be adduced as further support for an unknown source language. To summarize, there are several indications that the word for ‘thousand’ has been loaned independently into the individual Indo-European (and BaltoSlavic) branches, and the Finnic and other Uralic forms cannot be derived either from a common preform, or be explained as direct loanwords from Indo-European sources without accepting a number of awkward and poorly paralleled substitutions. As a result, it would seem that this word cannot be satisfactorily explained without assuming the involvement of an unknown language or unknown languages, and the word might have entered the IndoEuropean and Uralic languages independently from an unattested source. Given the distribution, we are perhaps dealing with a Wanderwort whose trajectory and original source are difficult to identify. However, we might also suggest some kind of connection with the so-called “West Uralic substrate”. In support of this idea, we can note that the phoneme *š has been considered characteristic of West Uralic words showing morphological and phonological irregularities (Живлов 2015; Aikio 2015a: 45–47). On the other hand, as the word is present already in Proto-Germanic, it must have spread into Europe fairly early, and drawing any conclusions on the basis of a single phoneme would be premature. 169 170 The regular outcome of word final *-m is apparently *-n, as shown by the 1sg.pres. ending EM -an (< *-Vm) (Bartens 1999: 50). In other instances, -m has been restored from oblique case forms, e.g. E uďem, M dial. uďəm ‘brain, marrow’ (?< *wVdəm; UEW 572–573). A development *-ńď- > *-ń- occurs in some grammatical morphemes in Erzya dialects (Paasonen 1903: 41), but is not common-Mordvin; therefore, the reconstruction *tušaNtə (Grünthal 2012: 335) cannot be correct. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 110 chapter 3 3.5.5 The word for ‘moss’ Thomsen (1890: 214) compared F sammal, E sammal, Li. sǭmal (< *sammal) ‘moss’ with Lt. sãmanos pl. ‘moss’, but considered the equation questionable. Although Vaba (2011: 757) still labels the comparison as possible, it has rarely featured in discussions of Baltic loanwords, being omitted from Kalima’s treatment (1936). Later on, without reference to the Baltic data, Ariste (1971: 10) labelled the Estonian word as a probable loan from an unknown substrate. The most obvious problem is that the geminate -mm- in Finnic cannot be explained on the basis of the Baltic evidence. A loanword in the opposite direction would in theory be possible, but the very existence of such loanwords has been considered doubtful (see 3.4). At any rate, there is no clear IE etymology (LEW 761; Smoczyński 2018: 1130). However the relationship with Baltic is interpreted, it is clear that the Finnic data cannot be divorced from a group of similar Sámi words referring to various mosses (cf. SSA iii: 151). Not only do none of these correspond regularly to the Finnic word, they also show irregular correspondences within Sámi. As many as four different groups must be distinguished: a) Sá. N seamul ‘spikemoss; house moss’, L sämol ‘(a kind of) peatmoss’ (< *seamōl) b) Sá. I siävŋul ‘a kind of peatmoss’ (< *seavmōl) c) Sá. Sk. sââu´ŋel ‘hairmoss’ (< *se̮vmēl) d) Sá. K sõvŋal ⟨sɛuŋaл⟩ (T.I. Itkonen 1958: 487) ‘hairmoss’ (< *se̮vme̮l) The surface cluster -vŋ- in Eastern Sámi could reflect a number of possible preforms,171 but -vm- seems to be the most suitable compromise with the Western forms. For *-vm- > *-vŋ-; compare Sá. S saajmie ~ I sävŋi, Sk. ⟨saṷ̄ ŋ́̄ ė⟩, K ⟨sà͕ɯ̭̄ŋ͕ᵉ⟩ (T.I. Itkonen 1958: 478; modern Sk. säu´nnj, K saa´vvn) ‘seam’, cf. Ic. saumur ‘seam’ (Kallio 2008b, fn. 3). This is a very interesting case, as the high level of irregularity within Sámi clearly suggests that our word belongs to a relatively recent palaeo-Laplandic substrate layer, entering the individual Sámi dialects independently (cf. Aikio 2004: 14–16; 2012a: 85). On the other hand, the word’s robust presence in Finnic and even as far south as Lithuanian brings the centre of gravity far away from the Arctic Circle. As a possible solution, we could speculate that the word was loaned into palaeo-Laplandic from further south (palaeo-Baltic?), and only from there into Sámi. On the other hand, as Sámi represents a centre of diversity, we might assume an ultimately Laplandic origin, in which case we would have to assume that the word was carried south. Given that we are hardly 171 Other possibilities are *-vŋ-, *-vń- or probably *-mŋ- (Eino Koponen p.c. May 2022). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic → finnic loans 111 dealing with a trade word, this would probably imply an actual (southward) migration, presumably by speakers of another unattested language, took place prior to the arrival of Finnic-speakers in the Baltic region. This could potentially suggest a genetic relationship between palaeo-Laplandic and at least one palaeo-Baltic language. 3.5.5.1 Conclusion In the above, I have attempted to substantiate the hypothesis that a certain proportion of the shared vocabulary between Baltic and Finnic may not represent mutual loanwords, but rather parallel borrowings taken by the two language families from an unattested source. For the most part, evidence adduced in favour of this hypothesis in the past cannot be further substantiated, as it depends primarily on the absence of an etymology. In theory, unusual morphology or phonology could favour a non-native etymology, but it is difficult to use this evidence to support a specifically non-Indo-European source. Nevertheless, in the course of this subchapter I have gathered some material which could provide some concrete linguistic support for the hypothesis. While I have tried to identify substitution patterns which could betray such parallel loanwords, a more robust argument can be built on etymologies for which there are simultaneously multiple indications of palaeo-Baltic origin. In this section, I have discussed three such cases, which I present in the Table 4, overleaf (the pre-forms correspond to the approximate time of Baltic-Finnic contacts). It is interesting to note that the three words point to a rather different contexts of borrowing. The word for ‘moss’ must be connected to the palaeoLaplandic substrate and with some kind of physical migration either into or out of Lapland, but the word for ‘herring’ shows a more localized distribution, and perhaps points to an autochthonous fishing community around the Baltic coast, similar to the one surmised already by Šmits. Finally, the word for ‘thousand’ is widely distributed, and must either be considered an old Wanderwort, or perhaps be associated with a group of other widespread loanwords identified in West Uralic. Although we should hesitate before drawing far-reaching conclusions on the basis of just a handful of words, the overall impression is of a rather complex language contact situation involving multiple donor languages. It seems unlikely that the pre-Indo-European and pre-Uralic languages of north-eastern Europe represented a monolith, and it is probable that multiple source languages contributed to the substratal lexicon of the attested languages. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 112 chapter 3 table 4 Probable shared substrate words in Baltic and Finnic Baltic Finnic Other comparanda Indo-European Uralic ‘herring’ *sile/ik-ē*silakka Gm. *sile/iT‘thousand’ *tûstant-(i)- *tušaNt(ə) Sl. *tū(t)sant-īGm. *tū(t)sn̥ t-ī‘moss’ *saman-ā*sammal Md./Ma. *tüšämSá. *semol Sá. ?*siwmal Some support for this argument could be the words for ‘seal’ in Baltic, Finnic and Sámi, which all appear to derive from different foreign sources: – Lt. rúonis, Lv. ruônis ‘seal’, which is clearly related to, but not regularly cognate with, OIr. rón, Breton reunig ‘seal’ (see pp. 266–267) – F hylje, E hüljes, Li. īlgaz (< *hülkes) ‘seal’, which seems to be connected to, but is hardly loaned from, ON selr, OHG selah (cf. Suolahti 1899: 64) (< *selha-; Iversen/Kroonen 2017: 519)172 – Sá. S nåervie (< *noarvē) ~ Sk. nue´rjj (< *nuorjē) ‘seal’, which are perhaps irregularly connected to Finnish norppa ‘ringed seal’ (Aikio 2004: 15) We could argue that these terms originally referred to different kinds of seal, but there is no indication that this should be the case, as they represent neutral terms in all of the languages where they are attested. On this basis, we might assume that Finnic and Baltic interacted with distinct fishing populations speaking potentially unrelated languages. Such a scenario can certainly not be ruled out, and perhaps more such cases could be identified with further research. As a final note, I would like to point out that the dearth of evidence adduced here cannot be taken as an indication that Finnic and Baltic have been only minimally affected by palaeo-Baltic languages, but simply that very little can be identified. Given that my methodology demands both the survival of the 172 Sadziński/Witczak (2016: 58–59) have additionally compared Norwegian Sámi (19ᵗʰ c.) dullja ‘(a kind of) seal’ (Stockfleth 1852: 694), for which they provide an arbitrary, and entirely erroneous, Proto-Sámi reconstruction *tüľɣa. This Sámi word is confined to older lexical sources, and looks to be an unexplained variant of Sá. N dealljá ‘harp seal’ (< PSá. *tealjā). Any kind of connection with Finnic *hülkes is more or less excluded on phonological grounds. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic → finnic loans 113 word in multiple branches, and the demonstration of identifiable irregularities, we cannot expect a huge amount of data to be available to us. Furthermore, it is naturally more difficult to substantiate a substrate origin based on words in unrelated languages, because apparently irregular substitutions in loanwords can often be accounted for by assuming different chronologies or dialectal differences, while such options are usually unavailable when dealing with exceptionless sound laws. It is merely a matter of fortune that enough material has survived in these three cases to allow us to make a case for a palaeo-Baltic origin. In fact, many more of the suggested Baltic loanwords in Finnic are of unclear ultimate origin, but with the tools currently available to us, this can only serve as a statistical argument. If this area of research continues to be pursued, I am confident that more hard evidence will be uncovered. 3.6 Analysis of contact relationship 3.6.1 Animal husbandry While no Baltic words related to cattle appear to have been loaned into Finnic, it is highly remarkable that two loanwords related to horse breeding seem to surface in Finnic as cattle terminology. Thus, Finnic *e̮hva ‘heifer’ and *härkä ‘ox’ can plausibly be analysed as loanwords from the Baltic words for ‘mare’ and ‘male horse’, respectively. The application of terminology for one domesticate to refer to another is trivial; a parallel can be seen in the adoption of the same Finnish härkä in North Sámi as heargi ‘draught reindeer’. However, as with the Sámi example, such a shift does most probably point to a difference in animal husbandry practices. The 9ᵗʰ century traveller Wulfstan of Hedeby remarked that the Balts consumed mare’s milk and ate the meat of their draught animals (Gimbutas 1963: 25–26).173 The milking of horses was potentially already practiced by early Indo-Europeans, as evidenced by Equus milk peptides identified in the dental calculus of two Yamnaya individuals from the western Steppe (Wilkin et al. 2021: 630). A possible analysis would be to associate the semantic shift from ‘horse’ to ‘cow’ with a transition from horse to cattle as milk animals. Remarkably, Proto-Finnic *lehmä ‘cow’ is the phonetically regular equivalent of the Mordvin word *lišmə (> E ľišme, M ľišmä) ‘horse’ (cf. Ojansuu 1908: 32), which might be understood in a similar mixed Finnic-Baltic cultural context.174 173 174 This tradition appears to have been continued by the Prussians until at least the 15ᵗʰ century, as shown by the gloss aswinan ‘kobilmilch’ in the Elbing Vocabulary (see further Топоров ПЯ i: 135–136). Note that this word is derived from the same Baltic *ešvā- which was loaned into Finnic. As another parallel for such a semantic shift, compare Ket kuʾs ‘cow’ as against Yugh Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 114 chapter 3 In this context, it is worth noting the remarkable absence of horse remains in Baltic Corded Ware material (Piličiauskas 2018: 186). This is typical of the Corded Ware culture in general, where the few extant horse remains belong to local wild populations (Librado et al. 2021). This is problematic to the identification of the Balts with Corded Ware. On the other hand, horse teeth are prevalent in barrow cemeteries from the Late Bronze/Early Iron age associated with the hillfort phenomenon (Merkevičius/Muradian 2016; Аллмяэ et al. 2018: 350; Legzdiņa et al. 2020: 1846). This must indicate a certain cultural significance of horses in the Baltic region, but admittedly does not directly inform us of their domestic status. In the Fatjanovo-derived Djakovo Culture, horse becomes the dominant domestic species during the Iron Age, contemporaneously to many cultural changes in the Eastern Baltic, while osteological evidence points to horse as a primary meat source (Кренке 2019: 43, 58). Most of the loanwords associated with animal husbandry concern sheep and goat. In this domain we can count *vōhi ‘goat’, *oinas ‘ram’, *vōtnas ‘lamb’, *villa ‘wool’, *paime̮n ‘shepherd’, and probably *karva ‘(animal) hair’. The earliest directly dated remains of domestic livestock in the Eastern Baltic date to the Middle Bronze Age, including a sheep/goat mandible from the mid-2ⁿᵈ millennium bce in central Žemaitia (Piličiauskas et al. 2016: 186; Motuzaitė Matuzevičiūtė 2018: 152). While similar chronologies have also been suggested for Estonia (Lõugas, Kriiska & Maldre 2007: 25), this dating is not certain as none of the finds have been radiocarbon dated. Evidence for large-scale sheep and goat farming is not found until the Late Bronze Age, or the mid-1ˢᵗ millennium bce (Rannamäe 2016: 23).175 Gimbutas (1963: 35) includes *hanhi ‘goose’ among her list of domestic species. According to Lang (2016: 17), the word must have referred to a wild species as, in his view, goose domestication took place no earlier than the 1ˢᵗ millennium bce in Southern Europe. However, recent research has established that 175 kuʾs ‘horse’ (cf. Fortescue/Vajda 2022: 268). Apparently, Proto-Finnic speakers were not introduced to milking by Indo-Europeans, as the word *lüpsä- ‘to milk’ appears to have been adopted from an unidentified source, from where it also entered Mordvin, Mari and Permic (Aikio 2015a: 46). Unworked bone remains may have been misdated due to layer mixing, while worked remains found in grave sites might be trade items (Lõugas, Kriiska & Maldre 2007; Rannamäe 2016: 23). While Rannamäe et al. claim that the earliest sheep bones date from 1200 bce, i.e. the Bronze Age, only one sample has been dated so early (1200–800 bce) by archaeological context. Furthermore, two bones from the same site which have been radiocarbon dated belong to the Late Iron Age and Modern Period, respectively, suggesting the possibility that the third bone has also been misdated. The oldest directly dated sheep remains from Estonia are found in Asva on Saaremaa, dating to 786–522bce. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic → finnic loans 115 the domestic and wild goose diverged as long as 5000 years ago (Heikkinen et al. 2015). In the Baltic, it has been claimed that the domestic goose emerged in the Middle Ages, but a recent study based on isotope analysis has identified potential domestic specimens in Estonia from the Late Iron Age (Ehrlich et al. 2021). The evidence is therefore not as conclusive as Lang would imply, but it must be admitted that concrete indications of domestic geese at a sufficiently early date appear to be lacking. 3.6.2 Agriculture Many agricultural loanwords from Baltic into Finnic constitute generic terms: *sēmen ‘seed’, *he̮ina ‘hay’, *pe̮lut ‘straw chaff’. More notable is the word *hernes ‘pea’, a plant which is first recorded in the Eastern Baltic in the mid-1ˢᵗ millennium bce (Pollmann 2014; Grikpėdis/Motuzaitė Matuzevičiūtė 2017: 6; Minkevičius et al. 2020). This coincides with a general diversification of cultivated crops in the Late Bronze Age/Early Iron Age, again associated with the hillfort phenomenon (Lang 2007; Motuzaitė Matuzevičiūtė 2018: 156). The loaning of the word for ‘pea’ specifically is striking, as while the plant is present throughout Northern Europe by the Late Bronze Age, it is comparatively infrequent, implying a low economic significance (Grabowski 2011: 488; Stika/Heiss 2012: 192). Etymologically, the Baltic word is a specialization of an inherited generic term for ‘grain’, which might point to the crop becoming a staple among Balts. This is not supported by the existing evidence from the East Baltic, however, where the pea is recorded with the lowest frequency of all crops, matching the situation in the rest of Europe (Pollmann 2014: 409). The Finnic words for other specific crops are not Baltic loanwords; *vehnä ‘wheat’ may only indirectly be connected with the Baltic word *avižā- ‘oats’ (for a discussion, see pp. 239–240). Considering that the first small-scale agriculture in the East Baltic appears to have been exclusively barley-based (Motuzaitė Matuzevičiūtė 2018), it is interesting to note that Finnic *osra (~ ?*ocra) and Baltic *mẹ̄žjai ‘barley’ are both of obscure origin.176 The above evidence appears to suggest that the Finnic speakers became acquainted with diversified agriculture by other means than through contacts with Baltic-speaking populations.177 176 177 For a discussion of various attempts to etymologize the Baltic word, see Kroonen et al. (2022: 15). The Finnic word has been derived from Indo-Iranian (Holopainen 2019: 155– 156), but since the proposed source does not correspond in sense, the etymology is doubtful. Finnic *rukis ‘rye’ and *kakra ‘oats’ have been adopted from Germanic, cf. Häkkinen/ Lempiäinen 1996: 167–173. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 116 chapter 3 3.6.3 The wheel Interesting from a cultural perspective is *rattas ‘wheel (of a cart)’, in the plural *rattahe̮t ‘cart’. Archaeological evidence for wheeled vehicles in the Bronze and Iron Age Baltic appears to be largely lacking, and the introduction of wheeled vehicles in the Late Bronze Age has only been inferred by indirect evidence (Viires apud Lang 2007: 252). There also appears to be a general dearth of evidence for wheels in Central Russia throughout the Bronze Age, aside from a pair of pottery discs discovered in a child’s grave in Balanovo, which has been interpreted as belonging to a model wagon (Piggot 1969: 302). According to Lang (2007: 252), *silta ‘bridge’ may also have been loaned in connection with wheeled vehicles (cf. von Hertzen 1973: 85), and may originally have referred to trackways across swampy areas, traces of which can be identified since the Roman Iron Age. Note the etymological connection of the Baltic source with Lt. dial. tìlės ‘planks (as paving)’ (cf. also F silta ‘wooden floor’) might further support such an original meaning. 3.6.4 Context Many scholars have characterized the Baltic–Finnic contact relationship as long-term, in some cases as having lasted millennia (Kallio 2008; Vaba 2011: 756; Lang 2016). In this context, it has been claimed that Finnic would have come close to being fully assimilated by Baltic, before eventually becoming dominant (Lang 2018a: 29). This scenario seems unnecessarily complex; furthermore, the structural influence of Baltic on the Proto-Finnic phonemic system appears to have been minimal, which contrasts strongly with other cases of intense language contact eventually leading to language replacement, such as Latvian and Livonian (Suhonen 1973: 53–66) or Veps and Russian (cf. Зайцева 2008: 79). The conservative phonology of Finnic from a Uralic standpoint makes it unlikely that it was almost replaced by Baltic, and rather speaks in favour of the assimilation of a Baltic dialect into Proto-Finnic (cf. Kallio 2015: 90; Kallio forthc.). Moreover, the most important linguistic evidence for a long-term contact relationship concerns the substitution of the Proto-Baltic long vowels *ē and *ā (Junttila 2012: 266); however, as argued in 3.5.2, the different substitution patterns need not necessarily be analysed as evidence of chronological differentiation. Even if they are, this would not necessarily imply continuous, long-term contact. A relatively large proportion of the Baltic loanwords constitute what Lang (2016: 17) has referred to as “luxury borrowings”, i.e. loanwords which cannot be connected with the transfer of cultural practices or material goods. It is highly remarkable that the Baltic loans in Finnic include several kinship terms, in particular *se̮sar ‘sister’, *tüttar ‘daughter’, *ne̮pat ‘nephew, niece’, *morcijan ‘bride’. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 baltic → finnic loans 117 The term *ate̮iva, which is recorded in the sense ‘marriageable woman’ in Veps (Зайцева 2010: 18) and as ‘married woman visiting her parents’ in Finnish, in combination with other borrowed words for female family members, is likely to suggest exogamous marriage practices (Gimbutas 1963: 36; Lang 2015: 72). Genetic studies of European populations have repeatedly referred to female exogamy as a driver of intercultural contact in the Corded Ware up until the Bronze Age (Knipper et al. 2017; Mittnik et al. 2018, 2019), although there have not been any studies investigating this phenomenon in more recent periods, or further east, where the contacts are most likely to have taken place, so it cannot yet be confirmed whether such a hypothesis is supported by the genetic evidence. As “luxury” loanwords, we can also consider terms connected with topography and nature, such as *me̮cca ‘forest’ and *halla ‘frost’, and the names of animals of low economic significance — here, we are largely dealing with those that have a negative connotation — *herhiläinen ‘hornet’, *kīli(l)äinen ‘botfly’, *vapsas ‘wasp’, *kärmes ‘snake’. In addition, we find the basic adjectives *ahtas ‘narrow’, *tühjä ‘empty’, *haljas ‘green’, *ke̮ltaine̮n ‘yellow’, and the body part terms *hammas ‘tooth’ and *napa ‘navel’. From a typological perspective, the last two are particularly remarkable: according to the WOLD database, both ‘tooth’ and ‘navel’ rank among the 400 least likely words to be borrowed.178 The above semantic clustering seems most coherent with a scenario involving a Baltic substrate in Finnic. Geographical terminology and words related to natural phenomena are frequently identified as characteristic of borrowings from linguistic substrates (e.g. Kalima 1919: 257–258; Bertoldi 1932: 94; Ariste 1971: 9–10; Saarikivi 2004; Aikio 2009: 41). Close semantic parallels for many of the borrowed animal names can be identified among the Finnic substrate words in Russian dialects, cf. R dial. па́ рма ‘botfly’, ки́ гачи ‘gnats’ (cf. Мызников 2019: 295), товка́ ч (Шахматов apud Куликовскій 1898: 119) ‘a kind of woodworm’, ши́ жлик ‘lizard’ (cf. Kalima 1919: 257; Мызников 2004: 113– 116). The strongest linguistic evidence for a mixed group involving bilingualism can be seen in the plurale tantum nouns *pe̮lut ‘straw chaff’, *nītet ‘heddle’ and *rattahe̮t ‘cart’, which correspond to Baltic nouns also used exclusively in the plural (in the relevant meanings). This implies that the Baltic words were identified as plural upon borrowing, which can only be understood if we assume a certain level of bilingualism. This is particularly remarkable in the case of Baltic 178 Note the Romance substrate word imlīq ‘navel’ (< *imbilicus, cf. Galician embigo) in Andalusian Arabic (Griffin 1959: 347). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 118 chapter 3 *nîtīs ‘heddle’, where the ending is morphologically ambiguous, and could only be understood as plural by a person well acquainted with Baltic grammar. The hypothesis of a Baltic substrate that was ultimately absorbed by Finnic would further be supported by the evidence that the source of the Baltic loanwords was not the direct ancestor of any attested Baltic language (see 3.3.4). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 chapter 4 Loanwords into Other Uralic Languages 4.1 Sámi Many of the originally Baltic loanwords have been loaned into Sámi through North Finnic, in several cases early enough be distributed throughout the entire family; compare the following: – Sá. S lijkie, N liigi, K li´jjg ‘surplus’ (< *lijkē) ← F liika (~ Lt. liẽkas) – Sá. S naepie, N náhpi, Sk. nää´pp, Ter nap̜p̜e (< *nāpē) ‘navel’ ← F napa (~ Lv. naba) – Sá. S daajvaj, N dávjá , Sk. täujja, Ter taj̄va (< *tāvjā) ‘often’ ← F taaja (~ Lt. tánkus) – Sá. N šaldi ‘bridge’, Sk. šâ´ldd ‘floor’ (< *še̮ltē) ← F silta, K šilta ‘bridge’ (~ Lt. tìltas) In the last case, a Finnic intermediary is proven by the initial consonant, which must be the result of the specifically Finnic change *ti > si (and further North Karelian > ši). The other cases also show vocalic substitutions indicative of borrowing rather than common inheritance; note that stem-final *ā is typical of younger loanwords (Aikio 2006b: 36).1 Since distribution is not a decisive factor, it is occasionally difficult to rule out a common Finno-Sámic proto-form. For instance, both F siemen ‘seed’ and Sá. N siepman could theoretically reflect a PU *sämən (compare F kieli tongue’ = Sá. N giella ‘language’ < PU *kälə), but the principle of parsimony speaks rather in favour of a Finnish transmission. Comparing the list of Baltic loans in Sámi given by Sammallahti (1999: 410– 411) with those accepted by Aikio (2012a: 107), it would appear that the latter’s revisions mainly involve the removal of words which could equally be borrowed through Finnic. Thus, examples which show the correlation F e, o ~ Sámi *ea, *oa have been omitted, since although such correspondences are found in inherited words, they are also common in Finnic borrowings of all ages (Aikio 2006b: 31–34).2 1 A younger age of Sámi *tāvjā might also be shown by the metathesis *vj > *jv in South Sámi, as South Sámi appears to have kept *vj and *kj distinct: cf. Sá. S vuevjie ‘clothing insert’ (= Finnish vaaja ‘wedge’ ← Baltic or Germanic; see p. 50); see Pystynen 2014b. 2 Specifically, PSá. *keartē (> N geardi) ‘time, layer, strand’, *seaprē (> L siebrre) ‘company, society’ and *loamē (> N loapmi) ‘gap, cleft’ could just as well be loans from F kerta, seura and © Anthony Jakob, 2024 | doi:10.1163/9789004686472_006 Anthony Jakob - 978-90-04-68647-2 This is an open access chapter distributed under the terms of the cc by-nc-nd 4.0 license. Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 120 chapter 4 However, not all of the etymologies originally accepted by Sammallahti are straightforward. The equation of Sá. N gohččut ‘call, order’, Sk. kåččad ‘call’ (< *koččō-) and F kutsu-, E kutsu-, Li. kutsõ, 3pres. kutsūb (< *kuccu-) ‘invite, call’ with Lt. kviẽsti (3pres. kviẽčia) ‘invite’ (Mägiste 1923: 35–36; Koivulehto 1986: 272–274) is doubtful, as the substitution of *vẹ̄ (or earlier *wei/*wai) with *u lacks phonological plausibility (cf. Junttila 2011: 107).3 Likewise, despite Nuutinen (1992), Sá. N bievla, Sk. piõull (< *pievle̮) = F pälvi ‘snow-free patch (in spring)’ cannot be compared with Lt. pal̃vė ‘wind-levelled plain among dunes’ (← Pr., cf. Sabaliauskas 1974), as the assumed metathesis in Sámi is ad hoc. The Sámi and Finnish words rather presuppose an earlier *päwlə. The old derivation of Sá. N giehka, Sk. ǩiõkk (< *kieke̮) and F käki, E dial. kägi (usu. kägu), Li. ke’g (< *käki) ‘cuckoo’ from Lt. gegė ̃ ‘cuckoo’ (Thomsen 1890: 172) can also not be accepted. As recognized by LEW (142–143), the Lithuanian form is a recent clipping of the inherited Lt. dial. gegužė ̃ (= Lv. dzȩguze; OR жегъзоулѧ, cf. СДРЯ 11–14 iii: 238; Николаев 2020: 593), and can hardly be dated to Proto-Baltic. Furthermore, the Finnic and Sámic words are most likely regularly cognate with Khanty *käɣ-əj (Vakh-Vasjugan köɣi, Surgut kȧ̆ɣʷi) ‘cuckoo’ (< PU *käkə).4 In addition, many of the etymologies involve serious obstacles on the semantic side. In the following cases, the incompatibility in meaning makes the equation highly improbable: – N duollji, Sk. tue´lˈlj (< tuoljē) ‘hide, skin rug’ = F talja ‘animal hide’ ~ Lt. dalià ‘fate’, Lv. daļa ‘portion, share’ — Koivulehto (1984: 12) attempts to bridge the semantic gap by comparing OE scearu* (attested scaru, obl.sg. sceare) ‘division’ (> MoE share) and R arch. скора́ ‘hide’, but both words must be loma (cf. SSA i: 348, ii: 90, iii: 172, respectively). In addition, the etymology PSá. *piemme̮- (N biebmat, Sk. peâmˈmad) ‘feed, rear’ ~ Lt. penìmis ‘fattening pig’ is explicitly rejected as phonologically problematic. 3 According to Koivulehto *ku̯ oit-ja- was adopted as *kut-ja- because “ein /j/ konnte [vor *cc] nicht bestehen”. However, a sequence *-jcc- seems to have been possible even in inherited vocabulary: F arch. seitsen ‘seven’ (< *säiccen < PU *ćäjćəmä, Aikio in prep. 109–110) and veitsi ‘knife’ (< *väicci ‘knife’ ?< *väjćə ~ Ko. dial. ve̮ź- ‘cut slantwise’, Hungarian vés ‘chisel, cut’, cf. UEW 565), where it results from a fortition *-jć- > *-jcc-. Koivulehto is led astray by the notion that *j in these stems derives from an earlier *ŋ (Koivulehto 1981: 169; compare PF *suicce̮t > F suitset, Võ. suidsõq, Li. Salaca suiksud ‘bridle’, from virtual *ćuwə-ŋćə-, cf. *suu > F, E suu, Li. sū ‘mouth’), for which there is no evidence (Aikio loc. cit.). As a result, there is no reason to suspect that a form *kuiccV- should have been phonotactically impossible. 4 Compare Vakh-Vasjugan wöɣ, Surgut wȧ̆ɣʷ, wɔ̈ɣ̆ ʷ ‘strength’ (< *wäkə, UEW 563). Note, however, that Aikio (2015b: 2–3) has suggested that *ä regularly yielded Khanty *ǖ before *k in Uralic ə-stems. He does not mention *käkə as a counter-example, presumably because he considers the Finno-Sámic word to be a Baltic loan (Aikio 2012a: 107). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 loanwords into other uralic languages – – – – – – 121 understood as parallel deverbal formations (cf. OE sceran ‘cut, divide’), and do not constitute a parallel for the semantic shift ‘share’ > ‘(animal) hide’. N faggi ‘(wooden) hook’, Sk. vâ´ǧǧ ‘wooden hook, pothook’ (< *ve̮ŋkē) ~ Lt. vìngis ‘bend, turn, bypass’ — A root etymology; the meaning ‘hook’ is unattested in Baltic (cf. Kalima 1936: 178–179).5 N johtit, Sk. jåå´tted (< *jotē-) ‘go, travel, migrate’ ~ Lt. judėt́ i ‘move (about), be restless’ — The original meaning in Baltic is certainly not ‘move’, but rather ‘be restless’, cf. the glosses ‘arguo, obiurgo’ in Szyrwid (cf. ALEW 491), Lv. dial. jũdît ‘unruhig machen’ (ME ii: 120) and the cognates Skt. yúdhyate ‘fight’, To. A yutk- (< *(H)ieudʰ-sk-) ‘worry’ (IEW 511–512; LIV 225–226). N luokta, Sk. luhtt (< *luokte̮) = F lahti, E laht (< *lakti) ‘bay, inlet’ ~ Lt. lañktis ‘yarn winder’ (Posti 1977: 267–268) — Rather a root etymology, comparing Lt. leñkti ‘bend’, from which other words for ‘bay’ have been derived, e.g. Lv. lìcis ‘bay, inlet’. However, the etymology is suspect since the right combination of form and meaning is unattested in Baltic (Saarikivi 2004: 200). N riessan ‘decorative fringe’, Sk. riõzzâm ‘collar band’ (< *riese̮mē) and the verb S rïesedh, N riessat (< riese̮-) ‘adorn’ ~ Lt. rìšti, Lv. dial. rist ‘tie (on, up)’ — The semantics are not compelling.6 Moreover, the Sámi vocalism is unexpected: *ie(–e̮) implies earlier *ä(–ə). Sá. I ruodâs, Sk. ruõddâs (< *ruonte̮s) ‘wrist of a glove’; F ranne, E ranne (< *rante̮h) ‘wrist’ ~ Lt. grandìs ‘Armband’ (Ruhig ii: 31) (Liukkonen 1999: 116–117) — LÄGLOS (iii: 125) has already questioned the plausibility of the semantic shift ‘bracelet’ > ‘wrist’, but the situation is in fact worse, since the sense ‘bracelet’ is limited to a single lexicographical source, while the usual meaning in Lithuanian is ‘(metal) link, ring’, cf. also Pr. E grandis · rincke ‘beam link on a plough’ (Trautmann 1910: 342). Sá. S saertie ‘reindeer heart (as food)’ (< *sārtē) ~ Lt. šerdìs, Lv. ser̂de ‘core, kernel’ (Koivulehto 1990: 150) — The South Sámi form is cherry-picked. The 5 Aikio (2009: 176–178) has previously suggested that this Sámi word is a palaeo-Laplandic substrate word in view of the variants *ve̮ŋe̮ (L vagŋa ‘hook, barb’, Sk. võŋŋ ‘snag, submerged tree stump’) and N vievgŋa ‘snag’ (< *vievŋe̮). 6 Thomsen (1890: 212, cf. SSA iii; 72–73) takes PF *rihma ‘thread, rope; snare’ from an mderivative of Baltic *riš-, citing Lt. rišìmas ‘(the process of) tying’, which is a productive derivative which cannot be blindly projected to Proto-Baltic and the Latvian obs. hapax (Valle apud Mancelius) riſẜamais ‘band’ (ME iii: 531; the definite form of the gerundive adjective). Liukkonen (1987: 9) has assumed an unattested source *rišma-. Grünthal (2012: 328–329) also analyses Md. E ŕiśme, M dial. ŕiśmä ‘chain’ as a Baltic loan, but this is more convincingly derived from Indo-Iranian, cf. Skt. RV raśmā́ inst.sg. (or nom.sg., Jamison Commentary vi.67.1) ‘rein’, Parth. rsn /rasan/ ‘rope’ (< *raćm̥ n-o-; Lubotsky 2001: 314), cf. Holopainen 2019: 207–208. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 122 chapter 4 other Sámi cognates: N sárdi ‘rib without meat, strip of reindeer liver’, Sk. sä´rdd ‘small piece of meat’, K saa´rrd ‘broad cut of meat’ clearly show an original meaning ‘piece of meat’.7 Next, there are a number of seemingly unproblematic Baltic loanwords with Finnic equivalents which did not feature in the above discussions due to their uncertain Indo-European background. Note, for instance, the following:8 – Sá. N gahpir, Sk. keä´pˈper (< *ke̮pērē) ‘hat, cap’ = F kypärä ‘helmet’, E kübar ‘hat (with a brim)’, Li. kibār (< *küpärä) ‘hat’ ~ Lt. kepùrė, Lv. cȩpure ‘hat’ (Thomsen 1890: 185) — The Baltic word has no plausible comparanda beyond Slavic (R чепе́ц, SCr. čèpac ‘kind of cap’, cf. REW iii: 316, ALEW 552).9 – N guovllas, Sk. kuvlâs (< *kuovle̮s) ‘wooden collar band’ = F kaula, E kael, Li. ka’ggõl (< *kakla) ‘neck’ ~ Lt. kãklas, Lv. kakls ‘neck’ (Thomsen 1890: 177) — The Baltic word is of uncertain etymology (cf. ALEW 502–503).10 – Sá. L muolos, Sk. muâlas (< *muolōs) ‘shore lead, i.e. strip of ice melt along the shoreline’ = F dial. malo ‘edge, flank’, K dial. (N) malo ‘shallows, shoreline’ ~ Lv. mala ‘edge, shore, boundary’, Lt. dial. lýg-malis ‘filled to the brim’ (Loorits apud Mägiste 1939: 68–69; Nuutinen 1987a) — The Baltic word is etymologically ambiguous.11 – N ruoida, Sk. ruõidd (< *ruojte̮) ‘shin, thigh’ = F reisi, E reis, Võ. arch. raiź (< *re̮ici) ‘thigh’ ~ Lt. ríetas, Lv. riẽta ‘thigh, loins; ham (of meat)’ (Thomsen 1890: 7 8 9 10 11 In a more geographically limited area, we also find a meaning ‘piece of fabric’, cf. N sárdi ‘piece of a tent’, L sárdde ‘strip of canvas’, which probably suggests a basic meaning ‘piece, strip’. Note, similarly, the above discussions of *luose̮ ‘salmon’ (= PF *lohi; p. 101), *luovē ‘raised platform’ (= PF *lava; pp. 101–102), *se̮snē ‘tanned reindeer leather’ (= PF *hihna; pp. 87– 88), *vuore̮s ‘old’ (~ PF *varas; pp. 100–101). On Sámi *kuojmē ‘companion’, see p. 50. The loan etymology requires the assumption of a metathesis *käpürä > *küpärä (Thomsen 1890: 96; Kalima 1936: 124). The alternative interpretations of the Finnic word as a native formation (Mikkola 1930: 442; Nilsson 1996) are unconvincing. The traditional etymology (Mikkola 1896b: 218; Trautmann 1923: 125; Derksen 2015: 220) compares Skt. cakrá- ‘wheel’, but this is semantically problematic, and the etymology is not taken up by Smoczyński (2018: 469–470). While Walde/Pokorny’s (i: 515) “ ‘Hals’ als ‘Dreher’” would have semantic parallels (cf. OR воротъ ‘neck’ beside воротитисѧ ‘return’, СДРЯ 11–14 i: 477; MP grdn ‘neck’ from grd- ‘revolve, turn’, Durkin Meisterernst 2004: 163), the word *kʷekʷl-(o)- was specialized in the meaning ‘wheel’ already in PIE, and the notion that it could have uniquely preserved an abstract meaning ‘turner’ in Proto-Baltic is far from trivial. Note that Grinaveckienė/Mackevič (1989: 74) have even suggested the Baltic word was borrowed from Finnic. More plausible than the comparison with Sln. molẹ́ti ‘jut, protrude’ (IEW 721–722) is the connection with OIr. mala ‘eyebrow’ < *mlH- (Pedersen 1913: 99; as a semantic parallel, OIr. brú ‘edge, shore’ < ‘brow’, see eDIL s.v.). Alternatively, compare ON mǫl ‘shingle, gravel bank’ < *malō- (but see de Vries 1962: 401). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 loanwords into other uralic languages 123 212) — The Baltic word has no certain comparanda beyond Slavic (SCS рить, Cz. řiť ‘anus’). The equation with Arm. eri ‘shoulder of animals’ (IEW 863) is uncertain (Martirosyan 2008: 263). – N suolu ‘island; isolated patch of forest’, Ter si̮elaj ‘island’ (< *suolōj) = F salo ‘dense forest; island; elevated spot in a swamp’, E salu ‘grove; area of woodland in an open landscape’ (< *salo)12 ~ Lt. salà, Lv. sala ‘island; elevated spot in a swamp’13 (Thomsen 1890: 214; Kalima 1936: 158) — The Baltic word lacks a satisfactory etymology.14 It cannot be excluded that the above words originated in Baltic, but depending on one’s assessment of the existing etymologies, alternatives cannot be excluded. Where the Baltic word is isolated within Indo-European, an early loan from Finnic into Baltic may still be on the table. Where Slavic equivalents are attested, this becomes far less likely; however, the possibility remains that the words in question are parallel loans from unattested source languages. This brings us to the unambiguous cases. Of the Finnic etymologies accepted in 3.2, five of them have a Sámi equivalent which cannot be explained as a recent loan from Finnic:15 – N sarvva, Sk. sõrvv (< *se̮rve̮) ‘elk’ = PF *hirvi ‘deer; elk’ ~ Pr. E sirwis ‘roe-deer’ – N suoldni, Sk. sue´lnn (< *suolnē) ‘mist over water in late summer; hoarfrost’ = PF *halla ‘frost, hoarfrost’ ~ Lt. šalnà, Lv. sal̂na ‘hoarfrost’ – N suorri, Sk. sue´rr (< *suorē) ‘branch, fork’ = PF *hara ‘branch, fork’ ~ Lv. zars ‘branch, prong’ 12 13 14 15 The Finnic and Sámi words reconstruct to a common proto-form *salaw (Kuokkala 2012: 78) with which we may compare Lt. obs. salavà ‘island, river island (German Werder)’ (Bezzenberger 1877: 320; Ruhig ii: 399), adduced already by Thomsen. In terms of word formation, salavà stands quite apart from other words with a suffix -ava, which usually have a collective meaning (cf. Skardžius 1941: 379–380). Despite the communis opinio (ME iii: 64; LEW 758; Smoczyński 2018: 1126–1127), it seems obvious that Lt. salà, Lv. sala ‘village’ is borrowed from Bel. сяло́ . It is hardly a coincidence that the Lithuanian word is practically limited to Vilniškiai dialects where s is regularly depalatalized (Zinkevičius 1966: 165 and Map 74; cf. Smalinskienė 1994: 178). Latin īnsula ‘island’ is hardly to be separated from OIr. inis ‘island’ (Ernout/Meillet 319– 320; de Vaan 2008: 306). Endzelīns (ME iii: 664) proposed that salà was abstracted from *api-sala ‘that which [water] flows around’, but such a form is unattested, and the verbal root *sal- ‘to flow’ is itself supported by doubtful evidence (Jakob forthc. b.). While a Baltic source is usually assumed (LEW 758; Sammallahti 2001: 411; Aikio 2012a: 107), Thomsen and Kalima both admit the possibility of a Finnic → Baltic loan (cf. also Bednarczuk 1976: 52), and others have suggested a loanword from an unknown source (Saarikivi 2004: 208; Aikio 2004: 24; J. Häkkinen 2009: 48; Holopainen, Kuokkala & Junttila 2017: 129). I can only imagine that *suolnē was an accidental omission in Aikio (2012a: 107). As for *vuossē, the omission perhaps follows from the fact that Sammallahti considered a Germanic etymology equally possible (see Chapter 3, fn. 37). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 124 chapter 4 – N duovli, Sk. tu´vll (< *tuovlē) ‘tinder (as a traditional remedy); tinder fungus’ = PF *takla ‘tinder’ ~ Lv. dial. dagla, daglis ‘tinder, fire sponge’ (ME i: 430, EH i: 313) – S vuessie, N vuosˈsi (< *vuossē) ‘handle (of a cooking pot, bucket)’ = PF *ansa ‘handle’ ~ Lt. ąsà, Lv. ùosa ‘handle, eyelet’16 Aikio lists *se̮rvēs (> N sarvvis, Sk. sââ´rves) ‘uncastrated reindeer buck’ as a separate loanword. In reference works (SKES 77–78; SSA i: 167), the Sámi word has been equated with F hirvas, K hirvaš ‘uncastrated reindeer buck’. However, given that this word is only known in the northern dialects of Finland and Karelia, combined with the exact semantic correspondence with Sámi, it seems much more probable that it is a partial calque resulting from a crossing of the native hirvi with Sámi *se̮rvēs (Junttila in prep. s.v. hirvas; cf. also the direct loanword sarvas; Aikio 2009: 276). Note that the interpretation of F hirvas and hirvi as independent loanwords from the Baltic masculine *širvas (= Pr. sirwis) and the feminine *širvē, respectively (Nieminen 1940: 378), could also be applied to Saami *se̮rve̮ and *se̮rvēs. However, it appears just as probable that *se̮rvēs as an inner-Sámi derivative with the same suffix as in Sá. S urries, Sk. åå´res (< *orēs) ‘male (animal)’.17 All the cited words in Finnic and Sámi can be given a common Uralic protoform. This can be interpreted in at least three ways: (a) the loans were adopted into a single Finno-Sámic proto-language; (b) the Sámi forms are very early adoptions from Finnic, predating most of the sound changes; or (c) the words were adopted independently by Sámi and Finnic, and the fact that they go back to identical proto-forms is due to coincidence. The main issue with option (a) is that the reconstruction of a Finno-Sámic branch is nowadays increasingly disfavoured, with the shared features being explained as the result of secondary areal diffusion (T. Itkonen 1997; Salminen 16 17 For the development *-ns- > *-ss-, compare Sá. N guosˈsi, K kū´ss ‘guest’ (< *kuossē = F kansa ‘guest’); cf. Sammallahti 1998: 54. The Inarilappisches Wörterbuch attests the form meččin /mečˈčin/ ‘im Walde’ which would reflect the inessive singular of a word *meaččē. The latter has been interpreted as an early loan from Baltic (Aikio 2012a: 107). The other Sámi languages attest a similar but irreconcilable *meaccē (> Sá. S miehtsie, Sk. meä´cˈc ‘forest’), pointing to a later loan. The Inari form corresponds formally to Lule miehttjen ‘against the wall of the tent, as far as possible from the hearth’, which is semantically aligned with the other West Sámi languages, e.g. South Sámi miehtjiedidh ‘move away, put by the wall’, Ume miehttjiedit ‘remove the pot from the fire’. South Sámi meahtsanidh ‘withdraw oneself to the wall (of the tent)’ = N meahccánit ‘stray too far (of cattle)’ might show the confusion of the two word families. The question, then, is whether Inari meččin ‘im Walde’ should also be explained as the result of a contamination of the Finnic loan mecci ‘forest’ and *meaččē ‘far away (from the hearth)’. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 loanwords into other uralic languages 125 1999: 20–23; 2002: 47–48; Saarikivi 2011: 106–109; Aikio 2022: 3–4). In view of this, Aikio (2012b: 73) opts for option (b), assuming that all of the Baltic loanwords in Sámi were mediated through Finnic, an opinion that was already held by Thomsen (1890: 28–29). A number of criticisms can be raised against this. First, there is no unambiguous evidence for Finnic loanwords in Sámi of a sufficiently early date. Very few identifiably Finnic loanwords predate the Sámi vowel shift, and even Sámi *puošē (> N buošši) ‘angry’, apparently an early loan from PF *paha ‘bad, evil’ (Sammallahti 1998: 183), evidently postdated *š > *s.18 It is unclear, however, exactly by what criteria such early mutual loanwords could be distinguished from common inheritances, and the existence of unidentifiable borrowings from this period cannot be excluded (Aikio 2012b: 72).19 Next, there are a number Baltic loanwords in Sámi which are unattested in Finnic. To my mind, there are three plausible examples: – N giehpa, Sk. ǩiõpp (< *kiepe̮) ‘soot’ ~ Lv. kvȩ̂pi, dial. kvêpji pl. ‘soot’ (cf. Sammallahti 1998: 127). – N loggut, K lå´ŋŋge (< *loŋkō- ~ *loŋkē-) ‘strip (birch bark); peel’ ~ Lt. dial. lùnkas, Lv. lûks ‘bast’; cf. further Pr. E lunkan; R лы́ ко, SCr. lȉko ‘bast’. A nominal form is attested in the Sámi loanword F lunka ‘bark which flakes off easily’ (Aikio 2009: 115–116) – N vietka, Sk. viõtkk (< *vietke̮) ‘adze’ ~ Lt. vedegà ‘adze’, Lv. vȩr̂ga, dial. vȩdga ‘ice chisel’; Pr. E wedigo ‘carpenter’s axe’ The latter two etymologies do not involve any significant formal or semantic issues,20 but cannot be considered unambiguous evidence of direct contacts between Sámi and Baltic, as the Baltic words themselves do not have reliable Indo-European etymologies. The word for ‘bast’ has a potentially irregular comparandum in Slavic (see p. 181), while the word for ‘adze’ contains an opaque suffix *-eg- otherwise found only in the equally obscure Lt. uodegà ‘tail’ (the derivation of the latter from úodas ‘mosquito’ per ALEW 1328 and Smoczyński 2018: 1563 wants semantic parallels). 18 19 20 Whether or not they were transmitted through Finnic, it is possible that *se̮rve̮ ‘elk’ (← *širvas) and *suolnē (< *šal̂nâ) ‘mist over water’ did not in fact predate the pre-Sámi change *š > *s, but merely predated the innovation of a new phoneme *š, as in the absence of such a phoneme, a substitution *š → *s would be in line with expectations (cf. Kallio 2009: 34). Although we might expect more such traces in the case of intense early language contact. For instance, one might anticipate evidence of the Finnic merger *t, *d, *ď, *č > *t in the Sámi material. Admittedly, Sá. *vietke̮ (< *wätkə) is perhaps not quite expected from Baltic *vedegâ. One might rather anticipate **wätäkä (> Sá. **vātēkē > Sá. N **váhttit). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 126 chapter 4 Things look more positive in the case of ‘soot’. Latvian kvȩ̂pi is clearly related to the verb kvêpt ‘smoke, smell; get covered in soot’, Lt. kvėp̃ ti ‘breathe, blow; smell’, kvãpas ‘breath, smell’. Although further connections are difficult,21 the inner-Baltic etymology seems solid. In Sámi, the substitution of *kv- with *k- is paralleled by Torne Sámi gierdnas ‘mill, grinder’ ← Nw. kvern and South Sámi gearhka ‘throat’ ← Nw. kverk (Quigstad 1893: 14; Koivulehto 1992a: 92). The question still stands as to whether this word for ‘soot’ might once have existed in Finnic but was subsequently simply lost (Aikio 2012a: 74). The Proto-Finnic word for soot (F noki, E nõgi < *noki) is itself of obscure origin, but it is possible that it replaced an earlier Baltic loan *kēpi. While this cannot be excluded, reconstructing unattested Finnic words to explain away evidence of direct contact is, of course, circular. Moreover, Aikio (loc. cit.) and Saarikivi (2022: 33) do still admit the possibility of direct contact on the basis of one example, namely Sá. S liejpie, N leaibi (< *leajpē) ‘(grey) alder’, which for phonological reasons could not have been adopted through Finnic *leppä ‘alder’, and whose preform *lejpä actually more closely resembles pre-Baltic *leîpâ than the Finnic word does. As I have discussed in detail above (see p. 89), there are several problems with this word family which make a simple Baltic loan hypothesis unsatisfactory; if there is any relation at all, it is most probably a shared substrate word in Balto-Slavic and West Uralic. Although the word for ‘alder’ may not be reliable, there is another word which provides evidence of direct, independent contact between Baltic and Sámi:22 – N suoidni, Sk. suei´nn (< *suojnē) ‘grass, hay’ ≠ F heinä, E hein, Li. āina ‘hay’ ~ Lt. šiẽnas, Lv. sìens ‘hay’ 21 22 Possibly here also R ко́ поть ‘soot’, but the loss of *u̯ is irregular. Contrary to Schrijver (1991: 260–263) and Derksen (2015: 268), the Latvian acute hardly warrants the awkward reconstruction *kh₂uep-, which cannot in any case account for Gr. καπνός ‘smoke’, Lat. vapor ‘steam, heat’ (IEW 596–597). We are probably dealing with metatony, as in Lv. dial. drêbt ‘beat; sleet’ vs. Lt. drėb̃ ti (cf. Chapter 7, fn. 78 and 79) and likewise Lv. têst ‘carve; shave’ vs. Lt. tašýti (cf. Chapter 3, fn. 8). South Sámi daktere ‘daughter (by marriage)’, Pite ⟨taktier⟩ (Lehtiranta 1989: 130) is presented by Sammallahti (1998: 127) as an example of a word which could not have been mediated by Finnic (where we find *tüttär). It is, however, more probable that the Sámi word is of Norse origin, especially in view of the limitation to the western edge of the family (cf. already Qvigstad 1893: 125). Although the word for ‘daughter’ in Old Norse is an assimilated dóttir, the older cluster -ht- is reflected as *-kt- in several loans, cf. Sá. S slikte, N livttis ‘smooth’ (< *liktēs ← *slihtaz; cf. ON sléttr); S raaktse ‘harness trace’ (*rākte̮s < *drahtuz; cf. ON dráttr ‘dragging’, Nw. dial. drått ‘trace’); N divttis ‘tight, watertight’ (< *tiktēs < *þinhtaz; cf. ON þéttr). See Posti (1953: 45). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 loanwords into other uralic languages 127 While the Finnic word, partly on the basis of the Sámi evidence, has usually been reconstructed *šaina, I have argued above extensively (see 3.3.1.4) that the word in question should be reconstructed *he̮ina for Proto-Finnic. If this is correct, then the Sámi equivalent cannot have been adopted through Finnic, but must rather represent an independent loanword. The difference can be explained by assuming an earlier chronology, namely a date before the Baltic monophthongization *ai > *ẹ̄.23 Nevertheless, it is possible to imagine an alternative account whereby preSámi *šajna was, after all, borrowed from pre-Finnic *še̮jna. The Sámi change *i > *e̮ is a recent innovation (Sammallahti 1998: 106), postdating at least some of the Norse contacts (Aikio 2006a: 15) and, more importantly, the Baltic loanword *se̮rve̮ ‘elk’ (?← pre-Finnic *širvə) ← Baltic *širvē (see above). Prior to this change, it is possible that pre-Sámi *a in fact represented the phonetically closest match to Finnic *e̮, meaning that the substitution *e̮ → *a would be perfectly intuitive. Therefore, at least the word for ‘soot’ and perhaps also ‘hay’ might offer some evidence for direct contact between Sámi and Baltic. Even so, the fact that five out of six of the Sámi loans from Baltic are shared with Finnic can hardly be considered coincidental. It seems, therefore, that some of the relevant material must have percolated through a Finno-Sámic dialect continuum, but that does not exclude a small level of direct contact taking place between Sámi and Baltic. 4.1.1 Earlier and Later Loanwords All of the examples mentioned above involving Baltic *š show a reflex *s in Sámi. Two other Sámi substitutions for this phoneme have been suggested in the literature: *č and *š. These examples have been used to support the idea of an older and younger layer of Baltic loanwords in Sámi, respectively. The former is supported by two etymologies (Kallio 2009: 32–33): – Sá. N čuorpmas, Sk. čuõrmâs (< *čuorme̮s) ‘hail’ ~ Lt. šarmà, Lv. sar̂ma ‘hoarfrost’ (Koivulehto 1983: 188–189) – Sá. N čohkut, Sk. čååkkad (< *čokō-) ‘to comb, currycomb’ = F suka, E suga, Li. sugā ‘currycomb; heckling comb’ ~ Lt. šùkos f.pl., Lv. suka or sukas f.pl. ‘comb, heckling comb’ Although the Baltic word for ‘comb’ has a probable cognate in Uk. щеть ‘bristle’, Sln. dial. ščę̑t ‘brush, thistle’ (Pronk/Pronk-Tiethoff 2019: 285), the further ety23 Similar argumentation can be made in the case of *ruojte̮ ‘thigh’ (cf. PF *re̮ici, Lt. ríetas), however, in this case there is no evidence for an original *-ai- diphthong in Baltic, and the word is of obscure origin. See above on pp. 122–123. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 128 chapter 4 mology is uncertain (ALEW 1216) so that the exact relationship between the words cannot be determined. In the other case, there is a semantic difference. Although not prohibitive (compare Lt. grúodas ‘frozen ground’ vs. OCS градъ ‘hail’), it is questionable whether this single etymology is of sufficient calibre to carry the weight of an entire loanword stratum, especially given that a term associated with a natural phenomenon such as ‘hail’ is an unlikely candidate for borrowing during incidental early contacts. As for the supposed ‘late’ loans, Sammallahti (2001: 401) has suggested two which would show Sámi *š- for Baltic *š- (or *ž-), and a third was later adduced by Koivulehto (apud Aikio 2009: 200). While Sammallahti assumed a ProtoBalto-Slavic source for these loanwords, the Sámi phoneme *š is of recent, probably post-Proto-Sámi origin (cf. Kallio 2009: 35), which suggests a later date. Even though the notion of Baltic loanwords in an already disintegrating Proto-Sámi seems a priori unlikely, we must keep our minds open at this point: – Sá. N šearrat, L sjerrat (< *šeare̮tē) ‘clear (of the sky)’ ~ Lt. žėrėt́ i ‘glow (e.g. of coals); shine, glitter’ — There is no precise formal or semantic match, so it is essentially a root etymology. Junttila (2015b: 477) has suggested that the Sámi word is rather a loan from Finnic *heretä (> E ere, dial. here ‘bright’).24 – Sá. S sealma ‘threshold; pass, ridge’, N šielbmá ‘threshold (of a tent)’ (< *šielmā) ~ Lt. šelmuõ ‘ridge (of a roof), eaves, gable’ (= SCr. sljȅme ‘mountain ridge; (dial.) roof ridge’) — There is no exact semantic correspondence. Aikio (2012a: 107) has tentatively suggested a loan etymology from Finnic *he̮lma (> F helma, E hõlm) ‘hem’. – Sá. S sjåavonje, N šūvon (< *šuovuńe̮) ‘well-trained shepherd dog’ ~ Lt. šuõ (obl. šun-) ‘dog’ — The comparison is phonologically problematic. Sammallahti (2001: 400) erroneously derives Lt. šuõ from *śou̯ on(i)-, and Kallio (2009: 35) reconstructs Baltic *šāvā on the basis of a Žemaitian form šova quoted in LEW (1023). The latter is evidently an untransposed Žem. šọva = šuvà, which is not a derivative, but a special development of the nominative singular (see Zinkevičius 1966: 256–257). All in all, we can reconstruct a Proto-East-Baltic *š(u̯ )ōn, obl. *šun-, neither of which can explain the Sámi word. None of the etymologies are convincing, and they can hardly serve as a basis for drawing the far-reaching conclusions which they would imply, namely that the Balts would have been in contact with the Sámi already after their migration into Fennoscandia. Of course, it is in principle possible that certain Baltic 24 Note that Holopainen/Junttila (2022: 103–105) and Junttila (in prep. s.v. *heräitt̆ äk) have suggested that this Finnic stem is after all ultimately of Baltic origin. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 loanwords into other uralic languages 129 populations crossed the Gulf of Finland, where enclaves could have interacted with Sámi populations until a relatively recent date. However, this is merely a speculation unsupported by any other evidence. Note in this connection that it has been suggested that the Sámi endonym PSá. *sāmē (> N sápmi) is derived from a Baltic source, cf. Lt. žẽmė ‘earth, land’ (Koivulehto 1993b). Koivulehto analyses *sāmē as a cognate to F Häme, the name of a historical region in Finland. The notion that a Baltic loanword would be used as a local toponym in Finland, which would later serve as the selfidentification of the Sámi people, would probably imply that the Balts settled in Finland, and most probably before the Sámi arrived there. This is an extremely bold claim. Furthermore, even though it seems attractive to compare F Häme with the self-designation of the Sámi, this encounters an important issue: it is hardly possible to separate Sámi *sāmē from Finnish Suomi ‘Finland’. These two forms together point to a common preform *sämä rather than *šämä (Pystynen 2018: 83; Holopainen 2021: 207–208). All in all, the idea that Balts should once have been present in the area of modern-day Finland is too bold a claim to base on a single toponym, especially as it remains possible, or even probable, that the ethnonym *sāmē was adopted from a palaeo-Lakelandic contact language of the type we know must have been spoken in this area before the arrival of the Sámi (Aikio 2012: 80–88).25 4.2 Mordvin The possibility that Baltic and Mordvin were in contact was already recognized in the 1880s in two articles by Wilhelm Tomaschek (1883: 704–705, 1889: 11–12). Of the 15 comparisons made in these works, many reappear in Thomsen’s work on Finnic-Baltic loanwords. In the latter’s opinion (1890: 154–155), these must mostly have passed through a dialect continuum from Finnic into Mordvin, although Thomsen does admit that a small number may have been borrowed directly. A similar conclusion was reached by Kalima (1936: 191–192). By contrast, as many as four of the seven loanwords accepted in the recent comprehensive study by van Pareren (2008) were classified as direct. The goal of this subchapter is to establish the degree of direct and indirect contacts between Baltic and Mordvin. 25 Reminiscent of *sāmē is Lv. ⟨Sa̷̓ hms⟩ ‘Finne; Oesulaner’ cited by Ulmann (1872: 244), but Endzelīns (ME iii: 803) attractively derives this form from Li. sārmā ‘Saaremaa’ (note Oesel = Saaremaa). Several parallels for the loss of /r/ after long vowels in Courland Latvian are provided in Endzelīns 1923: 159–160. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 130 chapter 4 4.2.1 Rejectable Comparisons The most recent detailed study (Grünthal 2012) shows a manifold increase in the number of accepted loanwords, with a total of 36.26 Unfortunately, a large proportion of the additional etymologies accepted and proposed by Grünthal involve hypothetical semantic shifts or anachronisms, and in my opinion should certainly be rejected: – Md. EM al (< *al / *alə) ‘egg’ ~ Lv. uõla ‘egg’, dial. ‘pebble’ (Joki 1973: 294) — The sense ‘egg’ is an extension of ‘pebble’ in only part of the Latvian dialects, displacing older pàuts (EH ii: 186 = Lt. dial. (S Aukšt.) paũtas, Pr. G paute ‘egg’), and is hardly to be reconstructed for Proto-Baltic (Lanszweert 1984: 38); compare Lt. uolà ‘whetstone, rock; cliff’. Furthermore, the substitution *(w)ō- → *a is phonologically unlikely (van Pareren 2008: 86).27 – Md. E dial. čonda (?< *šondə) ‘bride price’28 = F hinta, E hind, Li. īnda (< *hinta) ‘price’ ~ Lt. šim̃ tas ‘hundred’ (Uotila 1990) — The fact that šim̃ tas (usu. as pl. šimtaĩ ) can be used hyperbolically to mean ‘a lot’ (like English hundreds) cannot be considered a sufficient semantic bridge. – Md. E inže (pl. inšť), M inži (< *inžə : *inž-) ‘guest’ ?= F obs. inhiminen ‘person’, Vp. inehmoi ‘lazy or sickly person’, Võ. inemine, Li. (Salaca) imi ‘person’ (?< *inehminen) ~ Lt. įžymùs ‘notable, famous’ (Liukkonen 1999: 61– 62) — The Lithuanian word is a productive deverbal adjective from į-žymėt́ i ‘note, mark’, and cannot be reconstructed for Proto-Baltic. Furthermore, the semantic development is speculative. – Md. EM karks ‘belt, waistband; rope used to bind a sheaf’ ~ Lt. kárti ‘hang (up)’ (Grünthal 2012: 315) — EM -ks is a deverbal suffix, but the connection with ‘hanging’ is not evident. – Md. E penge (pl. penkť), M pengä (< *peŋgə : peŋg-) ‘(piece of) firewood’ ~ Lt. spiñgis ‘forest aisle’ (Grünthal 2012: 324) — The two words are semantically distant. – Md. E pusmo, M pusma (< *pusmə) ‘bunch, bundle’ ~ Lt. bùžmas (Grünthal 2012: 326) — Grünthal incorrectly glosses the Lithuanian word as ‘bunch’. 26 27 28 Grünthal explicitly marks as uncertain the Baltic loan etymologies for PMd. *ärkə ‘lake’ (on which see pp. 86–87), *kodər ‘twining plant stem’ (see Chapter 3, fn. 4), *mukərə ‘rump, rear’ (~ Lv. mugura ‘back’; against which see van Pareren 2008: 109–111), and Md. EM luv ‘space between the fingers’ ~ Lt. lomà ‘hollow, valley’ (the Md. word rather belongs with F lovi ‘cleft, notch’, Pystynen 2020b). These will be ignored in the following discussion. The alternative view is that EM al ‘egg’ is a semantic extension of EM al- ‘area under or below’ (Rédei 1968: 160; Keresztes 1986: 33), which is itself of Uralic origin (< *e̮la, Aikio in prep. 52–53). If čonda (Velikij Vrag) is the most archaic form, the metathetic variant čando could perhaps be explained as the result of contamination with čana ‘price’ ← R цѣна́ (cf. Вершинин i: 486). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 loanwords into other uralic languages 131 The word is only known from lexical sources; cf. bùźmas ‘eine Falte, Krause’ (Kurschat 1883: 66).29 – Md. E raško, M raška (< *raškə) ‘crotch, fork’ = F rahko ‘fork-shaped torch holder’ ~ Lt. dial. raškà ‘skeltu galu kartis obuoliams raškyti [a device with a forked end used to pick apples]’ (Skirsnemunė; Skardžius 1941: 41), rãškės ‘prietaisas obuoliams raškyti [a device used to pick apples]’ (Daugėliškis; LKŽ) (Liukkonen 1999: 114–115) — The basic sense of these rare Lithuanian words must be ‘picker’ (cf. rėk̃ šti, raškýti ‘to pick’), while the Finnic and Mordvin words would imply an original sense ‘fork, crotch’ (cf. Nilsson 2001: 185). – Md. E rudaz, M ərdaz (< *rudas) ‘dirt, faeces’ (Grünthal 2012: 329) ~ Lt. rùdas ‘chestnut brown’, Lv. ruds ‘red-brown’ — From a semantic perspective, the Russian data comes far closer, cf. Ru. dial. (Smolensk) руда́ ‘dirt, stain’, рудо́ й ‘dirty’.30 – Md. E ťeŕďe-, M ťeŕďə- (< *terďə-) ‘call over, invite’ ~ Lt. tìrdinti, Lv. tir̂dît ‘badger with questions, torment’ (Grünthal 2012: 335) — The Baltic words are frequentative derivatives of Lt. tìrti ‘question, examine’, Lv. obs. tirt ‘Ausfragen’ (Lange 1773: 351). The semantics are unconvincing. Wälchli (1997: 312–319) has suggested Baltic etymologies for a number of grammaticalized relational nouns. All of these are rejected by van Pareren, but accepted by Grünthal.31 Again, the semantic developments stretch the imagination: – Md. E lango, M langa (< *laŋgə) relational noun ‘on’, ‘surface’ ~ Lt. lankà ‘water meadow; swamp, valley’, Lv. lañka ‘low-lying meadow; river bend’. The basic sense in Baltic appears to be ‘river bend’, cf. R лука́ , Bg. лъка̀ ‘river bend; meadow in a river bend’ (~ Lt. leñkti ‘to bend’). 29 30 31 In Ruhig (ii: 53), we find buʒ́mas ‘Bauchbruch am Reße’; however, this is presumably a printing error for *bůʒ́mas; Mielcke (i: 31) lists the same word under boʒ́mas ‘das Bauchreß, der Bauch vom Reße’ and Nesselmann (1851: 333) has bůźmas = boźmas ‘das Eingeweidenetz, Bauchnetz’ (the word was not familiar to Kurschat 1883: 54). According to Nesselmann (but no-one else?), boźmas also = baźmas ‘eine große Menge, eine Masse von Menschen, Thieren, Körnern’, which must be where Grünthal’s ‘bunch’ ultimately originates (but note that all of the example sentences in LKŽ s.v. bãžmas refer to people, unlike the Mordvin words). But admittedly, Smolensk is geographically far removed from Mordvinia. Curious is the Russian dialectal form ру́ дос ‘swampy area where rusted water comes to the surface’, attested in the Komi Republic (СРНГ xxxv 235), cf. Komi rode̮g ‘dirt, stain; rust in standing water’, Mari E rüδaŋa-, W ərδäŋge- ‘to rust’ (Лыткин/Гуляев 1970: 241–242). For Komi and Mari, a common preform *rentV- could perhaps be reconstructed. According to P. Kallio (p.c. March 2023), Wälchli himself is now unenthusiastic about his older proposals. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 132 chapter 4 – Md. E potmo, M potma ‘insides, stomach, bosom’, E potso, M potsa iness.sg. ‘inside’ (< *potmə : *pot(m)-) ~ Lt. putmuõ ‘swelling (as an ailment)’ – Md. E turtov, dial. turtoŋ (< *turtəŋ lat.) ‘for’ ~ Lt. tur̃tas, ‘wealth, property’ In addition, a few etymologies must be rejected on formal grounds. Although Grünthal himself notes that sibilants in Erzya are not subject to palatal harmony (2012: 330; cf. Bartens 1999: 43), he still resorts to it in two cases where we find an unexpected sibilant reflex: – Md. E raśke (< *raśkə) ‘relative, kin’ ~ Lv. rads ‘relative, lineage’ — Grünthal posits a preform *radəs-kə, contradicting the Erzya evidence for *ś. Additionally, despite ME (ii: 463), it seems quite possible that the Latvian word is loaned from Russian; cf. OR родъ ‘lineage; birth, origin; relative (etc.)’ (СДРЯ x: 408–415) – Md. E dial. simeń, M dial. śiməń (< *siməń) ‘tribe, family’ ~ Lt. giminė ̃ ‘relative, tribe, family’ — The Erzya form proves an initial *s. In the latter case, it is assumed that *ś would substitute a palatalized allophone of *g in Baltic. Junttila (2018: 78), who also provides the erroneous reconstruction *śiməń, specifies this Baltic dialect as “Altlettgallisch”. He suggests two parallels: E dial. śive, M dial. śivä ‘salary, pay’ ~ Lv. dzîvuôt ‘live’, dial. +acc. ‘work, be occupied with’ and E śiŕe, occurring in collocation with paro ‘good’ in curses ~ Lt. gìrti ‘praise’, gẽras ‘good’. The evidence of these two words alone, both of which require additional assumptions, seems insufficient to support a substitution *g → *ś. Another supposed piece of evidence for an “Old Latgalian” source is the verb Md. E ŕeďa-, M dial. ŕäďa- (< *räďa-) ‘see, notice’ ~ Lt. regėt́ i, Lv. redzêt ‘see, discern’, refl. ‘seem, be evident’ (Wälchli 1997: 319–320; Junttila 2018: 79–80). Wälchli’s opinion is that *ď may have directly substituted *g, as Proto-Mordvin lacked a phoneme */g/. However, there is no reason to consider the loss of *g particularly ancient (Grünthal 2012: 328), and the possible Baltic loanword *lija ?< *läjkä ‘other’ (see below) must have predated it. In the opinion of van Pareren (2006: 49; cf. 2008: 120), we should expect *g → k in such a late loanword, which is indeed what we find in some borrowings from Tatar and Russian (Paasonen 1903: 17; Keresztes 1987: 67–68). According to Junttila, Mordvin *ď could directly substitute a Baltic palatalized *[g’], and a realization [d’] for /g’/ is indeed attested in South Aukštaitian (Zinkevičius 1966: 140–141). However, the hypothesis that a dialect in which the velars were palatalized was spoken in the necessary time and place remains unproven. As Junttila (2018: 80) himself admits, no evidence for a Latvian-type palatalization has been identified in Baltic substratal hydronymy. The evidence of Mordvin *räďa- alone is hardly enough to postulate such a feature for a hypothetical Baltic dialect. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 loanwords into other uralic languages 133 4.2.2 Turkic or Baltic? In a couple of cases, a Turkic origin appears just as likely or more probable than a Baltic origin: ‘honeycomb’. Md. E keŕaz (also as pl. keŕazt), M käŕaz (< *käŕas) ‘honeycomb’ ~ Lt. korỹs, Lv. kāre ‘honeycomb’ (Tomaschek 1883) — The Mordvin word cannot be separated from a wider group of Volgaic terms. On the one hand, we find Ma. E karaš, W käräš and Udm. karas (which is not regularly cognate with the Mari forms), and on the other Tatar käräz, Bashkir käräδ (< *käräz), all in the same sense. Due to its final sibilant, Chuvash karas ‘honeycomb’ cannot be cognate with the Volga Kipchak forms. Räsänen (1920: 245) has derived the Turkic words from Uralic, and these from an Iranian *kāras. As such a word is unattested in Iranian, this can hardly be accepted (cf. Joki 1973: 226–227; Holopainen 2019: 127). Disregarding the language-specific phonotactic limitations, Mordvin *käŕas is phonologically identical to Tatar käräz, and indeed already Paasonen (1897: 37) suggested that Mordvin borrowed the word from Tatar. If the Volgaic and Baltic words are indeed connected, one might speculate whether it was the Turkic words that were in fact loaned from Baltic. Indeed, this could potentially explain the unexpected front vocalism in Mordvin. In Turkic, *k was allophonically rendered as *[q] (> Chuv. x) in back-vocalic environments, resulting in an association of foreign /k/ with front vocalism, and leading to cases such as Chuv. kĕrpe, dial. kör̆ pe ‘grain’, Tat. körpä ‘bran’ ← R крупа́ ‘grain’ (see p. 256). Thus, we might anticipate a front-vocalic substitution in the case of a direct loan from Baltic. On the other hand, Volga Kipchak *-z would be difficult to explain starting from a Baltic nom.sg. *kârjas. Furthermore, Mari *käräš cannot be understood as a Volga Kipchak loan, as *-z should have been preserved in Mari, cf. e.g. Ma. E teŋə̑z, W taŋə̑ž ‘sea’ (cf. Tat. diŋgez, Kyrgyz deŋiz). As a result, the relationship between the various Volgaic forms is difficult to establish, and if there is any connection with Baltic at all, the exact route of borrowing cannot be recovered. However, it seems quite evident that Mordvin adopted this word specifically from Tatar. On the further relationship with Gr. κηρός, see pp. 248–249. ‘far’. Md. E talaj ‘quite a while (ago)’, cf. talajs ill.sg. ‘for long’, talajste elat.sg. ‘from a distance’, M talaj ‘quite a while, quite far’ ~ Lt. tolì, toliẽ, obs. tõl ‘far, distant’32 (Grünthal 2012: 333) — Grünthal suggests this Baltic source as an alternative to the older Turkic etymology (MdWb 2258–2259), which compared 32 The Lt. word generally refers to distance, but may also have a temporal reference (e.g. tolì priẽš ‘long before’). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 134 chapter 4 Kazakh talaj, Kyrgyz dalaj ‘a few, quite a lot; often’. Grünthal’s main criticism is that SSA (i: 138) does not mention any “corresponding words in the Turkic languages of the Volga region”. However, a glance in the Tatar and Chuvash dictionaries reveals that the word is indeed present there: Tat. talaj ‘quite a lot’ (ТРС ii: 302), Chuv. poet. ⟨талай хирне⟩ ‘to distant lands’ (Скворцов 440, s.v. талай ii; see also Федотов ii: 167). With regard to the semantics, we can note that the Turkic words can be used in certain case forms with a temporal and spatial reference; compare Tat. talajga ‘for (too) long’ (-ga dat), Kaz. talajdan beri ‘for a long time’ (-dan abl, beri ‘to here’), talaj žer ‘far away’ (žer ‘space’). As the Turkic etymology is phonologically trivial and raises no serious semantic issues, it should be preferred over Grünthal’s Baltic etymology. ‘yard’. Md. E kardaz ‘yard, stable’, M kaldaz ‘stable, pen’ (?< *kardas)33 ~ Lt. gar̃das ‘enclosure, stall’, Lv. dial. (SW) gãrds ‘pigpen’ (Tomaschek 1883) — The Mordvin word cannot be separated from Md. E kardo, M karda ‘stable, pen’, which Paasonen (MdWb 619), no doubt correctly, derives from Chuvash karta ‘stockyard, stable; fence’. The word is also found in the Volga Kipchak languages, viz. Tatar kirtä, dial. kärtä ‘pole; fence, enclosure’, Bashkir kärtä ‘pole, fence; stockyard’, and the Chuvash word was also borrowed into Komi karta ‘stable, barn’. These cannot be separated from a group of similar words in the Caucasus, cf. Oss. I kært ‘yard, estate’, D kært(æ) ‘stockyard’, Ingush kart ‘fence’ (Абаев 1958: 586–587). The question is whether Mordvin *-as can be seen as a suffix. While it is not a productive derivational element, such a suffix must be present in Md. E ńeŕgaz, M ńäŕgaz (< *näŕgas) ‘badger’, which is etymologically related to Mari E nerɣe, W nerɣə (< *nirgə) ‘badger’.34 There are numerous other Mordvin nouns ending in *-as, but very few can be reliably analysed (see Maticsák 2014). Nevertheless, as the word for ‘badger’ shows, the presence of final *-as is not sufficient to guarantee an Indo-European origin (pace Wälchli 1997: 307). 4.2.3 Acceptable Comparisons Despite the large number of rejected or doubtful comparisons, we are still left with a corpus of formally and semantically acceptable loan etymologies. A couple of these examples are also present in Finnic, and have therefore already 33 34 For a discussion of the Moksha -l-, see van Pareren (2008: 89–90). The relationship between Volgaic *närkä and Finnic *mäkrä (> F mäyrä, E mäger, määr, Li. mä’ggõrz) ‘badger’ is unclear, but a relationship looks possible: the irregular correspondence perhaps suggests a shared substrate word. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 loanwords into other uralic languages 135 been discussed elsewhere in this work. As a result, a simple list will suffice. Note that none of these examples can be considered certain evidence of direct contact between Baltic and Mordvin as the Baltic words themselves are of uncertain origin: – Md. E kšna, M šna (< *(šə)šna) ‘worked leather; leather strap’ ~ Lt. šikšnà, Lv. siksna ‘untanned leather; leather strap or belt’ (see pp. 87–88) – Md. E ľepe, M ľepä (< *lepə) ‘alder’ ~ Lt. líepa, Lv. liẽpa ‘lime tree’ (see p. 89) – Md. E dial. ťožań, M ťožäń (< ?*ťožən : *ťožəm-) ‘thousand’ ~ Lt. tū́kstantis, Lv. tũkstuôtis ‘thousand’ (see 3.5.4) – Md. E malaso, M malasa iness.sg. (< *mala-) ‘near’ ~ Lv. mala ‘edge, shore, boundary’ = F dial. malo ‘edge, flank’, Sá. L muolos ‘shore lead’ (see p. 122) The remaining cases are unique to Mordvin, and must be discussed separately. In doing so, it is important to evaluate not only the plausibility of the comparison, but also the etymological background of the suggested Baltic source. Only those with a clear Indo-European etymology can provide objective evidence in favour of a loanword from Baltic into Mordvin. Those of unclear ultimate origin are presented here in italics. ‘bast’. Md. E ľenge (pl. ľengt), M ľengä (< *leŋgə, obl. leŋg-) ‘bast’ ~ Lt. dial. lùnkas, Lv. lûks ‘bast’ — The Mordvin form could reflect an earlier *lüŋkV (cf. Aikio 2009: 116). Surprisingly, this semantically and formally attractive etymology is rejected by both van Pareren and Grünthal.35 The Baltic word is of unclear origin (see pp. 181–182). ‘bridle’. Md. E panсt, panst, M pandәz (< *pandəs) ‘bridle’ ~ Lt. pántis ‘hobble, fetter’ (= Pr. E panto; OCS пѫта pl. ‘fetters’) (Tomaschek 1889: 11) — While the semantic match is not exact, both bridles and hobbles are tools used to restrict a horse’s movement. The Lithuanian word is probably derived from the verbal root seen in Lt. pìnti, Arm. henum ‘weave’ (LIV 578–579). ‘knife’. Md. E pejeľ, M pejəľ (< *pejəľ) ‘knife’ ~ Lt. peĩlis ‘knife’; Pr. E kalopeilis ‘cleaver’ (Tomaschek 1883) — The scepticism of van Pareren (2008: 113– 114) is hardly justified, as the etymology appears formally and semantically straightforward. The Baltic word lacks an etymology; the older connections with Lt. pielà, R пила́ ‘saw’ are abandoned in recent sources (ALEW 862; Smoczyński 2018: 954). 35 Both suggest a native origin. Van Pareren (2008: 103–104) assumes a derivational relationship with Md. E ľejks ‘young alder (whose bark has been stripped)’ and ľevš ‘bast’, although a detailed morphological analysis is wanting. Grünthal (2012: 321) follows Mägiste (1962) in equating *ľengə with F niini, Komi ńin ‘bast’ (< *nijnə); this, however, leaves the stem-final velar unexplained. As an alternative, Grünthal adduces E luvoďe-, M luŋgə̑ďə- (< *luŋəďə-) ‘flake off; fade’ as a comparandum for the same Baltic word, but the unexpected substitution *nk → *ŋ and less obvious semantics makes this comparison less attractive. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 136 chapter 4 ‘millet’. Md. E suro, M sura (< *surə) ‘millet’ ~ Lt. sóros, Lv. dial. sûra², (17ᵗʰ c.) sāre ‘millet’ (Tomaschek 1883) — The Baltic word lacks a clear etymology. For a detailed discussion of this comparison, see pp. 261–263. ‘other’. Md. E ľija, M ľijä (< *lija)36 ‘other’ ~ Lt. liẽkas, Lv. lìeks ‘surplus’ (Paasonen 1909a: 89) — In Uralic terms, the Mordvin words could reflect *lekä or possibly *lejkä, which would both be reasonable substitutions of an East Baltic *lẹ̄ka-. An earlier form *laika- is more difficult, but I would not rule out a pre-Mordvin reconstruction *läjkä (compare Md. E śiśem, M śiśəm ‘seven’ < *ćäjćəmä; Aikio in prep. 119). The semantic difference seems to be bridged by the derived verb E ľijado-, M dial. ľijadə- ‘stay behind, remain’, which precisely corresponds in meaning to the Baltic verb seen in Lt. lìkti (3pres. liẽka) ‘remain, be left over’ (van Pareren 2006: 36, but sceptically 2008: 105). ‘soot’. Md. EM sod (< *sod) ‘soot’ ~ Lt. súodžiai, dial. súodys pl., Lv. obs. suods (EH ii: 610), dial. suôdri² ‘soot’ (Paasonen 1909a: 127) — A Baltic origin is rejected by van Pareren (2008: 122) and Grünthal (2012: 308) due to the existence of a native etymology. However, neither Mari E šüć, W sə̑ts ‘soot, coal’ nor Komi sa, Udm. su ‘soot’ (UEW 769) represent a phonological match,37 and they cannot be accepted as cognates. The substitution *ō → Md. o can be considered reasonable so long as the loanword postdated the pre-Mordvin change *o > *u. The Baltic word is cognate with R са́ жа, Sln. sáje ‘soot’ and further OE sōt ‘soot’.38 ‘thunder’. Md. E puŕgińe, dial. piŕgińe, M dial. puŕgəńä (< *puŕgəńə) ‘thunder’ ~ Lt. perkū́nas, Lv. pȩ̄r̀ kuôns ‘thunder’, also a theonym; Pr. E percunis ‘thunder’ (Tomaschek 1883) — The Mordvin vocalism must result from a metathesis, which could be motivated by the lack of rounded vowels in non-initial syllables in Proto-Mordvin (van Pareren 2008: 119). The palatalized suffix is probably to 36 37 38 The final -ä in Moksha is due to a secondary fronting of final -a after a palatal consonant (Bartens 1999: 63), cf. M pŕä (pŕa- in inflected forms) ‘head’ < PMd. *piŕa. Initial š- in the Malmyž dialect points to PMa. *š- (Wichmann 1906: 21; TschWb 740), suggesting the Mari word is instead cognate with Md. śeď ‘coal’ (< PU *ćüďə ‘coal’; for *d/*ď > PMa. *ć compare tić < *täwdə; Metsäranta 2020: 43; however, Aikio in prep. 147 adduces a different Mari cognate in this dataset: E šüj, W šü ‘charcoal’; perhaps *šü̆ć ~ *šü( j) is the result a paradigmatic split?). The correlation Komi a ~ Udmurt u does not usually occur in inherited words except where it is a reflection of *-eCə#, cf. Живлов 2013. Metsäranta (2020: 140–141) has attempted to substantiate a preform *setə by comparing the verb Komi so̮ t-, Udm. suti̮- ‘burn, set on fire’, allegedly < *set-tä- (differently on this verb see Aikio 2021: 169–173). Since I do not think that a lengthened grade yielded acute, the Balto-Slavic form (*sod-i-) cannot be directly equated with Germanic *sōta- (< *sōdo-), but both words probably derive from the root *sed- ‘to sit’ via the sense ‘sediment’. OIr. suide* ‘soot’, is to be derived from *sūdi̯ā- in view of Modern Irish súiche, Catalan sutge ‘soot’ and cannot be directly related (see Walde/Pokorny ii: 485; Zair 2012: 125, with lit.). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 loanwords into other uralic languages 137 be attributed to assimilation to the Mordvin diminutive suffix *-əńə.39 Despite REW (ii: 345–346), the Baltic words cannot be separated from OR Пероунъ ‘thunder god’, Pl. piorun ‘lightning’, which show an irregular correspondence with Baltic, most probably pointing to a foreign origin. ? ‘forest’. Md. EM viŕ (< *viŕ) ‘forest’ ~ Lv. dial. (Vidzeme) vẽris ‘spruce forest’, dial. ‘riverside meadow’ (Grünthal 2012: 336) — While the comparison seems attractive at first sight, the vowel substitution seems suspect if we assume Latvian ē continues Proto-Baltic */æː/. Some cases of Md. *i deriving from *ä do occur, but this appears to be conditioned by a preceding palatal (Aikio in prep. 114). No Indo-European etymology is suggested by Endzelīns (ME iv: 562),40 while Karulis’ (ii: 508) comparison with Gr. εὑρύς ‘broad, wide’ is neither semantically nor phonologically convincing. ? ‘left’. Md. E kerš, M kerži (< *kerš) ‘left’ ~ Lv. krèiss ‘left’ (Viitso 1990: 141; van Pareren 2008: 93; Grünthal 2012: 316) — Although semantically attractive, this etymology requires some assumptions on the phonological side. First, it must be assumed that the inadmissibility of initial consonant clusters resulted in a metathesis. While imaginable, reliable parallels are few (cf. Md. EM turba, dial. truba ‘horn’ ← R труба́ ). The second assumption is that Lv. -s- reflects Baltic *-š-. True, the traditional equation with Lt. kréisva ‘flaw’ (LEW 203) would imply Baltic *s, but due to the mismatch in accentuation, it is uncertain that the Lithuanian word belongs here. The Latvian word is apparently related (with metatony?; Derksen 1996: 190, 196–197) to kreĩlis, ķeĩris ‘left-hander’ (< *kreirīs) and further Lt. kreĩvas, R криво́ й ‘crooked’ (LEW 203; ALEW 523; Smoczyński 2018: 598). ? ‘salt’. Md. EM sal ‘salt’ (< *sal)41 ~ Lv. sā̀ls ‘salt’ (= Gr. ἅλς) — The Mordvin form cannot be directly equated with Finnic *sōla (despite Напольских 2015: 163–164), as the Finnic stem type *ō–a is of recent and secondary origin (cf. Plöger 1982), but it may be analysed as an independent loanword from Baltic with the vocalic substitution *ā → *a (Holopainen 2019: 215). On the other hand, the analysis as a direct Baltic loan is rendered somewhat uncertain by the Permic evidence (Komi sov, Upper Sysola so̯ l, Udm. si̮lal ‘salt’ < Proto39 40 41 Van Pareren posits a Baltic source *perku-, citing Narevian pjarkuſ (Zinkevičius 1985: 77). The controversies around the Narev glossary aside, this cannot be considered evidence of a shorter form; the loss of *n before final *-s is paralleled by garſ ‘stork’ ~ Lt. garnỹs. Endzelīns suggests a loan from Estonian veer ‘edge’, but only for the sense ‘riverside meadow’. Incidentally, this Estonian word has been considered cognate with Mordvin *viŕ (UEW 820–821). However, this is not phonologically acceptable; Aikio (2012b: 234) reconstructs the former as *wärə and equates instead Md. E veŕe, M väŕä ‘above, over’. The reconstruction *sal may be preferred over *salə in view of Sal · Zout in Witsen 1785; cf. Pystynen 2020b. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 138 chapter 4 Permic *so̯ l), which must also be related, but lacks a clear source. Apparently, we are dealing with an ancient Wanderwort of ultimately IE origin. For Mordvin, a proximate Baltic source is possible, but other possibilities are imaginable. ✝ ‘duck’. Md. E dial. šenže, šenš ‘duck’ ~ Lt. žąsìs, Lv. zùoss ‘goose’ — Due to its limited attestation, a Proto-Mordvin reconstruction is difficult. The only reliable attestations provided by Paasonen (MdWb 2227) derive from the Kadom and Kaljaevo dialects, which happen to be the same dialects which show a raising *a > e in šenžej, šenžij ‘spider’ (< *šanžəŋ; cf. Paasonen 1903: 81). Therefore, as well as *šenž- or *šänž-, a reconstruction *šanž- can also be considered. The latter would allow for a direct equation with Komi dial. (Udora) če̮ž, Udm. če̮ž ‘duck’.42 By contrast, Sammallahti (2001: 398) has reconstructed *šänšä and treated the Mordvin word as a regular cognate of PF *hanhi ‘goose’, which is a Baltic loanword. This preform was later substantiated by the sound law PU *ä(–ä) > PF *a(–e) (cf. Heikkilä 2014: 86). If this is correct, this loanword would have to predate the other Baltic loanwords in Finnic (cf. the preservation of *ä–ä in *mäntä ‘stirring stick’, *härkä ‘ox’). The most awkward aspect of this is that preFinnic *šanši is closer to Baltic *žans(i)- than the suggested West Uralic *šänšä. If we instead assume that Erzya šenže was an independent loan from Baltic (Grünthal 2012: 331), then we would have to assume that an identical assimilation *š–s > *š–š took place independently in Finnic and Mordvin.43 However, as with Finnic, this assimilation is potentially regular, as there do not seem to be any Proto-Mordvin words with *š–s. Nevertheless, the native etymology, on balance, seems more convincing. 42 43 The development *e̮(–ə) > Komi/Udmurt e̮ is possibly regular before a resonant (Aikio 2012b: 241), while the loss of the nasal is regular. Aikio (2015a: 57) instead compares the Permic data with the Ob-Ugric words for ‘mallard’, reconstructing *če̮čə. Of these, Khanty *čāč (> Vakh-Vasjugan čač, Kazym šɔš) could potentially also reflect *če̮nčə (the development *-nč# > *-č is not regular, but paralleled by *poč (> Vakh-Vasjugan poč, Surgut pŏč) ‘back (of the head)’ < PU *pončə). However, Mansi *šī̮šə (> West šē̮š, South sās) ‘mallard’ does indeed appear to rule out a nasal. While the simplification *-nš > *-š is regular in syllable coda (Pystynen 2020c: 256–257), the preserved long vowel in Western Mansi šē̮š implies a Proto-Mansi open syllable, and is thus not consistent with a nasal. Incidentally, a Permic word has been taken as a loanword from the Indo-European word for ‘goose’: Komi ʒ́o̯ʒ́eg̮ , (Jaźva) ʒ́u̇·ʒ́ok, Udm. ʒ́aźeg ‘goose’ (cf. Holopainen 2019: 377– 378, where either an Indo-Iranian or Baltic etymology are considered). The Permic forms show an irregular vowel correspondence: Komi *o̯ ~ Udm. *a is extremely rare (we expect Udmurt *u). Moreover, the Permic forms, if taken from *źans-, would presuppose yet another assimilation. Rejectable is Koivulehto’s (2001: 244) derivation of PSá. *ćuońēk (> N čuonjá) ‘goose’ from a hypothetical PIE “*ǵʰan-əd-”. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 loanwords into other uralic languages 139 ✝ ‘daughter’. Md. E ťejťeŕ ‘girl, daughter’ beside E śťiŕ, dial. stiŕ, M śťiŕ ‘girl, daughter’ ~ Lt. duktė ̃ ‘daughter’ — These two Mordvin forms appear to be found in almost complementary distribution across the dialects (in MdWb 2384, only the Gorodišče dialect attests both variants). It is generally assumed that the latter represents an irregular reduction of the former, although exactly how this works is unclear to me: particularly problematic is the Erzya form with unpalatalized s-. Since *-k- became *-v- or *-j- in Mordvin depending on vowel harmony, and there are examples of *-kt- > *-vt- in back-vocalic words (e.g. Md. E kavto ‘two’, avto- ‘set a trap’ < PU *kakta, *e̮kta), one might anticipate *-kt> *-jt- in front vocalic words, and reconstruct *tüktär for Mordvin. Unfortunately, this development is contradicted by Md. E ńevťa-, M ńefťə- ‘pluck, tear’ < PU *ńüktä (cf. F nyhtää ‘pluck’, Ma. E ńəkta- ‘skin’). Therefore, the relationship of this word to the Baltic data remains uncertain. ∵ In the above, we have identified 7 plausible and 3 possible loanwords from Baltic into Mordvin. Three of the plausible examples also have a clear IndoEuropean background (‘other’, ‘bridle’ and ‘soot’), which would appear to demonstrate direct, independent contacts between Mordvin and Baltic, and it is possible that some of the other words were also adopted from Baltic directly. Contrary to the conclusion of previous works on the subject, none of the Mordvin words could plausibly have been borrowed through Finnic. While Baltic *lẹ̄ka- ‘surplus’ has been borrowed into both Finnic and Mordvin, the two forms cannot be traced back to a common proto-form, with Mordvin pointing to front vocalism, and Finnic to back vocalism. The words for ‘belt’, ‘alder’ and ‘thunder’, noted at the start of this chapter, encounter similar issues (see the discussions in 3.4 and 3.5.4). With regard to semantics, words in the sense ‘knife’ or ‘bridle’ might well be understood as technological loans and be regarded as characteristic of an adstrate loan context, and this analysis seems most convincing given the small number of loanwords overall. In such a context, however, the words ‘soot’, and in particular ‘other’, are rather unsettling. Specifically, according to the World Loanword Database, ‘other’ ranks among the 300 least likely words to be borrowed. Of course, the loanword proposal presupposes that the word was borrowed in the sense ‘surplus’, with only a secondary shift to ‘other’. In this context, we could compare the Latvian suitāk ‘too much’, Pr. iii zuit ‘genug’, which I have suggested may have been borrowed from Slavic as trade jargon (see 1.1.8). On the other hand, given that Md. *lija could reflect a number of possible protoforms, one may ask whether the loan etymology is even correct. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 140 4.3 chapter 4 Mari By contrast to Mordvin, whose contacts with Baltic have never been doubted, the idea that there are Baltic loanwords in Mari has not been universally accepted. The staunchest opponent to the idea was Mägiste, who in a 1959 article provided alternative analyses for all suggested Baltic loanwords. However, despite his efforts,44 there still remain a small number of potential loanwords which have not been explained away by previous studies. Among those shared by other West Uralic branches, we can note the words for ‘belt’ and ‘thousand’ which have already been discussed (see 3.4 and 3.5.4). The following are exclusive to Mari: ? ‘house’. Ma. EW pört ‘house, cottage’ ← Lt. pirtìs ‘bath-house’ (Thomsen 1890: 208; Kalima 1936: 148) — Several scholars have assumed a Russian origin, instead (Nieminen 1953: 213; Mägiste 1959: 170; Вахрос 1963: 159; Bereczki 1994: 117; TschWb 541). However, this remains problematic, as there is no evidence that the Russian word was ever in use in the Volga region (see pp. 30–31).45 Starting from a Baltic source, the vocalism is not quite clear: note that the Mari vowel would be the usual reflex of PU *e̮ (cf. Aikio 2014a: 131–135). А. Дыбо (2008: 231–232) has suggested an alternative etymology. Contrary to the communis opinio (Räsänen 1920: 259; Федотов i: 462), which takes Chuv. pürt, dial. pört ‘house, cottage’ as a Mari loanword, she assumes a borrowing in the opposite direction, and compares the Chuvash word with Old Turkic barq ‘shrine, temple’46 (usu. in the collocation ev barq ‘house and home’; Clauson 1972: 359–360) and Yakut bïrt (Пекарский 625) ‘wellbeing, wealth’.47 44 45 46 47 Most of his explanations are unsuccessful. Even his claim that Ma. E šukerte, W šukerδə ‘for a long time’ must be segmented šuk-ertə (cf. Ma. E šuk ertak ‘for a long time’, TschWb 729) and therefore not contain a cognate of F kerta, Md. E kirda ‘time, -fold’ (cf. also Grünthal 2012: 317) is perhaps put into doubt by the compound Ma. W pülä-ɣerδə ‘quite a while ago’ (TschWb 573) which would appear to imply the former existence of a word *kirdә. Nieminen and Вахрос claim that the Mari word would prove that the Russian word used to be more widespread, but the sheer geographical distance from the actual Russian attestations makes this argument quite circular. Moreover, the Mari word is a general term for ‘house’, a sense unattested in Russian. Note that R dial. (Vetluga) перт ‘cottage’ (СРНГ 26: 294; Мызников 2019: 599) is a loan from Mari. For the translation, see Hao (2019), who points to a Chinese parallel text which would apparently prove the meaning ‘shrine’ for Old Turkic. As Hao points out, early texts show that a barq is something which can be built, so Clauson’s translation ‘moveable property’ (1972: 359) must be false, but Дыбо’s own gloss ‘здание, постройка’ also appears too general. The same Turkic comparison was also briefly mentioned in a slightly earlier contribution by Мудрак (2007). Yakut -rt is regular from *-rk (СИГТЯ v: 662); compare Yakut kïrt- ‘shear, Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 loanwords into other uralic languages 141 As for the vocalism, the correlation Chuv. ü ~ common Turkic a is found in other words after *b-, cf. Turk. parmak ~ Chuv. pürńe, dial. porńa ‘finger’; Old Turkic bāš (cf. Clauson 1972: 376) ~ Chuv. püśek ‘wound’. Мудрак (1993: 113) has plausibly analysed this as a reflex of the diphthong *ia, which elsewhere has a palatalizing effect in Bulghar.48 The main issue with the etymology is the final -t in Chuvash, which is not regular, but would have to be explained as due to the influence of the synonym śurt, dial. śort ‘house, building’; cf. the compound pürt-śurt ‘household’ (А. Савельев p.c. September 2021). If this etymology is accepted, the similarity with the Baltic forms must be considered coincidental. The picture is further complicated by a similar word in Sámi (N barta ‘hut, cabin’, Sk. põrtt ‘house, cottage, room’), which appears to show regular sound correspondences and is attested in all Sámi languages except South Sámi (Lehtiranta 2001: 96–97, who reconstructs *pe̮rtte̮). The word is also found in Finnish and in the north-eastern dialects of Karelian in the form pirtti ‘cabin, cottage’. The sense ‘bath-house’ is limited to some western Finnish dialects (cf. Вахрос 1963: 159) and is also found in Ume Sámi and in the Swedish loanword pörte (< obs. pyrte ← Finnish). In the more eastern North Finnic languages, we find an irregular e-vowel, cf. K pertti, Vp. perť ‘house, cottage’. This form was at first written off as secondary (Kalima 1936: 70), but later explained as due to Russian influence (Nieminen 1953: 216–217; SSA ii: 350). However, this explanation is quite uncertain, since the underived word is very rare in Russian, and is only recorded in the area of Novgorod and Pskov, which is too far south to have been in recent contact with Veps and Karelian. Furthermore, the usual Russian sense ‘bath-house’ is apparently not recorded for the form *pertti in Finnic. It is universally acknowledged that Sámi *pe̮rtte̮ is borrowed from Finnic (Thomsen 1890: 208; SKES 576; SSA iii: 350; Aikio 2006b: 29). But since the Finnic forms are so narrowly distributed and do not even reflect a common proto-form, one might even suggest that they were loaned from Sámi. The substitution Sámi *e̮ → Finnic *i is a well-attested form of ‘etymological nativization’ (Aikio 2009: 15–16). The substitution*e̮ → *e is less frequent, but also 48 trim’ < *kïrk- (ЭСТЯ vi: 238). Yakut ï (< *a) is a much-discussed issue that I will not enter ̄ ‘load (onto an into here, but I will note some more occurrences before *rt: Yakut dial. ïrt̄ ‘hawk’ (< *kārt-, ЭСТЯ v: 317– animal)’ (Пекарский 3822) (< *ārt-, ЭСТЯ i: 180–181), kïrt 319); note also the derivative sïrdā- ‘grow light’ beside arch. sarā- ‘to dawn’. As a couple of typological parallels for palatalization of a labial being expressed on the vowel, cf. Livonian käpā ‘hoof’ (< *kapja) as against paḑā ‘pillow’ (< *patja) (Kallio 2016: 45) and Tocharian B mit ‘honey’ (< *ḿətə < *medu-) as against śak ‘ten’ (< *ćəkə < *deḱm). This is apparently a result of the fact that palatalized labials are generally disfavoured cross-linguistically (Ohala 1978). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 142 chapter 4 attested; cf. F kelo ‘snag; dead tree’ ← Sámi *če̮le̮ (> Sá. S tjalle ‘tree stump’; Aikio 2009: 77). On the other hand, it is possible that the Finnic variant with /e/ vocalism has spread from Ingrian, where *ir > /er/ is regular, cf. Ingr. kerves ‘axe’ < *kirves (J. Pystynen p.c. June 2023). On balance, the latter explanation appears more likely, as otherwise the origin of the Sámi word remains unclear. Even if we assume a direct Baltic → Sámi loan, it is awkward that the Sámi word does not usually mean ‘bath-house’. To summarize, Baltic *pirt(i)- ‘bath-house’, a word of native origin, was borrowed into ONovg. *пьрть (see pp. 30–31), whence also F pirtti ‘house, cottage’ and (perhaps via Ingrian) Karelian pertti. Sámi *pe̮rtte̮ is most likely from Finnic. On the other hand, there is no clear way to connect Mari pört ‘house, cottage’ to the Baltic and Finnic data, and it has an alternative Turkic etymology which seems just as promising. As a result, this word cannot be considered to offer evidence of direct Baltic loanwords in Mari. ? ‘lynx’. Ma. E šurmaŋše, (Upša) šŭrmŏ, W sə̑rmə̑49 (< *šŭrmə) ‘lynx’ ← Lt. obs. šermuõ (modern šermuonėl̃ is), Lv. sȩr̂mulis ‘stoat’ (Топоров/Трубачев 1962: 248; Bednarczuk 1976: 46; Breidaks 1983: 47) — The Mari word is usually viewed as a Uralic inheritance (Collinder 1955: 8; UEW 490–491; Bereczki 2013: 258–259). However, the suggested cognates are mostly to be rejected.50 If we reconstruct *ćurmə for Mari, we might compare Khanty *ćōrəm (Irtysh ťurəm, Nizjamer śurəm) ‘weasel, marten, stoat’ (for *u(–ə) > *ō, cf. Aikio in prep. 141), although the Khanty affricate remains irregular. On the other hand, Komi dial. śer, Udm. śor ‘marten’ can be combined with the Mari word by reconstructing PU *ćirma (cf. Ma. E užar, W ə̑žar, Komi vež, Udm. vož < *wiša ‘green’, UEW 823; Aikio 2014a: 156).51 The reconstruction *ćirma does indeed bring us close to the Baltic forms. We may get even closer if we compare the apparent “zero-grade” formation širmuonėlis (Baranauskas, Ivanauskas), although since this variant is late and rare, it more likely represents a secondary development (e.g. contamination 49 50 51 On Western Mari s-, see Wichmann 1906: 23–25. Sá. Sk. čõrmm, K čirrm ‘evil spirit; wolf’ (which seem to be irregular even among themselves), on the one hand, and Forest Enets same, Tundra Nenets sarḿikᵊ ‘wolf’ (< *sårmå, Janhunen 1977: 136), on the other, do not match each other, or any of the other forms, in terms of vocalism. Whether *-rm- > *-r- in Permic is regular is uncertain. A parallel could be Komi jir (< *ji̮r, cf. Jaźva jər) ~ Sá. N jorbmi ‘deep spot in water’ (< *jurma, UEW 105). However, this etymology is (implicitly) rejected by Aikio (2002: 47). М. Живлов (p.c. October 2021) has suggested an alternative, and equally acceptable, etymology for the Komi word, comparing Khanty *jɔ̄ r (> Nizjamer jur, Kazym jǫr) ‘river bed’, also dial. ‘deep spot in water’ (OstWb 400), which would presuppose a Uralic *jurə. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 loanwords into other uralic languages 143 with šìrmas ‘grey, dapple-grey’). If the loan etymology is accepted, the directionality would have to be from Baltic to Mari: Lt. šermuõ has an almost perfect cognate in OHG harmo ‘stoat’. However, the potential Khanty or Permic comparanda mean that this can only be seen as one possibility among several.52 ? ‘mink’. Ma. E šaške, W šäškə (< *šäškə) ‘mink’ ~ Lt. šẽškas ‘polecat’ — As discussed on p. 85, the Mari word cannot be considered cognate with Finnic *hähkä ‘mink’; therefore, one might assume an independent loan from Baltic. Chuvash šaškĕ has been taken from Mari (Wichmann 1911: 25; Räsänen 1920: 264), but E. Itkonen (1953: 204; UEW 498) has suggested the opposite direction of borrowing in view of the existence of comparanda in Volga Kipchak, cf. Tatar čäške, Bashkir dial. šäške ‘mink’.53 Tatar č- is unexpected based on the Baltic original. One might assume it arose by dissimilation as in Tatar šešä beside dial. čiša ‘bottle; glass’ ← NP šīša (Ахметьянов 2015 ii: 442), dial. šišta ~ čišta ‘pole for climbing competitions’ ← R шест, gen.sg. шеста́ ‘pole’ (idem: 488). On the other hand, these parallels are inexact, as the variants with č- are in each case purely dialectal, while čäške belongs to the standard language. Moreover, there are instances of an assimilation *č–š > *š–š in Bashkir, including in the homonym šäške ‘cup’ (← R ча́ шка; cf. Ишкильдина 2018: 35). Further support for an initial affricate could be provided by Komi dial. (Udora) ćuš ‘mink’, which could reflect an earlier *ćaškV-.54 In addition, there is a clear resemblance with the narrowly distributed Sámi lexeme Sá. S tjetskie, Ume tjaskie ‘stoat’ (< *će̮ckē) (cf. Wichmann 1911: 25; Лыткин/Гуляев 1970: 314). In Uralic terms, the Sámi word could reflect *ći/üčkä (or *ćička). While the vocalism clearly rules out that the Sámi, Permic and Mari words are cognates, some kind of relationship is conceivable in the context of a shared Wanderwort or substrate word (cf. Junttila 2015a: 31). ? ‘stem’. E wurδο, Volga wŭrδo (?< *wŭrdə) ‘stem, handle’ ~ Lv. vãrde; Lt. vìrdis ‘cross beam for hanging or drying’ — The Baltic words have been compared since Būga (1908: 139) and Ojansuu (1921: 63) with F varsi, E vars, Li. vaŗž (< *varci, obl. varte̮-) ‘stem, handle’. At the same time, the Finnic words are almost 52 53 54 There is also a difference in semantics. Admittedly, Ruhig (i: 148) cites a meaning ‘eine wilde Katze’ for Lithuanian, but the reliability of this gloss is questionable. The “Kyrgyz” (more properly Kazakh) šeške cited in these works stems from Ильминский 1860–1861. Since Ильминский gathered his Kazakh materials in Orenburg and Bashkiria (cf. i: 109), we are probably dealing with a localized Bashkir loanword. The difference between Komi ć /tɕ/ and Tatar č /ɕ ~ tɕ/ is purely notational. Note that here, Proto-Komi *ć should be reconstructed. While a regular development *č–Š > *ć–Š has affected most Komi dialects, Udora has generally preserved č- in these words (Сорвачева/Безносикова 1990: 18). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 144 chapter 4 always equated with the cited Mari forms (Thomsen 1890: 237; E. Itkonen 1953: 159; SKES 1660), implying that the Baltic word has been adopted into Mari as well (thus explicitly Koivulehto 1979b: 142). There are several issues with this theory. First of all, the vocalic correlation between Finnic and Mari is not regular, so direct cognancy between these words is probably to be rejected.55 Secondly, Finnic *varci ‘stem, handle’ could alternatively be cognate with Sá. N veardi ‘mouthpiece of a pipe, handle of a rake’, I verdi ‘shaft’ < *wärtä (cf. E. Itkonen 1977: 6). The suggested Baltic source is semantically rather remote; this same Baltic word could rather be seen as the source of F orsi, E õrs, Li. vȯŗž (< *orci, obl. orte̮-) ‘beam; perch’, which could regularly derive from an earlier *wortə (Nieminen 1963: 238–240; Ritter 1993: 105–106). As the development *wo- > *wŭ- in Mari is regular, cf. Ma. E wurɣem, W wə̑rɣem ‘clothes’ (= Komi vur-, Hungarian varr ‘to sew’ < *worka-; Sammallahti 1988: 551), Ma. *wŭrdə ‘stem, handle’ — provided the reconstruction is correct — could be cognate with Finnic *orci. Thus, we are faced with the awkward situation that the Mari word corresponds phonologically to Finnic *orci ‘beam’, but semantically to Finnic *varci ‘stem, handle’. Since it is unlikely that both of these derive from the same Baltic word, the Mari word cannot be considered a certain Baltic loanword. † ‘rake’. Ma. E šor-wondo (cf. wondo ~ pondo ‘stem, stick’) ‘rake’ ~ Lv. zars ‘branch; prong’ (Aikio 2009: 149) — Above, I have accepted the Baltic loan origin of F haara ‘branch, fork’ (see p. 59) and Sá. N suorri, Sk. sue´rr ‘branch, fork’ (p. 123). At the same time, reference works have further equated the Finnic word with Mari šor-wondo (SKES 57; UEW 783). Semantically, there is no issue; the sense ‘rake’ is even attested in the Finnish derivative harava, and Aikio has previously accepted both the Baltic loan etymology and the Mari cognate. However, Bereczki (2013: 247) has pointed out forms with s- from the Malmyž dialect which would suggest a Proto-Mari *s- and rule out the etymology. Aikio (2015a: 56) agrees with Bereczki and instead proposes a comparison to Sá. N suorgi ‘fork, branch’. Therefore, this Mari word cannot be considered a Baltic loan. 55 Even within Mari, some of the dialects have reflexes of *u rather than *ŭ, e.g. Ma. W wurδə̑. A similar situation is found in the near synonym Ma. E wurɣo, Volga pŭrɣo, W wurɣə̑ ‘shaft’. In both words, we also find an irregular alternation of p- beside w-. This must be the result of decompounding (both words are frequent as second members of compounds; see the lists in TschWb 60–61). Either *p- or *w- could be primary: in the latter case, w- would be generalized from intervocalic position, and in the former, dial. p- would result from hypercorrection. А. Савельев (p.c. July 2023) sees in Mari *wŭrgə ~ *pŭrgə ‘shaft’ a loan from Turkic, cf. Chuvash părăx ‘tube, pipe’. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 loanwords into other uralic languages 145 As a result there is not a single Baltic etymology in Mari which does not have alternative explanations. In the case of the word for ‘mink’, a connection seems probable, but the nature of the relationship is far from clear. Here one might be inclined to side with Mägiste (1959: 176): “Wenn die Anzahl der evtl. balt. Lehnwörter im Tscher[emissischen] nur auf einen einzigen Fall begrenzt ist, dürfte kein Anlaß vorliegen, von balt. Lehnwörtern im Tscher. zu sprechen.” There certainly does not at this stage appear to be any solid evidence that would prove the existence of Baltic loanwords in Mari. 4.4 Permic The situation with regard to Permic is even less promising than with Mari. Here I will leave aside Koivulehto’s (1983: 122–127) proposal that certain ‘preBaltic’ loans (where the source forms are back-projections of Baltic data into Proto-Indo-European) may have been adopted into a common ‘Finno-Permic’ language. These loanwords, if reliable, would simply be too early to describe them as ‘Baltic’ per se. On the other hand, Живлов (2008) has suggested one direct Baltic loanword in Permic. According to him, Komi važ ~ Udm. vuž ‘old’ are derived from Baltic *vetuša- (> Lt. obs. vetušas, Lv. vȩcs) ‘old’. This seemingly attractive etymology has generally been well received (e.g. Pystynen 2016; Nikulin 2016). As Живлов notes, in inherited words, the correlation Komi a ~ Udmurt u is otherwise only observed as a reflex of the sequence *-etə-. Komi va, Udm. vu ‘water’ (< *wetə), Komi ma, Udm. dial. mu ‘honey’ (< *metə).56 It does not necessarily follow from these examples, however, that the conditioning factor was the lost *t; we might, for instance, rather be dealing with a special vocalic development in *CV-type roots (Лыткин 1964: 172).57 56 57 Живлов’s third example za ‘stem, stalk, shaft’, Udm. zu ‘stem of a pipe; axle of a cart’ < *setV is based on an equation with Ma. E šüδür, W šəδər ‘axle; spindle’ (UEW 757–758). However, since the Mari word has š- in the Malmyž dialect, the comparison is most probably incorrect. UEW reject the older comparison of the Permic word with Erzya dial. sad ‘stalk (of the hop plant or cucumber)’, yet this might be more promising. If we reconstruct *se̮tə for both forms, however, we will have to explain the difference between za, zu ‘stem’ and Komi vo, dial. (Upper Sysola) o̯ ‘year’, Udm. wa-pum ‘time, period’ (< *e̮də; e.g. was the lowering to Proto-Permic *å blocked by the w-prothesis?). As there appear to be no monosyllabic nouns in Komi -o̮ , we might entertain a regular development of *e(–ə) > Permic *o̮ in monosyllables followed by a further lowering *o̮ > *a in Komi. This two-stage analysis is supported by the fact that two verbal stems of the shape *Co̮ - have been suggested to derive from *e(–ə); viz. Komi lo̮ - ‘be, become’ (?< *lexə; cf. Metsäranta 2020: 327) and vo̮ - ‘come, arrive’ (< *wexə-; Metsäranta 2020: 146– 147). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 146 chapter 4 True, one of Живлов’s examples is indeed word internal: Komi tar, Udm. tur ~ F teeri, dial. tetri, E teder, Li. te’ddõr (< *tetri, obl. *tetre-) ‘black grouse’ (UEW 794); however, the Finnic–Permic equation is surely incorrect, as in other cases internal *-Cr- has given *-rC- in Permic.58 As the Finnic word has been considered a Baltic loan (Lt. tetervà ‘black grouse hen’, Lv. teteris ‘black grouse’),59 Nikulin (2016) interprets this Permic word as a loan from Baltic, too, setting up a pre-Permic *teδərə. However, it is more likely that the actual source is Iranian, cf. NP taδarv ‘pheasant’, Khot. ttara-, ttatara- ‘(Tibetan) partridge’ (?< *tataru, cf. Khot. pasa- ‘sheep’ < *paću).60 As Metsäranta (2020: 245) notes, the main weakness of this etymology is that the contact relationship depends on a single comparison. The Permic word has traditionally been etymologized as a cognate to F vanha, E vana, Li. vanā ‘old’ (< *wanša: UEW 813; Sammallahti 1988: 544), despite the irregular vocalic relationship. In passing, Aikio (2015a: 33) has mentioned a Samoyed *wåntå ‘old’ as a cognate to West Uralic *wanša. This form does not appear in the appendix to that article, but is apparently based on the Selkup stem *kuə̑ntə- attested in derivatives in Ket Selkup, viz. kwə̑ndəj ‘old’, kwə̑ndəga ‘old man or woman’ (Alatalo 1998: 20, 2004: 293).61 The Selkup word seems to be a phonologically regular equivalent of Finnic *vanha (compare Ket Selkup kwə̑dəgej ‘left’ ~ Estonian vasak ‘left’ < *wasa; Aikio 2015a: 66) and the semantics are ideal, so that a Uralic form *wanša can indeed be postulated. In this light, it becomes even more difficult to separate the Permic word for ‘old’, even if the Komi -a- remains unexplained.62 58 59 60 61 62 Komi bo̮ rd, Udm. burd ‘wing’ ?← Iranian *patra-; cf. Skt. pátra- ‘wing’ (Holopainen 2019: 180); Komi će̮rs, Udm. ćers ‘spindle; axis’ ← Iran. *častra-, cf. Pashto cā́x̌ay ‘spindle’ (Holopainen 2019: 378). The Baltic loan etymology is phonologically problematic. The Finnic word is rather of echoic origin like Eastern Mari küδər, Obdorsk Khanty kutər ‘black grouse’ (*kütrV ?) and Turkic *kürtük (> Shor kürtük, Khakas kürtkü) ‘black grouse’. The same vowel correspondence from Iranian *a is found in Komi dar, Udm. duri̮ ‘ladle’ (~ Skt. dárvi- ‘spoon’), and Komi taśti, Udm. tuśti̮ ‘cup, bowl’ (~ YAv. tašta-, MP (Pahlavi) tštˈ ‘bowl’; Rédei 1986: 68, 78). In addition, certain Iranian loans in Permic have predated the loss of intervocalic stops: Komi dial. gu- ‘steal’ (← *gada; cf. YAv. gaδa-, Pashto ɣal ‘thief’; Rédei 1986: 69); Komi ruć, Udm. ʒ́ići̮ ‘fox’ (← Iran. *ropāća-, cf. Parth. rwb’s /rōbās/, Oss. I ruvas ‘fox’; Palmér et al. 2021: 247). I thank Abel Warries for helping me track down this word. It is tempting to consider it a borrowing from another branch, probably Finnic (Saarikivi 2018: 312). However, the existence of Finnic loans already in Proto-Permic is doubtful, and Metsäranta has considered this proposal “anachronistic” (Metsäranta 2020: 245). From the point of view of vocalism it is possible to assume that Udmurt vuž is inherited, in which case we might limit the loanword proposal to Komi. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 loanwords into other uralic languages 147 Thus, as with Mari, I am led to a pessimistic conclusion as to whether there are any Baltic loanwords in Permic. In both cases, the evidence is very limited and alternative accounts are possible. It therefore does not certainly surpass the threshold of coincidence. 4.5 Conclusion The contacts between Baltic and the other West Uralic branches were by no means of the same calibre as those with Finnic. The evidence as regards Mari and Permic is inconclusive: all of the suggested examples have competing etymologies, and we cannot state with any confidence that any direct contact has taken place. In the case of Sámi and Mordvin, many of the etymologies previously proposed are formally or semantically dubious, and must be rejected. However, even if we limit ourselves to cases where the Baltic source has a clear etymology, there still remain a handful of convincing cases which cannot be rejected. Table 5, overleaf, illustrates the contact situation. Certain, direct loanwords are highlighted in bold. The majority of the loanwords in Sámi are shared with Finnic, and this appears to suggest that the contacts largely took place through Finnic mediation. Nevertheless, at least two direct loanwords have to be accepted. The situation with regard to Mordvin is quite different. In both cases where Mordvin shares a loanword with Finnic, the reconstructed proto-forms cannot be reconciled. Therefore, contrary to the claims of previous research, it does not seem helpful to assume that any of the words entered Mordvin through Finnic mediation. Given the small number of loanwords, we would expect the contacts to have been brief and incidental. However, as I have noted above with regard to Mordvin, the semantics are only partially consistent with this interpretation. Particularly remarkable is the loaning of a word for ‘soot’ into both Mordvin and Sámi, which is difficult to understand in an adstratal trade context. As there is no positive evidence for the presence of the Balts in Fennoscandia, it seems most parsimonious to assume that the Balts came into contact with pre-Proto-Sámi speakers before the latter migrated into the region (contrast the illustration in Aikio 2006a: 45). Kallio (2009: 39) has suggested that the Sámi had already arrived in the peninsula in the late 2ⁿᵈ millennium bce. Similarly, Lang (2018a: 26) has suggested that the Sámi may have begun their migration from the Upper and Middle Volga regions in the latter half of the 2ⁿᵈ millennium. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 148 table 5 Mordvin *kerš *lija *pandəs *sal *sod chapter 4 Baltic loanwords in Sámi and Mordvin Baltic *ansā*kvēpV?← *kreiš← *lẹ̄ka← *pantīs ?← *sāl← *sôd(i)*šal̂nā*žara*šẹ̄na*širvē*daglā- Finnic Sámi → *ansa → → → → → → → → = *vuossē ‘handle’ *kiepe̮ ‘soot’ ‘left’ *līka ‘surplus’ ‘bridle’ *sōla ‘salt’ ‘soot’ *halla = *suolnē ‘hoarfrost’ *hara = *suorē ‘fork’ *he̮ina ≠ *suojnē ‘hay’ *hirvi = *se̮rve̮ ‘elk’ *takla = *tuovlē ‘tinder’ However, there does not seem to be any certain evidence against a comparatively late migration; the earliest loanword evidence from Germanic can be dated as late as the first centuries ce (Aikio 2006a: 39–40; Kallio loc. cit.), and there is no other linguistic evidence that would necessitate such an early arrival of pre-Proto-Sámi speakers. Lamnidis (et al. 2018) have noted that an individual showing Siberian ancestry in Finland (dated 300–800 ce) correlates with modern Sámi populations, but there so far does not appear to be any genetic evidence which would support an earlier arrival of Uralic populations, the first individuals in the Baltic region showing Siberian ancestry being dated to the Final Bronze Age (Saag et al. 2019). There is currently very little ancient DNA evidence from Fennoscandia, however, so it is possible that such ancestry will later turn up. Linguistically, the single example of ‘hay’, if analysed correctly, would show that the independent contacts with Sámi took place at an earlier date than the contacts with Finnic, as the former would have predated the East Baltic monophthongization of inherited *ai. While this is an extremely tentative conclusion, it is possible that the contacts took place further east, closer to the Middle Volga region. Indeed, the contacts between the Balts and pre-Mordvin speaking populations have normally been located in the Volga-Oka region (cf. Grünthal 2012: 299–302), a proposal which has been encouraged primarily by hydronymic evidence (see the discussion on p. 36). The evidence of loanwords in itself is arguably a far stronger argument for a more eastern spread of the Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 loanwords into other uralic languages 149 Baltic languages. However, we should note that, in the absence of any back loans, there is no necessity in assuming that this source language was the direct ancestor of any modern or attested Baltic language. Rather we may be dealing with an eastern offshoot, which would permit us to place the ultimate Baltic homeland somewhere between this contact zone and the Baltic Sea region. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 part 2 Contacts with Unknown Languages ∵ Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 chapter 5 Introduction 5.1 Research History As soon as it became apparent that the Indo-European languages were intrusive to Central Europe, the question arose as to the region’s pre-Indo-European inhabitants. Schrader (1883: 161–162) admitted that a “vor- und nichtindogermanisch” lexical layer should probably be present in all Indo-European languages, yet conceded that it may never be possible to recognize it. In early research, words with a narrow geographical spread were typically explained as chain loanwords, their ultimate source being commented on only vaguely. For instance, Hehn (1870: 177–179) in treating the family of Gr. ἐρέβινθος ‘chickpea’, takes Lat. ervum ‘bitter vetch’ and OHG arawīz ‘pea’ as loanwords from Greek, and the latter as a “Fremdwort aus Kleinasien”. Words for other cultivated plants, like rye and hemp, are similarly taken back to unspecificied “eastern sources” (e.g. Schrader/Nehring i: 440, ii: 226). A more specific hypothesis of contact with an autochthonous European population emerged in the form of the Mediterranean Substrate Theory, which became commonplace in Romance linguistics during the early 20ᵗʰ century (for a prehistory of the concept, see Craddock 1969: 18–22). While the theory originated in Italy, it gained traction after garnering the support of Antoine Meillet, who in an influential article (1908–1909), suggested that a number of words common to Greek and Latin may represent parallel loanwords from another source (see also Hirt 1907: 568). Although Meillet still felt his hypothesis was “nécessairement une part d’arbitraire”, his implicit methodology was clear: if we can exclude cognation or a direct loanword relationship, our only option, aside from rejecting the relationship altogether, is to assume an unidentified source language. Around the same time, another theory was developed in Northern Europe by Sigmund Feist (1910: 350).1 Noting that a large proportion of the Germanic 1 Since the purpose of this chapter is to investigate contact with unknown non-Indo-European languages, theories of unattested Indo-European languages such as “Frühitalisch” (Haas 1960), Alteuropäisch (Krahe 1963; Schmid 1968) and Temematic (Holzer 1989), will remain outside of the scope of this work. Furthermore, I will not discuss theories of contact with other languages of “known” affiliation, such as the Vasconic substrate and Semitic superstrate theories of Theo Vennemann (e.g. Vennemann 2003). © Anthony Jakob, 2024 | doi:10.1163/9789004686472_007 Anthony Jakob - 978-90-04-68647-2 This is an open access chapter distributed under the terms of the cc by-nc-nd 4.0 license. Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 154 chapter 5 lexicon had not been etymologized, he hypothesized that the Germanic people were an “autochthone rasse” that became Indo-Europeanized secondarily. He famously estimated that 30% of the Germanic lexicon is of pre-Indo-European origin, a figure which has been much repeated (see Witczak 1996: 71 fn. 5 and 72 with lit.; Bichlmeier 2016: 319–324). It appears that Feist set the tone, as discussions of the Germanic Substrate Theory have continuously revolved around statistical measures of Indo-Europeanness with an emphasis on negative evidence (i.e. words without an etymology; see for instance Polomé 1986; Salmons 2004; Mailhammer 2008: 152–198; and the sceptical overviews in Bichlmeier 2016; Schuhmann 2016). More concretely, Hirt (1909: 69–70) drew attention to the large amount of seafaring terminology in Germanic lacking an etymology and Feist (1913: 187) noted the absence of widely-distributed fish names reconstructible for Proto-Indo-European (see also Schrader/Nehring i: 321). As a result of these early works, Seewörter have remained central to the discussion of the non-IE element in the Germanic lexicon (e.g. Sausverde 1996; Witczak 1996; Schuhmann 2014). Meanwhile, building more on the ideas of Meillet than those of Feist, the Slovene linguist Karel Oštir developed an eccentric theory which he termed “Alarodian”, following the orientalist Fritz Hommel. Adducing evidence from a dizzying array of languages, he proposed an equally complex system of “uralar[odischer] Stufenwechsel”, based on parallels from Uralic (1921: 24–33). Craddock’s assessment of Oštir’s work as “hopelessly obtuse” (1969: 32) may sound harsh, but as Oštir’s theories clearly did not stand the test of time, it is arguably fair.2 Nevertheless, Oštir holds an important position in the research history, in that he provides one of the most comprehensive catalogues of potential non-Indo-European components in Europe (including — what is relevant for our purposes — material from Balto-Slavic), as well as systematizing the alternations upon which this hypothesis was built in a way which has perhaps not been paralleled since (see Jakob forthc. a.). The Mediterranean Substrate took a methodological step forward with Bertoldi (1932), who approached the issue with a cautious hopefulness: while admitting the impossibility of achieving absolute certainty, in Bertoldi’s view, the way forward was methodological rigour.3 Yet Bertoldi’s austerity was short2 Oštir still gets a mention in the bibliographical notes to various Slavic etymological dictionaries (not only that of his compatriot, France Bezlaj (ESSJ), but also ЭССЯ and ESJS). 3 He warns “Ne pouvant presque jamais atteindre une certitude absolue, la nécessité à plus forte raison s’impose de ne jamais perdre de vue du moins les limites du possible” (Bertoldi 1932: 175). Contrast Schuchardt’s (1922: 21) criticism of Oštir: “er gibt sich keine Rechenschaft über die Grenzen der Erkenntnismöglichkeit”. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 introduction 155 lived, as his successor, Giovanni Alessio, once again embarked on a kind of “substrate expansionism”, speculating that the so-called “Mediterranean” substrate may have left traces as far afield as India (Alessio 1946: 142). With Alessio, the Mediterranean Substrate once again meets the Baltic, although mainly on the level of impressionistic equivalences between toponyms (such as Lt. Lietuvà ‘Lithuania’ ~ Gallo-Latin Letavia);4 and the “Veneti” theory (cf. also Feist 1932; for a discussion of this question, see Priestly 1997). True, he also adduced some more concrete lexical evidence (e.g. Gr. κηρός ~ Lt. korỹs ‘honeycomb’; see pp. 248–249). The first to apply the Mediterranean Substrate theory to Slavic, at least to any great extent, was the Czech etymologist Václav Machek. While he had already shown a willingness to push the Neogrammarian boundaries in the 30s (see Machek 1934), it was in a series of articles on Czech plant names in the mid-40s (Machek 1944–1946) that he began to refer specifically to a preIndo-European substrate.5 A few years later, Machek summarized his ideas on the subject (1950b), incorporating the existing views of the Italian school.6 He emphasizes the importance of comparing entire words, rather than resorting to vague root etymologies, and refers to several kinds of irregular correspondences which could point to a foreign origin (Machek 1950b: 148–151). He would later put his ideas to paper in a monographic treatment of plant names (1954),7 many of which he described as non-Indo-European. However, Machek did not limit himself to plants. He also, like Feist, commented that terms for fish tended to be “undurchsichtige, isolierte Wörter” (1947: 66). In fact, in his posthumously published etymological dictionary (1968), one finds the phrase “asi „praevropské“” (“probably pre-European”)8 so often that one might even be surprised at Kiparsky’s cautious optimism (1959a: 224–225; also 1975: 19), granted that the latter still considers Machek’s work a “kühner Flug der Phantasie”. It seems that the “Czech School” both started and ended with 4 Letavia is a Latinization of Old Breton Letau; for details, see Delamarre 2003: 204–205. 5 As Machek repeatedly stated (1944: 179; 1950b: 160; 1968: 10; Boček/Malčík 2011: 122, 304), he took his term “praevropský” from Josef Janko, who indeed did use the term significantly earlier to denote the pre-Indo-European inhabitants of Europe (e.g. Janko 1912: 140), but without reference to language. 6 This was his first publication on the subject outside of a Czech journal, and it apparently had some impact, drawing the attention of the Romance scholar Johannes Hubschmid and Indologist Manfred Mayrhofer (Boček/Malčík 2011: 303, 486). 7 According to his letters, Machek actually completed this book in 1944 (Boček/Malčík 2011: 485), although considering the relative caution of his contemporary articles on the subject, it seems likely that many of his appeals to substrate origin were added after this date. 8 This phrase also sometimes appears in Holub and Kopečný’s slightly earlier 1952 dictionary. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 156 chapter 5 Machek. While Havlová’s views on Machek’s work are largely positive (1994: 392), her own ESJS takes a considerably more sober stance on the issue, with Machek’s substrate proposals more often relegated to footnotes (as also in Трубачев’s ЭССЯ). At this point, our investigation runs into yet another strand of substrate research, namely the Pre-Greek Hypothesis. Although some wider connections for supposedly Pre-Greek words had been proposed long ago (cf. Kretschmer 1896: 405),9 Kuiper (1956: 221–225) was the first to draw direct parallels between pre-Greek and the Germanic Substrate, and in doing so approached the latter from a new angle. Kuiper’s key innovation was to explain the variation in stem-final consonants often found in Germanic — such as that between voiced and voiceless stops, and between geminates and singletons — to a substrate language. However, it would take his student Edzard Furnée to produce a monographic explication of this “consonant variation” theory, albeit not in connection to Germanic (Furnée 1972). The work of Furnée can in many respects be compared to that of Oštir, particularly the latter’s later work on bird names (Oštir 1930). Furnée’s monograph essentially constitutes a catalogue of consonant alternations in Pre-Greek, a mammoth task of twenty years, the results of which certainly have value in themselves (Dressler 1974: 736). However, his conclusions are marred by the frequent appeal to “exotic” comparanda like Basque, Berber and Caucasian languages (here, we often have Hubschmid to thank),10 even though he did not further develop Kuiper’s North European connections. Also like Oštir, Furnée’s work was generally ignored by later research (see, for instance, the negative reception in e.g. Georgiev 1971; Dressler 1974). The key exception was Beekes, another of Kuiper’s students, who reviewed Furnée’s work favourably (Beekes 1975), and cited the former systematically in his later dictionary (2010), characterizing the scholarly neglect for the author as “a major mistake in Greek scholarship” (idem: xiv). At the same time, he fundamentally disagreed with Furnée in the interpretation of these alternations. While Furnée preferred to see all the variation in the pre-Greek lexicon as the result of expressive alternations within the source language (1972: 89– 90), Beekes interpreted this variation as the result of different substitutions of foreign phonemes (1975: 71; see the similar reasoning already in Kuiper 1968; 9 10 For an extensive bibliographical treatment of the Pre-Greek Hypothesis, I refer to Furnée (1972: 29–79). For instance, following Hubschmid (FEW v: 173), Furnée (1972: 223, 285) connects Greek λάπη ‘scum, phlegm’ with forms in Basque and Berber, and even adds in Finnish lampi ‘pond’. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 introduction 157 Beekes 1969: 193–195). One can of course not help but agree that Furnée’s explanation is unsatisfactory: what is gained by positing an unattested source language when all variation is nevertheless deemed “expressive”? Yet Beekes’ own approach to the issue (as exemplified by Beekes 2010: xiii–xlii and passim; Beekes 2014) is certainly not immune to criticism, either (cf. e.g. Meissner 2013: 6–15; Garnier 2015). In the late 80s and early 90s, the substrate theory suddenly drew a lot of attention from American scholars. Although the most prominent voice was clearly that of Edgar Polomé (for example Polomé 1986, 1990, 1992), this wave of interest was apparently sparked by Eric Hamp’s article about the word for ‘apple’ published several years earlier (Hamp 1979). This word stood at the centre of the debate, largely revolving around the phonological features of the palaeo-European Substrate, with barely a single paper appearing on the subject that did not refer to it (cf. Markey 1989: 591; Hamp 1990: 296; Huld 1990: 398– 400; Polomé 1992: 77–78; Salmons 1992: 268–271). A return to Northern Europe once again represented a return to broad theoretical discussions with little data presentation, and to a large extent, the interest appears to have waned rather quickly. It was around the same time that Kuiper (1995) — without referring to any of the above authors — returned to the debate with a reiteration of his “consonant alternation” theory. Variations in stem-final consonantism (i.e. differences in voicing and gemination) are presented as important recurring features of European substrate words. It is after this publication that we start to see a new “Leiden school” emerge (although see already Schrijver 1991 passim). Kuiper’s “language of the geminates”, as Schrijver (2001: 420) would later christen it, has fed directly into the studies of Beekes (1996: 223–227) and Boutkan (1998, 2003a, 2003b) and the dictionary of Boutkan/Siebinga (2005).11 The key result is perhaps not so much a methodological shift, but more a normalization of the “substrate” concept within Leiden (see also Derksen 1999, 2000; Beekes 2000). In more recent years, several attempts have been made to formulate criteria by which substrate words might be identified. Polomé’s list (1989: 54–55), paraphrased by Salmons (1992: 267, 2004: 315), formed the basis for Aikio’s (2004: 8–9; 2012a: 83), while Schrijver (1997: 294–296) can be considered to have established the Dutch school of thought on the issue (cf. e.g. Lubotsky 2001: 301; 11 After Boutkan’s untimely death, Siebinga continued to pursue the former’s methodology as a “substrate word specialist” in the Amsterdam Etymologisch woordenboek van het Nederlands (see Philippa et al. i: 13). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 158 chapter 5 table 6 Suggested criteria for identifying substrate borrowings Salmons Absence of an etymology Limited geographical distribution Particular semantic fields Irregular correspondences Remarkable word formation Onomastic parallels ✓ ✓ ✓ Schrijver ✓ (✓) ✓ ✓ Aikio Beekes [ ✓] ✓ ✓ ✓ ✓ ✓ (✓) ✓ ✓ based on salmons (1992), schrijver (1997), aikio (2012a) and beekes (2010) Beekes 2010: xxiii). These attempts to formulate sets of criteria imply a fundamental recognition that the assumption of a loanword from an unknown source is better supported if it is backed up by multiple lines of evidence. Most explicit on this point was Schrijver (1997: 296): “If the IE origin of a word is rendered suspicious by a number of criteria, it is usually the cumulative evidence rather than an individual criterium that tips the balance.” However, Schrijver’s most important contribution, and something which forms a great part of this work, is his identification of recurring alternations, most significantly the “a-prefix” (see 7.1).12 It is Schrijver’s work that can be seen as having directly inspired the more recent studies by Kroonen (2012; see also Iversen/Kroonen 2017) and Matasović (2013; 2020). Many similarities between these lists can be observed, which are presented in Table 6, above. The differences in criteria partially derive from differences in the scope and research goals of the respective authors. For instance, Aikio’s criteria are designed as a test for the presence of a substrate layer within a language overall, whereas the other authors attempted to identify characteristics applicable to individual lexemes. The absence of a compelling etymology is, of course, a prerequisite for considering a word non-inherited, and therefore this criterion is implicitly present in the methodology of all authors, and Salmons (1992: 267) is explicit that the absence of an etymology in itself is the weakest criterion. These criteria will all be explored more deeply in the following section. 12 Although I value Schrijver’s methodological rigour, the extra-Indo-European comparisons he has drawn, for instance with Uralic (2001: 422–423), Hattic, Sumerian and Linear A (2018: 361–363), are rather too speculative for my taste. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 introduction 5.2 159 Methodological Considerations The task for much of the remainder of this monograph will be to produce a corpus of likely non-IE borrowings in and around the Baltic branch. It is therefore of vital importance to build a robust and consistent methodology to identify and reject potential data. Above, I have given an overview of the criteria identified by various authors for identifying loanwords from non-Indo-European sources. These are as follows: – Absence of an etymology – Limited geographical distribution – Belonging to particular semantic fields – Irregular correspondences – Remarkable word formation Not all of these criteria are equally strong, however. As discussed above, the absence of a compelling etymology, either as an inherited word or loanword, is a necessary prerequisite for a word to be considered a borrowing from a non-IE source, and this criterion need not be expressed explicitly in our methodology. A similar thing can be said of geographical distribution: words with comparanda on the eastern edge of the Indo-European language family (that is, in Indo-Aryan or Tocharian) can hardly come into question as non-IE loanwords in Baltic. While a geographically limited distribution does not prove a borrowing, as generally acknowledged (e.g. Schrijver 1997: 294), a broader distribution would essentially disprove it. Thus, geography constitutes a “negative criterion”. The formulation of what constitutes “broad” must remain vague, as any strict criterion not deriving itself from the data would be circular; however, it can be stated that the broader the distribution, the less probable it is that we are dealing with a non-IE borrowing. It also goes without saying that we cannot argue for a non-IE origin on the basis of semantics alone. Even words for local plants and animals which cannot have been known to Proto-Indo-European speakers may have native designations. A classic example is the application of the native term ‘elk’ to the indigenous American species Cervus canadensis (cf. Mallory/Adams 2006: 133). On the other hand, the chance of a word for a local species being borrowed is naturally significantly higher than for a basic vocabulary item.13 13 For instance, I have argued above (p. 112) and below (pp. 266–267) that Germanic, Finnic, Sámi, Baltic and Celtic have all borrowed their respective words for ‘seal’ from foreign sources. In addition, Russian не́рпа ‘ringed seal’ is borrowed from Finnic (cf. North Karelian ńorppi; REW ii: 214), and numerous other Sámi words for ‘seal’ have been suspected to be of Palaeo-Laplandic origin (Aikio 2004: 11). This is hardly surprising, given the Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 160 chapter 5 Unusual word-formation has also consistently been identified as a criterion for identifying non-Indo-European loanwords; according to Salmons (1992: 267) it is perhaps the “most powerful”. Yet as Schrijver (1997: 294) points out, it is often difficult to identify loanwords on the basis of affixes alone. Thus, while OCS сапогъ ‘shoe, boot’ does not have a compelling etymology (cf. REW ii: 578; ESJS 795) and contains a relatively infrequent suffix -огъ whose Indo-European background is uncertain, it would be circular to assume the word is of substrate origin purely on the basis of this suffix. After all, such a suffix (whether ultimately borrowed or inherited) has also been applied to native roots in Slavic.14 Similarly, Beekes has regarded οὐρά ‘tail’ as possibly pre-Greek in view of the “typically pre-Greek suffix” in the derivative οὐραχός ‘a foetal organ; apex of the heart, etc.’ (Beekes 2010: 1127). However, the suffix seems to have had some limited productivity: cf. στόμαχος ‘throat, gullet’ to στόμα ‘mouth’ (Chantraine 1933: 403), which Beekes himself accepts as Indo-European (2010: 1408). As a result, unusual affixation and specific semantics must both be considered insufficient indications of non-Indo-European origin. However, both may be used as an additional argument where this hypothesis is supported by other evidence. The only remaining criterion identified by multiple authors is that of irregular correspondence. It is clear that the presence of entirely plausible comparanda which do not regularly correspond to each other remains the most certain indication that a lexeme is of a non-Indo-European origin. Thus, in order to argue that a word is loaned from a non-IE source, we must identify comparanda that are both (a) plausible and (b) irregular. To this end, I have devised the following five-point test: 1. Is the data reliable? A word which is not reliably attested cannot be used as a basis for further analysis. This much is self-evident, but the question is not often explicitly asked, and it is remarkable how often big claims are made on the basis of doubtful data. To take a random example, the reconstruction of the IE word for ‘fire’ as *n̥ gnis might never have happened (or at least not as early) if it were not for the alleged Old Lithuanian “ungnis” (cf. Pedersen 1905: 395; Walde 1910: 377; Walde/Pokorny i: 323; for the form, see Bezzenberger 1877: 42). Yet since this 14 semantics, yet still, Dutch rob and MoLG Rubbe have been thought to represent languageinternal innovations (Philippa et al. iii: 671–672; Kluge/Seebold 770). Thus OR пирогъ ‘fine bread’ (СДРЯ 11–14: 391), Slk. piroh ‘dumpling’ is apparently derived from the root of OR пиръ ‘feast’ (cf. Vaillant 1947: 496–497); R острога́ ‘trident’, Sln. ostrǫ́ ga ‘spur; bramble’ evidently belong with R о́ стрый ‘sharp’ (REW ii: 287). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 introduction 161 word occurs in Bretke only once as against dozens of examples of ugnis (Būga 1923b: 399), there can hardly be any doubt that it was a simple slip of the pen. One of the fundamental flaws of Furnée’s work (and of Beekes’ continuation of it) is the perpetuation of doubtful forms which may represent scribal errors or late variants (for examples see Georgiev 1971; also cf. Nikolaev 2018: 2–4, 19– 20). Furnée makes a point to “resurrect” forms long discarded by philologists as evidence for a particular consonant alternation; the form χέλυμνα ‘tortoise’, attested once in Babrius’ fable The Tortoise and the Eagle, is now regarded a “wohl zu Unrecht angezweifelte Lesart” (Furnée 1972: 247; followed by Beekes 2010: 1623). But so long as an error is equally possible (see e.g. LSJ s.v. χέλυμνα), it is methodologically questionable to use the form as evidence. All things being equal, it is a far bolder claim that a form uniquely reflects a genuine dialect variant than to write it off as a simple error. As a general rule, the greater the importance of a form for the validity of a hypothesis, the more I have endeavoured to check its reliability. While a timeconsuming task, it is undoubtedly a fundamental requirement for any empirical investigation that the raw data used is of a good quality. 2. Do the words belong together? A potential weak spot in any etymological equation which is not kept in check by exceptionless sound laws is that what constitutes a “similar enough” comparandum is necessarily somewhat arbitrary (Schrijver 1997: 296). However, there are a couple of constraints which may be applied here to maximize objectivity. First, the comparisons should be semantically perfect — or at least almost perfect. An increase in semantic latitude leads to an increase in potential comparanda. If we apply the strictest semantic criteria, we essentially compare a group of synonyms in Language 1 with a corresponding group of synonyms in Language 2. In this context, the statistical significance of a potential “match” will vary depending on the level of synonymy exhibited by a particular seme. Matching terms for more specific concepts (such as ‘nose’ or ‘oak’), where synonymy tends to be minimal, will be more significant than those in more abstract semantic domains (‘strike’, ‘sad’, etc.). Moreover, since a relaxation of semantic criteria is likely to be accompanied by an increase in semantic abstraction, any loosening of semantic requirements will cause a disproportionate increase in our corpus of potential comparanda. Several scholars have compared the family of SCr. lȗb ‘outer bark’ with that of R лупи́ ть ‘strip (bark)’ in a substrate context (Beekes 1971, 1996: 221; Derksen 2008: 289, 2015: 296–297; Matasović 2013: 96; Šorgo 2020: 444–445). However, it should be noted that this verbal root in Slavic does not only refer to bark; cf. Slk. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 162 chapter 5 dial. lúpiť, Bg. лю̀ пя ‘shell, peel (of nuts, fruit, eggs)’. The two words also have plausible Indo-European comparanda: the word for ‘bark’ perfectly matches Go. laufs* ‘foliage’ (cf. OHG louft ‘bast’) and Alb. dial. labë ‘bark’,15 and also shows Indo-European ablaut: cf. Lat. liber ‘bark’ (< *luber, cf. Leumann 1977: 89–90) and Lv. luba ‘linden or fir-tree bark; roof shingle, board’ (ME ii: 509). On the other hand, the verb has plausible Indo-Iranian comparanda in Skt. lumpati (med. lupyáte) ‘tear’, Khot. rrv- ‘remove’ (Emmerick 1968: 117), MP rbʾy/rubāy-/ ‘rob, snatch’ (cf. Lv. làupît, Pl. łupić in the sense ‘rob, snatch’), whose semantics do not support an original connection to ‘bark’.16 As a result, treating the two words as variants of each other would seem unwarranted, and the partial semantic convergence within Slavic can be interpreted as secondary. A second constraint concerns what I will term “string length”. Not only is it important to compare entire words rather than abstracted “roots” (Machek 1950b: 148), the more linguistic material compared, the less likely it is that the similarity is coincidental (Holzer 1989: 22–26). This may provide an answer to what Simon (forthc.) has termed the “deus / θεός fallacy”: how can a methodology built around irregularity elevate itself above a pre-scientific collection of chance lookalikes? Although it is impossible to exclude chance entirely (as even in the traditional method), we might reduce the risk by applying a “string length” constraint. The following is what can be reconstructed for the words for ‘god’ based on internal evidence: table 7 deus θεός Comparison of deus and θεός *d *e *i̯ *u̯ *dʰ *e *? ∅ ≈ = ? ≠ The correspondence between Latin *i̯ and Greek -∅- could be regular, but since Greek -∅- can equally reflect IE *s, this can be labelled an ambiguous cor- 15 16 Çabej (1976: 307; cf. Demiraj 1997: 229) considers labë to be a variant of lapë ‘flap of skin, lobe’, but the assumption of irregular voicing is clearly ad hoc and not supported by the different meanings of the two words. While this etymology is formally and semantically flawless, it is hampered by the existence of another, equally acceptable, etymology for Indo-Iranian, namely the comparison with Lat. rumpō ‘break, burst’, ON reyfa ‘tear, rob’, which is in fact the more generally accepted one (cf. IEW 870; LIV 420). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 introduction 163 respondence. The initial consonant matches in place of articulation, but not manner, so we will call this an irregular correspondence. The only exact equation concerns the vowel *e. As a result, the comparison can be expressed in the form *De(i̯)-. As we know this comparison to be false, let us suppose that this is too little data to prove a relationship. What would be sufficient? Ideally, it would be desirable to mathematically quantify the similarities between words, but simply counting the number of correspondences could potentially give misleading results. It is important to factor in, for instance, the relative frequency of particular phonemes: as the number of possible vowels is far lower than the number of possible consonants, a correspondence in consonantism will in most cases be more significant. As a general, and somewhat arbitrary, guide, however, I would suggest that a comparison can be considered acceptable if at least three segments are equivalents or irregular equivalents, and of course, the more material compared, the more robust the etymology. 3. Is the correspondence irregular? The only positive linguistic evidence for cognancy is the existence of regular sound correspondences between phonemes, and therefore the possibility of reconstructing a common proto-form. Likewise, the only positive linguistic evidence for a non-IE origin must be considered the impossibility of reconstructing a common proto-form, which in most cases presupposes the presence of irregular sound correspondences. Such irregularities are the most important indication that a word could be of non-IE origin. Thus, the central pillar of my methodology can be called the “principle of irregularity”. Although irregularity has often been considered a criterion for identifying substrate words, in practice, it has not always been viewed as compulsory. Particularly in the context of the Germanic Substrate Theory, the absence of a plausible etymology has often been viewed as sufficient to substantiate a hypothesis of non-Indo-European origin (see the discussion in 5.1), This is exemplified, for instance, by the work of Boutkan/Siebinga, where we frequently encounter phrases such as “[t]he word has no outer-Gmc. cognates and must be of substratum origin.” (2005: 439, s.v. wepin). It seems clear, however, that a positive conclusion cannot be based only on negative evidence. The European word for ‘henbane’, represented by R белена́ , Cz. blín, Sln. blẹ̀n and OE beolone, OS bilina, OHG bilisa ‘henbane’, has come up several times in discussions of possible substrate words (cf. Polomé 1990: 334–335; Philippa et al. i: 316; Matasović 2013: 83). The same idea is also touched upon by Schrijver (1999: 25–26), before concluding that “the matter cannot be decided at present” (idem: 28). In my view, Schrijver’s ambivalence is indeed justified, as all of the evidence in the relevant languages can be explained in terms of IE morphoAnthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 164 chapter 5 logy (Derksen 2019). That does not mean that the word must necessarily be IE, but since it is circular to assume non-IE origin based only on the word’s limited geography, such examples will not be considered in this work. 4. Can the irregularity be explained? Naturally, wherever a potential irregularity is detected, it must first be excluded that we are actually dealing with a regular conditioned development. If this is not the case, then competing hypotheses are likely to involve analogy, contamination, or sporadic sound changes. Of course, such developments do occur, and ideally they should be excluded. In reality, with enough creativity, any kind of irregularity can be explained by such means, and only in exceptional cases will such an account be objectively superior to a loanword hypothesis. Proposing a loanword from an unattested source presupposes the presence of non-IE languages in the vicinity which became extinct before being written down. The more time we assume to have passed between this supposed language death and the start of historical records, the more plausible such a claim becomes. Irregular correspondences between reconstructed protoforms, which necessarily imply a certain time depth, are therefore more likely to point to non-IE loanwords than irregular correspondences between modern dialects. While seemingly intuitive, this has not been a major consideration in earlier works. Indeed, the methodology of Kuiper and Boutkan essentially relies on the uncritical back-projection of modern dialect forms. Thus, Boutkan (1998: 109–110) derives Middle Dutch dorpel, dreppel, drempel, and drumpel, all meaning ‘threshold’, from four distinct proto-forms, assuming these to be parallel loanwords from an unattested source. Whether these are assumed to have been borrowed into the individual Dutch dialects, thus suggesting the unattested source was still spoken during the historical period, or whether they are supposed to have been borrowed already into ProtoGermanic (coincidentally all being preserved into Middle Dutch), the flaw in this reasoning is obvious: whatever the explanation for these variants, it is unlikely to exclusively involve an unattested source language. Words of unclear derivation and unusual structure are particularly often subject to irregular “deformations” through folk etymology. This kind of development can affect both inherited words and loanwords: to take a random example, Uk. горобе́ць, Bel. dial. (Polesia) шворобе́й, шурабе́й (Журавлев 1980: 57) ‘sparrow’ irregularly continue Old East Slavic воробии (~ Pl. wróbel, Sln. vrábəc ‘sparrow’), yet such distortions, belonging to the historical period, can hardly be used as evidence of borrowing from an unattested language. In semantic domains such as bird names, one must also reckon with the influence of sound symbolism (Matasović 2020: 332–333). Irregular alternaAnthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 introduction 165 tions are generally common in the domain of expressive vocabulary, cf. the voiced-voiceless pairs R бры́ згать ~ пры́ скать ‘splash, sprinkle’, Pl. deptać ~ R топта́ ть ‘stamp, tread’, Lt. pam̃ pti ~ dial. bam̃ bti ‘swell, grow fat’ (cf. Liewehr 1956; Kiparsky 1968: 74). Similar ‘expressivization’ is presumably responsible for cases like Lv. šļaka beside slaka ‘drop (of liquid)’, MLG slagge ‘drizzle’ (pace Boutkan 2003b; cf. Endzelīns 1923: 137)17 and Lt. šmagóti ‘whip’ beside smõgti ‘strike’ (see LEW 647–648 and Fraenkel 1955: 12–13 with further lit.). An interesting case from a methodological point of view is the word for ‘lip’ attested in Lt. burnà ‘mouth, face’, Bg. dial. бъ̀рна ‘lip’ (only South Slavic),18 with a variant in p- limited to Latvian pur̂ns ‘face, snout’.19 In view of the Latvian evidence, this word has been considered a loan from an unknown source (Matasović 2013: 91; Derksen 2015: 106). However, we are faced with a similar question: does this mean the word was borrowed independently into Lithuanian and Latvian? Does this mean that non-IE groups were still present in the Baltic region after the break-up of Proto-East-Baltic? A more plausible explanation was provided by Kiparsky (1968), who attributes the Latvian word to a Finnic substrate. Indeed numerous examples of voiced-voiceless pairs occur in Latvian, and examples like Lv. pàtaga beside Lt. botãgas ‘whip, goad’ ← MBel. батогъ (ГСБМ i: 202) ‘cane (for punishment)’ certainly do lend themselves to such an explanation (cf. Li. pǭtõg ‘whip’; similarly ME iii: 190; note also Endzelīns 1923: 183). A similar alternation is found in Lv. dial. teĩba ‘chub, dace(?)’ beside dial. (Talsi) deĩbiņa ‘brown trout’, itself a back-loan from Li. teib ‘ide’ (< PF *stäipi, from a predecessor of Lv. obs. stiepats ‘dace’, see p. 97). As for pur̂ns, no Livonian equivalent is recorded, but an actual lexical loan in Livonian is not necessary for the assumption of substrate influence on a phonological level. This particular word belongs to a category of affective and 17 18 19 Although, as Boutkan (2003b: 246) himself admits, a German influence is difficult to exclude; cf. MoHG dial. (DWb xv: 254–255) Schlack ‘damp mass; heavy raindrop; mix of rain and snow’; cf. Lv. ſ̷chłahka ‘Regen und Schnee’ (Ulmann 1872: 296). Bg. бъ̀рна, dial. бъ̀рла, Mac. брна ‘lip (of an animal)’, SCr. bȑnjica ‘muzzle’, dial. ‘ring inserted into an animal’s snout’, (Čak.) brnjȕse f.pl. ‘moustache’ (cf. Boryś 1977), Sln. bŕna ‘a kind of carnival mask’, see ЭССЯ iii: 129–130. Slovak poet. perna ‘lip’, adduced by Machek (1961: 356), was accepted with enthusiasm by ЭССЯ (iii: 130; see also Derksen 2015: 106), but cannot belong here. The development *-r̥ > -er- before non-palatalized consonants is limited to a narrow group of East Slovak dialects (Krajčovič 1975: 129), while perna is only attested in western Slovakia (cf. SSN s.v.). Furthermore, one has to assume an additional ad hoc irregular development *-rn- > *-rto get the standard Slovak term pera ‘lip’. A derivation of the latter directly from *pъrna, per ЭССЯ, remains entirely fantastical. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 166 chapter 5 low-status vocabulary where substrate influence is common (cf. Aikio 2009: 47).20 In view of this, it is questionable whether the voicing alternation in this word can support the hypothesis of a non-IE source. However, such explanations are only rarely possible: the influence of a Uralic substrate may work for Latvian, but it can hardly apply to examples with a broader distribution (contra Schrijver 2001: 420–424; Andersen 2003: 68–71). 5. Is the irregularity paralleled? A final and very important step brings us back to the work of Oštir and Furnée: any kind of irregular correspondence is rendered considerably stronger if it can be supported by the existence of parallels. An important distinction between my approach and that of my predecessors, however, is a focus on the geography of an irregularity.21 Where a geographical distribution for the various reflexes can be identified, this strongly supports both the validity of this alternation, and the notion that it could reflect genuine dialectal variation in the source language. Considering that the various sub-branches of Indo-European cannot have been situated in the same time or place, we should also not expect the various non-IE substrates underlying them to be identical either (cf. Meissner 2013). In the Leiden substrate school, irregular correspondences have usually been explained as the result of different adaptations of a foreign phoneme (Kuiper 1968; Beekes 1969: 193–195, 1975; Schrijver 1997). For instance, in discussing examples of an alternation *ai ∞ *a in Celtic and Germanic, Schrijver (1997: 306–307) sets up a substrate phoneme */aə/, which is essentially a compromise between the two attested reflexes. The fact that the only other possibility considered by Schrijver is that *ai ∞ *a could represent a “morphophonemic alternation” within the source language illustrates that he took the homogeneity of the supposed substrate language for granted. However, so long as we are dealing with parallel loanwords, it is highly improbable that the source language in both cases was identical. 20 21 For a case study of Quechua loanwords in a variety of Bolivian Spanish, see Babel 2016. Compare also the Yiddish substrate words in (American) English: klutz ‘clumsy person’, schlep ‘haul, carry’, schmuck ‘contemptible person’, and, relevant here — schnozz ‘nose’. The role of affective words in the context of linguistic substrates is unfortunately not discussed in the recent handbook by Haspelmath and Tadmor (2009). This is partially dictated by the methodology of the World Loanword Database, which focuses on a fixed set of basic meanings, generally not extending to the realm of affective words. Attention has been paid to the geography of irregular alternations in works attempting to prove specifically Indo-European substrates, such as Holzer (1989). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 introduction 167 I would favour a more pragmatic approach. Since we cannot precisely identify the cause of irregular correspondences attributable to parallel borrowing, we can merely refer to the alternation itself and attempt to identify patterns in the material. A geographical distribution may favour the “dialectal” interpretation, but the reality may in fact be more complex, as we have next to no knowledge of the linguistic landscape of pre-Indo-European Europe. It is possible, for instance, that a loanword was mediated by yet another unattested language. Since these discussions will always remain on the level of speculation, they need not be pursued here any further. Due to the potential complexity, I would not consider the absence of a geographical distribution to disprove the validity of an alternation, but it may cause us to doubt the coherence of the material. Schrijver identifies a non-IE a-prefix in the Germanic words *amslōn- ‘blackbird’ and *arut- ‘ore’. These two examples fit together very well (see further 7.1), but it does not follow from this that all unexpected *a-’s in Indo-European should automatically be considered related. Schrijver’s further comparison of Greek (Cretan apud H.) ἄκαρα ‘legs’ with MW gar (pl. garreu) ‘leg, shank’ (1997: 310; 2018: 362) shows a very different geography, and it would be very risky to draw a direct parallel — this at least should not be our default assumption. 5.3 Excursus: Illegal Root Structures Although the impossibility of reconstructing a word for Proto-Indo-European normally implies the correspondences are irregular, in a few cases, this might be implied by the root structure itself. In this small excursus, I will discuss two structural issues which could serve as additional evidence of a non-IndoEuropean origin in certain cases. 5.3.1 *T_Dʰ Root It is generally accepted that Proto-Indo-European had a phonotactic limitation against roots containing both a voiced aspirate and a voiceless stop (e.g. Meillet 1912: 60; de Vaan 1999). Due to the merger of the voiced and voiced aspirate stops in Balto-Slavic, external evidence is sometimes required to demonstrate such a root structure, such as in the case of Lat. fax ‘torch’ ~ Lt. žvãkė ‘candle’ (whose vocalism is also problematic; cf. de Vaan 2008: 207–208 and 7.6).22 22 Similarly, Lat. fracēs ‘olive pomace’ and falx ‘sickle, scythe’ imply an illegal root struc- Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 168 chapter 5 If we do not accept a phoneme *b (see the discussion on p. 269), it follows that no Balto-Slavic root containing both a voiceless consonant and *b can be inherited. For instance, Pr. iii kaāubri (for *kiāubrin?) acc.sg. ‘thorn’ has been compared with OSw. hiupon pl.? ‘rosehip’, OHG hiufo ‘thornbush’ (cf. Stang 1972: 27). If *b is not reconstructed, the only possibility would be to reconstruct *keubʰ-nV- for Germanic (with Kluge’s law), but *keubʰ- is an illegal root. In this case, however, we must concede that the Prussian form, a hapax containing at least one obvious misprint, is hardly reliable enough to use. One possible case is the comparison between Pl. kobuz ‘hobby, Falco subbuteo’, USrb. kobušk ‘red-footed falcon’23 and ON haukr ‘hawk, falcon’, OHG habuh ‘hawk’, the first syllable of which implies *kobʰ-, which does indeed imply an illegal root structure. The suffix syllable is also curious. While almost all the Slavic forms continue a form *kab-ice- (trad. *kobьcь): Slk. kobec, Sln. skóbəc, dial. kóbəc ‘sparrowhawk’, R ко́ бчик ‘red-footed falcon’, this could be explained as the result of suffix replacement; compare for instance Slk. vrabec, Sln. vrábəc ‘sparrow’ as against the (probably older) Pl. wróbel.24 On the other hand, Polish kobuz seems difficult to explain as secondary. In theory, the -z could be seen as a direct reflex of *ǵ and be compared directly with the Germanic *-k-, but the implied ablaut pattern *kobʰouǵ- : *kobʰuǵ- does not look particularly Indo-European. As a result, even though a paper reconstruction is possible in Indo-European terms, both the root structure and suffix make it probable that we are dealing with parallel loanwords into Slavic and Germanic. 5.3.2 Clusters of Three Consonants in Roots It may also be put forward that Indo-European had a constraint against roots ending in three consonants (e.g. Schmidt-Brandt 1967: 14–15; Byrd 2010: 107). Beekes, in a discussion of non-IE vocabulary, states that “a root ending in three 23 24 ture. See the discussion on pp. 190–191. See also the discussion of OCS крѫгъ ‘circle’ (?< *krengʰ-) on p. 249. Uk. ко́ буз (Желеховский i: 353) is poorly attested and may well be a Polonism (Berneker i: 536). According to Schuster-Šewc (579), the Sorbian word might itself be loaned from Polish. ЭССЯ (x: 92) cite a variant “*kobъzъ” (= *kabuza-) on the basis of the Russian dialectal hapax кобе́з ‘a kind of small falcon’ (СРНГ xiii: 355; but I could not trace this form — the source given in СРНГ appears to be incorrect!) and the Polish hapax(?) ⟨kobzy⟩ inst.pl. in Mikołaj Rej (see SEJP ii: 303). This data is clearly too unreliable, not to mention that ЭССЯ’s reconstruction fails to account for the Russian form (and the latter could, incidentally, be *кобѣ́зъ). “Probably older” because it is more difficult to explain as secondary. I consider the similarity to Gr. (H) ῥόβιλλος · βασιλίσκος ὄρνις coincidental. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 introduction 169 consonants […] is very rare, but there are a few examples; so it is not a certain indication [of a] non-IE word” (Beekes 2000a: 22). The evidence for roots of this shape is indeed very slim. The following examples can be mentioned:25,26 – *bʰerHǵ-. Lt. béržas, R берёза; ON bjǫrk: Skt. bhūrjá-, Oss. I bærz, D bærzæ ‘birch’ This example seems fairly clear, but the widely accepted link with Skt. bhrā́jate ‘shine, beam’ is only possible if the full-grade in Balto-Slavic and Germanic is secondary.27 A zero-grade is indeed attested in Lt. dial. bìržis ‘birch grove’. – ?*bʰr[e]uHg-. Lat. fruor ‘enjoy’, Go. brukjan ‘need’, OE brūcan ‘use, enjoy; partake’ (LIV 96). The long -ū- in the Lat. participle frūctum is not probative, as *u would have been lengthened anyway by Lachmann’s law (cf. Weiss 1994: 39–40). The Germanic -ū- is most probably a secondary full-grade common in class-two strong verbs (cf. Kroonen 2011b: 112–117). Thus, we can equally reconstruct *bʰreug-. – *delh₁gʰ-. Gr. ἐν-δελεχής ‘perpetual’ ~ Skt. dīrghá-, YAv. darəɣa-, Alb. gjatë, Lt. ìlgas, OCS дльгъ ‘long’, possibly Go. tulgus ‘firm, sure’ The expected full-grade *dleh₁gʰ- is found in Skt. drā́ghīyas-, YAv. drājiiō ‘further’. Furthermore, it cannot be entirely excluded that the Greek form 25 26 27 In nominal roots, suffixation can often not be ruled out. Thus de Vaan (2003: 136) reconstructs ON ǫnd ‘vestibule, entrance hall’, YAv. ąiθiiā- /anθi̯ā-/ ‘door posts’ as *h₂enHt- in view of Skt. ā́tā- ‘door post’, but we may in theory be dealing with a t-stem. Compare similarly Gr. σκῦτος ‘leather’, MW eskit ‘boot, shoe’, OHG hūt ‘skin, hide’ (< *kuH-to-) beside Pr. E keuto ‘skin, leather’ (< *keh₁u-t- / *keuH-t-); cf. Lt. kẽvalas ‘shell’, and YAv. vaēiti- ‘willow’, Gr. οἶσος ‘chaste tree; osier’ (< *uoiH-t-) beside Gr. ῑτ̓ έα, Lv. vîtuõls, OHG wīda ‘willow’ (< *uiH-t-), which may be derived from the root of Lt. výti ‘weave, twine’, Lat. vieō ‘plait, weave’ (IEW 1120–1122); Skt. yū́ṣ-, Lat. jūs ‘broth, sauce’, Pr. E iuse ‘soup’ beside full-grade R уха́ ‘fish soup’, SCr. júha ‘soup’ (< *ieuH-s-), cf. Lt. jáuti ‘throw together, mix’, Skt. yuváti ‘bind’ < *ieuH- (LIV 314). Rejectable examples are: 1. Gr. ῥαιβός ‘crooked, bandy (of legs)’, Go. wraiqs* ‘crooked’ (< *ureh₂igʷ-), but there are plausible alternatives for Germanic (Kroonen 2013: 593); 2. OIr. cairem, MW cryd ‘shoemaker’ (< *kerh₁p-io-?, Matasović 2009: 189–190), on which see Chapter 7, fn. 61; 3. Skt. úpa-valhati ‘puzzle by riddles’, Gr. (Hom.) ἐλεφαίρομαι ‘deceive (vel sim.)’, Lt. vìlbinti ‘allure’ (< *uelh₁bʰ- per LIV 678), but Skt. -h- from -bh- is exceptional (cf. Lubotsky 1995: 127–128), the appurtenance of the Greek form might be disproven by the Myc. personal name erepa(i)ro (Beekes 2010: 409), and the Lt. form may well be of onomatopoeic origin, cf. ulbėt́ i = vilbėt́ i ‘warble, coo; flatter’; 4. Skt. ū́rj- ‘vigour’, Gr. ὀργή ‘disposition; anger’ do not reflect *uorHǵ- (pace Beekes 1969: 241) in view of YAv. vərəzuuaṇt‘invigorating’, OIr. ferg ‘anger’ (< *uerǵ-). The Sanskrit anlaut is probably regular as in Skt. ūrdhvá- = Gr. ὀρθός ‘upright’ (see van Beek 2011: 150–152). Go. bairhts ‘manifest, bright’ is unlikely to belong here, but is instead to be compared with MW berth ‘beautiful, rich’ (< *bʰerǵ⁽ʰ⁾-to-), which most probably rules out a laryngeal, and Alb. (i) bardhë ‘white’ (< *bʰorǵ⁽ʰ⁾-). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 170 chapter 5 directly reflects *-dlh₁gʰ- (see Rix 1976: 73–74; van Beek 2013: 561–563). The historical development of Gr. δολιχός ‘long’ is too obscure for us to base anything on it. – ?*h₁[e]uHdʰ-. Skt. ū́dhar/n-, Lat. ūber, OHG ūter, R вы́ мя ‘udder’ Van Beek has tentatively suggested that Gr. οὖθαρ ‘udder’ is regular from *uHdʰ- (2011: 153–154, fn. 48). If so, this would leave us only the initial glide of ON júgr ‘udder’ as evidence of an e-grade. If this could be secondary, the root may be *(H)ueHdʰ-.28 – ?*keh₂ik-. Lat. caecus ‘blind; invisible’, OIr. cáech ‘blind in one eye’, Go. haihs ‘one-eyed’; Skt. kekara- (late) ‘cross-eyed’ Mayrhofer (KEWA i: 264) considers the appurtenance of Skt. kekara“keineswegs sicher”. Without it, a reconstruction *kh₂eik- would be equally possible (Pronk 2019a: 139). – ?*(H)r[e]uHḱ-. Skt. rūkṣá- ‘rough, dry’, ?OAv. uruša- ‘meagre, emaciated’, OE rūh, gen.sg. rūwes (see Heidermanns 1993: 454–455) ‘rough’ A convincing explanation is not available. Compare, however, the proposed development *ur > *ru before a consonant (cf. Mayrhofer 1986: 161–162; Lubotsky 1994: 98–100). Could we start from a root *ureHḱ- with zero grade *urHḱ- > *ruHḱ-? Supporting evidence for a ban on roots ending in three-consonants, at least in pre-PIE, seems to be furnished by the Schwebeablaut in s-extensions to certain roots, a process which seems designed to avoid three-consonant clusters (Schindler 1970: 152; Ozoliņš 2015: 86–135): – *h₂eug- (Lt. áugti, Lat. augeō ‘grow, increase’, Go. aukan ‘multiply’) ~ *h₂ueg-s-29 (Skt. vavákṣa pf., Gr. ἀέξω, OHG wahsan) ‘grow, increase’ – *h₂elk- (Gr. ἀλκή ‘boldness, defence’, OE ealgian ‘defend, protect’) ~ *h₂lek-s(Skt. rákṣati ‘protect’, Gr. ἀλέξω ‘ward off, assist’) – *meiḱ- (Gr. μείγνυμι, Lt. miẽšti) ‘mix’ ~ *mieḱ-s- (Skt. myákṣati ‘sich festhalten; sich vereinigen’; Kümmel 2000: 388–389) 28 29 It is possible we are dealing with a compound; for instance, Garnier (2014: 149–150) has suggested a derivation involving the preverb *ud and the verbal root *dʰeh₁- ‘to suckle’ (with the ‘Kortlandt effect’, *ud-dʰh₁- > *uh₁dʰh₁-). A palatovelar might be implied by Lt. vešėt́ i ‘grow lush, thrive’. In view of the extreme rarity of the sequence *uḰ in reconstructed IE words, it is possible that there was a neutralization after *u (Meillet 1894: 292–293; Kortlandt 1979a: 58). If this is the case, we would expect this word to show an alternation *h₂ueǵs- : *h₂ugs-, and we could assume that the latter became generalized in Indo-Iranian and the former in Baltic. On the other hand, Smoczyński (2018: 1644; cf. also p. 1617 s.v. vãškas) sees this word as evidence that IE *-ks- regularly gave *š in East Baltic, rather than *kš as is usually assumed (cf. Stang 1966: 96). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 introduction 171 In conclusion, although there are a couple of unresolved issues, it seems highly probable that Proto-Indo-European did indeed prohibit roots ending in three consonants. Therefore, an implied root of this shape could again be used as an argument in favour of a non-IE origin. For instance, any root in Balto-Slavic containing a diphthong root with (a) acute accentuation not attributable to Winter’s law and (b) a final stop not analysable as a suffix (particularly *p, *b, *ś, *ź or *k)30 can be suspected to be of non-inherited origin. This applies to some of the examples discussed elsewhere in this work: – ? Lt. líepa, Lv. liẽpa; R ли́ па, SCr. lȉpa ‘lime tree’ (< *leiHp-); see p. 89. – Cz. labuť, SCr. lȁbūd ‘swan’ (< *HolHbʰ-); see pp. 176–177 and 234. – Lt. dial. lùnkas, Lv. lûks, Pr. E lunkan; R лы́ ко, SCr. lȉko ‘bast’ (< *l(u)nHk⁽ʷ⁾-); see pp. 181–182. – Lt. ríešutas, Lv. riẽksts, R орѣ́х ‘nut’ (< *(H)roiHs-); see pp. 238–239. It is not particularly difficult to find other potential examples. For instance Lt. slíekas, Lv. sliêka; Pr. E slayx ‘earthworm’ ~ Sw. dial. slå, Nw. dial. slo ‘slowworm’, OE slā-wyrm (translating Latin words for various kinds of serpent) (< *slaih(w)ō-, cf. Falk/Torp 1065; Stang 1972: 50). Yet I would hesitate to use the Balto-Slavic intonation alone as an argument to support a non-IE origin. There still remain a number of words containing an unexpected acute which is probably of non-laryngeal origin,31 and as long as this is the case, such evidence must be treated with care. As these instances remain very few, however, we may still consider intonation as supporting evidence for non-IE origin in cases where other evidence is available. 5.4 Preliminaries In the next three subsections, I will treat in detail all of the material which I consider to provide potential evidence for contact with pre-Baltic languages. I have restricted my material by the following criteria: (1) “pre-Baltic” words will be defined as those which are attested either in Baltic, or in both Slavic and one other “North European” branch (Germanic or Celtic); (2) the substrate proposal 30 31 Although there are some unambiguous examples of a deverbal suffix *-ka- (trad. *-kъ) in Slavic, cf. CS зна-къ ‘sign’ < OCS знати ‘to know’; зра-къ ‘sight, appearance’ < зьрѣти ‘to see’, it does not appear that there are any reliable examples of plain k-suffixes of BaltoSlavic age; in any case, the examples cited here are not readily analysable as containing a suffix. See for instance Lt. tánkus ‘dense’ (Chapter 3, fn. 24) and stíebas ‘stalk, trunk’, stámbas ‘stem, stalk’ (p. 60 and Chapter 3, fn. 55). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 172 chapter 5 must involve some kind of irregular correspondence and be trivial semantically (i.e. it should not contradict the criteria set forth in 5.2). I have not made any attempt to discuss every substrate word proposed where I do not consider there to be sufficient evidence of substrate origin. However, I do discuss certain words which have frequently been suggested in this context, or which require a more detailed rebuttal. These examples are marked with “†” and will not contribute to any further analysis. As my primary criteria for identifying substrate words is irregularity, and one of my goals in collecting the material is to identify geographical patterns, I have organized the lemmata according to the type of irregular alternation identified. The alternations have been organized into two main chapters — consonantism and vocalism — and each of these is divided into a number of subchapters. Each comparison is introduced by a word in bold, which normally represents the most frequent meaning present in the comparanda. Where two lemmata are discussed with the same meaning, these are disambiguated by a number in brackets. After this, forms are adduced, with “~” demarcating the forms showing the relevant alternation. After this, I have adduced any literature in which it is suggested that the given forms are of non-Indo-European origin. Where no literature is adduced, I am not aware of any existing proposals of that nature (although the comparison itself will usually have been made in an Indo-European context). I then go on to discuss issues concerning individual branches and reject incorrect comparanda, before making a judgement as to whether the given irregularity can be viewed as evidence that the word is of non-Indo-European origin. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 chapter 6 Consonantism 6.1 ‘Nasalization’, *-VNT- ∞ *-VT- Alternations of the type *-VNT- ∞ *-VT- have been noted particularly in the Greek material, where there are numerous compelling examples: φάρυγξ ~ φάρυξ ‘throat’, κόχλος ‘sea snail, Murex’ ~ κόγχη ‘mussel’, τέρμινθος ~ (Nicander) τρέμιθος ‘turpentine’ (cf. Kretschmer 1896: 403; Kuiper 1956: 213–215; Furnée 1972: 275–291). Words exhibiting such alternations have been used by the above authors to support theories of language contact with “pre-Greek”. This is supported by the obscure root etymology and suffixation of the relevant words. The interpretation of such alternations has varied. In the rendition of suspected Etruscan (Fiesel 1928: 60–61) and Thracian (Schrader/Nehring ii: 532) words, one has referred to ‘nasal vowels’ (cf. Huld 1990: 394; Kroonen 2012: 243), while in more recent literature, ‘nasal insertion’ has been the preferred option (see Furnée 1972: 269–270, with lit.). Kuiper (1956: 213; 1995: 68–69) suggested the term ‘prenasalization’ based on parallels he saw in the Munda languages. This has become the generally accepted term among Leiden scholars (see Kuiper 1956: 219–221; 1995: 68–72; Beekes 1996: 223–226; Boutkan 1998: 108–109; Schrijver 2001: 420–421). Beekes (2014: 14), albeit with hesitation, refers specifically to pre-nasalized stops. I would rather avoid the term ‘prenasalization’, particularly in the narrow sense of Beekes, as in theory, other interpretations of these alternations are possible. The above accounts, whether starting from nasal vowels and prenasalized consonants, both assume that the irregularities lie in synchronic phonological features of the donor language. However, it is not certain (or even likely) that the donor language was homogenous, and it would not be far-fetched to suppose the co-existence of sister languages or dialects where one has historically undergone a loss of syllable-final nasals. As discussed above (see p. 167), I find an agnostic approach most appropriate here. Outside of Greek, already Kretschmer (idem: 405) pointed out that Greek. ἐρέβινθος ‘chickpea’, and its irregular comparandum Lat. ervum ‘bitter vetch’ are quite possibly of non-IE origin. An equivalent to the Greek form without the nasal is OHG arawīz ‘pea’ (Oštir 1930: 14; Furnée 1972: 273; Kroonen 2012: 243; © Anthony Jakob, 2024 | doi:10.1163/9789004686472_008 Anthony Jakob - 978-90-04-68647-2 This is an open access chapter distributed under the terms of the cc by-nc-nd 4.0 license. Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 174 chapter 6 Thorsø forthc.).1 However, several more examples can be found in Northern Europe, and these will be the focus of our discussion here. 6.1.1 Alternation between *-VNT- and Short Vowel ‘grouse’. RCS ерѧбь (СДРЯ 11–14 iii: 219) ‘partridge’, Uk. оря́ бок, dial. о́ рябка, Pl. jarząbek ‘hazel grouse’, Sln. jerę̑b ‘partridge’ ~ Lt. jerubė ̃ ‘hazel grouse’; Lv. ir̃be, dial. (ME ii: 59) ìerube² ‘partridge’ (Derksen 2000, 2015: 212) — This bird name is characterized by what Derksen (2000: 80) has described as “spectacular stem variation”. The large number of variants has encouraged several solutions, e.g. Endzelīns (ME i: 708–709) tries to assume reduplication with dissimilatory loss of the r-, but abandons the connection with Slavic. It seems clear, however, that the Baltic and Slavic data should be kept together (cf. Fraenkel 1936a: 231; ALEW 481). Andersen, in a special article on the subject, attempts to reduce the material to four basic pre-forms, which he derives from two different roots (1996b: 75; 84–85). According to him, the forms containing a nasal are derived from the n-stem underlying Gr. ὄρνις ‘bird’ with a suffix *-bʰ- (on this pattern of derivation, see pp. 187–188), while those lacking the nasal should, in his opinion, be compared with Nw. jerpe, Sw. järpe ‘hazel grouse’.2 The most fundamental flaw in Andersen’s account is the failure to account for the standard Lithuanian form jerubė,̃ which shows a disyllabic stem but no nasal (Derksen 2000: 81–83). In fact, all the forms which would supposedly correspond to those in Norse probably result from syncope. Thus, Lv. ir̃be can be explained from an older *ierube (cf. High Latvian ìerube², ME ii: 59; EH i: 537), for which a convincing parallel may be found in il̃kss ‘carriage pole’ as against High Latvian ìelukši nom.pl. ‘carriage pole’ (Bezzenberger 1885: 169: ⟨ëḷukschi⟩ Zvirgzdene; cf. Endzelīns 1923: 47, EH i: 528).3 Variants in Baltic with a nasal are very rare. Juška (ii: 684) cites jerumbė ̃ as a variant of jerubė.̃ In addition to this, ‣ 1 Another example mentioned by Furnée is the word for ‘lynx’, which will be discussed in section 6.1.2. 2 These words must be derived from ON jarpr ‘chestnut brown (usu. of hair)’. The Norse adjective corresponds to OE eorp ‘dark, swarthy’, OHG erpfer · fus[c]us. Unlike with R рябо́ й (see fn. 7, below), it is by no means evident that the bird name is primary. 3 As already acknowledged by Andersen (1996a: 73), Bg. dial. (БЕР i: 73) ѐрбица is most likely an irregular reduction of ѐребица, and other alleged Slavic evidence is to be explained similarly (see Derksen 2000: 78). In Lithuanian, dial. jérbė is also most probably from jerubė ̃ (cf. the place name Jer̃biškiai < Jerùbiškiai cited in Zinkevičius 1966: 132). The Lithuanian evidence for a stem irb- possibly all stems ultimately from Latvian. Thus, ⟨Ýrbenis⟩ ‘Viburnum’ (Pabrėža 1834: 49) seems to be based on Latvian ir̃bene (cited by the author). Lt. vìrbė ‘hazel grouse’ (cf. HLv. dial. virbe, vìrba² ‘Rebhuhn’, ME iv: 603, EH ii: 786), for which LKŽ provides no dialectal attestations, was perhaps popularized by Ivanauskas’ Lietuvos paukščiai (the form is attributed to Ivanauskas in Elisonas’Zoologijos sistematikos terminų žodynėlis, 1920, p. 90, although Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 consonantism 175 LKŽ cites only the isolated arumbėl̃ ė (Palėvenė) and vierumbėl̃ ė (Marcinkonys). The explanation of these forms is rather unclear, but the limitation to some isolated dialects suggests they are secondary. Perhaps, in some areas, a certain role may have been played by Polish jarząb(ek). In fact, the Žem. *jėŕ umbė (in dialect notation ⟨jiêrộmbẹ⟩), recorded in Šateikiai (Papildymų kartoteka) may be a direct loanword from Polish, showing /ė/ regularly for Slavic /a/ after a palatal.4 Another fact left unaccounted for by Andersen is the initial je- in Lithuanian. As this variant is concentrated in Kauniškiai dialects, and not in dialects which show je- < *e-, the j- is most straightforwardly interpreted as original and cannot be taken back to an original *e-.5 While it is true that the correspondence between Lv. (dial.) ⁽*⁾ie- and Lithuanian je- is not regular (cf. Derksen 2000: 78–79),6 it still seems most parsimonious to assume that all the East Baltic data derives from a single proto-form, most probably *jerubē (with *ẹ̄rubē remaining a possibility). As for Slavic, the East Slavic forms with o- as against je- elsewhere suggest a Proto-Slavic form in *e- (Derksen 2000: 78); the variants with ja- attested in several Slavic languages may be secondary (pace Meillet/Vaillant 1933: 101; see Andersen 1996a: 74–76 for numerous parallels). Forms without an initial vowel are basically limited to East Slavic: e.g. MR рябь ‘partridge; ?hazel grouse’ (СРЯ 11–17 xxii: 281), dial. N рябь, ряб (cf. СРГК iv: 601), Bel. ра́ бчык, dial. рабо́ к ‘hazel grouse’. Beside this, they are marginally attested in Slovene: rę̑b, rebíca (Caf apud Pleteršnik ii: 412). The most likely solution is that we are dealing with instances of aphaeresis. Note that no such forms are found in West Slavic, where initial stress was generalized. As a parallel in a similar environment, compare R dial. лито́ ка, лито́ нья (and variants; СРНГ xvii: 73–74) ‘third stomach of ruminants’ ~ Pl. jelito ‘intestine’ (see also, in particular, ДАРЯ i: No. 33).7 It therefore seems that the Slavic words can probably be combined under a single preform *erę̄bi- (trad. *( j)erębь). 4 5 6 7 I have not found its original source). Note another Latvianism attributable to Ivanauskas: lestė ‘flounder’ (= Lv. dial. leste, see LKŽ). The variant ìrbė ‘hazel grouse’ is only known from Šlapelis’ dictionary (apud LKŽ). A potential parallel may be Žem. dial. munkà ‘suffering’, which has been analysed as a modification of mūkà ‘torment’ (← Bel.) under the influence of Pl. męka ‘torment’ (Zinkevičius 1966: 198). The Aukštaitian variants vėr̃ ūbė, jerū̃bė asserted by Būga (1923b: 402 and RR ii: 537), with a long medial syllable, seem otherwise to be unattested. I consider the variants with initial ja- and a- to be insignificant; cf. ãknos for jẽknos ‘liver’ (in Veliuona; see Juška i: 9), ái = jéi ‘if’ (LKŽ; see Zinkevičius 1966: 121–124). On the interchange of initial j- and v-, see Grinaveckis 1972: 74. Note also Lv. rubenis ‘black grouse’, which will be discussed on p. 241. The same distribution is found in the Slavic words for ‘rowan’ derived from the bird name Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 176 chapter 6 As a result, the modern dialects indeed show a great amount of variation, but the vast majority of this can be shown to be secondary. However, the second syllables of East Baltic *jerub- and Slavic *erę̄b- (trad. *( j)erębь) are not comparable in an Indo-European context, and the presence of a nasal in Slavic as against its absence in Baltic remains strong evidence of a non-IE origin. ‘swan (1)’. Pl. łabędź, Sln. labǫ́ d ‘swan’ ~ R ле́бедь, Bg. лѐбед; ON ǫlpt, OE ielfetu, OHG albiz, elbiz ‘swan’ (Oštir 1930: 14; Machek 1968: 316; Derksen 2000: 84; Kroonen 2013: 20) — The reconstruction of a single Proto-Slavic form seems impossible, but two widespread forms can be reconstructed: (1) Cz. labuť, Pl. łabędź, SCr. lȁbūd, Sln. labǫ́ d ‘swan’, which regularly reflect an acute *albǭdi- (trad. *olbǫdь; or *lābǭdi-, trad. *labǫdь);8 and (2) R ле́бедь, Uk. ле́бідь (gen.sg. -едя), CS *лебедь (attested дебель), Bg. лѐбед which reflect *lebedi(trad. *lebedь). The forms are almost in complementary distribution, although Pleteršnik (i: 503) cites a rather doubtful looking Sln. lebed from the dictionaries of Jarnik and Janežič,9 and some other forms in South Slavic, e.g. Mac. лабед and SCr. obs. lȅbūt (RJA v: 944) seem to show a confusion between the two forms.10 The mismatch between the second syllables *-bǭd- and *-bed- is difficult to account for in Indo-European terms. In Germanic, one has traditionally interpreted ON ǫlpt, OHG albiz beside OHG elbiz, OE ielfetu as reflecting two by-forms, *albut- beside *albit- (Noreen 1923: 151; Specht 1947: 114; IEW 30; de Vries 1962: 101; EWAhd 1033). The form *albut- would come close to Slavic *albǭdi- (trad. *olbǫdь), save for the nasal (cf. Meillet 1907: 377; Булаховский 1948: 118–119; Derksen 2008: 365). However, positing unmotivated by-forms is not an attractive solution. Since the u-umlaut in ON ǫlpt (gen.sg. alptar) can be attributed to the analogical extension of uumlaut to all feminine consonant stems (cf. Noreen 1923: 284–285; Kroonen ‣ 8 9 10 (the hazel grouse eats rowan berries in autumn; Cramp apud Andersen 1996b: 79; see the partial parallels adduced in Derksen 2000: 79–80 to which we may add German Vogelbeere ‘rowan’), as well as in the word for ‘mottled’ in East/West Slavic: R рябо́ й, Slk. jarabý ‘mottled’ derive from MR рябь, Slk. jarabica ‘partridge’ just as R голубо́ й ‘pale blue’ derives from го́лубь ‘pigeon’ (cf. Andersen 1996b: 78). On the final *t in some of the reflexes, which must be secondary, see the discussion in 6.2. This variant does not appear to be known dialectally (Tijmen Pronk p.c. October 2022). Despite ЭССЯ (vi: 19) and Николаев (2020: 39, fn. 6; cf. Зализняк 2019: 640), it seems incorrect to take the East Slavic forms from *lebę̄di- (trad. *lebędь). All of the Old Russian evidence suggests *-bed- (СДРЯ ii: 13–14), as does Ukrainian ле́бідь (gen.sg. ле́бедя). The modern Russian adjective лебя́ жий, is by all appearances a late creation, replacing earlier лебежий in the 17th century (СРЯ 11–17 viii: 183; cf. Булаховский 1968: 103). It can be considered a hypercorrection due to the widespread merger of /’a/ and /e/ in unstressed syllables (ДАРЯ i, No. 3). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 consonantism 177 2013: 26), the most straightforward solution would be to posit a t-stem *albet(slightly differently cf. Orel 2003: 13). In this case, the suffix would be more closely aligned with that of Slavic *lebedi- (trad. *lebedь; although note that Slavic requires *dʰ as against Germanic *d). The acute accent implied by the reflex la- throughout West Slavic would alone be sufficient reason to abandon the traditional comparison with Lat. albus ‘white’ (Miklosich 1886: 162; Osthoff 1898: 64–65; ЭССЯ vi: 19; and elsewhere; see Derksen 2000: 84), and when combined with the irregular alternation between *-eD- and *-onD- in the second syllable, the case for a loanword from a non-IE source appears very strong. For further discussion, see p. 234. ‘goosefoot’. Lt. balánda, Lv. dial. baluôda ‘goosefoot, Chenopodium’ ~ R лебеда́ ‘orache, Atriplex’, SCr. lobòda ‘goosefoot’; ?OS maldia, OHG melta ‘orache’ (Mikkola 1903: 46; Machek 1947: 66–67, 1950b: 149) — The prevailing view is that the Slavic term is somehow related to the word for ‘swan’ (above; cf. e.g. REW ii: 21–22; ЭССЯ vi: 18, xxxii: 50; Derksen 2000: 84, 2008: 366); however, as Vasmer and Derksen both admit, the alleged proto-form *albadā(trad. *olboda) could not possibly yield the attested forms. Practically all of the relevant evidence points instead to *labadā- (trad. *loboda): cf. unambiguously Slk. loboda, SCr. lobòda, Sln. lóboda (in SSKJ² stressed lobóda), Bg. ло̀ бода (and further R dial. лобода́ ) ‘orache’. Beside this, we find a variant *lebedā- (trad. *lebeda): R лебеда́ , Cz. lebeda, Sln. lebę́da ‘orache’, Pl. lebioda ‘goosefoot’. Some forms like SCr. labòda ‘goosefoot’ (РСА xi: 146) apparently show the secondary influence of the word for ‘swan’ (Derksen 2008: 366).11 The semantic relationship between ‘swan’ and ‘goosefoot’ is ostensibly paralleled by the English name for the plant,12 but Mikkola (1903: 46) has instead suggested we compare Lt. balánda, Lv. dial. baluôda, assuming that Slavic *labadā- (trad. *loboda) was derived via metathesis from *baladā- (due to the influence of ‘swan’?). He describes this as a “Kulturwort” and additionally adduces Greek βλίτον ‘purple amaranth’. Machek (1947: 66–67; 1950b: 149) mentions the same Balto-Slavic combination, but compares instead OHG melta ‘orache’ (< *maldjō-, Kroonen 2013: 251), which I consider more promising. In this case, we have to assume an additional alternation *b ∞ *m (see 6.4.2). On ‣ 11 12 The opposite direction of influence might explain the confusing variants in Bg. dial. ло̀ бод ‘swan’ (БЭР iii: 448), Sln. obs. ⟨lobòt⟩ ‘swan’ (17ᵗʰ c.; see Pleteršnik i: 526). The term seems first to be attested in the works of 16ᵗʰ century botanists (thus Philippa et al. ii: 167 quote Dodonaeus, dated 1554; OED cite W. Turner’s Names of Herbes from 1548). It is therefore, as stated in OED, most probably based on the form Chenopus, itself attributed to Pliny (see also Marzell i: 933; G. Hegi apud Kroll 1990: 46). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 178 chapter 6 the loss of the second syllable, see below on ‘oriole’. While the extra-Balto-Slavic comparanda are less certain, the comparison between Slavic *labadā- (trad. *loboda) and Baltic *balan̂ dā- looks tempting, and would be another example of the alternation *-VNT- ∞ *-V̆ T-. ‘pigeon (1)’. OCS голѫбь; R го́лубь, Pl. gołąb; Lat. columba ‘pigeon’ ~ OE culfre, culufre ‘pigeon’ (for refs. and a more detailed discussion, see p. 187) — Both Old English variants have been analysed as primary: Campbell (1959: 159, and already Pogatscher 1898: 98) considers culfre an example of syncope (cf. OE siolfor beside siolufr- ‘silver’), while Hogg (1992: 231–232) treats culufre as an instance of vowel epenthesis. I am inclined to side with Campbell on this issue,13 and reconstruct the preform as *kulubrō(n)-. Skeat (1882: 146) saw this as a ‘corrupted’ Latin columba, while Pogatscher (1898: 97) suggested the source could be found in a diminutive *columbula (cf. Old Occitan colombla ‘dove’, FEW ii: 930). An alternative was suggested by Holthausen (1899), who analysed the English word as cognate to Slavic *galǭbi- (trad. *golǫbь). The obvious issue with connecting the words either through borrowing or cognancy is the absence of a nasal in English.14 At the same time, it would be unattractive to separate *kulubrō- from Lat. columba. The correlation between Germanic *u and Italic *o is paralleled by OHG hulis (< *kulis-) against MW celyn (< *kolisno-) ‘holly’ (Kroonen 2013: 253; van Sluis et al. 2023: 216). This would be another example of an alternation involving nasals, and give support to the non-Indo-European origin of the word (see further pp. 187–189). ‘oriole’. Lt. volungė ̃ 3ᵃ, Lv. vãluôdze ‘oriole’ ~ Pl. wilga, Sln. vółga; R и́ волга, Bg. авлѝ га; ME wode-wale, MHG wite-wal ‘oriole’ (Oštir 1930: 101; Machek 1968: 694; Derksen 2008: 216–217; Kroonen 2013: 571; Matasović 2013: 87) — The Latvian form suggests an underlying *-ang- in the second syllable, which does not match the Lithuanian data. By way of a solution, ALEW 1469 suggests that the standard Latvian form is a hypercorrection based on a High Latvian dialect where *uo and *ū have merged. However, the typical development in ‣ ‣ 13 14 Judging by the examples provided in these sources, the epenthesis almost exclusively occurs before word-final _RC# (where C is usually a velar) or before the clusters -ht- or -gd-. In this context, the form culufre stands out as exceptional. Furthermore, as Hogg states, the epenthesis is typical of Northumbrian, while this form (according to the data in the Dictionary of Old English Web Corpus) appears to be more widespread. Compare siolufres gen.sg., attested in a West Saxon source, where the vowel is old. Pogatscher’s solution, involving a novel sound law *-mr- > -fr- has evidently not stood the test of time (see the alternative etymologies already in Holthausen 1934, s.v. ċealortún, hæf-ern, etc.). Paulus van Sluis (p.c. August 2021) pointed out to me that *kulumfrōnwould also be a possible preform, with regular loss of *m before *f, although in this case the syncope would be unexpected (Campbell 1959: 49; Hogg 1992: 230). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 179 consonantism High Latvian is in fact a chain shift, so that no merger takes place. Note in this context the South Aukštaitian hapax ulangėlė, cited in LKŽ, which might favour a reconstruction *-ang- and imply that Lt. volungė ̃ is originally an East Lithuanian form. ЭССЯ (xiii: 251–252) unites all of the Slavic forms under the reconstruction *ivilgā- (trad. *jьvьlga), but such a reconstruction is hardly possible, at least, for Sln. vółga and SCr. vȕga ‘oriole’, as initial *i- (trad. *jь-) is always preserved in these languages, while here no trace of the vowel can be identified. The status of Pl. wilga and Slk. vlha is less certain, as *i- (trad. *jь-) > ∅- is frequent here (see Derksen 2003 for the data). In any case, East Slavic clearly demands a reconstruction with *i-, as does Bulgarian авлѝ га, a form which is most easily explained by metathesis from CS ⁽*⁾ивлъга.15 The significance of this ‘prefix’ is unclear. It is hardly, with Трубачев (1972: 19–20), an irregular reduction of the prefix *iz- (trad. *jьz-; *z would not be lost before *v); neither is the parallel with R изю́ брь ‘Manchurian wapiti’ watertight; see p. 242. The Balto-Slavic comparison goes back at least to Miklosich (1865: 68; 1886: 379), but attempts to account for the relationship between the words in IndoEuropean terms (e.g. Mikkola 1897: 247) cannot be viewed with optimism. Moreover, treating the Baltic second syllable as a suffix (Endzelīns 1924: 123, citing the river name Bebrunga) does little to elucidate the relationship with Slavic.16 As a result, some recent works have rejected the relationship altogether (Smoczyński 2018: 1693; ALEW 1469). Nevertheless, the Baltic and Slavic words are semantically identical and share a consonantal structure: Baltic: Slavic: 15 16 v v â ĭ l l an ∅ gg- This CS form is attested among a list of birds in the Hexameron of John the Exarch; however, it is not entirely certain how it is to be read. The actual manuscript has “косыжє · ӥ соѥ̈ · ӥ влъгъı · ӥ жлъны · щурыжє”. Since the sequences ⟨ӥ соѥ̈ ⟩ ‘jays’ and ⟨ӥ жлъны⟩ ‘woodpeckers’ clearly both contain the word и ‘and’, it is natural to suspect that ⟨ӥ влъгъı⟩ does, too (thus Miklosich 1865: 68, and thence the CS form влъга usually encountered in the literature, e.g. ЭССЯ viii: 251). Aitzetmüller (1958: 38), on the other hand, reads “ивлъга” here, citing a variant ⟨и ивлъга⟩ and the modern Bulgarian evidence. This theory is supported by Bg. и́ волга attested in Геров (ii: 171; a dialectal form with *lъ > /ol/ like others recorded in Геров, e.g. мо́лзѫ ‘to milk’, мо́ рковъ ‘carrot’, iii: 78, 82). The etymological comparison (cf. Endzelīns 1914a: 126; LEW 1273–1274; REW i: 469) with YAv. vārənjana-, vārəɣna- ‘a bird of prey’ (cf. Sogdian wʾrɣnʾk, Khwarezmian wʾrɣnyk ‘falcon’, Hintze 1994: 198–199) is semantically weak. Note that Endzelīns and followers operate with Bartholomae’s non-specific translation ‘Name eines Vogels’, which might explain their enthusiasm. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 180 chapter 6 As far as the nasal alternation is concerned, it is clear that this word does not behave in quite the same manner as most of the above examples. Instead of an alternation between *-VNT- and *-V̆ T-, Slavic lacks the second syllable altogether. A potential parallel for this is found between Lt. balánda ‘goosefoot’ and OHG melta ‘orache’ (see above under ‘goosefoot’), provided a comparison between these forms is warranted.17 For the vocalic alternation, Oštir (1930: 22) has adduced Lat. taxus ~ R тис, Sln. tȋsa ‘yew tree’ as a parallel. While the latter is probably indeed of non-IE origin (see pp. 265–266), the parallel is imperfect due to differences in vowel length. For some other potential parallels, see 7.3.1. Endzelīns (1924: 123; similarly Machek 1950a: 49–50; Derksen 2008: 216–217; Kroonen 2013: 571, and others) compares this word with Germanic forms like ME wode-wale, MHG wite-wal ‘oriole’. A trace of the velar of the Balto-Slavic forms could be found in Swiss and Bavarian dial. Wiedewalch (attested since the 15ᵗʰ century, cf. Suolahti 1909: 170). Machek (loc. cit.) suggests the Germanic reconstruction *-walka in order to unite the material, but the loss of *-k- elsewhere would be irregular. A Germanic reconstruction *-walhō- might just work, however. The loss of *h in Low German and Dutch would be regular, cf. MDu., MLG male ‘bag’ (< *malhō-, Kroonen 2013: 351). While this development is more sporadic in Middle English and High German, the simplification of the cluster may have been supported by the word’s unstressed position as the second element of a compound. This would imply an additional alternation *g⁽ʰ⁾ ∞ *k (see 6.2.1). 6.1.2 Alternation between *-VNT- and Long Vowel ‘lynx’. Gr. λύγξ ~ Lt. lū́šis, Lv. lũsis, Pr. E luysis; R рысь, Sln. rȋs;18 OE lox, OHG luhs ‘lynx’; Arm. (hapax) lusann* ‘lynx’19 (Furnée 1972: 121–122; Martirosyan 2008: 317; Kroonen 2013: 342) — Strictly speaking, the East Baltic form for ‘lynx’ does not rule out an older nasal, and the word could therefore be identical, ‣ 17 18 19 The comparison of Lithuanian jerumbė ̃ : ìrbė ‘hazel-grouse’ (Derksen 2015: 510) is unlikely to be valid, as both are probably secondary variants of Baltic *jerub-. See above on this word. The most convincing explanation of the Slavic r- is contamination with the adjective in Cz. obs. rysý (Kott iii: 239), LSrb. obs. rysy ‘red-haired’ (see Śmieszek 1909: 408). One might argue that this adjective is itself derived from the name of the lynx, but certainly old is R ры́ жий ‘red-haired’, Pl. dial. rydzy ‘copper-red’, SCr. rȋđ ‘reddish, rust-coloured’ < *rȳdja(trad. *rydjь), from IE *h₁reudʰ-; see REW ii: 557–558. OIr. lug, translated as ‘lynx’ by Pedersen (1909: 186), apparently mainly on the basis of the formal similarity, is doubtful. For the interpretation as ‘warrior, hero’, see eDIL s.v. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 consonantism 181 aside from the acute intonation, to Greek λύγξ (see LEW 392; ALEW 696).20 However, at least Pr. E luysis ‘lynx’,21 Elfdalian luo ‘lynx’ (cf. Kroonen 2013: 342), and the West Germanic material are inconsistent with a nasal preform (Armenian is ambiguous; cf. Martirosyan 2008: 317). This nasal has been referred to as an ‘infix’ (e.g. Smoczyński 2018: 734; cf. Pedersen 1909: 188), but this remains ad hoc as there is no generally accepted morphological process of nasal infixation in nouns. Even granted this, the acute long vowel in Balto-Slavic as against the short *u in Germanic are still suggestive of parallel loanwords, as they preclude the reconstruction of a common proto-form (see 7.5.1). ? ‘bast’. Lt. dial. lùnkas, Lv. lûks, Pr. E lunkan ~ R лы́ ко, SCr. lȉko ‘bast’22 — The validity of this example depends on whether the loss of the nasal in Slavic is regular. It has been suggested based on pairs such as OCS исто* (nom.-acc.du. истесѣ) ~ Lt. ìnkstas ‘kidney’ and OCS въıкнѫти ‘learn, become accustomed’ ~ Lt. (pri-) jùnkti ‘get used to’ that high nasal vowels were denasalized in early Slavic under acute intonation (see Mikkola 1897: 246–247; Meillet 1907: 362; Arumaa 1964: 129–130; Kortlandt 1979b: 269). If the loss of the nasal can be considered regular in Slavic, then this example does not belong here. On the other hand, this Slavic sound law is far from certain (see in detail Pronk 2013). Lt. jùnkti itself contains a nasal infix (there is clearly no nasal in the causative OCS оучити ‘teach’). The nasal in Lt. ìnkstas has also been suspected to be secondary (LEW 188; Pronk 2013: 120).23 The clearest counter-evidence ‣ 20 21 22 23 In this connection, the Žem. variant lų́ nšis is usually mentioned (cf. also Specht 1947: 171– 172; Chantraine DELG iii: 648), but this form has no etymological value, resulting from a general sporadic nasalization of high vowels before sibilants (Būga 1922: 42; Trautmann 1923: 164; Zinkevičius 1966: 196–197). The Prussian form has long been problematized. Endzelīns (1943: 206) is undecided as to whether we are dealing with i-epenthesis or a spelling variant for /ū/ (see similarly Топоров ПЯ v: 389). Būga (1911: 41), on the other hand, read *lunsis. An important form is Lt. dial. luišỹs (Bartninkai), which supports the reality of the Prussian /ui/. Here, as Trautmann (1910: 145) already surmised, we are dealing with an epenthesis of -i- as also found sporadically in Western Žemaitia, particularly in ja-stems after rounded vowels (more examples in Bezzenberger 1887: 36, 1911: 311; Endzelīns 1914b: 102; Būga 1924a: cxxi–cxxii). Note similarly Pr. E girnoywis (where ⟨oy⟩ probably = */ui/) ‘quernstone’ < *girnuwīs,̆ cf. OCS жрънъı*. Older sources (cf. LEW 390–391; REW ii: 75) connect Pāli luñcati ‘pull out, pluck (a bird), tear, peel’ (CDIAL 642, KEWA iii: 105). This must be rejected on accentological grounds. The Pāli verb, provided it is inherited, could rather be connected with Lat. runcō ‘grub up, weed’, Gr. ὀρύσσω ‘dig (up)’. Deriving Cz. výheň ‘forge’ from *Hngʷni-o- (Hamp 1970a: 77; Kortlandt 1988: 388; Derksen 2008: 534) is not very satisfactory, especially since the difference in vocalism with Cz. oheň ‘fire’ is not well accounted for (compare Pronk 2013: 124–125). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 182 chapter 6 is the verb Pl. dąć (1sg.pres. dmę), SCr. dial. dȕti (1sg.pres. dmēm), Lt. dùmti ‘to blow’ where the reconstruction of *domH- for the Slavic infinitive (Derksen 2008: 114) is ad hoc (see already Meillet 1907: 366). I therefore see the word for ‘bast’ as a potential example of alternation between a sequence *-VNC- and *-V̄ C-. The foreign origin of this lemma is supported by the root-final cluster *-NHK-, implied by the acute accentuation (see 5.3.2). Note that this word has also been borrowed into Mordvin, although whether Baltic was the proximate source is uncertain (see p. 135). ? ‘elm (1)’. Lt. vìnkšna, Lv. vîksna; R вяз ‘elm’, SCr. dial. (РСА ii: 459) vȇz ‘field elm’ ~ OE wice ‘wych-elm’; Alb. vidh ‘elm’ (OED³ s.v. wych) — The Baltic forms reflect *vînž- + *-snā- > *vîn(k)šnā.24 OE wice is often assumed to contain a long vowel (Holthausen 1934: 392; IEW 1177), but OED³ (s.v. wych) argues that forms such as wiech (15ᵗʰ c.) would show the effects of northern lengthening in an open syllable, implying an original short vowel. On the other hand, a long vowel must be reconstructed for continental Germanic, cf. MoLG (obs.?) Wieke (= Prussian German Wieken ‘white elm; small-leaved lime’, Frischbier ii: 468), MoHG obs. Weiche ‘elm’ (< *wīkō(n)-). It is uncertain whether the Albanian form is consistent with a nasalized pre-form *uinǵ-. If Geg ãnkth ‘incubus, nightmare’ is derived from *h₂emǵʰ‘narrow’, it would imply satemization was blocked after a nasal (Demiraj 1997: 79; de Vaan 2018: 1745). On the other hand, this etymology is uncertain, and Huld (1981: 305) has pointed out a nasalized vĩdh in an early 20ᵗʰ century grammar, which, if reliable, would align Albanian with Balto-Slavic. There also appears to be some related Iranian data: Gorani wiz, Talysh vızm, vezm (Пирейко 1976: 46), Khunsari vizvā, Bakhtiari gzəm, Zaboli ɣuzbe (Henning 1963a: 71–72; Цаболов 2001: 214; В. Дыбо 2002: 469), all in the sense ‘elm’, are reconstructed by Henning as *u̯ izu̯ ā̆-, i.e. a virtual *uiǵ⁽ʰ⁾-uV-. Based on this reconstruction, the Iranian words could be cognate with the European forms, and confirm a broader distribution (cf. Polomé 1990: 334; Mallory/Adams 2006: 159). In a footnote, Henning (op. cit. 72) also admits the possibility of a reconstruction *u̯ inz-, bringing the Iranian words in line with Balto-Slavic (see again Henning 1965: 43). There are indeed potential examples of a nasal being lost ‣ 24 Since the -k- can be intrusive, the claim (in ALEW 1444) that the suffixation must predate the assibilation of *ǵ seems completely gratuitous; cf. the similar comment under Lt. añkštas ‘narrow’ (ALEW 34) which is rather < *aNž- (= OCS ѫзъкъ) + *-sta- (cf. áukštas ‘tall’ < áugti ‘to grow’ and Skardžius 1941: 324–325; LEW 11; Stang 1966: 108; Smoczyński 2018: 1671). The secondary nature of the velar might be proven by Zietela vyšnė ̃ ‘cross beam on a sledge’ (cf. Lt. dial. vìnkšna in the same sense), which might well stand for *vįšnė.̃ Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 consonantism 183 before a reflex of an IE palatovelar in Iranian,25 but also exceptions.26 The safer option, therefore, is to identify the Iranian with the Germanic forms. A radical solution is taken by ALEW (1444–1445), who do not mention the other Indo-European comparanda, and consider the Baltic and Slavic words independent derivatives of a root *ueiǵ- seen also in OCS вѧзати ‘bind, fetter’ and Lt. výža or vyžà ‘bast shoe’ (see already Būga 1922: 301). This can hardly be maintained, first of all, because the Slavic verbal forms show evidence of oxytone accentuation not consistent with Winter’s law (В. Дыбо 2000: 388; Derksen 2008: 521; РЭС ix: 235–236). More generally, it hardly seems attractive to separate the Balto-Slavic words from the synonyms in the other languages. OED³ (s.v. wych) have suggested that the formal problems could be accounted for by assuming the word originated in a non-Indo-European substrate language. Considering the parallels collected above, this possibility should be reckoned with. The question is whether the existence of Iranian cognates would rule out a non-Indo-European loanword (cf. ‘hemp’ on pp. 206–207). In this connection, we can remark that the Iranian cognates are late-attested and limited to a group of West Iranian languages spoken in a relatively compact geographical area. This might suggest the word is intrusive to Iranian; on the other hand, the fact that the word has apparently undergone satemization there would imply it is indeed very ancient. The only way out would be to assume the palatalization took place in the donor language (an IE satem language?). On balance, while a non-IE origin might help to explain the nasalized forms, it is difficult to account for all of the facts convincingly. † ‘nit’. Lt. glìnda (< *gnìnda?) ‘nit’ ~ R гни́ да, Sln. gnída; OE hnitu, OHG niz ‘nit’ (Beekes 1969: 290; Kroonen 2012: 247; van Sluis forthc.) — Kroonen has suggested this as an example of non-Indo-European nasalization. A nasal infix is also allegedly found in Latin lēns (usu. pl. lendēs) ‘nit’, but this form has so little in common with the other cognates (only the -d- poses no issue) that it is uncertain it belongs here (cf. van Sluis forthc.). Puhvel (1990: 366) posits a com- ‣ 25 26 As discussed by Martin Kümmel at the 2021 Österreichische Linguistik-Tagung. A partial parallel is the word for ‘twenty’, whereby against the remarkable parallelism of Oss. D insæj and Skt. viṃśatí ‘twenty’ (cf. Henning 1965: 43), the rest of Iranian shows *ī (YAv. vīsaiti, MP wyst /wīst/). However, the vowel turns up long here. As another possible example, note Parth. bz- ‘receive help’ as against YAv. bązaiti ‘support’ (cf. Cheung 2007: 72; however, Khot. baś- (Emmerick 1968: 94), Oss. ID bæzz- ‘be suitable’ reflect *bazi̯a- with probable zero-grade). Most notably MP hnzwg ‘narrow’ (→ Arm. anjuk), hardly to be separated from Skt. aṃhú‘narrow’ (Henning 1963b: 196–197). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 184 chapter 6 mon preform *lind- (< *nind-) for Lat. lēns and Lt. glìnda, assuming the initial guttural in the other languages is “moveable” (i.e. of secondary origin), which is clearly ad hoc. Note that even *lind- does not explain the Latin vocalism (de Vaan 2008: 334). As for the Lithuanian form, one is reminded of cases of secondary -ninC- < *-nīC-, which are particularly common in Žemaitian dialects: cf. bažnìnčia < bažnýčia ‘church’ (← Bel. бажнíца), dial. kningà < knygà ‘book’ (← Bel. кнíгa). The main issue here is that it is precisely in Žemaitian where we actually find a form without a nasal: dial. gnýda.27 However, this is not fatal, as such nasalized forms are only sporadic in Žemaitian. Furthermore, forms with a secondary nasal are also occasionally recorded in Aukštaitian; note in particular the agentnoun suffix -ininkas (beside dial. -inykas), which even belongs to the literary standard. The ablaut relationship between OE hnitu, OHG niz ‘nit’ (< *ḱnid-) and Alb. thëri, Gr. κονίς, pl. κονίδες ‘nit’ (< *ḱonid-) looks highly archaic, and is easier to explain in an IE context than through independent borrowings.28 It seems impossible to get away from the notion of taboo distortions here (cf. IEW 608): at least the initial gn- in Balto-Slavic must be explained in this way;29 in this context, we can note that many Slavic words starting in *gn- have a negative connotation, e.g. R гнус ‘gnats’, SCr. gnȏj ‘manure, pus’. It is possible that taboo also played a part in the replacement of earlier ⁽*⁾gnýda with glìnda in Lithuanian. On balance, due to the many difficulties with this word and its alleged existence in almost every European branch, I will leave it out of consideration here. 6.1.2.1 Conclusion The certain and possible evidence for nasal alternations is collected in Table 8, overleaf. The forms are presented as approximate quasi-Indo-European recon27 28 29 See LKŽ, where the word is marked as a Polish loanword. While this loan etymology cannot be ruled out, there is nothing in particular to suggest that the Žemaitian form is not simply regularly cognate with the Slavic forms. Latvian gnĩda ‘nit’ is of course ambiguous, and could reflect a preform with or without a nasal, or also be loaned from Slavic. The ablaut *ḱonid- : *ḱnid- seems to belong to a rather rare type, but compare *melit(Hitt. milit nom.sg., Gr. μέλι, Go. miliþ, Alb. mjaltë ‘honey’) : *mlit- (Hitt. maliddu- ‘sweet, pleasant’, Gr. βλίττω ‘cut out honeycomb’, ?Alb. (m)bletë ‘honeybee’). This is not the place to go into a discussion of Armenian anic (for *nic < *knid-s?, Martirosyan 2008: 86– 89) and Celtic forms pointing to *snida (why *s-?), although they may somehow belong here. Also note the voiced anlaut of Lt. blusà, R блоха́ ‘flea’ as opposed to Skt. plúṣi- ‘flea’. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 185 consonantism table 8 Possible examples of nasal alternations Baltic Slavic Germanic ‘grouse’ *i̯erubʰ- *i̯erembʰ- ? *rebʰ- ‘swan (1)’ – ‘goosefoot’ *bʰaland- *labʰadʰ- *malT- ‘pigeon (1)’ – *Golombʰ- *guluBr- ‘oriole’ *u̯ âlanG- *u̯ (i)lg⁽ʷ⁾- *u̯ alk- ‘lynx’ *lû(n)ḱ- *rûḱ- *luḱ- *albandʰ*lebʰedʰ Elsewhere *albʰed- ? ‘elm (1)’ *u̯ inǵ-sn- *u̯ inǵ- *u̯ (e)iǵ- ? ‘bast’ *lûnk- *lûk- – Lat. *kolombʰ- Gr. *lunḱIran. *u̯ iǵ-u̯ Alb. *u̯ i(n)ǵ- structions, but without the use of laryngeals. Long vowels which turn up as acute in Balto-Slavic are written with the caret ⟨ ˆ ⟩. Where the Indo-European reconstruction is ambiguous, cover symbols are used (e.g. *G in Slavic = *g⁽ʰ⁾ or *gʷ⁽ʰ⁾). Forms containing a nasal are presented in shaded cells. Where the presence of a nasal is ambiguous, the cell is shaded in a lighter grey. Several bird names occurring in Baltic, Slavic and Germanic show a conspicuous alternation in the suffixal syllable. It seems quite probable that these can be attributed to a related source. All of them show a morphologically similar structure involving a second syllable of the shape *VND alternating with *V̆ D. The distribution is fairly consistent, with the nasal being absent in Germanic, and Baltic and Slavic adopting an intermediate position. A couple of other European bird names can be noted with a similar structure, where irregularities also support the notion of borrowing. First, there is Lat. hirundō ‘swallow’, which should not be separated from Gr. χελῑδών ‘swallow’ (cf. Chantraine DELG iv: 1253), or from Alb. dallëndyshe ‘swallow’ (cf. Meyer Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 186 chapter 6 1891: 59),30 although they cannot go back to a common proto-form. Here, we find both a disagreement in terms of vocalism and between -r- and -l- (see also fn. 37, below). The variant without the nasal in Gr. χελῑδών strongly recalls the similar phenomenon in our northern European bird names. Another bird name with a similar structure is Lt. balañdis ‘pigeon’ (?~ Lat. palumbēs ‘wood pigeon’); for a detailed discussion of this word, see pp. 209–210. The word for ‘lynx’ is different in that the variant without a nasal occurs in the initial syllable and alternates with a long, acute vowel in Balto-Slavic. Although the word for ‘bast’ is superficially similar, as the nasal in the word for ‘lynx’ may be due to a phenomenon peculiar to (pre-)Greek, and the presence of a nasal in the Baltic word for ‘lynx’ is doubtful, there is no reason to suppose that these two words belong to the same loanword stratum. 6.2 Voicing Alternations Based on examples such as Lat. habeō (< *gʰ-) ~ Go. haban (< *k-) ‘to have’, the existence of Konsonantenwechsel or alternations between different consonant series in the Indo-European proto-language has repeatedly been suggested (e.g. Zupitza 1904: 387–391; Hirt 1927: 297–303; Machek 1934: 7–36; Otrębski 1939: 156–171). These proposals can be seen as reactions against rigid Neogrammarianism, with alternations invoked as an unexplained “mysterious force” awaiting later elucidation. Despite this, some of the comparisons were so tantalizing that the notion has not disappeared from the literature. Yet as the mechanism behind this alleged phenomenon has never properly been explained, it has never quite entered the mainstream, and remains incompatible with a strict application of the comparative method. Comparanda such as those collected by the above authors have also inspired other theories. Both Haas (1960) and Holzer (1989) have assumed the existence of a lost Indo-European language, which has undergone a consonant shift, underlying Latin and Slavic, respectively. While this remains a theoretical possibility, the heavy reliance on root etymologies, many of which often do not fare better than the traditional solutions (Аникин 1992 and in particular Matasović 2013: 77–82), has meant they have had little resonance among comparativists. 30 Alb. d- regularly corresponds to Lat. h- (Alb. dimër ~ Lat. hiems ‘winter’). It must be admitted that the alternative comparison with the Illyrian tribal name Ταυλάντιοι, reported by Hecateus of Miletus to have neighboured the Χελιδόνιοι(!), is tempting (Çabej 1976: 105– 106). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 consonantism 187 As noted above in 5.1 (see point 4), irregular voicing alternations are relatively frequent in words of an expressive character. Other alternations, particularly in final position, might be explained as the result of sandhi phenomena. Such an account is probably necessary for cases like Cz. labuť as against Pl. łabędź ‘swan’ (see above on pp. 176–177). The alternation between OCS (Euch.) дрѫгъ* ‘club, cudgel’, Slk. dial. drúh ‘thick branch’, SCr. dial. drȗg ‘pole, long sick’, against the dialectal variants Slk. drúk, SCr. druk (cf. RJA ii: 807) has been explained by positing a substrate origin (e.g. Derksen 2008: 121; Matasović 2013: 83–84), but given the existence of both variants side by side in the individual languages, this would imply the existence of non-IE groups in Europe practically until the modern period. Considering the improbability of this scenario, we are better off seeking an irregular motivation such as contamination or expressivization (REW i: 374; Liewehr 1956: 20; the Serbo-Croatian form might well originate in a dialect with word-final devoicing, T. Pronk p.c. March 2023). Despite this caveat, there are numerous examples of voicing alternations which, in my view, constitute plausible evidence for non-IndoEuropean origin. The examples below are organized into five groups based on the consonants involved. 6.2.1 Baltic *g⁽ʰ⁾ ∞ *k Elsewhere ‘pigeon (1)’. OCS голѫбь; OE culfre, culufre ‘pigeon’ ~ Lat. columba ‘pigeon’ (Oštir 1921: 49, 1930: 39; Machek 1951a: 103–104; Treimer 1954: 70; Machek 1968: 175; Bezlaj i [1977]: 159; Kleyner 2015: 53–54; ERHJ i [2016]: 284) — For the Old English word, see the discussion on p. 178. The identity of the Slavic and Latin words has long been recognized (already Bopp 1833: 336), but as the comparison is clearly irregular, it is generally rejected, having already been omitted from the fourth edition of Fick’s comparative dictionary (Stokes 1894: 92). Nevertheless, the similarity of the words has remained obvious. Leaving aside the ad hoc notion that the Slavic *g- is simply secondary (Shevelov 1964: 365; Lockwood 1990: 262), this word has been used to bolster theories of contact with unidentified Indo-European languages (cf. Haas 1960: 34; Holzer 1989: 161–162). Соболевскій (1914: 441) proposed that an unknown language had mediated a Latin loanword, while Szemerényi (1967: 20–21) insists on a Latin origin; however, only on the basis that the word cannot be explained within Slavic. What unites all these theories is the assumption that Lat. columba is inherited, of which there is no solid indication. Morphologically, the word is isolated in Latin, aside from the near synonym palumbēs ‘wood pigeon’. The traditional analysis sees these words as containing a compound suffix *-n-bʰ- (Brugmann 1906: 386, Meillet/Vaillant 1933; ЭССЯ vi: 216; Sławski SP viii: 46; Аникин ‣ Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 188 chapter 6 РЭС xi: 146), in which case Lat. columba would be derived from an underlying n-stem continued by Gr. κελαινός ‘black, dark’ (Prellwitz 1897: 102–104; Persson 1912: 169–171; IEW 547–548; Batisti 2021: 206–207).31 For Slavic, Derksen (2008: 175) comments that “the suffix *-(V)mbʰ- is frequent in bird-names”, noting the parallel in Slavic *erę̄bi- (trad. *( j)erębь) ‘grouse’ (on which see pp. 174–176).32 This morphological analysis is based primarily on the co-occurrence of Skt. vr̥ ṣ́ an- and vr̥ ṣabhá-, both appearing in the senses ‘manly’ and as a substantive ‘bull’. Except for the synonym r̥ ṣabhá- ‘bull’ (belonging with Av. aršan‘man, male’), other examples of this pattern of derivation within Indo-Aryan are quite uncertain.33 A close parallel to vr̥ ṣ́ an- beside vr̥ ṣabhá- is nevertheless found in Gr. ἔλαφος ‘deer’, which beside OCS eлeнь ‘deer’, Arm. ełn ‘doe, hind’, has traditionally been segmented ἔλ-α-φος (e.g. Prellwitz 1897: 100; Osthoff 1901: 305–308; Chantraine 1933: 263; Beekes 2008: 402). Despite this, other examples of this supposed compound suffix *-n-bʰ- are sparse, and appear to be limited to European bird names: Lat. palumbēs ‘wood pigeon’ and Arm. salamb ‘francolin’.34 While some productive suffixes in individual branches contain reflexes of *bʰ, their semantic function is not aligned (e.g. the Gr. diminutive -άφιον, the deadjectival OCS зълоба ‘evil’ < зълъ ‘bad, wicked’ and the Gothic adverbial suffix -ba). It therefore remains uncertain whether a suffix *bʰ can be reconstructed (most of the evidence adduced in Hyllested 2009: 202–205 is open to interpretation). Despite the potential derivational parallel in the words for ‘deer’, the separation of Lat. columba from OCS голѫбь feels artificial: the words mean exactly the same thing, and aside from the voicing of the initial stop, show an identical stem. It is a priori questionable that two branches would have used the same inherited suffix only in words for ‘pigeon’, and have independently innovated a word for ‘pigeon’ which happens to be virtually identical. The invalidity of the traditional morphological analysis would seem to be confirmed by other 31 32 33 34 See Batisti (2021: 207, fn. 4 with lit.) for other root etymologies, none of which are any more convincing. As for κόλυμβος ‘grebe’, I fully agree with Batisti that the word should be kept separate. Walde/Hofmann (i: 249) insist that the Slavic word must be native because of the colour term in R голубо́ й ‘light blue’, but this is rather a derivative of the word for ‘pigeon’ (Loewenthal 1901: 31–32; Machek 1951b: 103; Herne 1954: 91). Skt. śarabhá-, a kind of game animal, continued in Dardic and Nuristani in the senses ‘markhor, ibex, mountain goat’ (CDIAL 714) is supposedly connected to Lat. cornū ‘horn’ (EWA ii: 616; Nussbaum 1986: 6), but this is far from certain. Two words for ‘donkey’ — rāsabhá- and gardabhá- — are not well explained; the latter is probably not of IndoEuropean origin (EWA i: 473; cf. Pinault 2008: 393–394). Nothing can be said of Gr. κόραφος (H.), an unidentified bird. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 consonantism 189 irregularities, such as the missing nasal in the Old English form (see p. 178). It therefore seems entirely reasonable to explore the possibility of a non-IndoEuropean origin. Very curious is the Coptic comparandum adduced already by Oštir (1921: 49), cf. Sahidic ϭⲣⲟⲟⲙⲡⲉ, Bohairic ϭⲣⲟⲙⲡⲓ, Lycopolitan ϭⲣⲁⲙⲡⲉ ‘pigeon’, deriving from a Late Egyptian (~ 12ᵗʰ c. bce) form gr-(n)-p.t */kʰV̆ rámpV̆ / (cf. Allen 2020: 115).35 This form is written as ‘gr-bird of the sky’, and as a result has been viewed as a native formation by Egyptologists. Peust (1999: 280) has suggested that the Egyptian form may be the source of the Indo-European words (similarly Иванов 2002). On the other hand, Vycichl (1990: 249) has argued that the Egyptian spelling is folk-etymological (“la colombe n’est pas un «oiseau du ciel» comme l’aigle ou le faucon”), and supported the earlier suggestion of Worrel (1934: 67) that we are dealing with a borrowing from an unknown source. It is in any case clear that Egyptian cannot be the direct source of the European words, due to both a mismatch in vocalism (Latin -umb- requires a labial vowel, cf. Leumann 1977: 81), and Egyptian -r- vs. European -l-.36 The latter alternation is paralleled in the Mediterranean by Lat. hirundō ~ Gr. χελῑδών ‘swallow’.37 In principle, a North African source for a word for ‘pigeon’ would not be in contradiction to the facts of the bird’s domestication history (cf. Batisti 2021: 210), but it must be stressed that little is certain, except for the fact that the pigeon was domesticated extremely early (Gilbert/Shapiro 2013). ‘swan (2)’. Lt. gul̃bė, (Szyrwid, E dial.) gulbìs, Lv. gùlbis ‘swan’; Pr. E gulbis ~ MR колпь (СРЯ 11–17 vii: 254; R ко́лпица) ‘spoonbill’; Kash. kôłp, USrb. kołp (gen.sg. kołpja), SCr. dial. kȗf and kȗp ‘swan’ (Oštir 1930: 66; Derksen 1999; ALEW 432–433) — As to the rare SCr. gȗb ‘swan’, scholars are divided. Some reject it as an irrelevance (Vaillant 1929: 270 “douteux et sans intérêt”; Sławski 1960: 40), while others accept it at face value (Топоров ПЯ ii: 332; ЭССЯ vii: 190; Andersen 1996a: 124, 2003: 68; Derksen 1999: 72, 2008: 97). The SCr. form is indeed very poorly attested, going back to a form ⟨gūb⟩ in J. Stulli’s diction- ‣ 35 36 37 Allen actually reconstructs a final */-nipV̆ /, but apparently only because the Egyptian genitive marker ⟨n⟩ is reconstructed as */ni/. This might be anachronistic, as spellings with ⟨m⟩ are already attested in Late Egyptian (Allen op. cit.; see Erman/Grapow v: 181), suggesting that no vowel was present in at least some Late Egyptian varieties. The spellings with ⟨n⟩ may be etymological, or, as follows from the discussion below, folk etymological. On the nature of Egyptian ⟨r⟩, see Peust (1999: 127–129). And perhaps — if not a mere dissimilation — by Lat. līlium ~ Gr. λείριον ‘lily’. The latter are frequently also connected with Coptic ϩⲣⲏⲣⲉ ‘flower’ < Egyptian ḥrr.t */harīra.t/ (Worrel 1934: 67, Beekes 2010: 845; on the reconstruction, see Vycichl 1990: 94), but this etymology is suspect due to the absence of any reflection of the first syllable in the European languages, and the imprecise semantic match (cf. Vycichl 1983: 310). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 190 chapter 6 ary, where it is attributed to “Gjorg.” (apparently Ignazio Giorgi; cf. RJA iii: 484, where the form is explicitly labelled as doubtful) and a form guf in the 17ᵗʰ century dictionary of J. Mikalja (idem: 495). Such forms otherwise only appear to be attested in lexicographical sources.38 The difference between Baltic *gulb- and Slavic *kulp- (trad. *kъlp-) is already sufficient to suggest a non-IE origin (cf. Derksen 1999: 73 and passim).39 The distribution of the word in Slavic is remarkable, being limited to the peripheral dialects of West Slavic, an isolated pocket in South Slavic, and East Slavic in a secondary meaning. All of this seems suggestive of an archaism: this might be the older Slavic word for ‘swan’, which was later displaced by *albǭdi- (trad. *olbǫdь) in the West and *lebedi- (trad. *lebedь) in the East — a theory supported by the fact that no common Proto-Slavic form for the latter can be reconstructed (see pp. 176–177). ‘dregs’. Lv. (Kurzeme) dradži ‘dregs of melted fat’,40 Pr. E dragios pl. ‘dregs’; ON dregg ‘yeast, (pl.) dregs’ ~ Lat. fracēs f.pl. ‘olive pomace’ (Ernout/Meillet [1951]: 251; Schrijver 1991: 486; Derksen 2008: 121) — The traditional explanation (Walde/Hofmann i: 539; de Vaan 2008: 238; ALEW 248) that the stem-final /k/ in Latin fracēs was carried over from the nominative singular does not hold water, as the word was plurale tantum in Latin, the singular frax only being attested in glosses. Moreover, neutralizations on the basis of nominative forms are generally suspect, as the nominative usually occupies a weak position in analogical processes (see Niedermann 1918: 22–23; and note the discussion in Decaux 1966). As a result, Latin implies an illegal root structure *Dʰ_k- (see 5.3.1). The remainder of the words traditionally adduced here are uncertain. In Slavic, OCS (Ps. Sin.) дрождьѩ ‘dregs (Gr. τρυγίας)’, Pl. drożdże f.pl. ‘yeast, leaven’ suggests an underlying *drazg- or *drazdj-. The old explanation has been to posit *dʰragʰ-sk- for Slavic with the subsequent development to *-gsk- ‣ 38 39 40 I do not have access to all of the sources cited by РСА (iii: 721), but I suspect that most or all of the forms trace back to these two sources. One wonders whether there might have been a confusion among the lexicographers with the Latin loanword attested as dial. gȗb ‘goby’ (РСА loc. cit.) ← Latin gōbius (on this loan and variants, see M. Matasović 2011: 163–164 against ЭССЯ and Derksen loc. cit.). The Sln. dial. (Gorizia) golbica, which Bezlaj (i: 157) adduces in this connection, refers to the ‘skylark’, a tiny passerine bird which has absolutely nothing in common with the swan. Derksen’s inclusion of Pl. kiełb ‘gudgeon’ and OIr. gulban ‘sting; beak’ in a substrate context looks like an unnecessary stretch to me, as the semantic link between the three word families is not self-evident. Prussian Lithuanian drãgės (attested as dragges in Bretke) is most likely a loanword from Prussian (Žulys 1966: 151–152). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 consonantism 191 > *-zg- (e.g. Berneker i: 228; REW i: 371), for which cf. OCS про-брѣзгъ ‘dawn’ beside Skt. bhrā́jate. Regardless of whether this development is phonologically regular, the explanation is inadequate as there is insufficient evidence for a nominal suffix *-sk-. If we posit an underlying form *drazdj-,41 another reasonable etymology presents itself, namely a comparison with OE dræst, dærste ‘dregs, leaven’, OHG trestir ‘pomace’ < *dʰrosd-. Per tradition, the Germanic forms are derived from *drahstu- (i.e. *drag-stu-), to the root of ON dregg (Holthausen 1934: 69; IEW 251–252); cf. the semantically similar Go. maihstus*, OHG mist ‘dung’ < *mihstu-. However, the latter is no example of a suffix *-stu-, as it derives from a more primary *mihsa-, cf. OE meox ‘dung, filth’ (Kroonen 2013: 369), and is ultimately deverbal, cf. OE mīgan ‘urinate’. For *drag-, we have neither a verbal root nor a primary *-sa- derivative, which means that the alternative reconstruction *drastV- remains a clear possibility. Contra Meyer (1891: 72) and Demiraj (1997: 141), Alb. drā ‘dregs of melted fat’ cannot derive from *dragā, as *g was not lost intervocalically (cf. Schumacher 2013: 240). A possibility would be to posit a preform *drasā- < *dʰrHs-, and compare OE drōsna, Du. droes, droesem ‘dregs, sediment’ < *dʰroHs-. ‘scythe’. Lt. dal̃gis, dal̃gė, Pr. E doalgis ‘scythe’ ~ Lat. falx -cis ‘sickle, scythe’ (Alessio 1946: 165) — This rather self-evident comparison (cf. Mikkola 1899: 74; Hirt 1927: 299) has generally been disfavoured in view of the irregular Latin /k/ and irregular vocalic correspondence (Walde/Hofmann i: 449–450; LEW 81; against an analogical origin of the /k/, see above on ‘dregs’). The Latin word has been suspected to be of foreign origin, but the Baltic equivalent is rarely mentioned in this connection (e.g. Ernout/Meillet 214; Boutkan/Siebinga 2005: 75; de Vaan 2008: 200). For Latin, a non-IE origin is supported by the illegal root structure *Dʰ_k- (see 5.3.1). Walde/Hofmann (loc. cit.) claim that dal̃gis is “aus semasiologischen Gründen” better compared with OIr. dluigid ‘split, cleave’, ON telgja ‘carve, hew (wood or stone)’ (also Trautmann 1923: 44; IEW 196). This is rather a strange argument, since falx and dal̃gis mean exactly the same thing, and the cited verbs are semantically rather remote, belonging to the sphere of artisanry rather than agriculture.42 ‘rye’. Lt. rugiaĩ, Lv. rudzi, Pr. E rugis; R рожь, Sln. ȓž; ON rugr, OE ryge ‘rye’ ~ MW ryc ‘rye’ (Hoops 1915–1916: 509–510; Walde/Pokorny ii [1927]: 375; ‣ ‣ 41 42 Sln. drǫ̑zga ‘pulp, dregs of lard’ does not disprove the reconstruction with *zd, cf. Sln. drǫ̑zg ‘thrush’ < *drazda-; see p. 204. On the other hand, this Sln. word may not belong here, as it is highly reminiscent of the synonym trǫ̑ ska (cf. SCr. trȍska, arch. trȕska (RJA xviii: 829) ‘slag’). Note that, according to Schumacher (2004: 284–285), the Irish verb is rather to be reconstructed *dlug-, and a connection to the Germanic root is thus impossible. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 192 chapter 6 Schrader/Nehring ii [1929]: 265; Charpentier 1930: 71; Porzig 1954: 143; Markey 1989: 595; Polomé 1992: 70; OED³ s.v. rye)43 — All of the cited forms show a formant -i-. Divergent forms are attested in Continental Germanic, where we find OS rokko* (attested rogko), OHG rocko as against Old Frisian rogga, MDu. rogge. This vacillation in voicing is to be explained from an old n-stem (Kroonen 2011: 23). This is probably a localized innovation, however, and Kroonen (2013: 417) points out some possible West Germanic traces of *rugi-. MW ryc has generally been derived from Old English ryge (Schrader/ Nehring ii: 265; GPC iii: 3136; Witczak 2003: 110), but this is chronologically difficult, as Welsh /k/ could hardly be a substitute for the OE spirant /ʝ/, cf. MW pabi ‘poppy’ ← OE (Ælfric) papig (early ME papig), MW llidiat ‘gate’ ← OE hlidgeat ‘swing gate’ (cf. Parry-Williams 1923: 41–42).44 On the other hand, if the loan were of Proto-Celtic age, one would expect Celtic *g (> Welsh **∅).45 At face value, the Welsh data points to *ruki̯o- or *rukī-, showing a mismatch compared to the *gʰ elsewhere. Beyond Indo-European, a similar word is found in several Uralic and Turkic languages. Already Paasonen (1906: 2–3) recognized that the Mordvin and Permic words for ‘rye’ cannot be derived directly from Russian, as had previously been thought (e.g. Thomsen 1890: 213). For Mordvin roź, the problem is the final ź, which otherwise does not substitute Russian ž (pace Häkkinen/ Lempiäinen 1996: 169). In Permic, we have Komi ruʒ́eg̮ , Udmurt ʒ́eg, dial. ʒ́iźeg. Already the development *r- > ʒ́- in Udmurt speaks against a Russian loan, but the palatal affricate and suffix solidify this impression. The initial vowel in Udmurt reflects earlier *i̮ (< *u, Лыткин 1964: 215–218) which has become fronted by the following palatal, cf. Komi ruć, Udm. ʒ́ići̮ ‘fox’. The syncope in the standard language is attested dialectally in other lexemes, e.g. dial. slal 43 44 45 Here one has often included a form βρίζα (e.g. ΙΕW 1183), a crop which according to Galen was grown in Thrace and Macedonia, resembling τίφη ‘einkorn’, and from which a black and malodorous bread was made (cf. Schrader/Nehring ii: 265). The word is found in the sense ‘rye’ in modern Greek dialects, first resurfacing in a 16ᵗʰ century Macedonian–Greek glossary as ἄρжυ · βρίζα (cf. Mac. рж ‘rye’; Giannelli/Vaillant 1958: 32). Despite its meaning, it is perhaps better connected to a different Wanderwort, represented by Gr. ὄρυζα, Pashto wríže, Skt. vrīhí- ‘rice’ (Георгиев 1957: 55); for the shift to another kind of grain, cf. Kati wriċ ‘barley’ (CDIAL 708; Kümmel 2017b: 281). These could both be from Middle English according to GPC iii: 2663 (s.v. pabi) and GPC i: 1297 (s.v. fflodiart), but this is of little relevance if the spirantization of /g/ is dated to the continental Old English period (Campbell 1959: 173). Cf. MW meu-dwy ‘hermit’ (duw ‘God’), MCo. maw ‘boy, servant’ ← Germanic *magu > OE magu ‘boy, young man’ (van Sluis et al. 2023: 201, 212). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 consonantism 193 ‘salt’ beside literary si̮lal (Перевощиков 1962: 37–38). Thus, we can confidently reconstruct a Proto-Permic *ruʒ́ɛg ‘rye’.46 In Turkic, we find rather a similar situation, although little recognized. Ахметьянов (1981: 48–49) has argued that Tatar and Bashkir arïš ‘rye’ cannot have been adopted directly from Russian (thus e.g. Joki 1973: 162), as neither the prothetic a- nor root vowel can be accounted for (one would rather anticipate **ïruš or **erüš). Despite Федотов (ii: 474), a similar conclusion must be drawn with respect to Chuvash ïraš ‘rye’, as ï- never occurs as a prothetic vowel, nor is -a- for Russian stressed -o- the usual substitution.47 The correspondence of Chuv. ï- as against a- elsewhere is in fact more typical of inherited vocabulary, where it would reflect a Proto-Turkic phoneme notated *ạ (in the Russian school) or *ë (e.g. Doerfer 1971: 340–341). Reconstructing a word for ‘rye’ back to Proto-Turkic is suspect, however, as early rye cultivation is normally associated with Central Europe (cf. Hillman 1978: 157–158; Напольских 2006: 5–6; 2010: 56). Paasonen (op. cit.) assumes the word for ‘rye’ was adopted into the Uralic languages from Iranian, or more specifically, Scythian. Of course, as long as no Iranian equivalent is attested (cf. Kümmel 2017b: 283 on the alleged Pamir words), this remains purely hypothetical. Slightly better is Guus Kroonen’s suggestion (see Kroonen et al. 2022: 22) of an early Slavic loanword mediated by “steppe Iranian”. Although still hypothetical, this would obviate the need to reconstruct the word for Proto-Indo-Iranian. Furthermore, a couple of other agricultural Wanderwörter seem to have passed into Scythian from a Balto-Slavic dialect, most notably Oss. I ᵆxsyrf, D æxsirf ‘sickle’ ← Lv. sìrpis; R серп, etc. (Abaev 1965: 8–9; Gołąb 1992: 333).48 While the Slavic → Scythian route perhaps makes the most sense, a theoretical pre-Oss. *ruʒ(-æg) would hardly account for the Turkic evidence. If the Turkic forms belong here at all, then perhaps we can assume the initial *ạ/ëwas some kind of prefixal element or the like (see 7.1.2), but in the absence of parallels, the idea must be approached with caution. The Uralic and Turkic pal- 46 47 48 Mari E urža, W ə̑rža, rə̑ža ‘rye’ is indeed probably loaned from Russian. As a precise source, the final -a is best accounted for starting from gen.sg. ржа from R dial. рож (m.) (Orenburg, etc., see СРНГ xxxv: 146). Exceptions like Chuv. salat < со́лод ‘malt’ are rather to be explained from end-stressed forms (со́лод is originally accentually mobile, cf. Зализняк 2019: 541). The suffix *-ɛg seems to almost call for an Iranian origin and comparison with Oss. -æg (cf. Paasonen 1906: 4–5), as in Old Permic ⟨идог⟩ */ide̮g/ ‘angel’ (cf. Лыткин 1952: 65, line 27; also idem: 70, fn. 4 and idem: 130) ← Oss. D idawæg ‘angel, spiritual guardian’ (Абаев 1958: 348–349; Rédei 1986: 70). Not all cases of the suffix can be explained as Iranisms, however. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 194 chapter 6 atal affricates, by the way, do not necessitate a Slavic origin, as palatalization is a trivial change which could have occurred independently before a following *i or *i̯ in another (hypothetical) source language. It has long been suspected that the word for ‘rye’ is of non-Indo-European origin, although primarily on the basis of non-linguistic facts. The Celtic form (see above) can now provide some more concrete linguistic evidence in favour of this analysis. It is clear we are dealing with a cultural Wanderwort whose spread is difficult to precisely trace. Rye was first domesticated in Eastern Turkey and Armenia, but already arrived in Northern Italy in the Neolithic (Zohary/Hopf 2012: 63–66); however, the sporadic finds in later Polish sites are probably more consistent with the plant being tolerated as a weed than intentionally cultivated (Behre 1992: 142–143). Rye cultivation only really took off in Northern Europe during the Iron Age, and does not appear to have reached the Eastern Baltic until the common era (Grikpėdis/Motuzaitė Matuzevičiūtė 2016). On this basis, it would be tempting to interpret the East Baltic forms as Germanic loanwords, which would be phonologically unproblematic; compare similarly Lt. kviečiaĩ ‘wheat’, which I have interpreted as a Gothic loanword (see p. 40). A Germanic loanword cannot be ruled out on phonetic grounds for Slavic, although such an assumption would not be necessitated by the realia. ‘hornbeam’. R граб, Cz. habr, dial. hrab, SCr. grȁb ‘hornbeam’ ~ Lat. carpinus ‘hornbeam’. Here also belong Lt. skrõblas, skróblas, (S Aukšt.) skrúoblas ‘hornbeam’49 (Machek 1950b: 152; Holub/Kopečný 1952: 118; Matasović 2013: 84; ERHJ i [2016]: 291; Matasović forthc.) — The original Slavic form can be reconstructed *grābra- (trad. *grabrъ) with various dissimilations (Berneker i: 343; Skok i: 598), cf. SCr. dial. gràbar, Sln. dial. grȃbər.50 Perhaps here also belong ‣ 49 50 The literary standard is circumflex, although -ó- seems better supported dialectally. The variant with -úo- is in any case irregular (contamination with gúoba ‘elm’, úosis ‘ash’, glúosnis ‘willow’?). The initial s- in Baltic is also unclear. Otrębski (1955: 29; cf. 1939: 167) finds a parallel in Lt. strãzdas ~ R дрозд ‘thrush’, yet here we are probably dealing with anticipation of the second *s (see the discussion on p. 204). The generalization of /grab/ in East Slavic is not surprising considering the partial parallel in R брат ~ Cz. bratr, Sln. dial. brȃtər ‘brother’ (Holub/Kopečný 1952: 118). The two variants must clearly not be separated (despite Būga 1922: 82; LEW 176–177). Būga, followed by Boryś (2008: 176), has also adduced Pl. dial. gab, gabina ‘elm’, attested in transitional Polish-Belarusian dialects (= Bel. dial. габ, габíна). In view of their distribution, these words must no doubt be considered Balticisms (Лаучюте 1982: 43–44). Note that a loan directly from Lt. gúoba ‘elm’ is prohibited by the Slavic vocalism, so it would be preferable to posit a Prussian *gābas as the immediate source. The different form and semantics imply it should be separated from our words for ‘hornbeam’. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 consonantism 195 Lv. skābardis (skãbârdis² Dunika, EH ii: 503, skãbàrdis LVPPV) ‘beech’, Pr. E stoberwis (corrected to *sc-) ‘hornbeam’, assuming an original *skrābar- with dissimilatory loss of the first *r (Trautmann 1910: 439), although the formation of these words is obscure (cf. ME iii: 878; IEW 945). The comparison between Slavic *grābra- (trad. *grabrъ) and Lat. carpinus is obviously semantically attractive. The Latin word has no satisfactory etymology. The connection to carpō ‘pluck’ (supposedly < *‘cut’) based on the hornbeam’s crenated leaves (Walde/Hofmann i: 171; Schrijver 1991: 430) is hardly logical, as the plant’s leaves are neither sharp nor capable of cutting, nor for that matter, strikingly different to those of the elm or beech.51 The etymological equation involves multiple irregularities. First of all, the labial stop alternates in voicing along with the velar in a similar way to Lt. gul̃bė ~ Kash. kôłp ‘swan’, above. In addition, there is a metathesis of *r, which does not appear to have reliable parallels in my corpus. An alternative analysis is to assume that the Latin form goes back to an earlier *crarp- by dissimilation, as probably in prōcērus ‘lofty’ (~ crēscō ‘grow’; Leumann 1977: 315; de Vaan 2008: 491), prō portione ‘in proportion’ (with portione < *prō ratione, Ernout/Meillet 524). In this case, we would have a potentially more trivial metathetic relationship between *-Pr- and *-rP-.52 [‘oriole’. Lt. volungė ̃ 3ᵃ, Lv. vãluôdze ‘oriole’; Pl. wilga, Sln. vółga; R и́ волга, Bg. авлѝ га ~ ME wode-wale, MHG wite-wal ‘oriole’ — See the discussion on pp. 178–180.] † ‘many’. OCS мъногъ ‘many, numerous’; Go. manags, OHG manag ‘many’ ~ OIr. meinic, MW mynych ‘frequent’ (< *menekki-) (Boutkan 1998: 124; Schrijver 2001: 422; Boutkan/Siebinga 2005: 256–257; Philippa et al. iii [2007]: 334; Kroonen 2013: 352; Matasović 2013: 265; ERHJ i [2016]: 625; van Sluis forthc.) — The main question is whether the Slavic word can be interpreted as a Germanic loan (thus Hirt 1898: 355). The cost of this assumption would be an irregular raising *a > *u (trad. *ъ) in an unstressed syllable within Slavic (Младенов 1909: ‣ ‣ 51 52 A relationship with Umbrian krapuvi dat.sg., an epithet of Mars and Jupiter (Kretschmer 1921) cannot be demonstrated (cf. Untermann 2000: 309–310 with lit.). Comparing Hitt. karpina- ‘a kind of tree or bush’ (IEW 944; Puhvel 1997: 99) is also precarious in view of its uncertain meaning. I would also like to keep Lt. skir̃pstas ‘elm; alder buckthorn’, Pr. E skerptus ‘elm’ apart due to the semantic and formal difference (note also the Lithuanian circumflex; but it must be admitted that skir̃pstas has also been recorded in the sense ‘hornbeam’). Compare OSw. hagre as against MIr. corca ‘oats’ (van Sluis et al. 2023: 219; however, more sceptically: Kroonen et al. 2022: 20). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 196 chapter 6 85) for which a couple of parallels may be found.53 As the Germanic *-g- could just as well reflect *-k- in this position, this would then not be a certain example of a voicing alternation.54 Kortlandt (2007: 9) claims that мъногъ has developed from *min-aga- (trad. *mьnogъ) as a result of vowel assimilation, comparing Lt. minià ‘crowd’. However, it is doubtful that the Lithuanian word belongs here (LEW 453; ALEW 753 and Smoczyński 2018: 804–805 all accept a derivation from mìnti ‘tread, trample’; compare French foule ‘crowd’ < fouler ‘trample’, FEW iii: 846). Moreover, although the “normalized” form мъногъ is found in dictionaries, the OCS word is very frequently written мног-, showing an early reduction; it is uncertain how much the spelling мъног- can be relied on.55 While the geminate in Celtic is very difficult to explain in an IE context and could very well point to a non-IE borrowing, the difficulty in analysing the word within Slavic and the possibility of a Germanic mediation means that this word cannot be used here as an example of a voicing alternation. † ‘naked’. OCS (Supr.) голъ ‘naked’; OE calu, OHG kalo* ‘bald’ ~ Lat. calvus ‘bald’ (Philippa et al. ii [2005]: 593–594) — Despite the striking correspondence between the substantivized Lat. calva ‘bald head’ and OCS глава, Lt. galvà ‘head’ (Derksen 2008: 176), the comparison is probably false. West Germanic *kalwa- matches Latin calvus formally and semantically, and is therefore most easily explained as a loan from Latin (Senn 1933: 521; FEW ii: 106; cf. Philippa et al. loc. cit.).56 The Latin word must reflect *kalawo- < *klH-eu- (Schrijver 1991: 299) and can hardly be separated from Skt. kulva-, YAv. kauruua- ‘thin-haired’ (< *klH-uo-, on which see Lubotsky 1997: 144). In view of the Indo-Iranian com- ‣ 53 54 55 56 The only relatively clear example is the verb ‘to want’: cf. Pl. chcieć ‘to want’ vs. сhосiаż ‘although’ (= R хотѣ́ть, хотя́ ). A similar change has also been suggested in OCS съто ‘hundred’ if a loan ← Iranian *sata (cf. Vasmer 1913: 176; Шахматов 1916b: 29, Arumaa 1964: 130) and perhaps Pl. młyn, Cz. mlýn ‘mill’ if cognate with Pr. E malunis ‘mill’ (cf. Meillet 1907: 373–374; but compare Fraenkel 1951: 129). Suffice to say that neither of these parallels are uncontroversial. On a similar sporadic raising *e > *i (trad. *ь) before a palatal, cf. Kortlandt 1984–1985. It is unclear to me why Младенов (loc. cit.) and Kiparsky (1934: 75) after him insist that the Slavic and Germanic words must be cognate (cf. Viredaz 2020: 413–415). Compare the similar situation with regard to the inst.sg. мъноѭ, also seen as an example of this assimilation by Kortlandt, but essentially representing a “traditional” OCS form, not based explicitly on facts of the language (cf. Vaillant 1958: 446; Lunt 2001: 77). Leskien (1922: 109) has even considered the dative variant мьнѣ, on the contrary, to have arisen by assimilation from мънѣ. See also the discussion in Kapović (2006: 39–41), with lit., who problematizes the dative, but remarkably takes the form of the instrumental for granted. This possibility is denied by EWAhd v: 353, but without any argumentation. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 consonantism 197 paranda, it is more likely that the Balto-Slavic words are unrelated. They may instead be cognate with Arm. čeł ‘bald’ < *g⁽ʷ⁾el- (Olsen 1999: 206). 6.2.2 Baltic *k ∞ *g⁽ʰ⁾ Elsewhere ‘pear’. Pr. E crausy ‘pear tree’, crausios ‘pears’ ~ R гру́ ша, Cz. hruška ‘pear’ (Hehn 1870: 454; Trautmann 1923: 140; Machek 1954: 114; LEW [1962]: 296; Matasović 2013: 92) — According to Būga (1915: 342), Žem. kriáušė and Lv. dial. (Nīca) kraûsis² (ME ii: 264) are loaned from Prussian, which is supported by their marginal distribution (see also Топоров ПЯ iii: 168–169; Žulys 1966: 152; Аникин 2004: 380). The Slavic reflexes are as follows. In East Slavic, we find only гру́ ша,57 a form which is otherwise only known in West Slavic (e.g. Pl. grusza, Cz. hruška). Pleteršnik (i: 258) cites a Sln. grȗška, but this form is actually a normalization of dialectal /hrùːška/ (Karničar 1986: 153) and might represent a localized borrowing from a dialect with a realization /hr-/ < *xr-, cf. standard Sln. hrȗška (T. Pronk p.c. March 2023). South Slavic more typically shows *k(Bg. кру̀ ша, SCr. krȕška),58 and such a variant occurs as a relic in West Slavic, cf. USrb. arch. krušej (gen.sg. krušwje), LSrb. kšuška; Kash. krëszka (also Pl. dial., cf. Popowska-Taborska 1996: 154). As a result, the word for ‘pear’ in Slavic shows a partial complementary distribution. At first sight, the peripheral attestation of *k- in West Slavic would point to an archaism, suggesting that *g- has spread through this territory secondarily (although still in the preliterary period). If true, this would allow us to draw an earlier isogloss between East Slavic *g- and West/South Slavic *k-, which would be somewhat reminiscent of the situation with ‘oriole’ and ‘swan’ (see under 6.1.1). Since the 19ᵗʰ century, the consonantal alternation has been considered evidence that the word for ‘pear’ is a loanword from an unknown language.59 ‣ 57 58 59 It is tempting to see RCS хроуша as an early reflection of the Ruthenian ‘spirant g’, even though Shevelov (1979: 351) simply dismisses it as a scribal error. Curious is the form кроуша, which glosses the Greek name Апии (corrupt for Σινάπης!) in the Chronicle of George Hamartolos (СДРЯ 1338). A Bulgarian form? (cf. Пичхадзе 2002). Sln. hrȗška, SCr. dial. (NW) hruška presumably show a secondary spirantization (dissimilation?). Schrader (1901: 93; Schrader/Nehring i: 148) refers to a Kurdish “korêši, kurêši”, which has been routinely mentioned in later works (e.g. Berneker i: 358; LEW 296; REW i: 314; Sławski SP viii: 256). Schrader’s immediate source appears to be Rhea (1872: 145: “korēshī or kurēshī, n. pear”). However, I am unable to trace this form elsewhere. Noting that Rhea fails to distinguish /q/ and /k/ (see the editor’s note op. cit.: 120), Patrick Taylor (p.c. June 2022) has attractively suggested that the Kurdish word could represent the common surname Qureyşî, and that this would have referred to a cultivar associated with someone of that name. As a typological parallel, he notes the grape variety Kureyş (üzüm) found in eastern Turkey. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 198 chapter 6 This seems quite probable, and could perhaps explain the divergent forms within Slavic. Smoczyński (2018: 603) would rather posit an irregular “sonorization” in part of Slavic, which is ad hoc. No better is the native etymology in ЭССЯ (vii: 156–157) comparing MR крушити ~ R dial. груши́ ть ‘break up, crumble; destroy’ — an implausible suggestion from a semantic point of view (cf. Matasović 2013: 92). † ‘meadow’. Lt. lénkė ‘depression, marshy spot’;60 Pl. łąka ‘meadow’, Sln. lǫ́ ka ‘damp meadow by a river’ ~ R луг ‘meadow’, Pl. dial. (Sł. Warsz. ii: 805) łąg ‘flood meadow by a river’ (Derksen 2008: 288; 2015: 279–280) — Derksen further adduces Lt. líeknas, Lv. liẽkna ‘depression; marsh, swamp’ and a number of other forms.61 It can be difficult to tease the k-forms apart from derivatives of the root *lenk- ‘to bend’, cf. Lv. lañka ‘low-lying meadow; bend in a river’, R лука́ , Bg. лъка̀ ‘river bend; meadow in a river bend’, and if lénkė is metatonical (Derksen 1996: 200), there really seems to be no decisive argument against this internal etymology. Similarly, Slavic *lǭga- (trad. *lǫgъ) might be explained as an inner-Slavic derivative based on the present stem *lę̄g- (trad. *lęg-) ‘to lie’ (OCS 1sg.pres. лѧгѫ; see Loma 2012: 84); for the semantics, compare R лог ‘broad valley; (dial.) low-lying, damp spot; water meadow’ (cf. СРНГ xvii: 103), an undisputed derivative of лечь ‘lie’ (e.g. REW ii: 51). ‣ 6.2.2.1 Conclusion The certain and possible evidence for voicing alternations involving velars is collected in Table 9, overleaf. The principles used in this table are the same as for Table 8 (p. 185). In addition, forms which do not provide relevant data are presented in light grey. Shaded cells indicate forms containing a voiceless velar. The parallelism between these examples is quite striking. In all of the examples except ‘pear’, Balto-Slavic almost consistently shows a voiced velar, while Italo-Celtic shows a voiceless one. Seven examples not only showing a similar alternation, but also a largely matching distribution, can hardly be a coincidence. This correlation is most straightforwardly explained as a reflection of a genuine dialectal difference in the underlying source language. The 60 61 On léngė, see Chapter 1, fn. 61. I think it is going too far to include e.g. Sln. lúža ‘puddle’ and Lt. dial. liū̃gas (beside liū́gas) ‘puddle, marsh’, which lack the nasal. The Lithuanian word is always analysed as cognate, but I wonder whether it is rather a loan from Belarusian луг ‘meadow; swampy area’ — the initial /lʲ/ could be explained through contamination with the semantically similar liū́nas ‘swamp’. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 199 consonantism table 9 Possible examples of *g⁽ʰ⁾ ∞ *k alternations Baltic Slavic Germanic Elsewhere Lat. *kolombʰ- ‘pigeon (1)’ – *Golambʰ- *gulu[bʰ]- ‘swan (2)’ *Gulbʰ- *kulp- – ‘dregs’ *Dragʰ-i̯- – *dʰra[gʰ]-i̯- Lat. *dʰrak- ‘scythe’ *Dalgʰ- – – Lat. *dʰalk- ‘rye’ [*rugʰ-i̯-] *rugʰ-i̯- *ru[gʰ]-i̯/n- Celt. *rukī- ‘hornbeam’ *skrâB-r- *GrâB-r- – Lat. *k(r)arp- ‘oriole’ *u̯ âlanG- *u̯ (i)lg⁽ʷ⁾- *u̯ alk- ‘pear’ *kraus-i̯- *graus-i̯*kraus-i̯- Eg. *kʰVramp- – Slavic word for ‘swan’ is at first sight an exception, but we may consider this a function of the intermediate position of Slavic, which would enable contacts both with the Mediterranean and with Northern Europe. The Germanic evidence in this section is largely obscured by Verner’s Law: the words for ‘dregs’ and ‘rye’ could equally be taken back to an earlier *k and final stress. However, the word for ‘oriole’ appears unambiguously to imply *k. As will emerge from the following sections, examples of voicing alternations involving stops other than velars are relatively few. This might potentially be connected to the cross-linguistic tendency of /g/ towards lenition (Foley 1977: 25–35), exemplified by the Central European areal change *g > /ɣ ~ ɦ/. In languages which lack a phoneme /g/, such as Czech, foreign /g/ may be substituted with /k/ in loanwords, e.g. Czech dial. kuláš ‘goulash’, brikáda ‘brigade’ (ČJA v: 317). Thus, one possible explanation for a general trend towards the devoicing of velars in the south might be the mediation of an unattested language which lacked a phoneme */g/. Of course, this remains purely speculative, and one Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 200 chapter 6 might object that the word for ‘pigeon’ is already attested with a “fortis” */kʰ-/ in Late Egyptian, some two millennia before it emerges with a */g-/ in Northern Europe. 6.2.3 Alternations Involving Dentals ‘drone’. Lt. trãnas ‘drone’, R тру́ тень, Pl. arch. (Sł. Warsz. vii: 130) trąd, Sln. trǫ̑ t ‘drone’ ~ Gr. (Nicander) τενθρήνη ‘wasp’; OE dran, OHG treno ‘drone’ (Kuiper 1956: 221–222; Beekes 2010: 105; Matasović 2013: 96; van Sluis 2022: 12– 15, forthc.) — Kuiper and Beekes cite a plethora of variants for the Greek word, but it is all but impossible that each of these is equally old, and their sheer number would only support the notion of secondary developments. Most likely, we are dealing with multiple lemmata which have influenced each other (cf. Chantraine DELG i: 90). If we accept the derivation of ἀνθηδών ‘bee’ from ἄνθος ‘flower’ (Chantraine 1933: 361; cf. Frisk i: 108), then a collision with τενθρήνη ‘wasp’ would explain variants such as ἀνθρήνη (Aristotle) ‘a kind of hornet’ (differently Chantraine DELG loc. cit.). The apparent reduplication in τενθρήνη is reminiscent of (Epic) δένδρεον ‘tree’ (< *der-drew-, cf. Chantraine DELG i: 263). The only other form relevant for our purposes is the Hesychian gloss θρώναξ · κηφήν ‘drone (Laconian)’.62 OE dran and OS dran, drano (> MoLG Drahn) ‘drone’, most probably with short vowel (OED³ s.v. drone n.¹), differ in vocalism from OHG treno ‘drone’. As the OS variant drenon (acc.pl.) may be the result of a secondary development in the neighbourhood of /r/ (Cordes 1973: 137), the form with *e seems essentially to be limited to High German. MoE drone, attested since the 15ᵗʰ century, does not represent a regular continuation of the OE form and Kroonen (2013: 101) has argued that this, like MDu. dorne, could represent an additional ablaut variant *drunan-. While it is possible that the vocalic alternations could be explained by positing various ablaut grades in Germanic and Greek, the number of variants which have to be assumed makes this quite unattractive. Šorgo (2020: 437) rejects a non-Indo-European origin, prefering the traditional explanation that the whole family is of sound-symbolic origin (Walde/ Pokorny i: 861; Frisk i: 681–682, etc.). Certainly, some of the variants may be explained in this way; for instance the variant *drunan- could plausibly have ‣ 62 Unreliable is ἀθρήνη, only attested by Byzantine-period lexicographers. The forms θρήνη and θρηνῶδες are additionally cited by Beekes (2010: 105; evidently taken over from Winter 1950: 45). The former is a hapax in Eustathius (12ᵗʰ c. ce) and is probably a corrupt form, while the latter is a manuscript variant of τενθρηνιῶδες ‘honey-combed’ (“in der Überlieferung stark entstellt, z.T. zweifelhaft”, Frisk ii: 877). None of this evidence can be used to support pre-Greek origin. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 consonantism 201 arisen under the influence of MoE drone (since 16ᵗʰ c.), MDu. dronen ‘hum, buzz’ (OED³ loc. cit.). However, it is difficult to justify this analysis in detail. Other forms cited in this connection, e.g. Gr. θρῆνος ‘lament, dirge’ (: τενθρήνη), Pr. iii trinie 3pres. ‘threaten’ (: Lt. trãnas) (Frisk loc. cit.; Endzelīns 1943: 266; LEW 1110–1111), are semantically ambiguous and the assumption of an underlying Schallverbum remains without direct support. ‘reed’. Lt. dial. trùšis ‘reed’, Lv. dial. trusis ‘rush, bullrush’ (ME iv: 248; EH ii: 699); OCS тръсть, R arch. трость, OPl. treść (SSP ix: 184) ‘reed’ ~ Gr. θρύον ‘reed’ (Kuiper 1956: 224; Furnée 1972: 135; Beekes 2000a: 28; Matasović 2013: 88) — The OCS variant трьсть, along with R dial. (N) тресть are to be explained as cases of yer assimilation (Соболевскій 1910: 116–117).63 The Lithuanian variant triùšis has been seen as paralleling the Slavic variants (Булаховский 1958: 91), but is rather to be explained as a result of the frequent but sporadic dialectal change /Cr/ > /Crʲ/ (Zinkevičius 1966: 153–156). An innerBaltic derivative with ablaut is Lt. (Žem.) triaũšiai ‘horsetail, Equisetum’, Lv. (Stender apud ME iv: 227) traušļi ‘Flusskannenkraut’, i.e. Equisetum fluviatile(?); the further comparison with Lt. triáušėti ‘crack, split (usu. of hairs)’ (cf. Būga 1922: 288; LEW 1133) is semantically unattractive. The initial aspirate in Greek is not consistent with Balto-Slavic *t-. Smoczyński (2018: 1530) is willing to accept anticipatory aspiration due to *s (cf. Sommer 1905: 46–82; Chantraine DELG ii: 443). However, this development is assessed as highly doubtful by Frisk (i: 688), and is probably to be rejected. The Greek word is no longer mentioned by ALEW (1303), who leave the Balto-Slavic word without an etymology. It seems the Greek form can hardly be separated, but in view of the incongruent initial stops, the words cannot be directly cognate. Therefore, the suggestion of independent loanwords from an unknown source can be considered attractive. For a suggestion regarding Slavic *-st-, see under ‘furrow’ (p. 224). ? ‘lentil’. RCS лѧча, SCr. léća, Bg. лѐща (< *lę̄tjā-); Lat. lēns -tis ‘lentil’ ~ Gr. λάθυρος ‘grass pea’ (Hoops 1905: 463; Walde/Hofmann i: 783; LEW [1962]: 359; ЭССЯ xv [1988]: 63–65; etc.) — If the suffix *-jā- is an inner-Slavic innovation, it cannot entirely be excluded that the word was borrowed from Latin. However, it is difficult to explain OHG linsī ‘lentil’ as a Latin loanword (EWAhd v: 1323; Kluge/Seebold: 580).64 MDu. (15ᵗʰ c.) lins ‘lentil’ could phono- ‣ ‣ 63 64 Note that contra Соболевскій, Pl. trzcina, Cz. třtina ‘reed’ do not show a reflex of a front vowel, but have /r̥/ < *rs as in Cz. křtíti (OCz. krstiti), Pl. chrzcić ‘baptize’ < *kristītī (trad. *krьstiti), cf. Lamprecht, Šlosar & Bauer 1977: 71. A possible parallel for a borrowed nominative form is OS pavos, OHG bābest ‘pope’ (for the long ā, cf. bâbes in Notker, and also the loanword OCS папежь ‘pope’, ESJS 625). However, Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 202 chapter 6 logically be cognate, but has also been interpreted as a loan from German (de Vries 1971: 404–405, s.v. linze). As a result, a Proto-Germanic age is not ascertained, but the absence of the word in the other branches of Germanic may simply be due to the absence of the crop in northern Europe. A German origin is hardly possible for Lithuanian lę̃šis ‘lentil’ (cf. the recent loanword lìnzė ‘lens’).65,66 Due to the nasal vowel, a Slavic origin is also implausible (see 1.1.4). Thus, this form is rather a conundrum: as lentils do not emerge in the archaeological record for Lithuania until the Middle Ages (Grikpėdis/Motuzaitė Matuzevičiūtė 2020: 167), it is difficult to accept a non-Indo-European origin, but an exact source cannot be established with confidence. In Latvian, lȩ̃ca ‘lentil’ must be explained as a loan from East Slavic *ляча (cf. ME ii: 455; REW ii: 84),67 whence it has been adopted into Estonian as lääts ‘lentil’. The comparison with Greek has not been universally accepted (cf. e.g. Berneker i: 708; REW ii: 84; Frisk ii: 71), and indeed it is based on rather little linguistic material and depends on the ultimately unprovable assumption that Greek -ά- goes back to an earlier nasal vowel. Since the word also refers to an edible legume, it may well belong here, but the evidence remains uncertain. While it is probable that our word for ‘lentil’ is of non-IE origin, the clearest irregularity is between the Germanic sibilant on the one hand and the dental in Latin on the other. It is uncertain to what extent the Balto-Slavic evidence is relevant here. ? ‘lightning’. Pr. E mealde ‘lightning’ ~ ON poet. mjǫllnir ‘Thor’s hammer’; MW mellt pl. ‘lightning’ — ON mjǫllnir must reflect *melþuni- (cf. Noreen 1923: 199, 258; contra IEW 722). In view of the ambiguity of Lv. dial. milna ‘hammer of Pērkons’ (ME ii: 627) and OCS млънии ‘lightning’, where the dental has been lost before *-n- (cf. Endzelīns 1923: 162; Vaillant 1950: 90–91), the evidence for ‣ 65 66 67 the borrowing context is quite different; in the case of a title, the adoption of a nominative form is to be expected, cf. similarly Turkish papaz ← MGr. παπάς ‘priest’. Lt. /š/ is a poor phonological match for German /z/; furthermore, a computer-assisted search of the LKŽ did not yield any Germanic loanwords containing Lithuanian nasal vowels. West Žemaitian lẽ·išᵃs ‘lentil’ and łeyśiey ‘lens’ in Szyrwid (ALEW² s.v. láišis) apparently show sporadic dial. *ę > ei (Zinkevičius 1966: 137). The forms cited under láišis in LKŽ (the factual basis for the acute set up here is unclear) must partially reflect the same form with regular hardening of /ľ/ as in dial. (Zietela) lãšis ‘lentil’ . ЭССЯ (xv: 64) claims that the word is limited to South Slavic, apparently interpreting the Old Russian examples (cf. СДРЯ 100; СДРЯ 11–14 iv: 489) as Church Slavic loans. To my mind, it is very unlikely that an East Slavic scribe would ‘nativize’ CS лѧща as ⟨лѧча⟩ without actually being familiar with the word. Sergejus Tarasovas suggests to me that the dial. ля́ ща (Orjol, Kaluga) cited by Даль² (ii: 292) may be an incorrect transposition of a local *ля́ [ɕ]а (= */ля́ча/, cf. ДАРЯ i: No. 48) influenced by the Church Slavic spelling. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 consonantism 203 a voiced dental is limited to the Prussian Elbing Vocabulary.68 Since there are some other examples of unexplained voiced stops, such as girmis · made ‘maggot’ (= Lt. dial. kirmìs ‘worm’), this evidence must be treated with care. Taking it at face value, we may see it as evidence for a voicing alternation. ? ‘nettle’. Lt. notrė ̃ (acc.sg. nõtrę), dial. noterė ̃ (acc.sg. nóterę), Lv. nâtre; Pr. E noatis ‘nettle’; OIr. nenaid, MW dynat, danat ‘nettle’ (< *ninati-)69 ~ OSw. nätla, nätsla, OE netele, OHG nezzila ‘nettle’ (cf. underived Fårö Gutnish nate, nata) (Philippa et al. iii [2007]: 418; Matasović 2009: 291; Derksen 2015: 337) — Despite the difference in meaning, it is probable that Pl. nać, Sln. nȃt ‘vegetable tops’ also belong here.70 This semantic shift would imply that nettles were either eaten or given as fodder. On the basis of the East Baltic forms, Specht (1935: 253; followed by REW ii: 201) has reconstructed an archaic r-stem, but as the Slavic and Prussian i-stems cannot be explained on this basis (cf. ALEW 815), it is preferable to view the East Baltic forms as innovative (on the suffixation, see Skardžius 1941: 305–306). In principle, the Baltic forms could reflect a root *neh₂t-, while Celtic would be consistent with *nh₂t- (Zair 2012: 197). Parallels may also be found for the reduplication (see the discussion under ‘sedge’ on pp. 240–241). The Germanic dental is difficult to explain. Resorting to Kluge’s law would be ad hoc, since most of the evidence points to an original singleton *t. Kroonen (2013: 384) has suggested the Balto-Slavic forms were borrowed from Germanic, but this is unlikely in view of the formal discrepancy. Furthermore, Celtic clearly points towards an original *t (cf. Derksen 2015: 337). If the example is accepted as nonIE, the long vowel in Baltic can be compared with the other examples in 7.5.1. † ‘thrush’. R дрозд, Pl. drozd, SCr. drȍzd ‘thrush’ ~ Pr. E tresde; ON þrǫstr (attested in Þul Fugla, cf. Ic. þröstur ‘thrush’); Lat. turdus ‘thrush’; OIr. truit ‣ ‣ 68 69 70 Note that the Slavic reconstruction *muldnijā- (trad. *mъldni; Derksen 2008: 333 following ЭССЯ xx: 220) should be corrected to either *milnijā- (trad. *mьlni; Mikkola 1908: 123; Matasović 2008: 200) or *mulnijā- (trad. *mъlni) — the two are difficult to distinguish. The cluster *-dn- is based only on East Bel. dial. маладня́ ‘lightning’ (thus explicitly Мартынов 1985: 7), a form which is most certainly a hypercorrection in dialects with -dn- > -nn-, cf. Bel. dial. малання́ (ДАБМ No. 311; see Касаткин 1999: 124 and somewhat differently Wexler 1977: 149). The alternative reconstruction *nenati- (Pedersen 1909: 186; Schrijver 1995a: 49) is less probable, as this should have become **nanati by Joseph’s law. In East Slavic only the derived R dial. нати́ на (СРНГ xx: 219), Bel. нацíна, Uk. dial. нати́ ня. The usually cited Uk. dial. нать appears to be confined to the easternmost Carpathian dialects (АУМ ii: No. 324), so it is plausible that it represents a loanword from Slovak dial. nať. For other, less convincing, accounts of the Slavic word, see ЭССЯ xxiii: 186–187. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 204 chapter 6 ‘starling’71 (Ernout/Meillet [1951]: 708; Matasović 2009: 392; ERHJ i [2016]: 200; Matasović 2020: 335; Stifter forthc.) — With regard to the nature of the dental, Lt. strãzdas, Lv. strazds ‘thrush, starling’ are ambiguous, as they show an additional s-, possibly due to anticipation of the second sibilant (Walde/Pokorny i: 761; LEW 920).72 Apart from the initial *d- in Slavic, the correspondence with Norse is precise. The Prussian vocalism is surprising; there is an outside possibility that it nevertheless represents *trasdḗ (cf. Trautmann 1923: 327).73 Latin turdus is most straightforwardly explained as the reflex of the zerograde *trsd-. This zero-grade could also be continued in the Germanic diminutive OE þrostle ‘turdella’, MHG trostel ‘merula’ (< *trust-lō-; Kluge/Seebold 218), providing the position of the -r- has been restored on the basis of the full grade (Kroonen 2013: 545). Sln. drǫ̑zg (dial. drǫ̑zd) and SCr. dial. drȍzg ‘thrush’ (Skok i: 443) result from of a semi-regular dissimilation (Solmsen 1904: 578–579; Endzelīns 1911: 54–55, fn. 3). Dissimilation has also been suggested to account for the variant *trusk- attested in OE þrysce* (attested þryssce), OHG drosca (EWAhd ii: 803), which might be preferable to the suggestion of an innerGermanic suffixal formation (Kluge/Seebold; Kroonen loc. cit.).74 Problematic 71 72 73 74 Arm. tordik (Hamp 1978: 188, 1981: 88; de Vaan 2008: 634; Kroonen 2013: 545) is evidently a learned creation based on Italian tordo (V. Petrosyan on en.wiktionary.org, s.v. տորդիկ [8 April 2020]; Thorsø forthc.). As a parallel, note the Old English by-form strosle ‘blackbird’ (Kitson 1997: 485; OED³ s.v. throstle). The grapheme ⟨e⟩ in the Elbing Vocabulary only rarely stands for /a/, and usually in noninitial syllables (e.g. Pr. E pepelis ~ iii pippalins acc.pl. ‘bird’, E pirsten ‘finger’, cf. iii pīrstans acc.pl.). However, a potential parallel is found in Pr. E wessis ~ Lt. vãžis, dial. važỹs ‘one-horse sleigh’ (PKEŽ iv: 232; for the translation, see Trautmann 1910: 460). This is uncertain, however, as this word may also show the reintroduction of the vowel from the verbal root seen in Lt. vèžti ‘transport’. Old English þræsce is normally cited here, but as a hapax in the Corpus Glossary finding no concrete support in either later English or elsewhere in Germanic, its reliability is questionable. It seems more probable that the dialectal thresh (Oxfordshire, Berkshire), in which OED (s.v. thrush n.¹) would see a continuation of this *þræsce, contains a regional continuation of OE þrysce*. Perhaps it is a Kentish form (with y > e, Campbell 1959: 122–124) which has spread beyond its original geographical zone; compare similarly dial. (Sussex, Essex) sherve, sharve ‘service tree’ (< OE syrfe*, attested obl.sg. syrfan; OED³ s.v. serve n.¹) and perhaps SW dial. rex ‘rush’ (differently see OED³ s.v. rush n.¹). Note that the Old English word is glossed as truitius (cf. also the similar gloss þrisce · trutius) which Kitson (1997: 484) would see as a “corruption” of Latin turdus. Far more likely, this is a Latinization of Irish truit ‘starling’ (Suolahti 1909: 52, fn. 1). As for the semantics, it is worth mentioning that Lat. turdus is twice glossed as OE stær ‘starling’ (Lacey 2013: 66). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 consonantism 205 are the British forms OBret. tra[s]cl (modern drask, draskl), MoW tresglen ‘thrush’. Stifter (forthc.) equates the apparent suffix *-sk- in the above forms with that found in MW mwyalch ‘blackbird’ and alarch ‘swan’. In his opinion, this would favour a non-IE origin. As has long been recognized, OIr. truit ‘starling’ can reflect an earlier *trozdi(Zupitza 1900: 233; cf. Brugmann 1897: 691). However, the British equivalents MW trydw, MBret. tret ‘starling’ cannot, which has led them to be analysed as Goidelic loanwords (Walde/Pokorny i: 761; Stifter forthc.). The Welsh and Breton vocalism is consistent with the reflex of *u with i-affection (after the plural), while the unaffected vowel is preserved in the Old Breton gloss trot · strution.75 In any case, a Goidelic loanword seems preferable to assuming an ad hoc “expressive gemination” in Old Irish (de Bernardo Stempel 1999: 274; implicitly Matasović 2009: 392). Ernout/Meillet (p. 708) claim that it is futile to reconstruct the original form of this word. However, an ablauting *trosd- : *trsd- accounts for the Baltic, Germanic and Latin and Old Irish data without any serious problems.76 The remaining evidence for irregularity is the initial d- in Slavic, but it is possible that this has arisen through assimilation, as has undoubtedly occurred in MW drydw and MoIr. druid ‘starling’.77 As the Latin form is more easily explained starting from an Indo-European ablaut variant, while all of the irregular developments can be accounted for within the individual branches, I do not think there is any truly compelling evidence for a non-IE borrowing. 6.2.3.1 Conclusion The certain and possible evidence for voicing alternations involving dentals is collected in Table 10, overleaf (see p. 185 for help reading the table). Forms which do not provide relevant data are presented in light grey. Shaded cells indicate reflexes of a voiced or voiced aspirated dental. 75 76 77 The vocalism of OCo. troet is unclear as ⟨oe⟩ in other cases represents an inherited vowel sequence (e.g. hoern ‘iron’ = MW haearn; moelh ‘blackbird’ = MW mwyalch); we apparently have to assume contamination with e.g. OCo. hoet ‘duck’ (= MW hwyat). Hamp’s (1981: 88) insistence on *dʰ is not necessary, as Winter’s Law was blocked by an intervening *s (Kortlandt 1988: 394); and such a reconstruction is contradicted by Germanic. It has often been claimed that the word for ‘thrush’ is ultimately onomatopoeic (Suolahti 1909: 53; Булаховский 1948: 112; EWAhd ii: 803; Kluge/Seebold 218), but this does not seem certain to me. РЭС (xiv: 363) notes SCr. drsk!, representing the sound of the mistlethrush, but it is possible that this onomatopoeia partly derives from the name of the bird itself. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 206 table 10 chapter 6 Possible examples of alternations involving dentals Baltic Slavic Germanic Elsewhere ‘drone’ *tran- *trant- *dʰrV̆ n- Gr. *tʰrV̄ n- ‘reed’ *truḱ- *trust- – Gr. *tʰrus- ? ‘lentil’ ?*lenḱ- *lent-i̯- ?*lens- ? ‘lightning’ ? *meld⁽ʰ⁾- *mlT-ni- *melt-uni- Celt. *melt- ? ‘nettle’ *nât- *nât- *nad- Celt. *ninat- ?Gr. *lntʰLat. *l(e)nt- It is interesting that the examples in this section do not show a similar behaviour to the examples of *k ∞ *g⁽ʰ⁾ given in 6.2.1. Only the word for ‘lightning’ possibly shows the same distribution, with Baltic voiced *D contrasting with Celtic *t; however, as discussed above, this is based on rather tenuous evidence. The clearest examples here involve Greek; specifically, in two or three cases, we find Greek *tʰ as opposed to *t elsewhere. Since in Greek we actually find a voiceless stop, it is unclear whether back-projecting it to IE *dʰ would be anachronistic: perhaps, rather than a ‘voicing’ alternation, we are dealing with an ‘aspiration’ alternation. Such alternations are well-known in Greek words of presumed foreign origin, e.g. ἄνηθον ~ Aeol. ἄνητον ‘dill’ (Furnée 1972: 187–193). Against this conclusion, we can note that Germanic indeed does show a reflex of *dʰ in the word for ‘drone’. On the other hand, note the word for ‘turnip’, discussed in the following section, which might show a comparable ‘aspiration alternation’. 6.2.4 Alternations Involving Labials In two of the words discussed above (see 6.2.1), we have observed an alternation *b⁽ʰ⁾ ∞ *p occurring alongside *g⁽ʰ⁾ ∞ *k, cf. Lt. gul̃bė ~ Kash. kôłp ‘swan’ and R граб ~ Lat. carpinus ‘hornbeam’. Examples of an independent alternation *b⁽ʰ⁾ ∞ *p not associated with a parallel velar alternation are in fact comparatively few, and the only certain cases constitute rather widespread Wanderwörter: ‘hemp’. R конопля́ , Pl. konopie, SCr. kònoplja ‘hemp’ ~ OE hænep, OHG hanaf ; Gr. κάνναβις ‘hemp’ (Schrader/Nehring i [1923]: 441; Huld 1990: 406–407; ‣ Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 consonantism 207 Matasović 2013: 89; Kroonen 2013: 209; etc.) — In Baltic, we have Lt. kanãpės, Lv. kaņepes and Pr. E knapios, which are usually considered to be Slavic loans (e.g. Berneker i: 361; ME ii: 156–157; LEW 214; Levin 1974: 96; Smoczyński 2018: 482). On formal grounds, cognancy is equally possible (Būga 1913: 255– 256; PKEŽ ii: 231). While there is some evidence for hemp having been used in Lithuania during the 1ˢᵗ millennium ce (Gimbutas 1963: 117; Grikpėdis/ Motuzaitė Matuzevičiūtė 2020: 165), it is uncertain whether this evidence is early enough to rule out a Slavic origin for local hemp production. A form with *p is likewise widespread in Romance, cf. Italian canapa, Romanian cânepă, attested since Late Latin (cf. FEW ii: 213–214). On the authority of Herodotus, Greek κάνναβις is traditionally considered a loanword from Scythian or Thracian (cf. Schrader/Nehring i: 441; Frisk i: 779), although this has no concrete linguistic basis. In Ossetic, which would be closest to the supposed Scythian donor language, we find Oss. I gæn, D gænæ ‘hemp’, which probably implies *kanā- without the labial (cf. Abaev 1958: 513).78 Elsewhere in Iranian, a form *kanafa- seems to be suggested by Khotanese kaṃha- ‘hemp’ and NP kanaf ‘flax cord’ (Steingass 1892: 1055),79 while the NP variant kanab ‘hemp (seed); hempen rope’ (idem: 1052) would imply *kanapa-.80 The word for ‘hemp’ is widely recognized as a Wanderwort of indeterminate origin, and the precise source of the various p-forms in Europe is difficult to establish. The ultimate origin of the word has been seen in the Near East, cf. Syriac qnpʾ /qenpā/ ‘hemp (for making ropes)’, and Akkadian (Neo-Assyrian) qunnabu, ‘(possibly) the flower or seed of the hemp’. The latter would predate the Greek attestations, although it is hardly the original source (note that Sumerian *kunibu is a ghost, cf. Barber 1991: 38). [‘turnip’. Lt. rópė, Lat. rāpum ‘turnip’ ~ Gr. ῥάφανος ‘cabbage, radish’;81 ?MW erfin, Bret. irvin ‘turnip’ — See the discussion on p. 237.] ‣ 78 79 80 81 It seems at least possible that this could have developed via *kanapā- > *kanaba > *kanba (syncope, cf. Cheung 2002: 55–56), then by (irregular?) metathesis to *kabna > *kan(n)a (cf. Oss. I kʷynæg, D kunæg ‘meagre, small’ < *kabna, Cheung 2002: 30). In any case, the initial g- is irregular, and has no regular origin. The vocalism of Northern Kurdish (Kurmanji) kinif ‘hemp’ (cf. Цаболов 2001: 554) is unclear, but the form must be borrowed, as intervocalic *f has regularly given -v- in Kurdish, cf. nāv ‘navel’ < *nāfa- (Цаболов 2010: 32; M. Kümmel p.c. December 2022). A Proto-Iranian *p would also be suggested by Buddhist Sogdian kynpʾ (Gharib 1995: 203), perhaps meaning ‘hemp’ or ‘flax’, provided this is not an independent loan from Syriac (Henning 1946: 724). Bailey (1979: 51–52) quotes a MP (Pahlavi) kʾnb that I have been unable to verify. If reliable, it would seem to suggest *-b- (cf. Peyrot 2018: 270). Arm. kanepʿ, kanapʿ ‘hemp’ appears to be an Iranian loan, but its exact source is unclear. The Greek variant with -π- (cf. Beekes 2014: 61) rests on extremely doubtful evidence: (a) Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 208 chapter 6 ‣ ? ‘furrow’. Lt. bir̃žė ‘row, furrow; timber tract; border mark’, Lv. bìrze ‘furrow, strip of a field’ ~ Lat. porca, OHG furh, furuh, MW rych ‘furrow’ (on the Slavic equivalent, see p. 224) — The similarity of these forms has been noted by Machek (1968: 65) and Holzer (1989: 51–54), and the example remains one of the most attractive of Holzer’s “Temematic” etymologies, as the formal corres⁽⁾ pondence *prḱ∞ *bʰrǵʰ- is precise aside from the difference in voicing. What further speaks in favour of Holzer’s interpretation is that *prḱ- has some potential IE comparanda. LIV (475) sets up a verbal root *perḱ- ‘graben, aufreißen’. On further inspection, however, it turns out that the reconstructed semantics are based almost entirely on the word for ‘furrow’. The only comparandum attesting to a verbal root is Lt. peršėt́ i ‘to itch’, while the other nominal formations are uncertain. The Rigvedic párśāna- (3×) is of uncertain meaning: it probably refers to a low place, but may mean ‘valley’ or ‘plateau’ (cf. KEWA 228–229; Jamison Commentary vii.140.5). Aside from this, the only evidence is Lt. pró-perša (pra-peršà) ‘thawed patch in ice; break in the clouds; etc.’,82 but this, like núo-perša ‘infertile patch of land’, ìš-perša (Kupiškis) ‘deep rut in a road’ might well be derived from peršėt́ i in a secondary sense, cf. nu-, iš-peršėt́ i ‘go bad, spoil’. We may conclude that the evidence for the verbal root rests on the Lithuanian word for ‘to itch’, which is semantically remote. If the IE etymology can be abandoned, we may consider a non-IE origin for the whole group, which would eschew the need for a “Temematic” source or other IE substrate. In this case, ‘furrow’ can be considered an example of a *p ∞ *b⁽ʰ⁾ alternation. It is, however, a little troubling that none of the examples of a *k ∞ *g⁽ʰ⁾ alternation discussed in 6.2.1 were affected by satemization. If the word is non-IE, it must have been loaned extremely early, which may also be seen in the regular reflexes of syllabic *r̥ . 82 ⟨ῥαπάνια⟩, attested in a Hellenistic period papyrus. This, and other examples of confusion between stops in Egyptian papyri, can plausibly be attributed to Coptic first-language interference (Holton et al. 2020: 187); (b) Athenaeus (Deipnosophists ix, Chapter 8) tells us that Glaucus, apparently the author of a cookery book, wrote ῥάπυς for ῥάφυς (meaning βουνιάς ‘rapeseed’). Neither form is otherwise reliably attested (LSJ s.v.). Lt. praparšas, known only from Szyrwid, is typically adduced here (Walde/Pokorny ii: 46; Fraenkel 578; IEW 821; LIV 475). The gloss ‘Graben’ in all these sources (thus supporting the sense ‘to dig’), is based on Szyrwid’s row, fossa (SD 268b35). In the first edition of the dictionary, however, the word glosses Polish iaskinia, odchłan, prʒepáść (ALEW 102), suggesting a sense ‘chasm, abyss’. As none of these senses appear to have been recorded elsewhere, I am led to wonder whether Szyrwid was unsuccessfully attempting to render a sense such as ‘gap in the ice’ in Polish. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 consonantism 209 ? ‘pigeon (2)’. Lt. balañdis, Lv. baluôdis ‘pigeon’ ~ Lat. palumbēs ‘wood ‣pigeon’83 — Both words are traditionally explained as derivatives of colour terms (cf. Schulze 1910: 799–800). In Baltic, the root is assumed to be that of Lt. bálti ‘whiten’ (Skardžius 1941: 101; LEW 31);84 Karaliūnas (1993: 110) assumes an original colour adjective *balandas ‘whitish’ (cf. ‘white-marked one’, Levin 1992: 86). Derksen (2015: 78) questions the derivation from ‘white’ on semantic grounds as, according to him (after Levin loc. cit.), “whiteness is not a natural colouring in pigeons”. More importantly, an adjectival suffix *-anda- would be completely unparalleled and therefore ad hoc. Lat. palumbēs is usually derived from the root of palleō ‘be or grow pale’ (e.g. Walde/Hofmann ii: 242), cf. Gr. πέλεια ‘wild pigeon’ ~ πελιός ‘black and blue, livid’. The first syllable is also reminiscent of Pr. E poalis ‘pigeon’. If we start with ‘grey’, the semantic motivation makes some sense;85 the Latin second syllable could have been influenced by columba (Lockwood 1990: 262–263; de Vaan 2008: 126). Naturally, if we compare Lt. balañdis with Lat. palumbēs, both root etymologies would need to be abandoned. Due to the lack of morphological transparency on both sides of the equation, this might be justified. However, the irregularities are not limited to the initial stop; there is also a mismatch between the stem-final -b- in Latin as opposed to Baltic -d-. One way out is to assume, again, that the Latin word has been influenced by columba, although then one could question how exactly this etymology is preferable to the traditional explanation, which also demands the assumption of such a contamination. In defence of the new etymology, it seems more straightforward to assume contamination starting from a disyllabic *palond- rather than, with Lockwood, from a more basic *palēs.86 Klingenschmitt (1982: 165) compares Lat. palumbēs with Arm. aławni ‘pigeon’, reconstructing *plH-bʰ-nih₂- (in his notation *pl ̥h-bʰ-niə₂), implying the 83 84 85 86 Apparently here also Oss. I bælon, D bælæw ‘domestic pigeon’ (Абаев 1965: 17; Weber 1997). Due to the -l-, the Ossetic word is likely to be a loanword. It is unclear whether Baltic could plausibly be the source, as there is no other clear evidence of contact, and no obvious historical scenario. According to Sasha Lubotsky (p.c. April 2021), the Iron suffix -on (< *-ān-) might be equated with Baltic *-and- through regular *a > *ā before a consonant cluster and subsequent loss of *d. The Digor variant is of unclear formation. The existence of the frequently cited bãlas ‘white’ (known only from Juška) is perhaps questionable, see Jakob forthc. b. Cf. Russian сизя́ к ‘feral pigeon’ < си́ зый ‘dark bluish-grey’; Oss. ID æxsīnæg ‘wild pigeon’ < (Digor) æxsīn ‘dark grey’ (Абаев 1958: 220–221). Alternative, but no less ad hoc explanations would be to assume a dissimilation *b–b > *b–d in Baltic, or a suffixed Latin *palond-u̯ -. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 210 chapter 6 application of Thurneysen’s law of nasal metathesis in Latin (cf. Martirosyan 2008: 29). Such a preform would yield Latin *plamb- rather than *palomb-, and more importantly, would force us to disassociate the suffixes of columba and palumbēs, which seems quite unsatisfactory. Assuming Armenian has a derived n-stem, we may instead start from *palab-, i.e. a variant without a nasal (cf. section 6.1). This is speculative, however, as the Armenian word is open to interpretation (see Batisti 2021: 208–210 with lit.). † ‘hollow’. Slk. dúpä ‘den, burrow’, Sln. obs. (Pleteršnik i: 184) dúpa ‘die Erdhöhle’ (= Pl. dupa, Bg. ду̀ пе ‘arse’) ~ Lt. daubà ‘ravine; (PrLt.) den, burrow’ (Kuiper 1956: 223; 1995: 71–72; Schrijver 2001: 420; Philippa et al. i [2003]: 569 s.v. diep; Matasović 2013: 96; Derksen 2015: 144) — This word family is routinely quoted, mainly by members of the “Leiden school”, as an example of a substrate word. Kuiper’s main line of argument was built on the presence of numerous variants within Germanic, where root final *-b- seems to alternate with *-p-, *-bb-, *-pp- and *-mp- (thus ON dúfa, dýfa ‘dip (at a christening)’; Go. diups ‘deep’; MDu. dobbe ‘water pit, pool’; Nw. duppe and MLG dumpeln ‘dip’, respectively). This approach has been criticized by Kroonen (2011a: 255; 2011b: 127– 129), who has convincingly argued that the variation can be more plausibly explained as a result of various analogies after Kluge’s Law. It seems likely that the ‘nasal infix’ supposed for Lt. dum̃ blas, Lv. dum̃ bla ‘mud, sludge’ (LEW 108–109; Smoczyński 2018: 263; ALEW 276) is also an illusion.87 Rather, the -b- in these forms is epenthetic. This is possibly suggested by the forms dumłas (SD 64b17)88 and dumlelus (acc.pl., Daukantas 1846: 67; see LKŽ s.v. dumlas), and certainly by Lv. dial. (Vidzeme) dumûksnis ‘marsh’ (cf. Prellwitz 1909: 387; Schulze 1910: 791; ME i: 514). In general, there is a fairly consistent semantic distinction between the two word-groups. Almost all words containing a nasal mean ‘mud’ or ‘marsh’, while words lacking the nasal mean ‘valley, hollow’.89 The latter group are transparently derived from the verbal ‣ 87 88 89 The latter two sources point specifically to the Lt. 3pres. dum̃ ba as the source of the forms. The antiquity of this presentic formation cannot be proven, as nasal presents are productive in Lithuanian denominal verbs of the shape *TVT- (where T = any stop, see Villanueva Svensson 2010: 206–208), and moreover, ME (i: 509) reports a plain thematic dubu for Latvian. But note that Szyrwid also has ⟨dumbłas⟩ (SD 120b19). Compare, on the one hand, Lt. dum̃ blas, Lv. dial. dum̃ bla ‘mud’, Lv. dum̃ brs ‘boggy; marsh’ (the suffix in Lv. dial. dum̃ bêris ‘muddy pit; puddle’ is probably secondary), and on the other hand Lt. dubùs ‘hollow, concave’, Lv. dial. (Vārkava, ME i: 509) dubums ‘tree hollow’, Lt. daubà, Lv. dial. (ME i: 443) daũba ‘ravine’. The two roots do seem to have influenced each other, however, cf. Lt. dial. dumbrà ‘deep point in a river; pond’ vs. Lv. dial. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 consonantism 211 root in Lt. dùbti ‘sink down, become concave’, which is further cognate with Go. diups (< *dʰeubʰ-nó-) ‘deep’ and OIr. domain ‘deep’.90 In view of the large number of derivatives and extensive IE-like ablaut in the root *dʰeubʰ-, it seems more probable to me that it is inherited, despite the limited distribution. As a result of this, the Slavic forms with -p-, must either be unrelated or explained as the result of a secondary deformation. As I have identified no motivation for the latter, I would prefer to simply separate the forms. † ‘post (1)’. Lt. stul̃pas ‘post, pillar’, Lv. dial. stùlps ‘pillar, leg of a boot’; OCS стлъпъ ‘pillar, tower’ ~ R столб ‘post, pillar’, Sln. obs. (Caf apud Pleteršnik ii: 578) stółb ‘Pfahl’; ON stolpi ‘post, pillar’ (> ME stulpe ‘stake, post’, MDu. stolpe ‘small beam’) — Vasmer (REW iii: 18) rejected earlier proposals (Meringer 1909: 200; Stender-Petersen 1927: 279–281) to derive the Balto-Slavic words with -p- from Germanic, although he does not present any arguments. A point in favour of the loan etymology is that the Balto-Slavic p-forms are largely limited to the meaning ‘post, pillar’, while with -b- one finds archaic-looking derivatives such as Lt. stul̃bti ‘be stunned’, and Bg. стъ̀лба ‘staircase, ladder’, SCr. stȕba ‘step, stair’. On the other hand, the word is scarcely attested in Germanic, and one could seriously consider deriving the Norse word from Slavic (Tamm 1881: 31; dismissed, again without argumentation, by de Vries 1962: 551).91 The complexity of the analysis makes it difficult to draw any clear conclusions. ‣ 6.2.4.1 Conclusion The certain and possible evidence for voicing alternations involving labials is collected in Table 11, overleaf (see p. 185 for help reading the table). Shaded cells indicate reflexes of a voiceless labial. The cover symbol *B stands for *b⁽ʰ⁾. Forms which do not necessarily provide relevant data are presented in light grey. 90 91 duburs ‘deep and wide spot in a river’. Additionally, some Latvian words seem to belong with the former root, but lack an -m-: dubļi ‘muck; mud’, dubra ‘swamp, bog’. Might these reflect *dūb- < *dumb- with shortening before a -CR-cluster (cf. Derksen 2007: 44)? Often adduced are To. A tpär, B tapre ‘high’. However, the ‘Tocharian Grassmann’s law’ (Winter 1962), if valid, would predict To. B *tsapre. The original meaning ‘deep’ has been supported by the translation of To. A top, B tewpe as ‘mine’ (Adams 2013: 330). However, Imberciadori (2022) has argued that this word should instead be translated ‘heap’, which makes the comparison unattractive. In any case, the root connections with Nw. stelpe, MDu. stelpen ‘hinder’ or with Lt. stel̃bti ‘overshadow’ are not compelling. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 212 table 11 chapter 6 Possible examples of labial alternations Baltic Slavic Germanic Elsewhere ‘swan (2)’ *Gulbʰ- *kulp- – ‘hornbeam’ *skrâB-l- *GrâB-r- – Lat. *k(r)arp- ‘hemp’ [*kanap-] *kanap- *kanab- Gr. *kannab- ‘turnip’ *râp- *rêp- *rāP- Gr. *rabʰ?Celt. *arB- ? ‘furrow’ *Brǵʰ- ?*BorsD- *prk- It.-Celt. *prk- ? ‘pigeon’ *Balandʰ- – – Lat. *palomB- Beside the words for ‘swan’ and ‘hornbeam’, which show *p ∞ *b⁽ʰ⁾ alongside *k ∞ *g⁽ʰ⁾ in the same word, there are two other potential examples showing a similar distribution, although neither of these are certain. The word for ‘furrow’, if loaned from an unknown source, would be the only example of a voicing alternation predating satemization. As a result, whether it represents a manifestation of the same voicing alternation cannot be considered certain. The remaining words appear to show the opposite pattern: it is notable that both ‘hemp’ and ‘turnip’ are widespread words associated with agriculture, and it is likely that they spread as Wanderwörter. The word for ‘turnip’ might constitute an example of the ‘aspiration’ alternation observed in 6.2.3. On the other hand, the Celtic, and potentially also Germanic, comparanda point to an underlying *b⁽ʰ⁾, yet it is by no means certain that the divergent stops in Celtic and Greek can be equated with one another (as virtual *bʰ), and it is possible that they represent two unrelated phenomena — a specifically (pre-)Greek ‘aspiration alternation’ and a specifically (pre-)Celtic voicing. 6.2.5 Baltic *ž ∞ Slavic *s ‘oats’. Lt. ãvižos, Lv. àuzas ‘oats’ ~ R овёс, Sln. óvəs; Lat. avēna ‘oats’ (Ernout/ Meillet [1951]: 56; ?Pisani 1968: 14; Huld 1990: 404; FEW xxv [2002]: 1213; Oettinger 2003: 189; de Vaan 2008: 64–65) — The relationship between the Baltic and Slavic words is irregular, suggesting the word entered the two branches ‣ Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 consonantism 213 independently. Reconstructing a suffix *-s- in Slavic (thus Derksen 2008: 384) is ad hoc, as there not appear to be any other plausible cases of *-s- as a denominal suffix (cf. Vaillant 1974: 659).92 Moreover, the Latin vocalism also precludes the reconstruction of a common pre-form. De Vaan’s assumption of an underlying palatovelar and *aweksnā- for Latin is potentially anachronistic. As Huld points out, if we are dealing with a non-IE loanword, “a spirant of indeterminate voicing” would account for the facts. Lat. avēna ‘oats’ could equally reflect *au̯ e(T)s-n- (where *T can be essentially any stop, although *ts or *s would be most probable for our purposes). For further discussion, and on the question of Prussian wyse ‘oats’, see pp. 239–240. ? ‘fishing trap’. Lt. várža ‘fishing basket’, Lv. varzi ‘Setzkörbe’ (Lange 1773: 378), dial. var̂za² ‘fishing weir’ (ME iv: 481) ~ R ве́рша, Sln. vŕša ‘fishing basket’ (< *virs- + *-jā-; trad. *vьrs + *-ja) (cf. Pronk/Pronk-Tiethoff 2018: 295) — Existing etymological solutions either separate the two words, linking the Slavic forms with OCS врьхъ ‘top’ (REW i: 109), or assume a suffix *-si̯ā- for Slavic (Persson 1912: 505; Trautmann 1923: 355). The Baltic forms look related to the verb Lt. ver̃žti ‘tighten, tie up’, Lv. vir̃zît ‘direct, steer’ (cf. at-virzīt ‘untie’, ME i: 211), but this has not been generally accepted (cf. РЭС vi: 351; ALEW 1384).93 In view of the parallelism with the word for ‘oats’, above, it is tempting to derive these words from a non-Indo-European source. On the other hand, it is unclear to what extent it is justified to separate the words for ‘fishing basket’ from Latvian senses such as var̂za, var̃za ‘tangle, confusion’ (ME iv: 481–482), which clearly belong with the verbal root (cf. Lv. var̂zât ‘plait together, tangle’). In addition, the difference in vocalism is striking; this sort of vowel alternation is perhaps more easily explained as the result of Indo-European ablaut than through parallel borrowing (compare, with the opposite distribution, Lt. bir̃žė ~ OCS бразда ‘furrow’ on p. 224). ? ‘ploughshare’. Lt. lẽmežis ‘ploughshare’ ~ CS (Bes.) лемешь* (SJS ii: 112) ‘plough’, R ле́мех, dial. леме́ш, SCr. lèmeš ‘ploughshare’ — In view of its -s-, perhaps Lv. lemesis ‘ploughshare’ is a loan from East Slavic. The -s- could be a hypercorrection after the oblique cases (e.g. lemeša gen.sg.), cf. vìksne², gen.pl. ‣ ‣ 92 93 For the deverbal suffix, cf. OCS гласъ ‘voice, speech’ ~ глаголати ‘speak, proclaim’; CS кѫсъ* ‘bit, crumb’ ~ Lt. ką́ sti (kánd-) ‘to bite’; OCS смѣхъ ‘laughter’ ~ смиꙗти сѧ ‘to laugh’. Snoj (2003: 836) considers the word for ‘heather’ (see below) to be related, and the word for ‘fish trap’ to originally have meant ‘something woven (from heather)’. A fishing basket woven from heather does indeed appear to be found in the Highland Folk Museum, but I cannot verify whether such a tradition could have existed at an appropriate time in central Europe. See the doubts in РЭС (vi: 351–352), where all other etymological comparisons are also considered doubtful. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 214 chapter 6 vīkšņu ‘cherry tree’ ← R ви́ шня (cf. Būga 1922: 175–177 and also dial. lemešs, EH i: 733). Note that Lt. lẽmežis has itself been interpreted as a Slavic loan (Gołąb 1982: 130;94 LKŽ s.v.; ERHJ i: 546); and while its limitation to a narrow group of Šiauliškiai dialects rouses suspicion, this is not sufficient to confirm or deny this proposal. In Slavic, the most common variant is *lemeše- (trad. *lemešь), continuants of which are found in every Slavic language. In addition, forms are found with a final -ž, but these look secondary, being largely limited to South Slavic: Sln. lémež, SCr. dial. lèmež (РСА xi: 327), Čak. lemȅž (ERHJ i: 546), Bg. лемѐж (cf. the data in ЭССЯ xiv: 108–110). Perhaps one could assume the secondary influence of the deverbal noun suffix *-eže- (trad. *-ežь), which enjoyed a certain productivity in South Slavic (Berneker i: 700; cf. Vaillant 1974: 506). Furthermore, some forms seem to lack the initial *l-: CS емешь (Miklosich 1865: 1157), Bg. dial. емѐш (БЭР i: 495),95 SCr. dial. (Montenegro) jèmlješ (RJA iv: 587), R dial. (N) о́мех, оме́ш (and variants, СРНГ xxiii: 198–199, 201– 202; Мызников 2019: 556). Derksen (2015: 278) has considered the variant with *l- the result of a secondary contamination with the root *lemH- ‘to break’. This is rather difficult to accept: forms with *l- are much better represented in Slavic and the only forms found in Baltic. Provided the latter are not all Slavic loanwords, it would be highly improbable that the contamination could have occurred independently in both branches.96 An interesting proposal is put forward by Bańkowski (2000 ii: 19–20), who assumes contamination with a Proto-Slavic *lemę̄zi- (*-že-; trad. *lemęz/žь) represented by Pl. dial. lemiąże pl. (Sł. Warsz. ii: 714), OCz. lemiez, Sln. lę̑mez ‘rafter’. The assumption is that the latter would have been used in the sense ‘plough shaft’. The weakness of this theory is that neither word is attested in this meaning, but such a confusion does indeed appear to have occurred in some forms meaning ‘ploughshare’: cf. 94 95 96 Cited according to the Lithuanian Etymological Dictionary Database (available at etimologija.baltnexus.lt, accessed 9 November 2023), s.v. lẽmežis. БΕР claim that the development of /ľ/ to /j/ is a typical dialectal phenomenon. It is true that around Vraca (where емѐш is recorded), we also find e.g. пойѐ for полѐ ‘field’ (БДА Ф 109); however, here we are dealing with a reflex of older */lj/, and not */l/, and the authors of БΕР do not quote any evidence for this supposed dialectal change. It is notable that the given verb is attested (almost) exclusively in the o-grade in Slavic. Despite Schuster-Šewc (816; cf. ЭССЯ xiv: 113, 200), it seems unlikely that USrb. lemić ‘to break’, attested in some older sources beside łomić and corresponding to LSrb. łomiś, is a “Proto-Slavic archaism”. It is most probably due to internal analogical processes. Similar considerations apply to the Serbo-Croatian iterative lijèmati ‘beat, thrash’ (RJA vi: 64). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 consonantism 215 OPl. ⟨lyemyąszem⟩ inst.sg. (SSP iv: 19), Kash. lemiąż, Slk. dial. (apud ЭССЯ xiv 109) lemez. Kalima (1950) also considers the Slavic *l- to be secondary, and interprets the whole family as an Iranian loanword, comparing Persian dial. amāč, amāǰ ‘plough’ (cf. also REW ii: 267). This interpretation cannot be upheld, as the Persian word is itself a relatively recent loan from Turkic (cf. Turkish, Uighur amač ‘plough’; Doerfer 1965: 124). For the same reason, Komi ami̮ś, dial. (Upper Vyčegda) ameʒ́, Udmurt ameź, dial. omeź ‘ploughshare’ (Лыткин/Гуляев 1970: 32; Rédei 1986: 64) are likewise hardly of Iranian origin. The Turkic word is already attested since Kāşğarī (11ᵗʰ c. ce), but has a relatively limited distribution, being concentrated in Karluk Turkic and radiating from there into neighbouring sub-branches. Despite this, Starostin, Dybo & Mudrak (2003: 295–296) reconstruct the word for Proto-Turkic, offering the reconstruction *amač, and further comparing Manchu anja, Mongolian anjis ‘plough’. Regardless of whether one accepts the Altaic theory, the 6ᵗʰ millennium bce dating for Proto-Altaic supported by Starostin et al. (idem: 237) clearly rules out the possibility of a shared inherited word for ‘plough’ (cf. Vovin 2005: 75). The overall picture is nevertheless of a cultural Wanderwort “with a complicated history” (to quote Helimski 1997b: 121). At the least, it seems unattractive to separate Turkic (regional) *amač ‘plough’ from Permic *amεʒ́ ‘ploughshare’ and Slavic dial. *emeše- (trad. *( j)emešь) ‘ploughshare’. It does not look likely, however, that Turkic could have been the source of either word, as the Permic voiced affricate cannot be explained on this basis, and the Slavic front-vocalism is aberrant. As it is doubtful that the *l- in Slavic and Baltic can be considered folk-etymological, one may wonder whether this may also be attributed to non-IE borrowing. Rather than a phonetic motivation for an alternation between *l- and *∅-, a more reasonable account might be to assume the fossilization of a particle of some kind (such as in MDu. lomre ‘shade’ < Fr. l’ombre). However, no parallels of this alternation appear to be found within my corpus. ? ‘heather’. Lt. vìržis ‘heather’ ~ R dial. ве́рес (СРНГ iv: 131; РЭС vi: 284), Cz. vřes, SCr. vrȉjes ‘heather’ (Machek 1950b: 158–159; Smoczyński 2018: 1680) — Derksen (2008: 516), reconstructs a variant *verska- (trad. *verskъ) on the basis of R ве́реск, Uk. dial. (Makowiecki apud ЕСУМ i: 353) вереск, although these are most easily viewed as secondary. In Czech dialects, one finds a whole host of obscure variants, including ones with a final -k: vřesk, březek, etc. (see ČJA ii: 98; further on the initial b-, cf. ČJA v: 442–443). It is quite clear that these cannot all be old, and that we cannot explain the data without assuming convergence with unrelated plant names, cf. Cz. břečťán ‘ivy’, dial. ‘heather’, břest ‘elm’, dial. ‘heather’ (similarly R dial. ве́рест ‘heather’ after бе́рест ‘(field) elm’?), Cz. dial. ‣ Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 216 chapter 6 březa ‘birch’ beside březek ‘heather’. The source of the final /k/ in R ве́реск remains unclear (cf. берескле́т ‘spindle tree’? see РЭС vi: 284), but it is unlikely to date back to Proto-Slavic. The Balto-Slavic forms have long been compared with Gr. ἐρείκη ‘heather’ (assuming an earlier *wereikā), on the one hand, and OIr. fróich, MW gruc ‘heather’ (< *uroik-o-), on the other (e.g. Walde/Pokorny i: 273; REW i: 187). As this comparison is phonologically impossible in Indo-European terms, it has been suggested that these forms represent parallel loans from an unattested source (Machek 1950b: 158; Frisk i: 551; Matasović 2009: 431, 2013: 90; van Sluis forthc.). This would imply an underlying *ḱ and suggest a loan predating satemization, which is chronologically difficult, as there is no agreement even between Baltic and Slavic. Furthermore, the initial *w- is not ascertained for Greek, and the complete loss of the second-syllable diphthong in BaltoSlavic would be unparalleled. Thus, while the Celtic and Slavic forms potentially share three phonemes, the etymological equation of these forms is dubious. Standard Latvian vìrši ‘heather’ shows *-s-. Considering the variation within Slavic, one may argue that the choice of Lt. vìržis (and Lv. dial. vir̂ži², ME iv: 620) as a comparandum amounts to cherry-picking. Smoczyński (2018: 1680) suggests that -ž- may have arisen due to assimilation, or alternatively result from a folk-etymological connection with ver̃žti ‘tighten, tie up’ (thus also T. Pronk apud Matasović 2013: 90). Neither of these explanations strike me as convincing, but at the same time, this cannot be classed as a certain example of a voicing alternation. On Žem. birzdžiai ‘heather’, see p. 223. 6.2.5.1 Conclusion The certain and possible evidence for the alternation *ž ∞ *š is collected in Table 12, overleaf (see p. 185 for help reading the table). Forms which do not provide relevant data are presented in light grey. In Slavic, the cover symbol *S₁ may stand for quasi-IE *ḱ or a cluster *(T)s. The cover symbol *S₂ may also reflect quasi-IE *s directly. At first sight, there appear to be a number of striking parallels for the irregular alternation between *ž and *š found in the word for oats (Pronk/PronkTiethoff 2018: 295). However, after examining each case on its individual merits, the picture is somewhat less optimistic. Although we do indeed find a similar distribution between Baltic *ž and Slavic *s, the words for ‘heather’ and ‘fishing basket’ are ambiguous, and it remains uncertain that the word for ‘ploughshare’ is directly comparable as we seem to be dealing with a Wanderwort showing a broad Central Asian distribution. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 217 consonantism table 12 Possible examples of an alternation *ž ∞ *š Baltic Slavic Elsewhere ‘oats’ *au̯ iź- *au̯ iS₁- Lat. *au̯ e(T)s-n- ? ‘fishing trap’ *u̯ arź- *u̯ rS₁-i̯- ? ‘ploughshare’ *lemeź- *(l)emeS₂-i̯- ? ‘heather’ 6.3 *u̯ rź*u̯ rś- Tur. *amač *u̯ erS₁- Sibilant Clusters 6.3.1 *Cs ∞ *sC Some studies into non-Indo-European loanwords have drawn attention to doublets showing the metathesis of s-clusters (Oštir 1930: 5–6; Furnée 1972: 392–393; Šorgo 2020: 459). Of course, irregular metatheses do occur, and one might ask exactly what feature of a suggested substrate language could lie behind such an alternation (cf. Beekes 2014: 18). Here, it is worth remembering that our non-Indo-European source language was probably not a monolith, and that regular metatheses do occur. For instance, compare the regular developments *#ks- > *#sk- in Baltic (Stang 1966: 95), *-ps- > -sp- in Latin (Leumann 1977: 202; cf. Hamp 2003), and *-sk- > *-ks- in Ob-Ugric (Aikio 2015b: 2) and (often but sporadically) in late West Saxon (Hogg 1992: 298). Thus, one way in which such an irregularity could be explained would be to assume that one of the donor languages underwent a (regular) metathesis. Collecting examples of metathesis is therefore not necessarily irrelevant to the question of language contact. ‘wax’. Lt. vãškas, Lv. vasks; OCS воскъ ‘wax’ ~ ON vax, OHG wahs ‘wax’ (Machek 1968: 697; Polomé 1986: 661) — The Lithuanian -šk- is in itself problematic, as outside of a RUKI environment, it is difficult to derive it from any Indo-European cluster (Villanueva Svensson 2009: 15–16). The most frequent solution is to suggest a proto-form *uoḱs-ko- (Lidén 1897: 28; Kiparsky 1934: 96; Kortlandt 1979a: 59; Derksen 2008: 529), but what does not seem to have been noted is that *-ḱsk- would hardly have yielded Germanic *-hs- in the first ‣ Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 218 chapter 6 place; compare OHG misken ‘to mix’ (< *miḱ-ske-; LIV 428–429). The alternative reconstruction *-ḱk- (ALEW 1386) equally fails to explain the Germanic evidence (cf. Arumaa 1976: 98).97 Unless we assume an irregular, and apparently unmotivated, metathesis (thus e.g. Endzelīns 1911: 57; Smoczyński 2018: 1617;98 РЭС viii: 286), the Balto-Slavic and Germanic forms cannot be regarded as regularly cognate, and the disagreement between the two words might best be accounted for by assuming parallel borrowings from a non-IE source. A number of examples of an alternation *-ks- ∞ *-sk- have been identified elsewhere in Europe. First of all, we can mention the comparison of OHG dahs (< *þahsa-) ‘badger’ with the name of the Middle Irish legendary figure Tadhg (< *tazgo-) mac Céin, who was associated with a taboo against eating badger meat (on which see Mac an Bhaird 1980) (Kroonen 2013: 531; van Sluis et al. 2023: 212). More reliable examples can be found between Greek and Latin: Gr. ἰξός ~ Lat. viscum ‘mistletoe; birdlime’, Gr. ἀξίνη ~ Lat. ascia ‘axe’ (Furnée 1972: 393; de Vaan 2008: 57). ‘sturgeon’. Lt. erškẽtas; Pr. E esketres ‘sturgeon’; Lat. excetra ‘sea serpent’ ~ R осётр, Cz. jeseter, SCr. jèsetra ‘sturgeon’ (Pisani 1968: 20–21; for further refs. and discussion of the Germanic comparanda, see pp. 236–237) — The correspondence between Baltic and Slavic is quite irregular. Although Prussian esketres · stoer99 and Slavic *esetra- (trad. *( j)esetrъ) are hardly to be separated, the -k- in Baltic remains a problem. A change *ešetras ≫ *ešketras due to the influence of Lt. erškėt̃ is, eršketỹs ‘wild rose’ (Būga 1922: 195; Endzelīns 1943: 171; Топоров ПЯ ii: 88–91) is hardly plausible; an association with this word can only come into question to explain the later metathesis *ešketras ≫ erškẽtas but not the ‘intermediate’ form attested in Prussian and as eſchketras ‘walfisch’ in Bretke.100 In principle, if the Slavic -s- goes back to *-ḱs-, the relationship between the Baltic and Slavic words could be understood as metathetic.101 ‣ 97 98 99 100 101 Stang (1972: 61) does not see any need to comment on this irregularity; likewise Vasmer (REW i: 231). Fraenkel (LEW 1207) refers to Endzelīns (1911: 57), who operates with an unexplained sporadic alternation already in Proto-Indo-European (cf. Būga 1922: 176; Otrębski 1939: 133). Smoczyński assumes an ad hoc metathesis only for Slavic, but overcomplicates the Baltic evidence through the assumption of an unattested reflex *vašas (for a suggestion on Finnish vaha, see Chapter 3, fn. 163). To be read /esketrĭs/? Compare erßketris · Wallfisch in Lexicon Lithuanicum (ALEW 303). Žulys (1966: 152–153) plausibly interprets this word in Bretke as a Prussianism. Kortlandt (2000: 125), on the other hand, who expects *e- > a- in Prussian, takes the initial e- as evidence that the word was loaned from Lithuanian (also ALEW 303). The etymological connection with Pl. obs. (Sł Warsz. ii: 171) jesiory pl. ‘fishbones’ and Lt. ešerỹs ‘perch’ (Brückner 1927: 206; REW ii: 281–282, Derksen 2008: 144) is morphologically problematic (*es-et-r- beside *es-er-?). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 consonantism 219 The comparison with Lat. excetra ‘sea serpent’, however (e.g. Trautmann 1910: 331), suggests an original cluster *ḱsk, similar to the one traditionally reconstructed for ‘wax’ (see above). In fact, the Latin word could even be analysed as a regular cognate of Lithuanian ešketras, assuming a reconstruction *eḱsketr-. Nevertheless, to connect the Slavic word, we need to assume an ad hoc simplification *-ḱsk- > *-ḱs-, which is without parallel. The main issue with the comparison is semantic. The oldest attested meaning of the word in Latin is a kind of mythological sea serpent (Pisani 1968: 21; TLL vi: 2165). As sturgeons are a particularly large fish, such a semantic shift is quite imaginable. Compare, for instance, Bretke’s use of the word ešketras to render the biblical Walfisch (Žulys 1966: 153), or Finnish sampi ‘sturgeon’, dial. ‘fish god’ (Liukkonen 1999: 124). A loanword from Greek ἔχιδνα ‘viper’ through Etruscan mediation (Walde/Hofmann i: 425–426), as noted by Pisani, is phonologically problematic. As the similarity between the Balto-Slavic and Latin forms is so striking, and the semantic difference is easily bridgeable, it seems plausible that these words belong together. ? ‘aspen’. Lv. apse; Pr. E abse; R оси́ на, LSrb. wósa, Sln. jesíka ‘aspen’ ~ ON poet. ǫsp (cf. Ic. ösp ‘aspen, poplar’), OHG aspa ‘aspen’ (Meillet 1909: 70; Machek 1954: 132; Skok ii [1972]: 759; Boutkan/Siebinga 2005: 94; Kroonen 2013: 39; Matasović forthc.) — Arm. dial. opʿi ‘poplar’ most probably belongs here, too. The Armenian word can reflect *Hops- (Friedrich 1970: 49–50; Witczak 1991; on the phonology, see also Clackson 1994: 99–100; Kümmel 2017a: 442), although a reconstruction *Hosp-, matching Germanic, cannot be ruled out, either (Normier 1981: 24, fn. 23). It is usually assumed, however, that the metathesis was a Germanic-internal phenomenon (cf. IEW 55).102 This metathesis would be irregular, but it could quite reasonably have been motivated by an association with *aska- ‘ash’ (see Normier 1981: 25–26 with lit.; note also the discussion in Chapter 7, fn. 83). This example of metathesis is therefore uncertain. For a detailed discussion of the Lithuanian forms and Turkic comparanda, see pp. 278–279. ‣ 6.3.2 Baltic *sT ∞ Slavic/Germanic *(T)s In a footnote, Endzelīns (1911: 43–44) has enumerated some examples of apparent alternations between *st and *ts in the Indo-European material. Although he does not make any claim as to the regularity of such a metathesis, Kroonen/ Lubotsky (2009) have proposed that the development *ts- > *st- was indeed 102 Contra Kluge/Seebold (p. 189) and Kroonen (2013: 39), a Proto-Germanic variant *apsōcannot be posited on the basis of the OE variant æpse*, which is the result of an internal development (Campbell 1959: 185). Contrast OHG aspa with wefsa ‘wasp’ (< *waps-jō-). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 220 chapter 6 regular in Germanic on the basis of the equation of Skt. tsárati ‘sneak’ and Go. stilan ‘steal’. To this we may add another compelling example adduced by Endzelīns: – Skt. tsárati ‘sneak’, Go. stilan ‘steal’, Lt. selėt́ i ‘lurk, sneak’, Arm. sołim ‘crawl, creep’ – Skt. tsáru- ‘handle, hilt’,103 Gr. στελεά ‘handle’, ON stjǫlr ‘butt, rump’, OE stela ‘stalk, stem’ (cf. ME stele ‘the handle of a tool or utensil’), ?Arm. stełn ‘stalk, branch’ The amount of data is quite limited, and Armenian shows conflicting reflexes of the initial cluster. Nevertheless, by assuming that selėt́ i shows the regular Baltic reflex of *ts-, we can also account for a few other unexpected cases of sin Baltic: – Lt. sárgas, Lv. sar̂gs ‘guard’ ~ OCS стражь, R сто́ рож ‘guard’, cf. Gr. στέργω ‘feel affection’ (cf. REW ii: 20; Derksen 2008: 467)104 – Lt. síena, Lv. siêna ‘wall’ ~ OCS стѣна ‘(defensive) wall, barrier’, metaphorically ‘rock face’ (Brückner 1927: 529; Kalima 1934: 552, who reject — perhaps unnecessarily — the old comparison with Go. stains ‘stone’) – ? Lt. súolas, Lv. suôls ‘bench’ ~ Go. stols ‘seat, throne’. The Germanic word has alternatively been derived from *sd-ōl- to the root *sed- ‘sit’ (Kerkhof apud Kroonen 2013: 481; cf. Martirosyan 2008: 610–611), but this can be viewed as a serious alternative.105 On the other hand, there are a couple of examples which show the opposite correlation, and which therefore cannot be accounted for with any Indo-European reconstruction. It is possible that these represent parallel loanwords from nonIE sources: ‘bison’. Lt. stum̃ bras ‘bison’, Lv. stum̃ brs ‘aurochs’ ~ Pr. E ⟨wissambs’⟩ · ewer; R зубр, Pl. obs. (cf. Sł. Warsz. viii: 374) ząbr ‘bison’; OE wesend, OHG wisunt ‘bison’ (Schrader/Nehring ii: 261; Machek 1968: 719; Kroonen 2012: 253; Šorgo ‣ 103 104 105 This tsáru- is hardly the same word as tsáru- RV ‘ein schleichendes Tier’ as maintained by EWA i: 687. A change *ts- > *st- in Slavic and Greek is perhaps unexpected, typologically speaking, as *ps- and *ks- are both preserved word-initially in Greek, and we have *ks- > *kṣ- > *x- in Slavic (cf. Pl. dial. chybać ‘rush; sway’ ~ Skt. ví kṣobhate ‘stagger’). But we should not a priori assume that *ts- (in which the two phonemes have the same place of articulation) should have behaved similarly to other *Cs-type clusters. Petri Kallio (p.c. March 2023) points out, for instance, Western Finnish -tt- < *-ts- (e.g. mettä < metsä ‘forest’) beside preserved -ps-, -ks-. In any case, the Baltic word, already in view of its acute intonation, is not, with Būga (1922: 280), to be compared with OCS село ‘field, estate, settlement’ (which might be ?< *sedla-, trad. *sedlo; Brückner 1927: 491–492; Stang 1972: 47) or Lat. solium ‘seat, throne’ (probably with *d > l, de Vaan 2008: 571). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 consonantism 221 2020: 455–456) — The Prussian attestation is abbreviated in the original, and is normally restored to wissamb[ri]s (Trautmann 1910: 464; Endzelīns 1943: 276). If the -b- in Balto-Slavic is epenthetic in a cluster *-mr- (Būga 1912: 45),106 then the correlation between Germanic *-und- ~ Baltic *-umr- could reflect the same *d ~ *r alternation as in Lt. sidãbras ~ OCS сьребро ‘silver’ and Pr. E wobsdus ~ Lt. opšrùs ‘badger’ (see pp. 225–227). However, note that it is in principle not possible to rule out a reading wissamb[i]s for Prussian. Problematic are the Latvian variants in s-: sũbrs (ME iii: 1129; EH ii: 606) and sum̃ brs (ME iii: 1120; LVPPV). The preserved -m- and accentuation of the Latvian forms seem to point towards borrowing. It is tempting to interpret sũbrs as a loan from East Slavic (Petersson 1921: 39; with secondary s-?), in which case sum̃ brs might be a Polonism. In any case, it seems obvious that the cited words for ‘bison’ cannot be separated from one another (cf. REW ii: 107; Būga 1912: 44–46). In view of the numerous problems with reconstructing a common proto-form, it seems most probable that we are dealing with a word of non-IE provenance. On the element *wi- in Germanic and Prussian, see 7.1.3. Note also the mismatch in vocalism between East and West Baltic (see 7.3.1).107 ‘roe’. Lt. stìrna, Lv. stir̃na ~ OR сьрна́ (Зализняк 2019: 205), Sln. sŕna ‘roe deer’ — Endzelīns (1909: 378; cf. EH ii: 489) has pointed to a form ⟨ẜirnos⟩ acc.pl. ‘roe’, attested in Rehehusen’s 16ᵗʰ century Manuductio ad linguam Lettonicam. If this is not merely an error (cf. Fennell 1982: 339), then it perhaps results from a contamination with the Slavic word. Despite Endzelīns and many who have followed him, I doubt it should be considered a unique archaism (but compare ‘bison’, above). Most agree that stìrna is of IE origin and related to Pr. E sirwis ‘roe deer’, Lat. cervus ‘deer’ and further the root for ‘horn’ (Trautmann 1923: 260; Nussbaum 1986: 8, fn. 16; Derksen 2015: 429). The initial st- has been subject to numerous explanations. Early scholars suggested a loan from Slavic (e.g. J. Schmidt 1895: 37; Mikkola 1908: 14; also Mayer 1990: 102), assuming a pre-Slavic *ć was adopted as Baltic *st. There is no other evidence from early Slavic loans, however, that would support an affricate pronunciation at such a recent date. Alternatively, Andersen (2003: 53–54) has suggested a loan from an unknown IE dialect.108 ‣ 106 107 108 Compare Lt. dum̃ blas ‘mud, sludge’ ~ Lv. dumûksnis ‘marsh’ (see p. 210). The involvement of the pan-Caucasian term for ‘bison’ (Oss. ID dombaj, Karachay dommaj, Bzyp Abkhaz a-domp’éj, Georgian domba; Иванов 1975; Абаев 1996: 206; Kroonen 2012: 253) in this equation is less certain, as the initial d- and the suffix both need to be accounted for. Another issue with the traditional etymology is the accentual difference between Baltic and Slavic (cf. Meillet 1905: 446). Assuming vr̥ddhi per Petit 2004: 184; Villanueva Svensson 2011: 31 seems like an ad hoc solution, see Pronk 2012: 11–13. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 222 chapter 6 In my opinion, it is worth asking whether the IE etymology might be wrong; after all the roe, compared to the red deer and the elk, has far less prominent horns.109 ‘thousand’. Lt. tū́kstantis, Lv. tũkstuôtis ‘thousand’ ~ OCS тъıсѫщи, тъıсѧщи ‘thousand’; Go. þusundi, ON þúsund ‘thousand’ (Stang 1966: 282; 1972: 49) — Note that the *-s- in Slavic cannot reflect a simple *s (which should have become *x by the RUKI law), but would be quite elegantly be explained from *ts, a reconstruction which could also work for Germanic. See 3.5.4 for a detailed discussion of this word and the Uralic comparanda. ? ‘fast’. Lt. dial. bruzgùs, Žem. bruzdùs ‘quick, agile’ ~ OCS (Supr.) брьзо adv. ‘quickly’, MR борзы́ и ‘fast (of horses)’, OCz. brzý (Gebauer i: 111), SCr. bȓz ‘fast’ — Much has been made of the variant бо́ рзды, attested in Middle Belarusian since the 15ᵗʰ century (ГСБМ ii: 148–151). As support for the latter’s antiquity, Ильинский (1910: 324) has adduced the SCr. dial. (Montenegro) brzdica ‘rapids’ from Vuk (RJA i: 695; РСА ii: 157; Skok i: 222) and modern Polish barzdo (replacing OPl. barzo in the 16–17ᵗʰ centuries). Despite a general consensus, I consider the doubts voiced already by Потебня (1881: 1) still valid. SCr. brzdica is curiously paralleled by dial. brzdar (РСА ii: 156) for br̀zār ‘a kind of leather bag’, in which Skok (i: 222) would see a contamination.110 Perhaps Derksen (2008: 70) is correct in blaming the Belarusian variant on Baltic influence (but see below). Although these variants present some problems, I doubt that the evidence is sufficient to support a Proto-Slavic variant *burzda- (trad. *bъrzdъ). The comparison of the Baltic and Slavic data implies multiple irregularities. First, there is the irregular correlation between Slavic *-ur- (trad. *-ъr-) and Baltic -ru-.111 Secondly, there is a disagreement between Baltic -zd- ~ -zg- and Slavic *-z-. If -zd- can be set up as original in Baltic (which should not be taken for granted; the Aukštaitian -g- would in any case be left unexplained), we might ‣ ‣ 109 110 111 Incidentally, I would also keep apart the words for ‘cow’, Lt. kárvė, R коро́ ва, as neither the acute nor the initial velar are well accounted for. Pr. E kurwis ‘ox’, for what it is worth, would in my opinion suggest a labiovelar. For Pl. barzdo, see Łoś (1922: 148), who also adduces Pl. smardz ‘morel’ ≪ OPl. smarsz (SSP viii: 318). Sln. brzdit ‘stolz (von Pferden)’ (Murko apud Pleteršnik i: 68) which Bezlaj (i: 50) included here, is derived from bŕzda ‘bridle’ (Furlan 2013: 119). I am hesitant to put any weight on the variant burzdùs, which seems only to have been recorded by Kurschat (1883: 65) who himself marks it as an unfamiliar word with the note “in Südlitt.”. Even more doubtful is the variant burzgùs. In the LKŽ, it is equated with bruzgùs, with a single illustrative sentence: “Mūs mergaitės tokios bur̃zgios”. Yet a very similar example found in the Papildymų kartoteka, “Kõ tà mergáitė tokià burzgì?”, is glossed as “niurzgùs” = ‘grumpy’! Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 223 consonantism table 13 Possible examples of sibilant metathesis Baltic Slavic Germanic Elsewhere ‘wax’ *u̯ o(ḱ)sk- *u̯ o(ḱ)sk- *u̯ oks- ‘sturgeon’ *e(ḱ)sket-r- *e(ḱ)set-r- ?*(ḱ)str- Lat. *eksket-r- ? ‘aspen’ *op(u)s- *osp- Arm. *ops- (or *osp-) ‘bison’ *stum(bʰ)r- *(d)zam(bʰ)r- *u̯ i(t)snT- ‘roe’ *st(i)r̂n- *(t)s(i)rn- – ‘thousand’ *tûstant- *tûts(a)nt- *tū(t)snT- ? ‘fast’ *Bur(dʰ)z- – ?*BruzD- *ops- PF *tušantMd./Ma. *tüšäm be able to set up an irregular correspondence between Baltic *zd and Slavic *(d)z, parallel to the examples of *st ∞ *(t)s, above. This correspondence could potentially find a parallel in the word for ‘heather’. As against the standard vìržis, Mielcke (ii: 270) cites birʒdʒei ‘heydekraut’. The reality of this form seems to be confirmed by the form brizdei ‘Calluna’, attributed by Pabrėža (1834: 60) to Prussian Lithuanian (admittedly, this is perhaps simply miscopied from Mielcke). This would also show *b ∞ *v (see 6.4.2), but in view of the large amount of variants shown by the word for ‘heather’ in Slavic (see pp. 215–216), it would seem hasty to draw any dramatic conclusions on the basis of such scanty data. 6.3.2.1 Conclusion The certain and possible evidence of ‘sibilant metathesis’ is collected in Table 13, above (see p. 185 for help reading the table). Forms which do not provide relevant data are presented in light grey. Shaded cells indicate sibilant-initial clusters. The clearest pattern concerns the alternation *sT ∞ *(T)s: here we consistently find a sequence *st in Baltic. In Germanic and Slavic, the surface realization is just a sibilant; however, in the word for ‘thousand’, there is indirect Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 224 chapter 6 support for the reconstruction *ts, as only this reconstruction can unite the Slavic and Germanic data and explain the absence of the RUKI law in Slavic. Since a *t could have been present in the other examples, and they show a comparable pattern, it seems reasonable to assume they result from the same substratal phenomenon. 6.3.3 Other Alternations Involving Sibilants (a) ? ‘furrow’. Lt. bir̃žė ‘row, furrow; timber tract; border mark’, Lv. bìrze ‘furrow, row’ ~ R борозда́ , Cz. brázda, SCr. brázda ‘furrow’ — The Slavic word is traditionally compared with Skt. bhr̥ ṣṭí- ‘point, peak’ (< *bʰrḱ-ti-, cf. EWA ii: 273). The implied suffix *-d- in Slavic is difficult to set up (see Vaillant 1974: 490 for some doubtful examples), and the semantics are hardly compelling, in any case. The main disadvantage of the etymology is that we would need to abandon any connection with the Baltic synonym (cf. Holzer 1989: 53).112 The inclusion of the Slavic evidence implies an additional alternation between Baltic *ž and Slavic *zd. Perhaps this can be compared with Lt. triùšis ~ OCS тръсть ‘reed’ (see p. 201). On the comparison with Lat. porca ‘furrow’ etc., see p. 208. (b) ‘beard’. Lt. barzdà, Lv. dial. bā̀rzda; OE beard, OHG bart ~ OCS брада, R борода́ (acc.sg. бо́ роду); Lat. barba ‘beard’ (Schrijver 1991: 448; Kuiper 1995: 66; Derksen 2015: 82; Pronk 2019a: 147) — Kroonen (2011b: 150–151) has presented a native etymology for this word. He assumes that the Germanic word for ‘beard’ is connected to ON broddr, OE brord ‘point, tip; shoot’ (< *bruzda-; thus already Pedersen 1895: 73) and ON borð, OE bord ‘board, plank; side of a ship’ (< *bur(z)da-). He opts for the reconstruction *barzda- (cf. Kroonen 2013: 54), which, being the result of a reshuffling of ablaut within Germanic, would imply that the Latin and Balto-Slavic words are Germanic loanwords. The reconstruction *barzda- for Germanic (likewise e.g. Kluge/Seebold 93) would provide a natural explanation for Lt. barzdà. We may interpret the Baltic and Germanic words as cognate or, following Kroonen’s model, view the Baltic word as a loan from Gothic. The Germanic loan etymology might be supported by the absence of the RUKI law in Lithuanian. It would incidentally be attractive to see Crimean Gothic bars, which has previously been considered a transmission error or a unique retention of nom.sg. -s (Lehmann 1986: 62– 63), as a direct reflection of this preform.113 ‣ ‣ 112 113 Note that the Baltic word is left unmentioned by e.g. Berneker (i: 75), Vasmer (REW i: 109) and ЭССЯ (ii: 220). This would require a return to the more traditional view that the words for ‘board’ (cf. Go. fotu-baurd ‘footstool’) are unrelated, for which something can indeed be said; the partial semantic convergence in Norse may be secondary. Latvian bā̀rda is in any case due Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 consonantism 225 The main problems arise when considering the Slavic and Latin evidence. Despite the claim to the contrary in ALEW (102–103), the loss of -z- in Slavic would be irregular (cf. Pedersen 1895: 72–73). This could be remedied by assuming a Slavic loanword from West Germanic; however, mobile accentuation is generally thought to be atypical of Germanic loanwords (Meillet 1909: 69; Pronk-Tiethoff 2012: 242–244 with lit.). A Latin loan from Germanic faces chronological issues as the change *-rð- > -rb- belongs to the preliterary period (e.g. Weiss 2020: 208). Moreover, it is likely that a preceding sibilant would have blocked the frication of inherited *dʰ, whether inherited *sdʰ merged with *st (Meiser 1998: 119; Weiss 2020: 161) or with *sd (Lubotsky 2004). As a result, the Latin form is only consistent with a pre-form without *s. ‘in calf’. Lt. ber̃gždžias ‘barren, fruitless’ ~ R dial. берёжая ‘in foal’, SCr. dial. brȅđa ‘in calf; pregnant’ (< *berdjā; trad. *berdja) — Lat. forda ‘in calf’ is ambiguous, and could reflect either *bʰrd- or *bʰrsd- (cf. Leumann 1977: 210– 211). Despite ALEW (116–117), the Baltic and Slavic forms are not formally identical, not only because the loss of -z- in Slavic would be irregular (see above), but also because the Slavic form exhibits an acute. The difference in intonation could be accounted for by reconstructing *bʰerd- for Slavic and *bʰersd- for Baltic (with *-s- blocking Winter’s law). The morphological function of this *s would be unclear, however, and the parallelism of this example with the word for ‘beard’ makes it rather tempting to view both in the context of parallel loanwords. ‣ 6.4 Other Irregularities 6.4.1 Alternations Involving Dentals (a) ‘silver’. Lt. sidãbras, Lv. sudrabs, dial. sidrabs ~ Pr. iii sirablan acc.sg.; OCS сьребро, Cz. stříbro, Sln. srebrọ̑ ‘silver’; Go. silubr, ON silfr, OHG silabar* ‘silver’ (Ipsen 1924: 229–230; Stang 1972: 47; Huld 1990: 409–410; Boutkan/Kossmann 2001; Mallory/Adams 2006: 242; Kroonen 2013: 436; Šorgo 2020: 448; Thorsø et al. 2023: 108; van Sluis et al. 2023: 221) — This word has widely been considered an ancient Wanderwort. The original form must probably be reconstructed with *r–r, with different dissimilations in Germanic and Prussian. Nevertheless, the East Baltic -d- is difficult to write off as dissimilatory, ‣ to an internal development, as implied not only by the Lithuanian equivalent, but also by the Latvian dialect data (Kregždys 2004: 20–21; ALEW loc. cit.). Perhaps it is Germaninfluenced: cf. Pr. E bordus ‘beard’ = */bārdus/ which is probably from MLG bart, nom.pl. bārde (Smoczyński 2000: 178). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 226 chapter 6 since such a dissimilation would be entirely without parallel. Further comparanda are found in Celtiberian ⟨śilaᴘuṙ⟩ ‘money, ?silver’ (K.H. Schmidt 1977: 55) and Basque zilhar ‘silver’ (Boutkan/Kossmann 2001; Thorsø et al. 2023: 108). An additional issue within Balto-Slavic is the fact that the medial -a- in Baltic does not match Slavic *-e- (see 7.2.2). ? ‘badger’. Pr. E wobsdus ‘badger’ ~ Lt. opšrùs ‘badger’ (Apſchro gen.sg. in Bretke implies an a-stem) (cf. Bellquist 1993: 344) — The reality of the Prussian form is confirmed by the gloss wobsdis ‘quod dicitur eyn luchs’ (probably to be corrected to *eyn dachs, Töppen 1867: 155; Gerullis 1922: 205) as well as perhaps Kashubian jôpsc (< *āpsti-; trad. *apstь)114 ‘badger’ (Лаучюте 1982: 78). In view of this, Smoczyński’s dismissal of the Prussian form as having “no explanatory value” (2018: 885) is too hasty.115 A similar form also seems to occur in the Lexicon Lithuanicum: ⟨opßcʒus⟩ ‘fisch otter’ (ALEW 721; the same form is also given beside ⟨ůbßrus⟩ in ClG 663). This comes particularly close to the Kashubian form, and one might suspect that both have been borrowed from Prussian. However, the difference in voicing remains to be explained. The alternation between Lt. -r- and Prussian -d- is unlikely to be due to different suffixation, as -d- is not a productive suffix (LEW 517–518; Smoczyński SEJL² s.v.). In theory, one may compare the similar alternation in the word for ‘silver’, with the caveat that the distribution does not match. However, we must note that Lv. âpsis, dial. (Vidzeme) âpša ‘badger’ does not appear to contain either “suffix”. ALEW suggests that Lv. âpša may derive from an earlier *âpsćā(< *âpštjā-), thus coming close to the marginally attested Lt. opščius (ALEW 721; see above). It does indeed seem likely that the development *stj > *š was regular in Latvian (Endzelīns 1923: 125–126), but as the reconstruction of *t remains hypothetical, it is uncertain whether the irregularities in this word can be used to support foreign origin. However, the relationship between the forms is also ‣ 114 115 The derivation seems acceptable so long as the voiceless auslaut can be attributed to wordfinal devoicing. Alternatively, Boryś (SEK ii: 341) suggests the Kashubian word is cognate with Polish jaźwiec ‘badger’ through a development *jazvc > *jasfc > *japsc. Indeed, this might better explain the variant jôlsc (?< *javzc). Since both etymologies require an irregular development, it is difficult to decide between them. Perhaps the two options could be combined if we assume that the inherited word for ‘badger’ was influenced by the Prussian word. I fail to comprehend Smoczyński’s problematization of the initial w- in Prussian, especially since no such issue is taken with deriving Pr. E wosux ‘he-goat’ from *āž-uk- (Smoczyński 2018: 886). A prothetic w- is regular before o- in the dialect of the Elbing Vocabulary, as has long been recognized (Trautmann 1910: 158); this is confirmed by the complete absence of words starting with o-. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 consonantism 227 difficult to explain in an IE context. On the possibility that the word for ‘bison’ also shows a *d ∞ *r alternation, see p. 221. (b) ? ‘bee’. Lt. bìtė, dial. (Žem.) bitìs, Lv. bite; Pr. E bitte, TC bete ‘bee’ ~ OCS (Ps. Sin.) бьчела, Cz. včela, Sln. čəbę̑la ‘bee’; OIr. bech ‘bee’, MW begegyr ‘drone’ (Hamp 1971: 187; Vennemann 1998; Philippa et al. i [2003]: 308; Matasović 2009: 65; van Sluis 2022: 6–10, forthc.; note also Machek 1968: 679) — The relationship between the Baltic and Slavic words is difficult to account for. A reconstruction *bit-kelā- (Knut-Olof Falk apud REW ii: 471) could theoretically work, but the analysis of the second element remains unclear. It is therefore usually assumed that we are dealing with different suffixes, *bʰi-t- beside *bʰi-k- (Specht 1947: 46; IEW 116), added to the zero-grade of a root *bʰei-, which is indeed attested with ablaut in Germanic: OHG bīa (MoHG dial. Beie), Du. bij against OHG bini (with short vowel confirmed by Notker), MoHG Biene, MLG bene ‘bee’ (Kroonen 2011b: 228–231). Vennemann (1998: 478–479; cf. Takács 2001: 109–110 with some older macbjt rocomparatavist refs.) has drawn a further comparison with Egyptian ‘bee’ (Erman/Grapow i: 434; cf. the derivative in Coptic ⲉⲃⲓⲱ ‘honey’, Vycichl 1983: 38). The similarity is indeed striking, especially if the -t can be considered a feminine suffix (which is not certain; Takács 2001: 109). Such a suggestion is also historically plausible, since the first depictions of hive beekeeping derive from Egypt (Crane 1999: 162), although there is admittedly a great geographical distance between Egypt and the Northern Europe, to which our word is restricted (see van Sluis 2022: 7). The main obstacle to uniting the European forms is the Celtic vocalism. Although *biko- has sometimes been reconstructed (e.g. Berneker i: 116; Heiermeier apud LEW 1329; Matasović 2009: 65), the broader consensus among Celticists favours *beko- (Stokes 1894: 166; Pedersen 1909: 88; LEIA B-24–25; van Sluis 2022: 8). This is supported by North Occitan bèca ‘wasp’, which is most probably a Gaulish loanword (Delamarre 2003: 70).116 As noted by Pedersen (loc. cit.), the Slavic form could potentially reflect an earlier *bečelā(trad. *bečela) with the raising of unstressed *e before a palatal; compare, ‣ 116 The Celtic etymology has been rejected by A. Thomas (editor’s note, Romania 35, 139) and FEW (xiv: 344). They note that Creuse bièco would imply an earlier *bęsca, which they assume was metathesized from *guespa, deriving ultimately from Lat. vespa ‘wasp’. However, the /s/ ought to have been preserved in Limousin, cf. crespa ‘kind of pancake’ (= French crêpe). In addition, the word was probably originally masculine (note the Limousin variant bèc), which would explain the preserved -c (Occitan lac ‘lake’ < lacus; cf. FEW v: 126; old -ca should have yielded -cha in Limousin, as in pescha ‘fish’ < *pisca; cf. Thomas loc. cit.). The diphthong in Creuse may be explained as due to contamination with gyepo (cited apud FEW xiv: 344) ‘wasp’, with which it is in competition in this area. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 228 chapter 6 in a very similar environment, OCS вьчера ‘yesterday’ beside вечеръ ‘evening’ (cf. Kortlandt 1984–1985). As the comparison between Baltic *bʰit- and Slavic/Celtic *bʰek- would rest on the first phoneme alone, it is not entirely clear that there is enough material to draw a reliable comparison.117 6.4.2 Alternations Involving Labials (a) ‘bean’. Pr. E babo; R боб, Pl. bób, Sln. bòb; Lat. faba ‘bean’ ~ ON baun, OHG bōna ‘bean’ (Machek 1950b: 158; Kuryłowicz 1956a: 194; Schrijver 1991: 488; РЭС iii [2009]: 283; Kroonen 2013: 55; Matasović 2013: 83; Pronk/Pronk-Tiethoff 2018: 282; Šorgo 2020: 435) — According to Walde/Pokorny (ii: 131), the Germanic forms developed from *babnō- by dissimilation, an ad hoc suggestion that has gained few serious proponents (ЭССЯ ii: 149; Bańkowski 2000 i: 69 and with hesitation Kluge/Seebold 96). Instead, one has tended to keep the Germanic words apart (Kretschmer 1896: 146; Petersson 1909: 390; de Vries 1962: 29; implicitly Trautmann 1923: 23; REW i: 180). If the words are indeed related, the disagreement between Slavic and Latin *bʰ and Germanic *w would favour independent borrowings from a non-IE source.118 Note also in this context the Latin a-vocalism (see 7.6). However, one must remain cautious due to the small amount of phonetic material compared.119 Kretschmer (1896: 146) has assumed a connection with Lt. pupà, Lv. pupa ‘bean’, positing a loan from Slavic through a Finnic intermediary (Berneker i: 65; Walde/Pokorny loc. cit.). This can hardly come into question: Livonian pubā ‘bean’ is a Baltic loan (Thomsen 1890: 100; Petersson 1909: 390; Sabaliauskas 1959: 235), while Finnish papu is a loan from Slavic (Kalima 1956: 102). Pronk/Pronk-Tiethoff (2018: 282) have instead suggested that pupà is a loan from the same foreign source as the above forms. While the resemblance (a stem consisting of two labial stops) is indeed striking, neither of the implied alternations are precisely paralleled in my material ‣ 117 118 119 But note further van Sluis (forthc.), who presents some potential parallels for an alternation between *k, *t and *∅ in possible substrate words. In this connection, note Berber *a-baw ‘faba bean’, which Kossmann (1999: 113–114) states cannot be a direct loan from Latin. His current opinion (cf. Kossmann 2021: 16) is that we are dealing with a Wanderwort which has “spread over the Berber territory in post-protoBerber times”. If a Latin origin is ruled out, this begs the question as to whether it has been adopted from a related non-IE source. An additional argument for foreign origin could be provided by the gloss haba ‘faba’, attributed to the “Falisci” by Terentius Scaurus. If this word really did belong to Faliscan proper, the absence of the change *-bʰ- > *-f- would imply a Proto-Italic *-b-. This would not match the *-bʰ- required by Balto-Slavic, and rule out the reconstruction of a common proto-form. However, since the development of initial f- > h- was probably not limited to Faliscan, and the reliability of glossators’ attributions is often questionable, it is difficult to base much on this form (cf. Bakkum 2009: 83, 209). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 consonantism 229 (for *p ∞ *b⁽ʰ⁾ see 6.2.2 and for *u ∞ *a see 7.3.1), which makes the inclusion of this form somewhat precarious. It is also difficult to find reliable parallels for the alternation *b⁽ʰ⁾ ∞ *w. One relatively clear case without Balto-Slavic comparanda is the word for ‘pea’ (Gr. ἐρέβινθος ‘chickpea’ ~ OHG arawīz ‘pea’; see e.g. Kroonen 2012: 242–244; Thorsø forthc. with lit.). In addition, Machek (1950b: 152–153; 1968: 132) has suggested we compare the Slavic word for ‘oak’ (OCS дѫбъ, etc.) with OHG tanna ‘fir’ (thus a virtual *dʰonbʰ- ~ *dʰonu̯ -). However, the semantic distance makes this comparison very uncertain.120 (b) ‘carrot’. R морко́ вь, SCr. mȑkva ‘carrot’; OE moru ‘edible root’, OHG moraha ‘carrot’ ~ Lv. bur̃kãns ‘carrot’121 (Machek 1950b: 158, 1954: 167; Kroonen 2013: 378; Matasović 2013: 88; ERHJ i [2016]: 639; Pronk/Pronk-Tiethoff 2018: 282; Šorgo 2020: 446) — The status of the Latvian word is somewhat problematic, as it may also have been loaned from Baltic German Burkane (see the discussion on pp. 31–32). However, it can be noted that while the word in Russian and German is clearly a late replacement for an older term for ‘carrot’, there is no other candidate for an old term within Baltic (Bentlin 2008: 247). A possible trace of this word in Lithuanian can be found in Szyrwid’s burkuntay · pasternak, with an unclear -t-,122 while the usual form is the very recent loanword morkvà, morkà ‘carrot’ (← Bel. мо́ рква). Moksha puŕʿkä ‘carrot’ has been derived from Russian (Mikkola 1894: 91; Helimski apud РЭС ii: 223). However, the Russian form is not attested anywhere in the vicinity of Mordovia, being limited to the area adjacent to the Baltic-speaking territory (see pp. 31–32). There are also phonological obstacles ‣ 120 121 122 Within Baltic, one might also cite Lt. kalavìjas ‘sword’ as against Pr. E kalabian (= iii kalbīan acc.sg.). Yet it seems more attractive to explain this disagreement by assuming a Lithuanian loanword in Prussian. There are several cases of German /v/ being substituted as Prussian /b/, such as Pr. iii ebangelion acc.sg. ‘gospel’, burwalkan acc.sg. ‘yard’ (← MHG vorwërc ‘estate’), which implies that the Prussian still had a bilabial /w/ until recently, and therefore substituted a foreign /v/ with a labial stop. One wonders if a similar solution might be on the cards for the Lithuanian form birzdžiai ‘heather’, attested in Prussian Lithuania (see p. 223). Although the attested word for ‘heather’ in Prussian is E sylo (~ Žem. dial. šìlas), this does not rule out the existence of dialectal synonyms. North Žemaitian bur̃konas is a loan from Latvian (cf. Sabaliauskas 1960b: 261). Perhaps the same can apply to the rather aberrant burkúnas given by Juška (i: 254), the geographical origin of which cannot be ascertained. Another trace of this word could be found in Lv. burkan̂ ts² (attested in Snēpele, EH i: 254) if this was borrowed from Lithuanian (differently Sabaliauskas 1960b: 261). The stem-final -t- has a curious parallel in Estonian porgand -i ‘carrot’ (cf. Būga 1925: 771), but this has been analysed as an excrescence within Estonian (Blokland 2005: 298–299). I will leave these forms out of consideration. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 230 chapter 6 to a Russian source. Firstly, the loss of the final -n is unmotivated, the final sequence -an being known even in inherited words (cf. Md. M kućkan ‘greater spotted eagle’ < PU *kočka), and -n having been preserved in other Russian loanwords (cf. Moksha dial. karman ‘pouch, handbag’ ← R карма́ н; praban ‘drum’ ← R бараба́ н, etc.). Second, the palatal ŕ (with subsequent fronting *-a > *-ä; Bartens 1999: 63) is not easily explained on the basis of the Russian data.123 Both of these issues are equally prohibitive to the derivation of the Mordvin word directly from Baltic (thus Donner 1884: 266–267; Иллич-Свитыч 1960: 18). According to Junttila (2019: 51), any word for ‘carrot’ must be recent, as carrot cultivation only became widespread in northern Europe in the Middle Ages.124 Archaeologically, the evidence is “deplorably fragmentary” (Zohary/Hopf 2012: 160), so it seems difficult to draw firm conclusions. Part of the reason for this is the difficulty or impossibility of distinguishing between wild and cultivated specimens, and because cultivated varieties were usually harvested before going to seed (Karg/Robinson 2000: 137; Mueller-Bieniek 2010: 1725). The German word Möhre ‘carrot’, having regular cognates in Old English, must date at least to Proto-Germanic. However, it did not necessarily originally refer to the domesticated carrot. In several glosses, OHG moraha is given specifically as pastinaca silvatica ‘wild carrot’ (see AWb s.v. mor(a)ha); the exact referent of OE moru is not known except for the fact that it was distinct from the foreign wealh-more (glossed pastinaca, daucus; cf. Dictionary of Old English Plant Names,125 s.v. more (1) with lit.); in Middle English, more referred to both inedible and edible roots. In conclusion, there is a close resemblance between Slavic/Germanic *murkand Baltic *burk-. If they go back to parallel borrowings from another source, then we might be dealing with an original term for ‘edible root’ which has become specialized in the sense ‘carrot’ in the individual languages. The word has spread into the Finnic languages (see p. 32) and Mordvin, but the route or 123 124 125 This argument is perhaps not as convincing, as Paasonen (MdWb) records dialectal variants of Erzya morkov ‘carrot’ (which is borrowed from Russian морко́ вь) with a similar palatal — moŕʿko·v, ḿiŕkou̯ . This does not appear to be a general phenomenon, however, so we might assume an exceptional solution, such as transfer of the palatal feature from the Russian final /v’/ to the previous syllable. Junttila proposes a novel etymology (2014: 131; 2019: 51–52), deriving the Baltic German word for carrot from the place name Burgundy in the context of Hanseatic trade. As a parallel, he offers Hungarian burgonya ‘potato’, which is of the same origin. The obvious problem with this etymology is that Baltic German Burkane differs in consonantism, vocalism and place of stress from MoHG Burgund; there seem to be too many missing links in this etymology for it to be accepted. Online database, accessed at http://oldenglish‑plantnames.org/. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 consonantism 231 even trajectory of its spread is difficult to reconstruct. Since the word seems to be old in Germanic, having predated Grimm’s law, we are probably dealing with an originally European word, which may have entered Germanic, Slavic, and Baltic, and ultimately Mordvin, independently from related non-IE sources.126 [? ‘goosefoot’. Lt. balánda, Lv. dial. baluôda ‘goosefoot, Chenopodium’ ~ OS maldia, OHG melta ‘orache, Atriplex’ — See the discussion on pp. 177–178.] (c) ‘oven’. Pr. E Vumpȋs ‘oven’ ~ Go. auhns*, OHG ovan ‘oven’ (for refs. and more discussion, see p. 254) — In view of Pr. E umnode ‘bakehouse’, the word Vumpȋs (for ?*umpins < *umpns) probably stands for an underlying */umnəs/, see also “[monticulus], qui dicitur Vmne prutenice, id est clibanus” 1331 (Gerullis 1922: 33; PKEŽ iv: 267; for the -p-, compare Pr. G kampnit ~ kamnet ‘horse’). It is tempting to compare Prussian *umna- with Germanic *ufna- directly, which would imply an irregular correspondence *p ∞ *m. For further discussion, and on the possible connection with Gr. ἰπνός ‘oven, furnace’, see p. 254. Fraenkel (1936c; see also LEW 1156–1157) attempts to derive both the Prussian word and Lt. dial. ùblas ‘indoor oven for producing tar’ from Germanic. He assumes the Lithuanian word was adopted “von der Weichselgermanen”; however, the attested Gothic auhn acc.sg. ‘oven’, which shows a dissimilation *f > *h (see Kroonen 2013: 557), is hardly a suitable source, and Fraenkel’s *ubnas does not appear to be continued by any Germanic language.127 For Prussian, he points to Sw. dial. (Rietz 486) omn and suggests a possible Scandinavian origin. However, there is no certain evidence of Scandinavian loans in Baltic (see Chapter 2). The Lithuanian word is phonologically rather difficult to compare with the other forms due to the need to assume a “suffix replacement”, and its appurtenance remains uncertain. A possible parallel for the alternation *P ∞ *m is found in the comparison of the Slavic word for ‘oak’ (OCS дѫбъ, etc.) with Finnic *tammi (> F tammi, E tamm, Li. täm) ‘oak’, suggested by Machek (1968: 132). The Finnic word has regular cognates at least in Mordvin (E tumo, M tuma ‘oak’) and probably also in Mari (E tumo, W tum ‘oak’; on the vocalism cf. Живлов 2014: 125; Metsäranta ‣ ‣ 126 127 Guus Kroonen (p.c. September 2021) points me towards some similar North-East Caucasian forms: Lak marχ̄, Dargwa marqʷa ‘root’. I remain agnostic as to whether these could be somehow related. In view of the substitution of Gothic lowered /ę̆/ in Lt. pẽkus ← Gothic faihu (see p. 41), we might expect Gothic */ǫ̆ / to turn up as Baltic */a/, although the existence of an East Germanic dialect which did not undergo *u > au is conceivable. In any case, a later West Germanic origin is out of the question, as German *o with open syllable lengthening is never adopted as Lt. ŭ (see Alminauskis 1934, passim and e.g. 144–145). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 232 chapter 6 2020: 81).128 The possibility that the Slavic and West Uralic words could go back to a shared substrate has been suggested again recently by Живлов (2015) and Aikio (apud Matasović forthc.). While the similarity between the words is curious, it is difficult to imagine a plausible way to bridge the geographical distance between Proto-Slavic and Volga Uralic. (d) ? ‘aftermath’. Lt. atólas, Lv. atãls; Pr. E attolis ‘aftermath’ ~ R dial. ота́ ва, Sln. otȃva ‘aftermath’ — Vasmer (REW ii: 289) compares the Slavic words with Sln. otáviti, Cz. otaviti ‘revive, strengthen’, and considers them a derivative of the verbal root seen in OR тъıти ‘grow fat’ (СРЯ 11–17 xxx: 257–258). It is equally possible that the verb in question is denominal, however: cf. Lv. dial. (ME i: 149, EH i: 133) atãlêtiês ‘recover, get one’s breath back’ < atãls (cf. Thomsen 1890: 159; Gāters 1953: 113). Various root etymologies for Baltic are summarized in LEW (p. 22), but the similarity of the Baltic and Slavic words encourages a direct comparison (cf. Miklosich 1886: 228; Trautmann 1923: 16; Witczak 2001: 44–45). A segmentation of the Baltic word as *atâ-la- and reference to the nominal prefix Lt. ató- is unlikely, as this prefix is unknown elsewhere in Baltic and is probably a Lithuanian innovation on the model of nominal pó-, pró- (etc.). As the Baltic and Slavic words are so similar, and the suffix *-âla- would be unusual, one may consider an alternation *l ∞ *w, which is phonetically plausible, although not paralleled. On the question of the Finnic comparanda, see 3.5.3. ‣ 128 J. Häkkinen (2009: 37–38) considers the West Uralic term a probable substrate word, but without mention of the Slavic comparandum. The comparison between Slavic and West Uralic was already made by Топоров/Трубачев (1962: 246; see also Tomaschek 1883: 704), who saw the Slavic word as a loan from a dialect of Proto-Finnic. This can hardly be seriously considered due to the probable geographical distance between Proto-Slavic and Proto-Finnic and in the absence of convincing parallels. Напольских (2002: 143–145) rather sees the Uralic word as a borrowing from a lost Baltic dialect. Finally, Witczak (2020: 75–76) has interpreted the Slavic word as a loan from a West Uralic compound of *tammə ‘oak’ + *puwə ‘tree’. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 chapter 7 Vocalism 7.1 Initial Vowels In an important article, Schrijver (1997: 307–310) suggested the existence of a morpheme *a-, which he supposed appeared in a number of non-IndoEuropean lexemes with a European distribution. He observed that the presence of the ‘morpheme’ in a couple of cases correlated with a ‘reduced’ stem. The most convincing example of this phenomenon is found in the word for ‘blackbird’: *mesal- Lat. merula; MW mwyalch, Bret. moualc’h (< PCelt. *mesal-(s)kā-) *a-msl- OE ōsle, OHG amsla (< PGm. *amslōn-) ‘blackbird’ An excellent parallel is found in the comparison of Lat. raudus ‘piece of copper or brass’ and ODu. arut (attested in Latin context), OHG aruz ‘ore’ (Schrijver 1997: 308; Kroonen 2013: 37). As well as showing a similar correlation between the presence of *a- and a ‘reduced’ stem (*raud- ∞ *a-rud-), the two variants also show an identical geographical distribution. This distribution is, however, disturbed by the addition with Sumerian uruda (< aruda; Jagersma 2010: 61) ‘copper’ (see Thorsø et al. 2023: 109). Although Schrijver’s (2018: 363) suggestion that the language of Europe’s first farmers could have been related to Hattic would somewhat resolve the geographical issue, there is also a huge time difference involved. It would be quite a stretch to assume that such morphological alternations as found in Hattic1 would have been preserved in Europe intact for millennia after its colonization by farming populations. Several more suggested examples of the morpheme *a- have been collected by Iversen/Kroonen (2017: 518) and Schrijver (2018: 361–363; cf. also Matasović 2020: 338–342), although not all of them show the expected pattern of stem reduction. I have divided my evidence into those which do and those which do not follow this pattern. 1 In fact, it seems hardly possible to rule out that the relevant vowel reductions (associated with the definite article) constitute a young development unique to Hattic. © Anthony Jakob, 2024 | doi:10.1163/9789004686472_009 Anthony Jakob - 978-90-04-68647-2 This is an open access chapter distributed under the terms of the cc by-nc-nd 4.0 license. Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 234 chapter 7 7.1.1 *a- ∞ *∅- with ‘Reduced’ Stem ‘swan (1)’. R ле́бедь, Bg. лѐбед ~ Pl. łabędź, Sln. labǫ́ d; ON ǫlpt, OHG elbiz (for refs. and further discussion, see pp. 176–177) — In East Slavic and Bulgarian, one finds the form *lebedi- (trad. *lebedь). This is the reconstruction given by early authors (e.g. Miklosich 1886: 162; Osthoff 1898: 65; cf. Andersen 1996a: 124),2 although *elbedi- (trad. *elbedь) is nowadays more popular (Булаховский 1948: 118; ЭССЯ vi: 19; Sławski SP vi: 40). The older reconstruction is preferable, as the idea that *eRC- should develop to RĕC- throughout Slavic is doubtful: corroborating examples are lacking, and one would expect a parallel treatment to *aRC- (trad. *oRC-; cf. Vaillant 1950: 160–161; Derksen 2008: 143; Jakob forthc. a.). Derksen (2000: 84) has suggested to account for the different Slavic reflexes by assuming a ‘prefix’ *a- (thus *a-lb- ∞ *leb-). Although this idea was not taken over in his dictionary (2008: 143), it does seem a plausible way to account for the two forms. The irregular alternation between *-bǭd- and *-bed- in the second syllable already strongly suggests a non-IE origin. Furthermore, the geography, with the a-forms restricted to Germanic and the western part of the Slavic area, seems quite consistent with the examples adduced by Schrijver. ? ‘elm (2)’. OR ильмъ, OPl. ilem* (hapax, attested ⟨Ylem⟩ 1472; SSP iii: 15), dial. ilmak (Sł. Warsz. ii: 78), Sln. dial. (Carinthia) lìm < *ilm (Erjavec 1883: 293; Karničar 1990: 51); OE elm, OHG elm; Lat. ulmus ‘elm’ ~ MIr. lem; MW llwyf ‘elm’ (Machek 1954: 90; Polomé 1990: 334; Schrijver 1997: 311; van Sluis forthc.; Matasović forthc.) — Latin ulmus can probably reflect *elmo- with *e- > *obefore velarized /l/, as in olor ‘swan’ ~ MW alarch (< *elar- with Joseph’s law; Schrijver 1995a: 76), followed by regular *olC > ulC (cf. Weiss 2020: 150–151). Quite alone stands ON almr ‘elm’: perhaps ths initial a- has been carried over from other tree names (cf. ON askr ‘ash’, OSw. asp ‘aspen’, al ‘alder’). Matasović (2009: 237), like Pedersen (1905: 313–314), has made an attempt to explain the words in terms of IE ablaut, but has later favoured a non-IE origin (Matasović forthc.). The Slavic words have often been derived from Germanic (Miklosich 1886: 95; Berneker i: 424; Kiparsky 1934: 148), or more specifically, MHG ilme (attested since the 13ᵗʰ c.). In view of (1) the early attestation in Russian (already the Novgorod First Chronicle)3 and its widespread appearance in Russian toponymy (Vasmer 1938: 452; В. Васильев 2012: 427–429), and (2) the non-trivial ‣ ‣ 2 Osthoff reconstructed Pl. łabędź etc. as *lōbʰ-, comparing the Hesychian gloss ἀλωφούς · λευκούς, which occurs beside a parallel gloss ἀλφούς · λευκούς. The former is most probably a mere transmission error (Beekes 2010: 77; Gippert 2017: 184–185), meaning that Osthoff’s reconstruction has no real basis. 3 See Folio 113b (under the year 6738) of the Synodal Codex. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 vocalism 235 development of Bel. dial. лём, LSrb. lom, showing the regular outcomes of a vocalized yer (cf. Schaarschmidt 1997: 62) and loss of *i- (trad. *jь-), a late Germanic loan is out of the question (thus also Friedrich 1970: 81–82; ЭССЯ viii 222–223; Derksen 2008: 211).4 Despite Machek’s (1954: 90) claim to the contrary, all the forms can be united under a single proto-form, *ilima-.5 The irregularity essentially depends on the Celtic evidence, namely MIr. lem (< *lemo- / *limo-) and MW llwyf (< *leimo-) ‘elm’. These words fail to correspond with each other even within Celtic. Schrijver (1997: 311) characterizes the relationship between the Celtic and other European words as *V -lm- ~ *lVm-. Nevertheless, no precise parallels can be identified, and since the comparison only consists of two consonants, there is always a possibility that the similarity is due to chance. Curiously, a very similar word is also found in several Turkic languages, cf. Chuv. jĕlme, Tatar elmä; Kumyk elme ‘elm’, Noghai elmen ‘aspen’. On the basis of these forms, CИГТЯ (i: 126) offers the Proto-Turkic reconstruction *elmen. If the final -n in Noghai is secondary after emen ‘oak’, the Caucasian Turkic forms could be combined under *elme; however, the Volga Turkic forms imply a reconstruction *ilmä (cf. А. Дыбо 2007: 129–130). The initial j- in Chuvash is irregular and would suggest a reconstruction *jilmä, but it is perhaps secondary; Савельев (p.c. August 2021) has informed me of a form ⟨ѣльмя̀⟩ in an 18ᵗʰ century source, which would imply */ĕlmä/. As the Volga Turkic vowel shift can be dated to the 15ᵗʰ or 16ᵗʰ centuries (Doerfer 1971: 329), even a Middle Russian origin could be considered, although in view of the lack of parallels, we may be inclined to date the loanword earlier. The reality of linguistic contact between Turks and early Slavs is 4 It is interesting to consider the possibility of a loan from early West Germanic, however. West Germanic *ŏ may well in some cases have been adopted as *u in Slavic, cf. OCS хлъмъ ~ OS (Heliand) holm* ‘hill’, where attested Gothic uses a different word for ‘hill’, hlain(s)*. There are no certain examples of the development WG *e → Slavic *i (trad. *ь; the word for ‘radish’? cf. Sabaliauskas 1960b: 258), but such a substitution might be anticipated. In this case, one could envisage an earlier Slavic loan from West Germanic *elma- (cf. OHG elm). For the insertion of *i (trad. *ь) after *l as a reflection of the Germanic ‘clear l’, compare OR Ольга < ON Helga. 5 Pace ESJS 448, there is no reason to reconstruct a Proto-Slavic variant *lima- (trad. *lьmъ). The loss of *i- (trad. *jь-) is semi-regular in West and South-East Slavic (Derksen 2003; the ‘Russian’ form лём cited in ESJS is in fact Belarusian, cf. СРНГ xvi: 346). The other forms, Pl. dial. lim (Sł. Warsz. ii: 743), R dial. (Siberia) лим (СРНГ xvii: 47; ?cf. и́ лим, Даль² ii: 39), Sln. lom (Cigale 1860: 1306), do not show regular reflexes of *lima- (trad. *lьmъ) and must be explained otherwise. The CS form льмъ (Bes.), found twice on a single page (cf. SJS iv: 636), is evidently a scribal error for the Latin loan оульмъ ‘elm’, which is attested only in this text, and was apparently unfamiliar to the copyist. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 236 chapter 7 proven by the existence of common Slavic borrowings from Turkic, cf. R arch. толма́ чь, Cz. obs. tlumač ‘interpreter’6 ← Turkic *tïlmač (> Old Turkic tilmač, Tatar tïlmač ‘interpreter’, Yakut tïlbās ‘translation’; cf. REW iii: 115–116; Doerfer 1965: 662–665; ЭСТЯ iii: 233–235). Moreover, a couple more early Slavic loanwords have apparently found their way into Volga Turkic, most strikingly Tat. dial. könǯälä, Chuv. kĕnčele ~ dial. kănčala7 ‘flax prepared for spinning’ ← Sl. *kǭželi- (trad. *kǫželь; > R dial. ку́ жель, Bg. dial. къ̀жел ‘flax prepared for spinning’, Cz. kužel ‘distaff’). From a phonological and geographical point of view, however, it seems difficult to derive the Caucasian Turkic forms directly from Slavic. Although the exact source of the Turkic words remains elusive, it is more probable that these are ultimately of Indo-European origin rather than representing independent witnesses of a non-IE Wanderwort. ? ‘sturgeon’. R осётр, Cz. jeseter, SCr. jèsetra; Lt. erškẽtas; Pr. E esketres ‘sturgeon’ ~ Ic. styrja, OHG sturio ‘sturgeon’ (Oštir 1930: 6; Machek 1950b: 150; Bezlaj i: 228; Kroonen 2012: 240, 2013: 488; Iversen/Kroonen 2017: 513) — Both Baltic and Slavic point unambiguously to an initial *e- (with regular development to o- in East Slavic, pace Andersen 1996a: 147). Despite the difficulties with interpreting initial vowels in Balto-Slavic (see 7.2), it might still be wise to keep this example apart from other examples of the *a-prefix. Kroonen (2013: 488), who reconstructs *asetr- for Balto-Slavic against *str- in Germanic, would interpret the Germanic u-vocalism as resulting from a ‘reduced stem’ with a vocalic *r̥ . As Kroonen notes, based on the other examples of a-prefixation, we should expect *a-str- beside *setr-. He argues that the original “ablaut” may have been “reshuffled”, although since we do not have a clear understanding of the mechanisms underlying the alternations in the attested material, such an interpretation is risky; Šorgo (2020: 449–450) rejects the example altogether. An alternative non-IE analysis would be possible in the context of the *e ∞ *u alternation seen in *klen- ~ *klun- ‘maple’ (see 7.3.2), although an additional metathesis of *r would have to be assumed. Above (see pp. 218–219), I have argued in favour of a comparison of the BaltoSlavic words with Lat. excetra ‘sea serpent’. If this is accepted, the comparison ‣ 6 Czech-Slovak shows a strange ‘ablaut’ between tlumač ‘interpreter’ : tlumočiti ‘interpret’. This opens up the possibility that Slavic originally had a short vowel in the second syllable (in agreement with Turkic), preserved here in the derived verb, while in the noun it was secondarily assimilated to the agent noun suffix *-āče- (trad. *-ačь; on which see Vaillant 1974: 321–323). 7 This dialect form shows the expected Chuvash reflex with the development *küN- > *kuN(as in Chuv. kăn = Turk. gün ‘day’; kămpa = Tat. gömbä ‘mushroom’). In this light, the more common front-vocalic form is perhaps loaned from Tatar. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 vocalism 237 with Germanic *stur- becomes rather more impressionistic, even though the semantic correspondence with Germanic is perfect. If we follow Kroonen and reconstruct a ‘reduced stem’ without the initial vowel, we could set up a preform *ks(k)tr-. While this could well develop into *stur-, the fact that so much material has to be lost to achieve the Germanic forms makes the suggestion rather dubious. ? ‘turnip’. Lt. rópė, OHG ruoba, MDu. rove, Lat. rāpum ‘turnip’; Gr. ῥάφανος ‘cabbage, radish’; R рѣ́па, SCr. rȅpa (dial. rȉpa) ‘turnip’ ~ MW erfin, Bret. irvin ‘turnip’ (Oštir 1930: 64; Machek 1954: 57; Walde/Hofmann ii [1954]: 418; Furnée 1972: 163; Čop 1973: 29; Chantraine DELG iv [1977]: 968; Bezlaj iii [1995]: 171; Kroonen 2013: 415; Pronk/Pronk-Tiethoff 2018: 301) — Lt. rópė along with the Germanic and Latin forms support a common reconstruction *reh₂p-. Nevertheless, the irregularities in the other cognates clearly point to a foreign origin, which has long been recognized. These irregularities include (1) the vocalic alternation between *-ă- in Greek, *-ē- or *-ai- in Slavic, and *-ā- elsewhere; (2) the mismatch of Greek -φ-, Celtic *-b- as against *-p- elsewhere; (3) the correspondence Gr. ῥ- ~ MDu. r-, in IE terms suggesting an initial *r-, which is not typical of inherited vocabulary (e.g. Lehmann 1951). If the Celtic forms are to be segmented *a-rb-, then they may reflect a ‘prefixed’ variant of the pan-European word for ‘turnip’. Despite the close formal parallel with the other cases of ‘stem reduction’, I still find it difficult to entirely rule out chance resemblance, given that very little material (i.e. *-rP-) is being compared here.8 Considering the broad geography, we must in any case be dealing with a cultural Wanderwort. In view of the narrow distribution of the a-prefixed variant, it seems more probable that it was formed locally on the basis of material loaned from elsewhere. In this case, perhaps this word could provide an indication of the productivity of a-prefixation among the pre-IndoEuropean languages of Northern Europe. † ‘heron, stork’. SCr. róda ‘white stork’ (uncommon in the dialects; Skok iii: 153) ~ Gr. ἐρῳδιός (since Homer) ‘heron, egret’, Lat. ardea ‘heron’; ?ON arta (attested in Þul Fugla; for the meaning, cf. Ic. urt ‘teal’, Sw. årta ‘garganey’) (Beekes 2000a: 27; Лигорио 2012; Kroonen 2013: 36; Iversen & Kroonen 2017: 518; Pronk 2019a: 154; Matasović 2020: 339) — The appurtenance of the Norse word seems doubtful to me for semantic reasons. The comparison between the others is obviously attractive. It should be noted, however, that Latin ardea ‣ ‣ 8 In addition, Celtic rarely shows a prefixed form in such alternations. In this respect, note Schrijver’s comparison of Gaul. alauda (apud Pliny et al.; see TLL for attestations), Old French aloe and OE lāurice, lāwerce ‘lark’, but I must admit that I am not entirely convinced by this equation (cf. Matasović 2020: 340). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 238 chapter 7 could be syncopated < *arVdejā- (cf. Walde/Pokorny i: 146–147). In this case, the comparison with Greek need not be considered an unambiguous example of ‘stem reduction’, although the correlation between Latin a- and Greek ἐremains irregular. The supposition of an *a- ∞ *∅- alternation, on the other hand, depends entirely on the Serbo-Croatian form.9 Schrijver (1991: 65) has suggested that SCr. róda is a Romance loanword. Лигорио (2012: 23–25) rightfully criticizes this theory, since (1) Romance *-dja would give SCr. -ža (cf. lòpīž ‘earthen cooking pot’ < ⁽*⁾lapideum, FEW v: 160; M. Matasović 2011: 165); (2) the supposed Romance word *arda is nowhere attested; (3) the expected reflex of *arC- in South Slavic would be *raC-. At the same time, the word’s isolation certainly does give us cause for doubt (Matasović 2020: 339). In view of the widespread European and Slavic association of the stork with childbirth, one might, for instance, envisage a connection with SCr. ròditi ‘give birth’ (cf. Liewehr 1954: 90). Due to the doubtful Proto-Slavic status of this word, this example is too uncertain to be used here. 7.1.2 *a- ∞ *∅- with No Stem Reduction ‘nut’. Pr. E buccareisis ‘beechnut’; Lt. ríešutas,10 Lv. riẽksts ‘nut’ ~ R орѣ́х, Sln. órẹh ‘nut’ (Fraenkel 1950b: 238; Polák 1955: 55; LEW [1965]: 731; Bezlaj ii [1982]: 253; Matasović 2013: 93; Blažek 2018: 5–6) — Practically everyone who has accepted a non-IE origin has adduced additional forms such as Gr. κάρυον ‘nut’, ἄρυα (H.) ‘walnuts’, Alb. arrë ‘walnut’ in support of this claim. I am not optimistic about these comparisons; the only thing shared by ἄρυα and ríešutas is the phoneme *r. Nevertheless, even without this evidence, the inexplicable initial vowel in Slavic allows us to make a fairly convincing case for foreign origin. It cannot contain the verbal prefix *ab- (trad. *ob-) ‘around’ because *-bwould not have been lost before *-r- (cf. OCS обрѣсти ‘find, devise’).11 If we ‣ 9 10 11 The hapax ῥῳδιός in Hipponax is probably due to aphaeresis, cf. μάσθλης (Hippocrates) ‘leather’ < ἱμάσθλη, and other examples in Strömberg (1944: 44–45), as well as θέλω < ἐθέλω ‘want, wish’. The oft-cited form riešas seems to derive from Miežinis (1894: 206), who has ⟨rieszas, rieszutas⟩. It seems doubtful that this is a genuine dialectal variant, and if genuine, it is probably a back formation. The suffix -utas is rare, but it is shared by ãšutas ‘horse hair’ and degùtas ‘tar’. Note that both ãšutas and ríešutas decline as a consonant stem in East Vilniškiai dialects, as has been repeatedly pointed out (cf. Būga apud Trautmann 1923: 241; Specht 1947: 62; Fraenkel LEW 731; Ambrazas 1993: 57; Derksen 2015: 328), but this is of little value, as consonant stems became productive in this region (Zinkevičius 1966: 263). This phonological issue is not even mentioned by ЭССЯ (xxix: 71), who list the word under *obrěxъ, and consider it (following Ильинский 1916: 153; Трубачев 1971: 65) a derivative of *rěšiti ‘to untie’. Note the criticism on this point by Крысько (2014: 104). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 vocalism 239 assume an ablauting *Hor- : *Hr- (Specht 1947: 62), this would imply that the element *-ois- is a suffix, of which there is no indication.12 A non-IE origin might also be favoured by the root structure (virtual *(H)roiHs-; see 5.3.2). ? ‘oats’. Pr. E wyse ‘oats’ ~ Lt. ãvižos, Lv. àuzas; R овёс, Sln. óvəs; Lat. avēna ‘oats’ (Pronk/Pronk-Tiethoff 2018: 294–295; Kroonen et al. 2022: 19–20; for more refs. and discussion, see pp. 212–213) — Whether this example belongs here depends heavily on the analysis of the Prussian form, the only one in which initial *a- is lacking. The interpretation of the Prussian data is unfortunately not straightforward. A similar word for ‘oats’ is attested in Simon Grunau’s Prussian vocabulary: wisge. While we might be tempted to read /wizje/, the word must rather be identified with wìzges in Daukantas (Leskien 1891: 274; see LKŽ s.v. vizgė for additional data). On the other hand, wyse in the Elbing Vocabulary cannot be corrected to *wysge, as its reality is confirmed by the grain tax craysewisse found in 15ᵗʰ century East Prussian documents.13 The result is that we have little choice but to accept the existence of two dialectal synonyms for ‘oats’ in Prussian and Žemaitian. In light of this, it is plausible that the two forms influenced one another, and that pre-Prussian *awizē lost its initial *a- due to the influence of wisge (Pronk/Pronk-Tiethoff 2018: 294–295). The latter may be cognate with Latin virga ‘shoot, twig, rod’, OHG wisc · [ faeni] strues ‘bundle [of hay]’ (LEW 1269). A further potential piece of evidence for a form without an initial vowel can be found within Uralic. Although the comparison of Lat. *awe(T)sna ‘oats’ with Karelian and East Finnish vehnä ‘wheat’ would seem almost obvious, I am not aware of any relationship having been proposed before.14 The Karelian word cannot be separated from Md. EM viš, and Ma. E wiste, wiśte, NW wištə, meaning ‘spelt’. Although the Mari sibilant is somewhat unexpected, the reflexes ‣ 12 13 14 Other Slavic examples are difficult to adduce here, as *ab- (trad. *ob-) can usually not be excluded on formal grounds. A case in point is OCS оскръдъ, Pl. oskard ‘pickaxe’ as against Pr. E sturdis ?*/skurdəs/ · bicke, where the Slavic prefix does not appear semantically motivated, but on formal grounds could represent *ab-. Moreover, one is reminded of CS омлатъ beside OCS млатъ ‘hammer’ (cf. Mikkola 1898: 302). I leave such examples aside. e.g. “schessel habir von czenden, phlughabir und craysewisse” 1431. For the data, see Töppen (1867: 151–152), who convincingly interprets the word as ‘Heuhafer’ (cf. Pr. E crays · hew). I have now published the suggestion in Kroonen et al. (2022: 20). In addition, note that R. Matasović independently offered the same comparison during the workshop Sub-IndoEuropean Europe in August 2021. Koivulehto (2002: 592) has suggested a far less attractive loan etymology starting from IE *ḱueitnó- ‘white’, a back-projection of Germanic *hwīta‘white’, in turn the source of *hwaitja- ‘wheat’. Not only is the back-projection of this Germanic form to IE unwarranted, the suggested source also leaves the Uralic *š unaccounted for. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 240 chapter 7 seem largely consistent with a reconstruction *šn.15 However, the vocalism in Mari is irregular, and in Uralic terms would rather suggest *wäšnä (cf. Aikio 2014a: 157 and passim). In this context, the very narrow distribution within Finnic seems to almost call for a loanword origin (cf. von Hertzen 1973: 90; Häkkinen/Lempiäinen 1996: 159). In combination, all of these facts solidify the impression of a word intrusive to Uralic. Although there is a semantic discrepancy between ‘oats’ in Indo-European and ‘wheat, spelt’ in Uralic, it nevertheless seems quite possible that all of these forms derive from an earlier agricultural Wanderwort. If these words belong here, they would be a clear example of a form without initial a-. However, it is by no means clear that we are dealing with a morpheme *a- or some other phenomenon, such as aphaeresis, which might be a symptom of the borrowing process resulting from the more restrictive phonotactics of Uralic. ? ‘sedge’. OE secg, Du. zegge, OHG sahar ‘sedge’; OIr. seisc, MW hesc coll. ‘sedge’16 ~ R осо́ ка, Uk. осока́ ‘sedge’ (Kroonen 2013: 421; Iversen/Kroonen 2017: 518) — The relationship between Germanic *sok- and Celtic *se-sk- exactly parallels the semantically similar Lt. néndrė ‘reed’ (< *ne-nd-) beside Hitt. nāta‘reed’, MP nʾy ‘pipe, flute’ (< *nod-; de Bernardo Stempel 1999: 68).17 These words therefore permit an analysis in terms of Indo-European morphology. The Slavic word formally corresponds to Lt. ãšaka ‘husk of grain; small fish bone’, which is ‣ 15 16 17 Compare Võ. pähn, Md. M pä(k)šä, Ma. E pište ‘lime tree’ (< PU *pä(k)šnä) and F dial. hähnä, Md. M dial. šäkši and Ma. E šište ‘woodpecker’ (?< *šä(k)ćnä). It seems possible that the *k in these Mordvin words is secondary (P. Kallio apud Holopainen 2019: 249, but note the different reflex in the word for ‘spelt’), which would be supported by the Mari reflex of *pä(k)šnä, where we fail to find the otherwise regular development *ä > *ü before *kš (cf. Aikio 2014a: 155). Traditionally (e.g. UEW 716), the word for ‘lime tree’ is reconstructed as *päkšnä with a three-consonant cluster. Note that UEW (p. 772) rejects the appurtenance of šäkši altogether. Perhaps as a result, the Mordvin word for ‘woodpecker’ is not even mentioned by Aikio (2015a: 44; in prep. 108–109), but it seems that it must belong here and that we should return to UEW’s reconstruction *šäćnä / *ćäšnä? with various assimilations (as opposed to Aikio’s preferred ćäćnä). The Celtic forms have been adopted into Romance, cf. Occitan sesca ‘bulrush’, while IberoRomance (Spanish and Catalan sisca, xisca ‘reed’) suggests a divergent preform with *ī. Coromines/Pascual (v: 264) attempt to solve this by assuming a borrowing through Mosarabian (see also FEW xi: 551). Reduplication is also shown by OIr. nenaid as against Lt. notrė ̃ ‘nettle’ (see p. 203). Compare similarly Lt. papar̃tis (dial. papártis) ‘fern’ against MIr. raith ‘ferns, bracken’ < *prH-ti(Schrijver 1995a: 178; Zair 2012: 76), if this does not contain the prefix pa- (Gliwa 2009: 82). Typologically speaking, one may assume that the reduplication seen in these plant names had a collective function, although there is no actual evidence for this (awkwardly, the unreduplicated raith is in fact a collective, but this fact is not decisive due to the productivity of collectives in Celtic). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 vocalism 241 generally derived from the root of aštrùs ‘sharp’ (cf. ME i: 142 with lit.). Such a derivation makes sense for ‘sedge’, which has sharp leaves. On the other hand, the suffix *-akā has no close parallels within Slavic (Vaillant 1974: 543; though cf. Варбот 1984: 167), which remains an argument against an internal derivation (the suffix is also rare in Baltic, cf. Skardžius 1941: 125–126). As a result, although a non-IE origin seems possible, this example remains uncertain. ? ‘grouse’. Lt. jerubė ̃ ‘hazel grouse’, Lv. ir̃be ‘partridge’; Pl. jarząbek, Sln. jerę̑b ‘hazel grouse’ ~ Lv. rubenis ‘black grouse’; OHG reba-huon, MLG rap-hōn ‘partridge’ (Derksen 2000; see pp. 174–176 for more discussion) — The Baltic evidence implies an initial *je- (or possibly *ẹ̄-; see p. 175), but Lv. rubenis ‘black grouse’ lacks the initial syllable altogether. At first sight, the latter is a close match to ON rjúpa, Nw. rype ‘ptarmigan’ (Walde/Pokorny ii: 360; Derksen 2000). Jóhannesson (1942: 223) has called the appurtenance of the Norse words into question, however, deriving them instead from the root of Ic. ropa ‘belch; brag’, synchronically used to describe the sound the ptarmigan makes during the breeding season (see also IEW 871; Kroonen 2013: 411). This explanation chimes well with the alternative etymology deriving Lv. rubenis from rubinât ‘kollern, falzen (von Birkhähnen)’ (ME iii: 552). On the other hand, it is reasonable to suspect that rubinât is denominal in origin (LEW 744 refers us to Fraenkel 1937: 362, where the parallel Cz. dial. křepeliti (Kott vi: 727) ‘twitter (of a quail)’ < křepel ‘quail’ is cited; cf. Derksen 2000: 81). This would be supported by the verb’s isolation within Baltic.18 In Germanic, the verb is unlikely to be denominal, as the primary meaning of Ic. ropa appears to be ‘belch’, cf. MDu. ruppen in the same sense, and the derived OHG ropfezzen, MDu. op-ruspen. The Germanic word is presumably of imitative origin (Kroonen 2013: 411). Even if the position of ON rjúpa ‘ptarmigan’ remains unclear, it still seems attractive to compare the Balto-Slavic data with the West Germanic words for ‘partridge’ (Derksen 2000: 77, 79). While MLG rap-hōn can be explained as a folk etymological distortion after rap ‘fast, agile’ (DWb xiv: 334; Suolahti 1909: 256; Kluge/Seebold 2011: 750), OHG reba-huon ‘partridge’ does not have a convincing ulterior etymology (Suolahti mentions the call: ‘zirrep’). The comparison with the Balto-Slavic forms presupposes (a) the irregular loss of the first syllable *i̯e- and (b) an irregular vocalic relationship *e ∞ *u. On the latter, see 7.3.2. ‣ 18 Its supposed Lithuanian cognate, rubėti, is apparently attested just once in a list of Švėkšna dialect words sent into the newspaper Viltis by a K. Jazdauskis: “Rubėti, brazdėti, grumėti, bildėti — “sinonimai”” (see Viltis 1908, No. 114 [1 Oct.], p. 3). Rubėti (if not simply a printing error!) is evidently a secondary variant of the synonymous rabėt́ i (LKŽ). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 242 chapter 7 7.1.3 *wi-? ‘bison’. Pr. E ⟨wissambs’⟩ · ewer; OE wesend, weosend, OHG wisunt ‘bison’ ~ R зубр, Pl. obs. ząbr, Sln. zǫ̑ bər; Lt. stum̃ bras ‘bison’ (for refs. and further discussion, see pp. 220–221) — The element wi- in Prussian and Germanic is unexplained and difficult to account for without recourse to ad hoc contaminations (cf. LEW 932 with lit.).19 Compounded with the other irregularities (pp. 220–221), a non-inherited origin looks probable. As a potential, although speculative, parallel for the prefix *wi-, Kroonen (2013: 457; cf. also Šorgo 2020: 449; van Sluis forthc.) has adduced the Gaulish uisumarus · trifolium (apud Marcellus of Bordeaux; Delamarre 2003: 322–333), as opposed to MoIr. seamair, Ic. smæra ‘clover’ (this suggestion in fact goes back to Oštir 1930: 26). Kroonen (2012: 254, 2013: 571) has also suggested that the initial i- in R изю́ брь ‘Manchurian wapiti’ might be identified with the *wi- in Prussian and Germanic. While at first sight attractive, this suggestion is probably to be rejected. First, it is suspicious that the given word is limited to Russian, and that the species it refers to is only present to the east of Lake Baikal.20 It can be noted that there are a couple of other words in Siberian dialects which show an epenthetic /i-/ before /z-/, cf. Siberian dial. иза́ боль ‘indeed’ (СРНГ xii: 84) = dial. за́ быль (Аникин 2003: 201) and изу́ фрь = MR зуфь, зуфрь (СРЯ 11–17 vi: 70) ‘a kind of woollen fabric’, of Turkic origin (cf. Turk. sof ‘woollen fabric’; see Аникин 2000: 215, 220). This is probably to be explained by assuming the interference of a substrate in which initial /z-/ is not permitted, cf. Khakas izep ‘pocket’ ← R dial. зепь (itself borrowed through a Turkic language, ultimately from Arabic, cf. Räsänen 1969: 124; Аникин 2003: 216), Yakut dial. ïhïr̄ ‘fat’ ← R жир (Аникин 2003: 199).21 Without the support of изю́ брь, it is also difficult to assess whether this initial *wi- can have anything to do with the initial /i/ in R и́ волга ‘oriole’ (see p. 179 for a discussion of this element). ‣ 19 20 21 For example, Petersson (1921: 39–41) assumes the Prussian form has wi- after German Wisent, and connects the latter with Skt. viṣā́ṇa- ‘horn’ (thus also van der Meulen apud Derksen 2015: 433). Ильинский (1926: 56) assumes instead that the Prussian word itself is cognate with Skt. viṣā́ṇa-, with a second element *bʰr- ‘bearing’. Young (1998: 204–205) sees the element *wis- in OHG wisa ‘meadow’. The word’s earliest attestation is in the derivative изубрина (1495, СРЯ 11–17 vi: 209) in a report from a Moscow delegation about a mission to the Grand Dutchy of Lithuania. Logically, the meaning must be ‘bison meat’, but the specific context (“три бочки изубрины”) makes it tempting to assume a transmission error. An epenthetic initial i- is also found in some Turkic loanwords already in Middle Russian: изарбавъ (17ᵗʰ c.; СРЯ 11–17 vi: 92) ‘brocade’ (~ Ottoman Turkish zerbaf ), изумрутъ (15ᵗʰ c.; idem: 212) ‘emerald’ (~ Turk. zümrüt), which might suggest transmission through a similar substrate. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 243 vocalism table 14 Possible examples of ‘prefixation’ Baltic ‘swan (1)’ – ‘turnip’ *râp- ? ‘sturgeon’ *e-(ḱ)sket-r- Slavic *a-lbandʰ*lebʰedʰ Celt. *a-rb- *e-(ḱ)set-r- *(k)st(u)r- Lat. *e-ksket-r- *e-lm- Tur. *e-lm- *a-lm- Celt. *lē̆m- *i-lim- ‘nut’ *rais- *a-rais- *u̯ iź- *a-lbʰed*rāp- – ? ‘oats’ – *a-uiS- – ? ‘sedge’ – *a-sak- *sak-i̯- ? ‘grouse’ *ie-rubʰ- *(i)e-rembʰ- *rebʰ- *zam(bʰ)r- *u̯ i-(t)snT- *u̯ -epr-i̯- *epr̥- bison ? ‘boar’ *u̯ i-sam(bʰ)r*stum(bʰ)r? *u̯ -epr-i̯- Elsewhere *rêp-/*raip- ? ‘elm (2)’ *a-uiź- Germanic Lat. *a-wesnUral. *wešnä Celt. *se-sk- Lat. *aper- ‘boar’. OCS (Ps. Sin.) вепрь ‘boar’; Lv. vepris ‘castrated boar’ ~ Lat. aper, ‣OE? eofor, OHG ebur* (attested epur, eber) ‘wild boar’ (Machek 1968: 684; Polomé 1990: 335; Kroonen 2013: 114; Šorgo 2020: 438) — The comparison of these forms is obvious, although an explanation of the initial *w- in BaltoSlavic is lacking (Walde/Hofmann i: 56; Kluge/Seebold: 226; Derksen 2008: 515). Perhaps this *w- can be identified with the element *wi- found in the word for ‘bison’, discussed above. True, the distribution of this ‘prefix’ in the two Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 244 chapter 7 examples is almost diametrically opposite, and without further examples, it is difficult to draw any solid conclusions. 7.1.3.1 Conclusion The result of this section is a rather heterogenous group of mostly uncertain examples, which are collected in Table 14, above (see p. 185 for help reading the table). Shaded cells indicate forms containing a ‘prefix’. In view of the rather different distributions and behaviours of these words, it is unlikely that they represent a single phenomenon. Most reminiscent of Schrijver’s ‘a-prefix’ is the word for ‘swan’, which meets three criteria: (a) an alternation *a- ∞ *∅-, (b) an apparent ‘stem reduction’ and (c) a geographical distribution similar to that of the European words for ‘blackbird’ and ‘ore’. The word for ‘turnip’ also appears to fit this pattern quite well, although it is more widespread, and must have partially spread as a cultural Wanderwort. A similar phenomenon has also been proposed to occur in the words for ‘elm’ and ‘sturgeon’, but both of these involve a number of issues and cannot be considered entirely certain. The remaining cases do show an initial vowel in some of the continuants, but do not show the expected pattern of ‘stem reduction’. While it cannot be ruled out that such alternations derive from a related substrate alternation, it is difficult to rule out that they result from an unrelated phenomenon, such as aphaeresis, either resulting from the borrowing process or taking place within the source language. 7.2 Alternations between Front and Back Vowels Although there are some words in Balto-Slavic which appear to show an unclear alternation between *a- and *e- in initial position (see, for instance, pp. 276– 277 on the word for ‘alder’), the value of this alternation is unclear as a result of ‘Rozwadowski’s change’ (cf. Rozwadowski 1915; Andersen 1996a: 102–104 and passim; Derksen 2002) — the observation that Balto-Slavic *e- sometimes occurs in place of *a- under as of yet unclear conditions. Since the development is also found in inherited words, such as in Lt. erẽlis ‘eagle’ < *h₃er- (cf. Gr. ὄρνις ‘bird’, Hitt. hāran- ‘eagle’), I agree with Andersen (1996a: 105) that little is gained by invoking non-IE substrates. This applies to examples such as Lv. dial. (Kurzeme) ẽrcis ~ Gr. ἄρκευθος ‘juniper’ (Beekes 2000a: 27; Derksen 2015: 533–534), even though the unclear Greek suffix quite possibly suggests at least the Greek word is of non-IE origin (cf. also κέλευθος ‘way, journey’, Chantraine 1933: 366–367). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 vocalism 245 In other positions, the evidence may also be ambiguous. Cases like Lt. vãsara ‘summer’ ~ OCS (Ps. Sin.) весна ‘spring’ (= Gr. ἔαρ ‘spring’) and Lt. vãkaras, Lv. vakars ~ OCS вечеръ ‘evening’ must result from a combination of assimilation and neo-ablaut (Otrębski 1955: 24–26; Hamp 1970b). One may in principle suggest a similar explanation for examples such as Lt. sidãbras, Pr. iii sirablan ~ OCS сьребро, even though, in this case, there is plenty of other evidence for a non-IE origin (see pp. 225–226). In the following examples, the irregular vocalism is supported by examples beyond just Baltic. 7.2.1 Baltic *e ∞ Elsewhere *a ‘hellebore’. RCS чемерь ‘hellebore, hemlock; poison’; Bel. чэ́мер ‘white hellebore’, Mac. dial. чемер ‘Veratrum lobelianum’ (cf. ЭССЯ ix: 52–53);22 OHG hemera ‘hellebore’ ~ Gr. κάμαρος, κάμμαρος ‘a poisonous plant: ?aconite, ?larkspur’ (Furnée 1972: 343; Huld 1990: 405–406; Beekes 2000a: 28; Kroonen 2013: 219; Derksen 2015: 236; ERHJ i [2016]: 125) — The Greek vocalism and geminate -μ- suggest a foreign origin. Šorgo (2020: 440) has doubted the appurtenance of the Greek word due to its different meaning. However, since it also refers to poisonous meadow plants, this doubt hardly seems unjustified. The underived word in Slavic has undergone various semantic shifts connected to the plant’s poison — Bg. dial. чѐмер ‘distress; demon’, SCr. čȅmēr ‘bitterness; distress; venom’, Slk. čemer ‘a kind of disease associated with blood clots’ — while the botanical sense has been recharacterized with suffixes: R чемери́ ца, Pl. (dated) ciemierzyca; Sln. čmeríka, Bg. чемерѝ ка. The usual sense appears to be Veratrum (‘white’ or ‘false’ hellebore). Similarly, the sense Veratrum album is recorded in German for the Carinthian dialectal form hammer (Grimm DWb x: 316), matching the gloss of hemern, hemer-wurz in the 18ᵗʰ c. Polyglotten-Lexicon der Naturgeschichte (op. cit. 983).23 ‣ 22 23 As I argue in detail elsewhere (Jakob forthc. b.), Lt. kẽmeras ‘hemp-agrimony’, although belonging to the modern standard language, originated as a ghost word. It was the result of Nesselmann’s misinterpretation of the form Kiemerai. Alpen (ClG i: 73) as ‘Alpkraut’. As Szyrwid translates the same word as mára, incubus ephi[a]ltes (see ALEW² s.v. kiemerai), Alpen is clearly to be understood as the plural of Alp ‘daemon, incubus’. Nesselmann’s Kėmerai was misinterpreted by Kurschat (1883: 177) as kemerai, whence it found its way into botanical reference literature and finally into the standard language. The use of the word with reference to true Hellebores (Helleborus sp.; cf. Marzell iv: 1016) is perhaps due to the influence of classical nomenclature. For instance, the 13ᵗʰ c. Breslau Arzneibuch distinguishes the white and black hemern (MWb s.v. hemere), an obvious calque on Lat. helleborus albus and niger. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 246 chapter 7 ‘ramsons’. Lt. obs. kermušis ‘wild garlic’ (ClG i: 1088, see ALEW 553; Ruhig i: ‣59);24 OCS чрѣмошъ acc.sg. (Rosenschon 1993: 150), R черемша́ , Pl. trzem- ucha, SCr. dial. (Lika) crìjemuša ‘ramsons’ ~ OE hramsa, MLG ramese (MoLG Rāmsche, Marzell i: 211); MW craf ‘ramsons’ (Machek 1950b: 158; Beekes 2000a: 29; Matasović 2009: 222, 2013: 89) — Further, with o-vocalism, Gr. κρόμμυον ‘onion’. The Greek geminate is difficult to explain from an Indo-European perspective (Chantraine DELG ii: 586; the Epic variant κρόμυον may be metrical; LSJ s.v.) and itself already points to foreign origin. As a result, the reconstruction of an ablauting u-stem *ḱremH-u- : *ḱrmH-eu- (Matasović 2005: 369; and already Hamp 1965: 232) is beside the point. The original Slavic form is difficult to reconstruct: while R черемша́ suggests an underlying *čermušā- (trad. *čermъša), the OCS hapax чрѣмошъ — which appears to match Sln. črę̑moš (Pleteršnik i: 109) and SCr. dial. cremoš (also widespread in toponymy, cf. Skok i: 273) — is not consistent with a medial *u (trad. *ъ; despite Sławski SP i: 154; ЭССЯ iv: 68). The word admittedly appears in a whole host of corrupt forms in South Slavic (cf. e.g. SCr. srȉjemuž,25 Sln. čẹ́maž (SSKJ²) ‘ramsons’), but if the South Slavic form *čermaše- (trad. *čermošь) is old, then the suffix syllable does not match that of Lt. kermušis. The evidential value of this alternation is of course low, but there are other irregularities which make a hypothesis of non-IE origin probable. Chantraine (DELG ii: 586) explains the variant κρέμυον, attested in Hesychius, along with MoGr. κρεμμύδι ‘onion’ as the result of a labial dissimilation *o–u > *e–u. Some similar cases are indeed known from Modern Greek (cf. αλεπού ‘fox’ < MGr. ἀλωπού; cf. Holton et al. 2020: 68), but the development is by no means regular. MIr. crem, also crim, ‘wild garlic’ has also been interpreted as secondary for *kramu- (> MW craf ), like OIr. tel beside taul ‘forehead; boss of a shield’ (Thurneysen 1946: 52; Bernardo Stempel 1987: 101). In favour of this, one has cited the personal name Craumthann, which is a rare variant of Crimthann (see eDIL s.v.). The absence of spellings in -au- or -u- for the plant name itself is 24 25 The form *kermùšė (Kurschat 1883: 178, marked as an unfamiliar word; also the citation form in Trautmann 1923: 128; LEW 243; Derksen 2015: 239–240; etc.) is dubious and seems to derive ultimately from Mielcke (1800 i: 116) who has Kermußės f. wilder Knoblauch, apparently miscopied from Ruhig (the German–Lithuanian part still has Kermußis in this sense, cf. idem ii: 303). The dial. kermušė,̃ kérmušas (Juška iii: 85) ‘tip of a drill’, as shown by kiáurmušis grąžtas, literally ‘through-beating drill’ (Gegrėnai, LKŽ), is an unrelated compound consisting of mùšti ‘beat’ (cf. LEW 243) and kiáurai ‘through’. For the phonological development, compare dial. kel̃tvartis Veliuona, Seredžius < kiaũl-tvartis ‘pig sty’ (Skardžius 1941: 427). ЭССЯ (iv: 68) and ALEW (1175) seem certainly correct to dismiss this form as evidence for a Proto-Slavic variant with *s-. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 vocalism 247 disturbing, but as the word is not attested particularly well or early, this might be put down to chance. According to Stifter (1998: 227, fn. 2), such fluctuation in vocalism is ultimately the result of the i-mutation of a second-syllable *u. As such an explanation does not appear to be viable here, whether crem can be accounted for in this way remains uncertain.26 While the explanations in both cases are admittedly shaky, it must be acknowledged that the evidence for a stem *kremu- outside of Balto-Slavic is of an uneven and marginal nature, and we may tentatively operate with an original distribution of *e in Balto-Slavic, *a in Celtic (and Germanic) and *o in Greek. Purely on the basis of the formal similarity, it hardly seems possible to separate Uk. чере́мха, Cz. střemcha, Sln. črę̑msa, obs. čremha (i.e. ⟨zhremha⟩ in Jarnik) ‘bird cherry’.27 Berneker’s interpretation (i: 145) that the common factor is the strong smell (the bark of the bird cherry has a strong, acrid smell) has generally been followed by later authors (e.g. Walde/Pokorny i: 426; REW iii: 321; ALEW 553–554). In Baltic, one finds Lt. šermùkšnis, Lv. sȩ̄r̀ maûksis (~ sȩ̄r̀ mûkslis, etc.; ME i: 829–830) ‘rowan’, whose initial *š- has been attributed to Gutturalwechsel (e.g. LEW 243; Derksen 2015: 240).28 In view of the consistent meaning of the word outside of Balto-Slavic, the transference to a tree name must be considered secondary. It is therefore not of direct relevance to the word’s ultimate origin. ‘garlic’. R чесно́ к, Cz. česnek, Sln. čésən ‘garlic’ ~ OIr. cainnenn ‘garlic; leek’, OW cennin ‘leeks (coll.)’ (< *kasnīnā-) (Schrijver 1995b: 16–18; Derksen 2008: 86; Matasović 2009: 193, 2013: 89; van Sluis forthc.) — Schrijver challenged the old interpretation of the Slavic word as a derivative of the root for ‘scratch, ‣ 26 27 28 It is interesting, however, that eDIL (s.v. tul) cites a nom.sg. taul beside dat.pl. telaib ‘boss of a shield’ from the Middle Irish Lebor na hUidre. The latter form would actually contain an i-mutating factor in its ending *-bi, and one might wonder whether this could represent the original distribution of the variants. There is no benefit in treating R dial. (Tver’, Даль² iv: 610) чере́ма (= черёма, Даль³ iv: 1312) as the oldest variant (pace Sławski SP: 153; Matasović 2005: 369); this dialect variant is evidently back-formed from черёмуха or from dial. черёмка (e.g. СРГК v: 773). The latter is itself probably a corruption of черёмха after the diminutive suffix -ка; compare similarly Pl. dial. (Sławski op. cit.) trzemka, Cz. dial. (Machek 1968: 586) střenka. ЭССЯ (iv: 68) would rather see the Baltic words as the result of an assimilation *k–š > *š–š and Matasović (2005: 369–370) assumes contamination of two originally distinct words. Strangely, Matasović assumes that it is the word for ‘ramsons’ that had *ḱ-, which is precisely the opposite of what is found in Baltic. Note that the claim (ALEW 1175) that Lv. cȩ̄r̀ maûksis (with variants, ME i: 377–378, EH i: 268) ‘rowan’ agrees with Lt. kermušis is probably an illusion in view of the numerous other examples of secondary c- for *s- cited by Endzelīns (1905: 183–185, 1923: 130–131). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 248 chapter 7 comb’ (cf. R чеса́ ть; Miklosich 1886: 35; Meillet 1905: 453; REW iii: 330; ЭССЯ iv: 89–90), and attractively compared the cited Celtic material. He pointed to the mismatch in vocalism as evidence that we are dealing with parallel borrowings from a substrate language. Falileyev/Isaac (2003) have questioned Schrijver’s appeal to substrate, and argue that Celtic could reflect a preform *ksno- with a-epenthesis as in MW adar ‘birds (pl.)’ (< *ptarV-; op. cit. 8, fn. 25; see also Zair 2012: 185, fn. 27 with references to earlier literature). On the other hand, it is questionable that epenthesis should have occurred in a cluster *ksn-, especially in view of OIr. sine ‘teat, nipple’ (< *sfeni̯o-), which reflects a similar cluster *pst- (cf. YAv. fštāna‘breast’, IEW 990). Moreover, the semantic association with ‘scratch, comb’ is tenuous. Berneker (i: 151) adduced OHG kloba-louh* ‘garlic’ ~ klioban ‘split’ as a supposed parallel, but the Slavic forms do not mean ‘split’; occasional senses like ‘pluck (leaves, feathers)’ (RJA i: 946) are unusual and clearly secondary. The basic meaning is ‘comb’ ≫ ‘scratch’ (see also SCr. kòsa ‘hair’; Hitt. kis-ᶻⁱ ‘to comb’, IEW 585–586; Kloekhorst 2008: 481–482).29 7.2.2 Balto-Slavic *ā̆ ∞ *ē̆ Elsewhere [‘cottage’. Lt. trobà ‘peasant house; room’ ~ Oscan trííbúm acc.sg. ‘house’, OIr. treb ‘residence, estate’ (etc.) — See pp. 270–271.] [‘ground elder’. Lt. garšvà ~ OHG gires — See p. 277.] ‘honeycomb’. Lt. korỹs, Lv. kāre30 ‘honeycomb’ ~ Gr. κηρός ‘wax; honeycomb’, Lat. cēra ‘wax’ (Chantraine 1933: 371; Alessio 1946: 161–162; Ernout/Meillet [1951]: 114; Deroy 1956: 190; Pisani 1968: 19; Beekes 2010: 689–670; van Sluis 2022: 17–18; Kroonen forthc.)31 — The Baltic word can only be connected ‣ ‣ ‣ 29 30 31 Falileyev/Isaac (op. cit. 5–6) also adduce some forms from Uralic languages: Komi (Permjak) komiʒ́, (Jaźva) ku·mić, Udm. kumi̮ź ‘wild garlic’ (< Proto-Permic *ko̯ mi̮ʒ́, cf. Лыткин 1964: 47); Mansi (West) kɔśśm, (North) χōsman (< *kāšmā-) ‘onion’, Hungarian hagyma ‘onion’. These forms possibly reflect a common proto-form *kaćmə (cf. UEW 164–165; Живлов 2014: 130), although the non-canonical phonotactics (*-CR- cluster) make it improbable we are dealing with an inherited word in Uralic (Holopainen 2022: 106). One might be tempted to consider the Uralic words continuants of the same pre-European source word, but in view of the only approximate formal match and geographical distance, it is more probable that the similarity is coincidental. LVPPV has kàre, which ME (ii: 195) cite from Plāņi. This would be consistent with kâre² Dunika (EH i: 602) and High Latvian kàre². The latter, however, could also correspond to dial. kãres (ME loc. cit.). Establishing the original intonation is difficult. Adams (2013: 694) tentatively compares the Tocharian B hapax śeriye, but admits that his gloss ‘± wax, honeycomb’ is based entirely on the comparison with Greek. The word is attested in a list of medical ingredients in a broken context, and no translation is attempted in the recent critical edition by Tatsushi Tamai (2020). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 vocalism 249 with the Greek if the latter was borrowed from Ionic-Attic into the other dialects, and Lat. cēra from Greek (Boisacq 1916: 450; Walde/Hofmann i: 202), but this possibility is generally viewed with scepticism (Ernout/Meillet 114; Chantraine DELG ii: 527).32 The suffixation in the derivative κήρινθος ‘propolis’ is usually mentioned as a key indicator of the word’s non-Indo-European origin. A possible Wanderwort is supported by the unclear Volgaic comparanda, on which see p. 133. In addition to these, we may also note Estonian kärg (dial. käri, Saaremaa kärv) < *kärjeh33 ‘honeycomb’ (cf. Vaba 1990b: 176–177), another form with front vocalism. In view of the difference in vocalism, a direct loanword from Baltic appears unlikely (compare 3.5.2), and for geographical reasons, a borrowing from a Turkic or Volga Uralic language is also extremely improbable. In conclusion, it must be admitted that the exact routes of movement of this word remain quite unclear, but the mismatch in vocalism between Greek and Baltic may nevertheless be used to support the analysis as a non-IE loanword. ? ‘circle’. OCS крѫгъ ‘circle’ ~ ON hringr, MDu. rinc ‘circle, curve’ (Philippa et al. iii [2007]: 132 s.v. kring) — The Germanic and Slavic words could in principle be combined by assuming apophonic variants *e : *o. However, this lacks a clear motivation, and the implied root *krengʰ- would violate the IndoEuropean root constraints (Kroonen 2013: 247; see 5.3). Philippa et al. point to the Germanic variant *kringa- (ON kringr, usu. í/um kring ‘(all) around’, MHG krinc, kranc ‘circle, vicinity’, MDu. crinc ‘circle, curve’) and suggest the possibility that the word was borrowed from a substrate. However, it seems more attractive to interpret this variant as a Reimbildung based on the verbal root seen in MDu. cringhen ‘turn (back)’ (which is cognate with Lt. grę̃žti ‘turn; bore, drill’; Stang 1972: 24). A non-IE origin might still be supported by the *e ∞ *a alternation, but this is naturally rather meagre evidence. ? ‘people’. Lt. tautà, Lv. tàuta; Pr. E tauto ‘people’ ~ Go. þiuda, OE þēod, OHG diot, diota ‘nation, people’ (Beekes 1998: 461–463; de Vaan 2008: 618; Derksen 2010: 38, 2015: 461) — The Germanic forms point to *teut-, which could also account for Oscan touto ‘civitas’, OIr. túath, MW tut ‘people, country’. This reconstruction would be supported by Venetic (Làgole) teuta ‘civitas(?)’ ‣ ‣ 32 33 Although denying the possibility of Proto-Greek *ā, Frisk (i: 843–844) is still inclined to view the Latin word as a loan from Greek. There really seems to be no positive evidence for this (see already Osthoff 1901: 22) and the Lat. -a remains a potential obstacle to the loan etymology (Ernout/Meillet 114). An original käri, gen. kärje (which probably reflects *kärjeh, like puri gen. purje ‘sail’ < *purjeh, cf. p. 90) has undergone various analogical reshufflings. The standard form kärg is built after cases such as jälg, gen. jälje ‘trail, track’ = F jälki. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 250 chapter 7 (Lejeune 1974: 110–111) and the Gaulish *teuta reconstructed on the basis of onomastic evidence (Delamarre 2003: 295).34 The Baltic forms have often been interpreted as showing the regular reflex of *teut- (e.g. Brugmann 1897: 202; Stang 1966: 73). According to the formulation of Zubatý (1898: 396) and Endzelīns (1911: 82–83; cf. also Vaillant 1950: 123) the development *eu > *au was regular before a consonant if there was a back vowel in the following syllable, but the only decent example of this is precisely Lt. tautà (cf. Endzelīns 1911: 83).35 All other clear evidence points to the preservation of *euC in Baltic (Berneker 1899; Pedersen 1935; Kortlandt 1979a: 57; Derksen 2010: 37). If this rule is rejected, then one might assume apophonic variants (e.g. Endzelīns 1911: 82); however, nouns with e/o ablaut were rare in PIE (cf. recently Kloekhorst 2014: 151–161; van Beek 2018). Petit (2000: 143) has suggested that the unexpected Baltic vocalism might be due to neo-ablaut on the basis of other feminine a-stems, but then one wonders why other nouns of a similar structure (such as Lt. žiáunos ‘gills’ and liaukà ‘gland’, cited by Petit himself) were not subject to this analogical pressure. In sum, the Baltic vocalism still lacks a satisfactory explanation, so that we might consider Beekes’ account as a nonIndo-European loanword a possible option. It should be admitted, however, that the words in all the other European languages can go back to a common proto-form. † ‘bull’. Lt. tauras ‘buffalo, aurochs’ (Bretke, Morkūnas; see ALEW 1248), dial. taũris ‘calf, bull (vel sim.)’ (cf. Arumaa 1930: 19, No. 2; LKŽ),36 Pr. E tauris · wesant (for ‘aurochs’?, PKEŽ iv: 186; Young 1998: 201–203); OCS тоуръ ‘bull’, Cz. tur ‘bovid’, OPl. tur · bubalus (SSP ix: 227–228) ~ ON þjórr ‘(young) bull’ (Ipsen 1924: 227–228; Beekes 2000a: 30; Kroonen 2012: 250, 2013: 478, 540; Šorgo 2020: ‣ 34 35 36 Van der Staaij (1975: 197–198) has argued that Venetic eu is a secondary, dialectal phenomenon due to its geographical distribution, although the absence of early examples with -eu- may simply be due to the absence of early evidence from the relevant regions (cf. Lejeune 1974: 111). Matasović (2009: 386) has suggested that Gaul. Teut- is “just a spelling variant”, but there seems to be other evidence for the preservation of *-eu- in Gaulish. (cf. Leucetius epithet of Mars ~ OIr. lóichet ‘lightning; gleam’, Go. liuhaþ ‘light’, Delamarre 2003: 200; Neuio-dunon, placename in Pannonia, cf. Nouio-dunum ‘Neu-Châtel’, op. cit. 236). As for Lt. laũkas, Lv. làuks ‘blaze-faced’, there is no reason to prefer a direct equation with Gr. λευκός ‘bright, clear, white’ over an o-grade adjective of the type Lt. raũdas ‘reddish brown’, Go. rauþs* ‘red’ (cf. Berneker 1899: 164; Petit 2000: 120). The accented form taũras of the standard language was apparently introduced by Būga (1912: 40–44). Interestingly, Būga had previously (RR ii: 718) labelled *taũras as erroneous; but has later defended it on the basis of toponymic evidence. Since the aurochs went extinct in the 17ᵗʰ century, this accented form must in any case be regarded a learned creation. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 vocalism 251 453; van Sluis forthc.) — With a-vocalism, cf. also Gr. ταῦρος, Lat. taurus ‘bull’ and Alb. ter ‘steer’ (with analogical umlaut from the plural; cf. Demiraj 1997: 384; Orel 2000: 224; Matzinger 2006: 56; Schumacher 2013: 228).37 OIr. tarb, MW tarw ‘bull’ reflect *tarwo-. The metathesis is generally explained as the result of contamination (with *karwo- ‘deer’, cf. MW carw, or less likely OIr. poet. ferb ‘cow’, Walde/Hofmann ii: 651; LEIA T-31; de Bernardo Stempel 1999: 214–215).38 By contrast, Latin lacks the expected metathesis to *tarvus, which suggests a relatively recent loanword (cf. Трубачев 1960: 7; de Vaan 2008: 607; Weiss 2020: 170).39 A possibility would be to assume a Sabellic origin, as with Lat. bōs ‘cow’, cf. Oscan ταυρομ acc.sg. ‘a sacrificial animal(?)’ (Untermann 2000: 777–778). This would imply that the metathesis did not occur in Sabellic, although the argument is admittedly circular (there is no certain evidence either way).40 In Germanic, two similar words for ‘bull’ are found, both of which show evocalism. ON þjórr ‘bull’ (< *þeura-) and Go. stiur ‘ox, calf’, OHG stior ‘(young) bull’ (< *steura-). Ipsen, and several others (see refs. above), have pointed to the fluctuation between *þ and *st, as well as the vocalic alternation, as evidence of a non-IE origin. The older theory (Brugmann 1906: 353; Petersson 1921: 40–41; Walde/ Pokorny i: 711; Mallory/Adams 1997: 135) separates the two Germanic words, taking *steura- together with YAv. staora- ‘pack animal; cattle’, Parth. ʿsṯwrʾn /istōrān/ pl. ‘cattle’ (Durkin Meisterernst 2004: 91)41 and assuming *þeura- pro 37 38 39 40 41 Since a development *eu > Alb. e has been widely assumed (e.g. Huld 1984: 155; Demiraj 1997: 46) one may be tempted, with Mallory/Adams 2006: 136, to equate Alb. ter directly with ON þjórr. However, Matzinger (2006: 57; cf. de Vaan 2018: 1739) has put this sound law into doubt. Furthermore, the expected vocalism appears to be preserved in Alb. taroç, tarok ‘young bull’. The regularity of the change *wr > *rw (Matasović 2009: 371) is uncertain. It is contradicted at least by OIr. gúaire ~ Lt. gauraĩ ‘animal hair’ and MIr. glúair (< *glauri-) ‘clear, bright’ ~ ON gløggr (< *glawwa-) ‘clear, distinct; clever’ (Zair 2012: 237). De Vaan’s treatment of this development as regular in Latin (cf. alvus ‘belly’, nervus ‘sinew’, parvus ‘small’ ~ Gr. αὐλός ‘pipe, hollow tube’; νεῦρον, παῦρος) is certainly preferable to older notions of a ‘sporadic’ fluctuation (often assumed for PIE itself, thus Pedersen 1909: 176; Specht 1947: 35; Leumann 1977: 101). However, details need working out: cf. caurus ‘northwest wind’, caulis ‘stalk, stem’, īnstaurō ‘repeat, restore’. Sabellic shares *eu > *ou with Latin, and the development has usually been assumed for Proto-Italic (e.g. Brugmann 1897: 197; Leumann 1977: 61). Since nervus ‘nerve, sinew’ < *neuro- must predate this change, this would force us to assume the metathesis was ProtoItalic, too. On the other hand, if the evidence for the preservation of *eu into early Latin is taken seriously (see Weiss 2020: 112–113), this would imply the vowel development was independent in the two subfamilies (note also fn. 34, above, on Venetic). Usually seen to be a derivative of the adjective in Skt. sthūrá- ‘big, strong’, Khot. stura‘thick, large’, MDu. stuur ‘strong, fierce’, OHG stiuri ‘strong, proud’ (e.g. IEW 1009–1010). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 252 chapter 7 **þaura- to be the result of a contamination with the former. It does indeed seem suspect to disregard this semantically and formally convincing match with Iranian in favour of an irregular comparison. If *steura- and *þaura- were in competition in Proto-Germanic, such a contamination seems quite imaginable. A frequent argument in favour of a non-native origin is the existence of a similar word in Semitic, cf. Akkadian šūru, Aramaic twr */tōr/, Arabic ṯawr, Geʿez sor ‘bull, ox’ (< *ṯawr-, Militarev/Kogan 2005: 307–309). While some have assumed a direct loan from Semitic into Indo-European (J. Schmidt 1891: 7; Vennemann 1995: 88–89) or in the other direction (e.g. Walde/Pokorny i: 711; Walde/Hofmann ii: 651), several scholars have argued that the word is better derived from some other, third source (Feist 1913: 411; Ipsen 1924: 227–228; Schrader/Nehring ii: 261; Гамкрелидзе/Иванов 1984: 519–520).42 Blažek (2003) has proposed a plethora of Afroasiatic comparanda, some of which were taken over by Militarev/Kogan (2005: 309–310). The latter additionally cite some Chadic words for ‘elephant’ and an isolated Kachama (Omotic) word for ‘rhinoceros’. Leaving aside these semantically questionable comparisons, and since most of the Cushitic forms cited by Blažek reflect an unrelated *tsawadu (Kießling/Mous 2003: 293), what we are left with is Ma’a/Mbugu churú ‘bull’ (cf. Militarev/Nikolaev 2020: 205–206). The evidence is therefore scarce, and if we add that Mous (1996: 202, 210) specifically warns against the reconstruction of proto-forms on the sole evidence of Ma’a, a mixed language with a complex history, the situation looks even less favourable. Thus, while the existence of Afroasiatic comparanda would more or less confirm a specifically Semitic → Indo-European loanword (Militarev/Nikolaev loc. cit.), the external evidence can hardly be relied on.43 To give better support to the Indo-European status of the word, Mallory/ Adams (1997: 135; 2006: 140) adduce a Khotanese ttura- ‘mountain goat’, first included here by Bailey (1979: 132). True, as a late Khotanese word, ttura- could in theory reflect *taura- (pace Simon 2008: 299) beside several other options (cf. Emmerick 1989: 212). However, the word is a hapax in a difficult passage 42 43 Delamarre (2003: 292), by contrast, considers the similarity coincidental. Maarten Kossmann (p.c. March 2023) considers it possible to compare the Semitic word with Berber *a-zgăr ‘bull’, providing (1) the *z is assimilated from *s, in which case the initial consonant correspondence would match Semitic *ṯn- ‘two’ vs. Berber *sin (vel sim.), and (2) the *g is derived from *w, which can be related to the known (but poorly understood) alternation between *w and *g in some Berber lemmas. For more detail, I refer to the original Twitter discussion between Kossmann (@ait_kisou) and @irzastan posted 8 February 2022. If this comparison is indeed correct, it would again tip the balance in favour of a Semitic → Indo-European loan. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 vocalism 253 whose meaning is far from certain (F. Dragoni p.c. November 2021),44 and is therefore too unreliable to use. Абаев (1958: 390–391) has additionally mentioned Ossetic I ʒæbīdyr, D ʒæbodur ‘West Caucasian ibex’, but it is unclear whether -dyr, -dur can be segmented, as the first element remains obscure.45 Kroonen (2013: 540) associates the Germanic *eu with the similar vowel attested in Etruscan θevru-mineś ‘Minotaur’. The Etruscan -ev- is indeed problematic, and like the initial θ-, appears to rule out a direct loanword from Greek Μῑνώταυρος (cf. Fiesel 1928: 80–81; Kretschmer 1940: 266). On the other hand, due to the distance between the Germanic and Etruscan homelands (wherever the latter may be), it is unlikely the developments can be associated with each other. It is difficult to use the Etruscan evidence to support a non-IE origin, but it is not entirely clear where the word was adopted from. In conclusion, as far as Indo-European goes, the word is limited to Europe, but the arguments in favour of its foreign origin are somewhat circumstantial. The irregular form in Germanic could be explained as resulting from a fairly well-grounded contamination, and therefore the word cannot be classed as a certain case of an *e ∞ *a alternation. A reconstruction *th₂eu-ro- could account for most of the other data. Perhaps the strongest evidence for a non-IE origin remains the irregular Celtic form, for which yet another contamination must be assumed. How exactly the Etruscan and Semitic words fit into the picture is unclear. Due to the complexity of this example, I will leave it out of consideration here. † ‘poppy’. R мак, Sln. màk; OSw. val-mogha; Gr. μήκων (Doric μάκων) ‘poppy’ ~ OHG maho, mago ‘poppy’ (Beekes 2000a: 29; Boutkan 2003a: 15; Matasović 2013: 89) — For a discussion of Lv. maguône (etc.) ‘poppy’, which is most likely of Germanic origin, see p. 41.46 On account of MoHG Mohn, the OHG variant maho has often been attributed a long vowel (e.g. Kluge/Götze 396). As Kluge/Seebold (484) point out, an OHG ā (< PGm. *ē) would hardly be compatible with the Doric Greek -ά-. ‣ 44 45 46 Skjærvø (2002: 35) does not attempt a translation. In an earlier publication (1949: 49), Абаев considered the word to be of native Caucasian origin, adducing Karachay ǯuğutur, to which we can add Kabardian šəqʷłtər in the same sense. He segments the Karachay word ǯuğu-tur (with the second element assumed to be ← R тур?), comparing Georgian ǯixvi ‘West Caucasian ibex’. Just how a form of this shape could be borrowed into Ossetic as ʒæb- is unclear to me. Bailey (1979: 132) sees in the first element the word Oss. ID ʒæbæx ‘good’. This species of Ibex is referred to as тур in Russian, which is quite a surprising semantic shift: perhaps it was encouraged by the similar-sounding Karachay word? It seems much less probable that these Baltic forms could show evidence of a *g ∞ *k alternation (Oštir 1929: 107). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 254 chapter 7 However, as argued by Schaffner (2001: 358–361), the apparent MoHG evidence for a long vowel may be explained as the result of an early contraction over *h (cf. MHG mān beside mahen).47 The evidence is therefore consistent with an ablauting *meh₂k- : *mh₂k- (cf. Kroonen 2011b: 311–314, 2013: 371). The word for ‘poppy’ has often been suspected to be a prehistoric loanword on the basis of cultural facts (Būga 1924a: 18; Kluge/Götze 396; Machek 1950b: 158; Sabaliauskas 1960a: 71, 1990: 261). On the other hand, some scholars have seen the distribution and evidence for ablaut as a clear indication of an inherited origin (Frisk ii: 225; Kluge/Seebold 484). As we cannot prove a loanword on formal grounds, this word cannot be considered here. 7.2.3 *u ∞ *i ‘oven’. Pr. E Vumpȋs ‘oven’; Go. auhns*, OHG ovan ‘oven’ ~ Gr. ἰπνός ‘oven, furnace’ (Kroonen 2013: 557) — For a discussion of the Prussian and Germanic forms, see further p. 231. The comparison between Gr. ἰπνός and Germanic *ufna- seems nigh inescapable. The old, traditional, equation with Skt. ukhā́‘cooking pot’, on the other hand, is phonologically impossible and must be rejected (see Frisk i: 732–733; EWA i: 210; Kroonen 2013: 557). For the Greek word, Vine (1999: 19–23) has suggested an alternative etymology starting from *sp-no- with i-epenthesis to the root of Gr. ἕψω ‘boil’, Arm. epʿem ‘cook’ < *seps-, yet there is no other evidence that the final *s in this root is suffixal.48 According to Vine (1999: 22), the inscriptional hιπνε[ύεσθαι] is “exceedingly difficult to explain away”, and while indeed Threatte (1980: 494) writes that non-etymological h- is “virtually unknown in fifth-century Attic texts”, some of the examples accepted by him as etymologically justified are perhaps not, e.g. hακόσια ‘unwilling’ (before 460 bce) < *n̥ -u̯ ekontia (cf. Beekes 2010: 400), and a handful of examples are still acknowledged as irregular (Threatte 1980: 495) so that a single attestation can be considered insufficient to prove an initial aspirate. It seems more likely that the equation with OHG ovan and Gr. ἰπνός should be maintained, and that we should assume a non-IE alternation *i ∞ *u. Unfortunately, this alternation does not appear to be supported by other certain examples. ‣ 47 48 This seems to be paralleled by Middle German stol ‘steel’ (Elbing Vocabulary; see Chapter 2, fn. 4) < OHG stăhal (Swiss Stăchel, cf. Schw. Id. x: 1197). As for the widely acknowledged constraint against two like stops in an IE root, this evidently did not apply to fricatives, cf. Skt. sásti, Hitt. ses-ᶻⁱ ‘sleep’ (< *ses-, LIV 536–537; Kloekhorst 2008: 746), Hitt. huhha-, Lat. avus ‘grandfather’ < *h₂euh₂- (op. cit. 352), possibly *h₃neh₃-mn- ‘name’ (see Beekes 1987; van Beek 2011: 52–53). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 vocalism 255 ‣ ? ‘cod’. R треска́ ‘cod’ ~ ON þorskr, MLG dorsch ‘cod’ — The other Slavic forms (Uk. трiска́ , Bel. траска́ ; Cz. treska, Sln. tréska, SCr. trȅska, etc.) are all regarded as recent loanwords from Russian (cf. ЕСУМ v: 645; Machek 1968: 650; Bezlaj iv: 220), so we must base our conclusions on the Russian evidence alone. In СРЯ 11–17 (xxx: 131–132), the forms are normalized under трѣска², consistent with the traditional etymological equation with RCS трѣска ‘splinter’ (REW iii: 137). However, none of the citations are actually spelled with ⟨ѣ⟩49 and the MR variant троска (СРЯ 11–17 xxx: 180; cf. dial. (Karelia) тро́ ска, СРГК v: 518) might instead imply an earlier *трьска (cf. Pedersen 1895: 72) with yer umlaut (as in тресть ~ трость ‘reed’, see p. 201). However, note similarly MR трѣска ~ троска ‘sharpened stick; stake’, which I cannot explain. Although the word must be old in Norse (cf. the early loanword in F turska, E tursk, Li. tūrska ‘cod’; LÄGLOS iii: 322–323), MLG dorsch ‘cod’ and MDu. dorsch* (attested dorssch) ‘a kind of fish’ need not be inherited, and have been interpreted as Norse loanwords (Philippa et al. i: 615; Kluge/Seebold 212). However, on formal grounds, cognancy cannot be ruled out. Likewise, R треска́ has been interpreted as a loanword from Germanic (Tamm apud de Vries 1962: 618; Machek 1968: 650, allegedly from “an unattested northern variant”; Kluge/Seebold 212), but this is phonologically implausible. If we set up an original *tresk-, this could be combined with Germanic *þurska- by reconstructing an ablauting *trsk- : *tresk-. However, if the Russian form goes back to an earlier *trisk-, the connection cannot be maintained in Indo-European terms. Without other Slavic cognates, it is difficult to make a convincing case for a non-Indo-European origin. Even though reconstructing old ablaut for a noun in the sense ‘cod’ is questionable, this word cannot serve as certain evidence. 7.2.3.1 Conclusion The certain and possible evidence front ∞ back vocalic alternations is collected in Table 15, overleaf (see p. 185 for help reading the table). Forms which do not provide relevant data are presented in light grey. Shaded cells indicate forms with back vocalism. 49 Зализняк (2019: 185) keeps треска́ ‘cod’ and трѣска́ ‘splinter’ apart. The word first appears in 16ᵗʰ century North Russian monasterial accounting books, but it is possible it had been in use earlier among illiterate fishing populations. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 256 chapter 7 table 15 Possible examples of front ∞ back alternations Baltic Slavic Germanic Elsewhere ‘hellebore’ – *kemer- *kemer- Gr. *kam(m)ar- ‘ramsons’ *kermus- ‘garlic’ – *kesn- – Celt. *kasn- ‘cottage’ *trâB- – *t(u)rb- It.-Celt. *trē̆b- ‘honeycomb’ *kâr- – – ‘ground elder’ *Gârsu̯ - – *gʰerVs- ‘oven’ *umnV- ? ‘circle’ – *krangʰ- *krengʰ- ? ‘people’ *taut- – *teut- ? ‘cod’ – *tri/esk- *t(u)rsk- *kermus? *kermas- *kra/omus- *upno- Gr. *krommusCelt. *kram- Gr. *kērTur. *käräs Gr. *ipno- It.-Celt. *teut- It seems that we can identify two main groups. In the former, Balto-Slavic shows *e as against *a (or *o) elsewhere. It is remarkable that all three examples of this alternation show an initial *k-. If this is not mere coincidence, we might assume a phonetic solution. It is reminiscent of the situation in Turkic, where due to the allophonic change *k > */q/ in back-vocalic contexts, loanwords with /k/ are automatically adopted with front vocalism, as Chuv. kămpa, dial. kŏmpa, Tat. gömbä ‘mushroom’ (← Sl. *gǭbā-, trad. *gǫba), Chuv. kĕrpe, dial. kör̆ pe ‘groats’, Tat. körpä ‘bran’ (← R крупа́ ).50 50 In this respect, one might note the unexplained East Baltic words for ‘marten’ — Lt. kiáunė, Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 257 vocalism The second group shows the opposite distribution, with the back vocalism being limited to Balto-Slavic. The coherence of this group is less certain, as in each case the comparanda show a distinct pattern of correspondences.51 In this context, compare also the potential examples of front/back alternations among the Finnic-Baltic isoglosses discussed in 3.5.3. 7.3 Alternations between Low and High Vowels 7.3.1 Baltic/Slavic High Vowel ∞ Low Vowel Elsewhere [‘bison’. Lt. stum̃ bras ‘bison’, Lv. stum̃ brs, OHG wisunt ‘aurochs’ ~ Pr. E ⟨wissambs’⟩ · ewer — See the discussion on pp. 220–221.] [‘oriole’. Pl. wilga, Sln. vółga ~ Lt. volungė ̃ 3ᵃ, Lv. vãluôdze, ME wode-wale, MHG wite-wal ‘oriole’ — See the discussion on pp. 178–180.] ? ‘heel’. Lt. kulkšnìs ‘ankle; hind heel of an animal’, Lv. dial. (ME ii: 307) kulksnis ‘ankle joint (in animals); leg’ ~ Lat. calx ‘heel (of a person or animal)’ (Matasović 2013: 89) — Because of kulſchnis nom.pl. ‘ankles’ in Bretke, it is attractive to view the -k- as intrusive (Berneker i: 660; Derksen 2015: 262; ALEW 618); compare Lt. kùlšis, kùlšė; Lv. dial. (ME ii: 308) kulša ‘hip, loins’, Pr. E culczi ‘hip’.52 With this, one would like to compare SCr. kȕk, Bg. къ̀лка ‘thigh, hip’,53 although this would require an unmotivated Gutturalwechsel (Berneker loc. cit.; Trautmann 1923: 145). The vowel in Latin calx is difficult to derive from any reasonable IE preform (Ernout/Meillet 89; against Schrijver’s rule *ke- > ‣ ‣ ‣ 51 52 53 Lv. caûna — which point to e-grade, as opposed to the other Balto-Slavic forms, which would imply an old *a or *o, cf. Pr. E caune, MR куна́ (Зализняк 2019: 206), Cz. kuna, SCr. kúna ‘marten’. The word is of unclear derivation (Derksen 2015: 242; ALEW 560). True, in the above examples, the e-vocalism was found throughout Balto-Slavic, so whether the phenomenon is the same is uncertain. Here one could also mention Lt. pélkė; Pr. E pelky ‘marsh’ ~ Gr. (H.) παλκός · πηλός ‘earth, mud’ (cf. Alessio 1946: 160; van Beek 2013: 548, fn. 21; Derksen 2015: 349–350), but it is difficult to base much on a mere gloss. A potential additional example within Slavic would be the word for ‘swan’ (see pp. 176– 177), cf. Sln. labǫ́ d ~ R ле́бедь, although here the additional nasal in the former must be factored in. It is tempting to further compare Lt. kul̃nas, kulnìs, Lv. dial. (Rucava) kulna ‘heel’, which could reflect earlier *kulkna- (IEW 928) as in Lt. bal̃nas ‘saddle’ (= Pr. E balgnan) with the blocking of palatalization before *n (cf. Pr. E balsinis ‘cushion’). However, the difference between Pr. E kulnis ‘ankle’ and balgnan ‘saddle’ speaks against this. Only South Slavic. Hardly here belong R (hapax?) колкъ (“коло́ къ?” sic. Даль² ii: 139) ‘bony stump of a bovine horn’ and Cz. kelka ‘(arch.) stump of a limb; (hunting term) tail of a deer’. The latter does not show a regular reflex of *kulk- (trad. *kъlk-). Similarly, USrb. kulka ‘ankle’ cannot be from *kulk- and is probably merely a diminutive of kula ‘bulge’ (cf. Schuster-Šewc 723). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 258 chapter 7 *ka-, see Meiser 1998: 82–83), but due to the difficulties in analysing the BaltoSlavic data, attributing *a ∞ *u to borrowing from a non-IE source language may be premature. A potential parallel can be found in the equation between RCS громѣждь, гремѣждь (СРЯ 11–17 iv: 129)54 and Lat. grāmae (TLL vi: 2165) ‘rheum in the eye’ (cf. de Vaan 2008: 270; Matasović 2013: 84). The original form of the RCS word is uncertain, but would be consistent with *grim- (trad. *grьm-; СДРЯ 602; Berneker i: 360; Sławski SP viii: 267; РЭС xii: 78) or *grum- (trad. *grъm-; ЭССЯ vii: 159; Derksen 2008: 194). In the latter case,55 we would be dealing with a similar *a ∞ *u alternation. The long vowel in Latin56 could also be accounted for if we assume an underlying *gra(k)sm-. In Slavic, *x (< *ks or *s + RUKI) would have disappeared without a trace before a resonant as in *lūnā- (trad. *luna) ‘moon’ < *louksneh₂- (Pronk 2018: 300); thus, Slavic could theoretically reflect an earlier *gru(k)sm-.57 ? ‘salmon’. Lt. lašišà, Lv. lasis ‘salmon’ ~ Pr. E lalasso */lasasā/, R лосо́ сь, Pl. łosoś ‘salmon’. Further ON lax, OHG lahs ‘salmon’ (cf. Laumane 1973: 116; Ariste 1975: 468)58 — A Lithuanian variant lãšis was recorded by Nesselmann (1851: 350, “bei Memel”). Similar forms have been recorded all along the Western coastline,59 as well as near the Latvian border (lãšė Ylakiai, Kivyliai; Vanagienė 2014), and these correspond formally to Lv. lasis. Generally, Lt. lašišà has been ‣ 54 55 56 57 58 59 Further Slavic forms have an unclear initial *k-: Sln. krmẹ́žəlj, in addition to which Sławski (SP xiii: 267) adduces SCr. Čak. dial. kȑmež, Kajk. dial. kr̀meželj (cf. kr̀mežalj in the dictionary of Popović apud РСА x: 216). Both languages also attest a shorter form: Sln. krmę́lj (lexicographically recorded), SCr. dial. kr̀melj (a kind of haplology?). In favour of *grum- (trad. *grъm-), we can note that the form with -o- is attested some two centuries earlier (contrary to the 14ᵗʰ century date usually given (e.g. СРЯ 11–17 loc. cit.), Жолобов (2007: 35) has convincingly argued that the RCS Parenesis of Ephrem the Syrian should be dated to the 13ᵗʰ century), and that -e- would be more easily explained as a secondary assimilation. The word is rare, but the long vowel is metrically secured in Plautus. The derived adjective grammō(n)sus would therefore show the littera rule. However, if we are willing to permit an alternation *l ∞ *r (see p. 189), it would seem obvious to compare Greek γλαμυρός ‘bleary-eyed’ (cf. H. γλάμος ‘mucus’, glamae ‘rheum’ in Paulus ex Festo). This would speak against a reconstruction such as *gra(k)sm-. In view of its initial l-, Oss. D læsæg (recorded only lexicographically, cf. “Не документировано”, Абаев 1973: 32–33) must be a borrowing (Абаев 1965: 37–38). Similarly, Arm. losdi ‘salmon’, attested only in the Armenian-Latin dictionary of Stefanus Roszka (V. Petrosyan on en.wiktionary.org, s.v. լոսդի [16 September 2019]; cf. Martirosyan 2008: 312) is most likely a local adoption of Romanian lostrița, lostița ‘Danube salmon’. The latter ultimately derives from Slavic (cf. Diebold 1976: 368). cf. lašìs Rusnė (on the Neman), lãšė Kintai, lãšis Kukuliškiai (Papildymų kartoteka), lašìs Palanga (LKŽ). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 vocalism 259 considered a derivative of lãšis (Skardžius 1941: 317; Specht 1947: 31), but it seems more probable that laš- (and Latvian las-) derives from lašiš- (> *lašš-) by syncope. Note, with a similar distribution, the syncopated Žemaitian vẽčas (Daukantas, Juška) and Lv. vȩcs as against Lt. obs. vetušas ‘old’. In Latvian, the development is closely paralleled by tacis ‘fishing weir’ (< *tacsis < *tacisīs, cf. Lt. takišỹs; see Chapter 3, fn. 136). Compare also, with a different distribution, Lt. lopšỹs ‘cradle’ as against dial. Žem. lopišỹs. Due to the fluctuation in stem type, Baltic *lašiš- may be an original root noun. The difference between East Baltic *lašiš- and Slavic/Prussian *lašas/š- is difficult to account for in Indo-European terms. True, there are a couple of East Baltic forms which might show -a-, which could suggest that the stem *lašiš- is a recent development. Szyrwid has łaſaßa ‘salmo, łosoś’, and Lexicon Lithuanicum has laßaßa (ALEW 647). However, neither of these forms are entirely reliable: the form in Szyrwid might have been influenced by the Polish equivalent (the first -s- seems to imply this), and the other form might be a Prussianism (cf. PKEŽ iii: 31). Nevertheless, the value of this alternation is not entirely certain. The Balto-Germanic word for ‘salmon’ has almost universally been considered cognate to Tocharian B laks ‘fish’ (Walde/Pokorny ii: 381; IEW 653; Derksen 2015: 274–275), and this has been seen as important in discussing the Indo-European homeland (e.g. Diebold 1976; Mallory/Adams 2006: 146). However, the comparison is phonologically irregular, as the Tocharian form is only consistent with an earlier *-u- (Ringe 1992: 92). I must agree with Pinault (2009: 241, fn. 74), that the only correct solution is to reject the Tocharian form as a cognate. We are therefore dealing with a circum-Baltic term for local fauna which might plausibly be explained as a loanword from an unknown source. [? ‘grouse’. Lt. jerubė ̃ ‘hazel grouse’, Lv. rubenis ‘black grouse’ ~ OHG rebahuon, MLG rap-hōn ‘partridge’ — Note the reverse distribution. See the discussion on p. 241.] † ‘shoe’. Lt. kùrpė ‘clog, shoe’, Lv. kur̃pe; Pr. E kurpe ‘shoe’; Cz. dial. (Kott i: 827) krpě ‘Schneereifen’, Sln. kŕplja ‘snow-shoe’ ~ Gr. καρβάτιναι ‘shoes of undressed leather’ (Furnée 1972: 146; Beekes 2000a: 28; Derksen 2008: 263) — This example could possibly show the alternation *u ∞ *a,60 but the Balto-Slavic ‣ ‣ 60 Another potential example is Pr. E spurglis ‘sparrow’ as opposed to Go. sparwa, OHG sparo ‘sparrow’ (Matasović 2013: 87), but it cannot be excluded that these represent an ablauting *sprgʷʰ- : *sporgʷʰ-. The reconstruction remains uncertain in view of numerous, but all doubtful, Greek comparanda: σπέργουλος, πέργουλον (H.) ‘a wild bird’, σπαράσιον (H.) ‘a bird resembling a sparrow’ (Frisk ii: 1130; Schrijver 1997: 304) and ψάρ ‘starling’ (Walde/Pokorny ii: 666; Kroonen 2013: 466). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 260 chapter 7 stem might just as well be identical to the first syllable of Gr. κρηπίς, pl. κρηπῖδες ‘kind of half boot’ (?< *kʷrh₁p-, with Greek dissimilation *kʷ–p > *k–p; on which recently see van Beek 2022: 466).61 7.3.2 Slavic *e ∞ Germanic *u ‘maple’. R клён, Pl. klon ‘maple’, SCr. klȅn ‘field maple’ ~ ON hlynr (attested in kennings; cf. Ic. hlynur ‘sycamore maple’), MLG lönenholt ‘maplewood’ (MoLG Löhn ‘maple’, see Marzell ii: 73; whence probably MoHG obs. Lehne, cf. DWb xii: 1137) (Oštir 1930: 22; Machek 1950b: 154; Matasović 2013: 85, forthc.)62 — The lexicographically attested SCr. dial. kȗn ‘maple’ is usually taken to represent an old *klina- (trad. *klьnъ; Berneker i: 512; REW i: 567; ЭССЯ ix: 195).63 Considering the isolation of this dialect form, it seems difficult to justify reconstructing it for Proto-Slavic. Skok (ii: 95) remarks on the similarity to the “Macedonian” κλινότροχον quoted by Theophrastus,64 referring to a kind of maple, and suspects a localized borrowing. The position of Lt. klẽvas, Lv. kļava, kļavs ‘maple’ is unclear. The analysis *klen-u̯ o- (e.g. Oštir 1930: 68; Bańkowski 2000 i: 706) would be in contradiction to Žem. tę́vas, Lv. tiêvs ‘thin’ < *tenh₂-u̯ o- (~ Lat. tenuis ‘thin, fine’; cf. also ЭССЯ ix: 194). In fact, the Latvian evidence appears to suggest an earlier *kljawa- (Endzelīns 1911: 94; Stang 1972: 28–29), which could be supported by the absence of l-hardening in Lithuanian dialects (Zinkevičius 1966: 160). It is difficult to account for the alternation *kliou- ∞ *klen- even within a nonIE context — as ALEW (p. 584) points out, all we are left with is a common element *kl-. Therefore, I am inclined to leave the Baltic terms out of the comparison. ‣ 61 62 63 64 OIr. cairem, MW cryd ‘shoemaker’ have been compared and derived from *kerh₁pio(LEIA C-21; Matasović 2009: 189–190); however, this reconstruction is unlikely to yield the attested forms (Zair 2012: 83) and the Celtic words may more attractively be compared with Lat. corium ‘leather’ (de Bernardo Stempel 1987: 93; compare Gr. σκῡτεύς ‘cobbler’ < σκῦτος ‘leather’). It almost seems a sheer accident that Schrader (1901: 33 and Schrader/Nehring i: 38) included here OCo. kelin ‘holly’. He left the word without a gloss, and in the same unglossed form it was repeated by Trautmann (1923: 136), Berneker (i: 512, adding Welsh celyn) and Vasmer (REW i: 567); cf. Friedrich (1970: 64), where Welsh celyn is incorrectly glossed ‘maple’. The etymology, with the correct gloss, is explicitly defended by Specht (1947: 60) and Fraenkel (LEW 270–271), while it is explicitly — and surely correctly — rejected by Stang (1972: 29; implicitly e.g. IEW 603). Although *kluna- (trad. *klъnъ; Miklosich 1886: 118) or *kulna- (trad. *kъlnъ) would also be possible. Alongside a variant γλεῖνος, of unspecified dialectal affiliation (cf. Meyer 1892: 325–326). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 vocalism 261 Germanic *hlun- could theoretically be analysed as a secondary zero-grade to *hlen-, thereby matching the Slavic form; however, since such a full-grade variant is unattested, this hypothesis is rather circular. It seems quite possible that the irregular correspondence could be explained as the result of borrowings from a non-Indo-European source. A similar explanation might also account for the “Macedonian” κλινό-, mentioned above, although too little is known about Ancient Macedonian for this form to be used here. [‘silver’. OCS сьребро, Cz. stříbro, Sln. srebrọ̑ ‘silver’ ~ Go. silubr, ON silfr, OHG silabar* ‘silver’ — See the discussion on pp. 225–226.] ? ‘frogspawn’. Lt. kurkulaĩ, Lv. kur̂kuļi (LVPPV: kùrkulis); Pl. skrzek, Cz. dial. (Kott v: 752) žаbо-křеkу ‘frogspawn’ ~ ON hrogn, OHG rogo ‘(fish) roe’ (Polomé 1986: 661) — Germanic *kruk- and Baltic *kurk- seem to show an irregular metathetic relationship. The analysis of the Balto-Slavic data is difficult, however, due to repeated contaminations with words for ‘croak’ (cf. Machek 1924: 128– 130). First and foremost, the Baltic forms look like derivatives of Lt. kur̃kti, Lv. kùrkt ‘to croak’ (Būga 1923–1924: 139; cf. Nesselmann 1851: 212). In Slavic, compare R dial. кряк ‘frogspawn’ (hardly with -я- < *-ę̄-, pace REW i: 674) beside кря́ кать ‘croak’ (СРНГ xv: 365–366), and further Pl. skrzek ‘croaking; frogspawn’. Since ‘frogspawn’ as ‘the croaker’ does not make much sense, I assume these contaminations are secondary. In this case, one is tempted to give preference to the Lithuanian *krekulai (Miežinis 1894: 118),65 dial. krekùčiai (LKŽ) ‘frogspawn’, which are closer to the Slavic forms. This could perhaps support the connection with Lt. krèkti ‘coagulate, congeal’ (Berneker i: 613–614; LEW 293) and suggest that the Germanic evidence is unrelated. On the other hand, the parallelism with *klen- : *klun- ‘maple’, discussed above, opens up the possibility of analysing *krek- : *kruk- as a non-IE borrowing. ‣ ‣ 7.3.3 ? Lithuanian *ā ∞ Latvian *ū ? ‘millet’. Lt. sóros ~ Lv. dial. (Kurzeme) sûra² ‘millet’ (Pronk/Pronk-Tiethoff 2018: 293; Kroonen et al. 2022: 22) — The inner-Baltic correspondence is irregular, but Elger’s 1683 dictionary has Lv. ⟨sâre⟩, ⟨sare⟩ = */sāre/ ‘milium’ (see ME iii: 806; Nieminen 1956: 164–165). This might suggest the Latvian variant with -ū- is secondary, although only ad hoc accounts can be given for it (cf. Nieminen 1956: 175–176). The best explanation is to assume a contamination with South Estonian suurmaq, Li. sūrmõd pl. ‘groats’, which is supported by ‣ 65 Attested as ⟨krakulai⟩, but its alphabetical position after ⟨krekinties⟩ implies a misprint. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 262 chapter 7 the Salaca Livonian compound kriev sūrmed ‘millet’ (kriev ‘Russian’; Winkler/Pajusalu 2009: 87, 182). The existing etymological explanations of the Baltic word are all unsatisfactory. A derivation from an older *psārā and comparison with Slavic *prasa- (trad. *proso) ‘millet’ (Hirt 1927: 309; Otrębski 1939: 137), or Skt. psāti ‘chew, consume’ (Nieminen 1956: 170) remains highly hypothetical, especially in view of the contradictory development of initial *ps- observed in Lt. spenỹs ‘teat’ < *psten-.66 The Baltic word is obviously related to Md. E suro, M sura ‘millet’.67 Although some have derived the Mordvin word from Baltic (Thomsen 1890: 219; SSA iii: 201; Kallio 2008a: 268), the opposite direction has often been preferred (Ojansuu 1921: 57–60; Kalima 1936: 210; Топоров/Трубачев 1962: 248). This was supported by the claim that the Mordvin lexeme is cognate with F sora, E dial. sõra ‘gravel, coarse sand’. As an alternative, Toivonen (1928: 233) has adduced Komi ze̮r, (Jaźva) zu̇ ·r ‘oats’, Udmurt ze̮r ‘bromegrass’, an equation which was taken up by UEW (766) and Лыткин/Гуляев (1970: 106, with hesitation). Should the Permic words belong here, the semantic shift could be explained as a result of a migration beyond the northern limits of millet cultivation (about 57° N in the Eastern Baltic, cf. Grikpėdis/Motuzaitė Matuzevičiūtė 2020: 163). If we admit a secondary voicing in Permic,68 then we could suggest a Uralic reconstruction *sorə. Such a reconstruction might also work for Mordvin, although admittedly, due to the PMd. final *-ə, Mordvin is more consistent with a reconstruction *sora rather than *sorə (Pystynen 2020b). Van Pareren (2008: 124; cf. Junttila 2015a: 23) is inclined to reject any relationship between the Baltic and Mordvin words because Baltic *-ā- is difficult to square with Mordvin *-u-. According to Nieminen (1956: 173), the comparison would only be possible should we assume an original *ō in preMordvin; however, the traditional reconstruction of Proto-Uralic *ō has since 66 67 68 Witczak’s (1997: 30–32) comparison with MP xwʾr /xwār/ ‘food’, Oss. I xor, D xwar ‘cereal, barley’, allegedly from *sueh₂r- is impossible. Not only is the loss of *u in Baltic suspect, but the Iranian root is clearly *hwăr- < *sueR- (Cheung 2007: 147–148), i.e. not consistent with a laryngeal. See Kroonen et al. 2022: 22. I cannot accept the suggestion of Nieminen (1956: 175) that the similarity is coincidental. Sporadic initial voicing is attested in Permic, compare e.g. Komi be̮ž, Udm. bi̮ž < *pončə ‘tail’ (Sammallahti 1988: 547), but there remains a question as to whether this also applies to sibilants. The few possible equations, e.g. Komi ziľ ‘diligent, hard-working’ ~ Hungarian ügyes ‘skilled, capable’ (UEW 442–443) and Komi za ‘stem, stalk, shaft’ ~ Erzya sad ‘stalk’ (see Chapter 4, fn. 56), all involve some additional phonological obstacles. For *o–ə > Komi/Udmurt e̮, cf. Komi-Permjak e̮s-, Udmurt e̮ski̮- ‘vomit’ < *oskə- (cf. Metsäranta 2020: 103). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 vocalism 263 been shown to be faulty: most examples of *ō should instead be reconstructed *a(–ə) (see Aikio 2012b). If we set up an earlier *sarə for Mordvin, this would bring it closer to the Baltic data, but force us to separate the Permic evidence; moreover, the Mordvin change *a(–ə) > (*o >) *u has been dated very early and considered to be part of a chain of vowel shifts shared by Sámi (Живлов 2014: 116–117), which would make the reconstruction of a pre-Mordvin *sarə anachronistic. In conclusion, it is unclear whether such a hypothesis would shed any light on the Latvian -ū-, as (especially in view of the evidence from Elger’s dictionary) the time depth is probably too shallow to assume influence of a substrate. However, a direct loan relationship between Mordvin and Baltic words for ‘millet’ remains doubtful, and it is quite possible that these represent parallel loanwords from a third source. Since the correlation between Baltic *ā and Mordvin *u could be characterized as one of height, this example may still belong here. 7.3.3.1 Conclusion The certain and possible evidence high ∞ low vocalic alternations is collected in Table 16, overleaf (see p. 185 for help reading the table). Forms which do not provide relevant data are presented in light grey. Shaded cells indicate forms containing high vowels. The evidence falls into two broad groups: those which show a high vowel in Balto-Slavic, and those which show a high vowel in Germanic. The former group consists mainly of more or less uncertain examples, and since there is little coherence in terms of distribution and co-occurring phenomena, it is quite unlikely that all of the examples can be directly compared. Those which show a high vowel in Germanic form a much more promising group — all three show *e (= *[æ]) in Slavic and *u in Germanic. Since the vowels *æ and *u could hardly be more different, representing opposite extremes of the vowel triangle, it may seem dubious to derive them from a common source. However, it is not too difficult to find such correlations between related languages, such as between e.g. Erzya Mordvin kenže (dial. känǯä) and Obdorsk Khanty kuns ‘nail’ (< *künči), while the regular development *e(–ä) > Permic *ǫ leads to examples such as Finnish pesä ‘nest’ ~ Udmurt puz ‘egg’. Thus, the problem may be resolved by assuming distinct but related donor languages. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 264 chapter 7 table 16 Possible examples of high ∞ low alternations Baltic *u̯ i-sam(bʰ)r- ‘bison’ *stum(bʰ)r- Slavic Germanic *za/um(bʰ)r- *u̯ i-s(u)nT- Elsewhere ‘oriole’ *u̯ ālanG- *u̯ (i)lg⁽ʷ⁾- *u̯ alk- ? ‘heel’ *kulḱ-ni- ? *kulk- – ? ‘salmon’ *laśiś- *laSaS- *laks- ? ‘grouse’ *i̯e-rubʰ- *(i̯)e-re/imbʰ- *rebʰ- ‘maple’ – *klen- *klun- ? Mac. *klin- ‘silver’ *sidʰabʰr- *sirebr- *silubr- Celt. ?*silapr- ? ‘frogspawn’ ? *krek- *krēk- *kruk- ? ‘millet’ – – 7.4 *sār- Lat. *kalk- Md. *sur- Alternations between Monophthongs and Diphthongs 7.4.1 ? Baltic *a ∞ Slavic *ai/ei Schrijver (1997: 304–307) has adduced several examples in which Germanic *ai appears to correspond to Celtic *a, the clearest being MW baed (< *basio-) against OE bār (< *baiza-) ‘boar’.69 None of his examples involve Balto-Slavic, but a few cases can be identified in which Slavic potentially contains a diphthong. Although it is admittedly not possible to rule out an original long vowel based on the Slavic-internal evidence, in each case, a diphthong would be more easily reconcilable with the Baltic evidence. 69 Note, however, that van Sluis et al. (2023: 231) consider the Celtic word a possible West Germanic loan. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 vocalism 265 OCS листъ, R лист, Cz. list ‘leaf’ ~ Lt. dial. lãkštas ‘large flat leaf; leaf ‣used‘leaf’. as a baking sheet’, dialectically and in older texts ‘leaf’ (see ALEW² s.v.), Lv. laksti pl. ‘leaves of herbs; vegetable tops’ — The Slavic word is usually compared with a different Baltic lemma, namely Lt. láiškas ‘narrow leaf (e.g. of an onion), blade; green shoot’, as a neologism ‘(postal) letter’, Lv. laiska (hapax? cf. ME ii: 410–411) ‘leaf on a flax or cereal stalk’. According to this theory, the Slavic word would show a different ablaut grade and a derivation in *-to-. If we assume a non-inherited origin, the comparison with Lt. lãkštas appears more straightforward; moreover, this word is semantically a better fit, as it is well attested in the more general sense ‘leaf’. It seems attractive to compare both the Baltic and Slavic words with a group of West Uralic words for ‘leaf’: Sá. N lasta, Sk. lõstt (< *le̮ste̮); F lehti, E leht, Li. lē’ḑ (< *lehti), Ma. E ləštaš, W ləštäš (< *lĭštäš), all of which can regularly reflect PU *leštə (UEW 689). Incidentally, the similarity to both the Baltic and Slavic words has already been noted: Sammallahti (1977: 123–124; cf. SSA ii: 58–59) has assumed a direct loan from the Baltic *la(k)šta-, while Viitso (1992: 189) and Напольских/Энговатова (2000: 229) have posited an early loan from Slavic. Finally, Blažek (2019: 216) has suggested a loan from an unattested Baltic *lišt-, a zero-grade equivalent of Slavic *līsta- (trad. *listъ). М. Живлов (p.c.) has noted that the similarity between these words might be better accounted for by assuming parallel loans from a substrate language. As the above solutions provide a convincing account of only part of the data, this is certainly worthy of consideration. A similar vocalic relationship is found between R тис, Cz. tis, Sln. (Pleteršnik ii: 670) tȋs, (SSKJ²) tȋsa ‘yew tree’70 and Latin taxus ‘yew tree’, which has widely been regarded a non-IE borrowing (Schrader in Hehn 1911: 532; Schrader/Nehring i: 225; Oštir 1930: 22, 90; Machek 1950: 152; REW iii: 107; Sławski SEJP i: 103),71 a suggestion which, in principle, seems attractive: the Latin and Slavic words are semantically identical and 70 71 RCS тисъ translates Gr. κέδρος ‘cedar’ (cf. СДРЯ 960; СРЯ 11–17 xxix: 350; cf. also OCS (Ps. Sin.) тіса glossing the Greek loanword кедрі, SJS iv: 457). Most likely, this is merely a localization of a Mediterranean dendronym and does not attest to a genuine currency of the word in this sense (pace Blažek/Janyšková 2015: 91). In any case, all of the modern languages are in agreement in meaning ‘yew’. Blažek/Janyšková (2015: 87) have suggested that the Slavic word may have been loaned from a Dalmatian *tis, which would be the regular reflex of Lat. taxeus ‘made of yew’. While an interesting suggestion, it is hampered, as the authors note, by the fact that this Dalmatian word is hypothetical (and the adjective taxeus does not appear to be continued in other Romance languages). More importantly, in a common Slavic loanword, one would undoubtedly anticipate a substitution *ĭ → *ĭ (trad. *ь), as in e.g. Cz. mše, SCr. dial. mȁša ‘mass’ (← ML missa, M. Matasović 2011: 114–115). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 266 chapter 7 formally similar (Slavic *s may derive from *ḱs, or perhaps from foreign *ks in a borrowing postdating the RUKI law). If Slavic *ī (trad. *i) in these cases derives from an earlier *ei, then we are dealing with an underlying alternation *a ∞ *ei. At face value, this could be understood as a combination of a ‘diphthongal’ alternation (like the one described by Schrijver) and a front–back alternation (as in 7.2). Naturally, it is possible that both *a and *ei derive directly from some other source like *ai; as usual, any explanation of this alternation will remain in the domain of speculation. † ‘hazel (1)’. Lt. dial. lazdà ‘hazel’, usually ‘cane, stick’, Lv. lazda; Pr E laxde ‘hazel’ ~ Uk. dial. лíска (usu. лiщи́ на), Cz. líska, SCr. lijèska ‘hazel’ — The comparison with Alb. lajthi ‘hazel’ (Meyer 1891: 234; Jokl 1923: 203–206; Huld 1990: 401) is suspect. As proven by the form lakthi in Dalmatian Albanian, an original cluster *-kθ- is to be reconstructed (see Demiraj 1997: 231–232). As *-kθ- is not the known reflex of any inherited cluster, -th- must be a suffix, while the stem *lak- cannot easily be compared with the Balto-Slavic data.72 The Baltic and Slavic forms were already compared by Miklosich (1886: 167), but the comparison is generally viewed with scepticism (cf. REW ii: 34; ALEW 652). Derksen (2008: 274), however, states that “there can hardly be any doubt” that the comparison is correct. I am not entirely convinced: even if we are willing to accept an alternation *a ∞ ai, for which a partial parallel could be the word for ‘turnip’ (cf. Lt. rópė ~ R рѣ́па, p. 237), we are still left with the obscure relationship between Baltic *-zd- and Slavic *-sk-. In principle, given the lack of old attestations, the Slavic forms do not appear to exclude a reconstruction *lēs-ukā- (trad. *lěsъka), i.e. a derivative of the noun in OCS лѣсъ ‘copse, thicket’ (REW loc. cit.; cf. also ЭССЯ xiv: 241 with lit.). Considering the vocalism and consonantism are both irregular, it is uncertain, despite Derksen, whether the similarity is sufficient to warrant a comparison at all. ‣ 7.4.2 ? Baltic *ā/ō ∞ *au Elsewhere ‘seal’. Lt. rúonis, Lv. ruônis ‘seal’ ~ OIr. rón, MW moel-rawn, Bret. reunig ‘seal’ (Ariste 1971: 10; Wagner 1981: 26; Sausverde 1996: 139; Stifter 2023, forthc.) — Considering that the similarity between the Baltic and the Celtic words seems obvious (Būga 1911: 37, 1922: 279), it is remarkable that the Celtic data is not even mentioned in most Baltic etymological dictionaries (LEW 746–747; Karulis ii: ‣ 72 The direct equation of Lt. lazdà with OIr. slat ‘rod, twig’ (Kroonen 2011b: 217–218; ALEW 652) is not possible, as the Celtic form must be reconstructed *slattā (Schrijver 1995a: 431; Matasović 2009: 345), cf. Modern Irish slat as against nead ‘nest’ (< *nisdo-). The connection with OCS лоза ‘vine’ (Berneker i: 736; REW ii: 43–44) is phonologically and semantically implausible. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 vocalism 267 129; Smoczyński 2018: 1115), which instead offer speculative root etymologies. The reason for this omission is that the Irish form was long ago explained as a Germanic loanword (Bezzenberger in Stokes 1894: 235; Pedersen 1909: 21; ME iii: 581). However, the suggested source, Old English hran ‘(a small kind of) whale’, is not phonologically suitable. The variant spelling hron shows the rounding of short /a/ before a nasal (Hogg 1992: 14), and demonstrates that we are dealing with a short vowel (see also the considerations of Stifter 2023: 183). The Celtic reconstruction is problematic. Stifter (2023: 183) has stated that the British and Goidelic forms cannot be united under a common proto-form, but he has later (Stifter forthc.) suggested the reconstruction *rauno-, comparing the homonym MoIr. rón ‘horse-hair’, MW rawn ‘coarse animal hair’ < *rauno- (cf. R руно́ ‘fleece’).73 If this reconstruction is valid, then the only way to compare the Baltic and Celtic forms would be to reconstruct *reh₂u-no- for Celtic and *roh₂u-n- for Baltic. However, even then, the development of *-oh₂u> Baltic *-ô- is highly suspect (Villanueva Svensson 2015). Since it is a priori questionable that Proto-Indo-European could have had a word for ‘seal’, such manipulations feel superfluous. Instead, it is more probable that Celtic and Baltic loaned their respective words for ‘seal’ from related sources. Considering that seals are marine animals, this must have occurred relatively late in both branches. ? ‘palate’. Lt. gomurỹs 3ᵃ ‘palate’, Lv. gãmurs ‘larynx; windpipe’ ~ OHG goumo · facia ‘gullet; throat’, MoHG Gaumen ‘palate’ (Derksen 2015: 184) — Almost all of the Germanic evidence points to *gōma(n)-, cf. ON gómr ‘palate; floor of the mouth’, OE gōma ‘palate; gullet’, OHG guomo ‘palate; throat’ (cf. Kroonen 2011b: 302), which would harmonize nicely with the Baltic data, allowing for the reconstruction of a shared proto-form *gʰeh₂m-.74 However, this fails to account for the High German evidence, for which various solutions have been proposed. In OHG, we find guomo beside goumo. The alternation -ou- ~ -uo- has been attributed to various reductions of a long diphthong *āu (Winter 1982: 183; Kluge/Seebold 336). A possible condition for this variation was suggested by Kroonen (2013: 185): in his opinion, the pre-Proto-Germanic diphthong *ōu ‣ 73 74 The a-vocalism raises problems for Slavic here, though, since the clear evidence for oxytone accentuation (cf. also Cz. rouno, SCr. (Vuk) rúno ‘fleece’; Зализняк 1985: 135; Derksen 2008: 440) is hardly consistent with an internal laryngeal. See the detailed discussion in Stifter forthc. The received connection with Gr. χάσκω ‘yawn, gape’ (IEW 449) appears to be contradicted by the initial g- in Baltic as opposed to the ž- in Lt. žiótis ‘open one’s mouth’. On the relationship of the Greek and Baltic forms, see Lubotsky 2011. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 268 chapter 7 developed to *ō in open syllables, while being shortened to *au in closed syllables (cf. also idem: xv–xvi).75 He therefore suggests that an earlier paradigm *gōumōn (> *gōm-), obl. *gōumn- (> *gaum-) could account for both OHG variants. However, it should be noted that OHG spellings in goum-, gaum- are rare,76 and a MHG continuation is uncertain (see MWb s.v. guome). MoHG Gaumen, which only becomes common in the 16ᵗʰ century (see DWb iv: 1576–1578), is usually assumed to continue the OHG by-form goumo. However, a number of dialect forms appear to suggest a prototype *gūman-, cf. Swiss (16ᵗʰ c.) gūme (Schw. Id. ii: 308), Cimbrian gaumo (cf. Schmeller/ Bergmann 1855: 39–40), Upper Saxon gaumen (DWb iv: 1577 under 3b), Prussian German gūmə (PrWb ii: 261), as well as MLG gume (Schiller/Lübben i: 165) and Lower Saxon gūmen (NdsWb ii: 135). The standard German form may in principle be derived from this preform, too. In that case, we might instead assume an old ablauting *gōman- : *gūman-, somewhat comparable to that observed in Go. fon, gen.sg. funins ‘fire’, reflecting an earlier *peh₂ur, obl. *puh₂n- (< *ph₂u-n-) (for more potential examples, see Kroonen 2011b: 319– 324). Either interpretation of the Germanic evidence appears to require a root containing *u, a reconstruction which is excluded by the Baltic evidence. This word may therefore possibly show evidence for an alternation *ā ∞ *āu. Note, however, that the interpretation of the Germanic ablaut alternation in laryngealistic terms may be anachronistic if we are actually dealing with a post-PIE loanword. Alternatively, we might interpret the continental Germanic evidence for *ū as indicative of a non-IE alternation *ā ∞ *ū. Whatever the solution, the Baltic and Germanic forms are difficult to combine in an Indo-European context. Due to ambiguities in interpreting the Germanic evidence, this cannot be considered a certain example of a diphthong alternation. 7.5 Length Alternations 7.5.1 Baltic Long ∞ Elsewhere Short ‘apple’. Lt. obuolỹs 3ᵃ, Lv. âbuõls ‘apple’ (beside Lt. obelìs, Lv. âbele ‘apple tree’, an old consonant stem); Pr. E woble; R я́ блоко, Cz. jablko, Sln. jábołkọ ‘apple’ ~ OE æppel, OHG apful; OIr. ubull, MW aval ‘apple’ (Kluge/Götze [1948]: 20– ‣ 75 76 Compare ON nór ‘ship (as a kenning); tempering trough (= Ic. nó-trog)’ (< *neh₂u-; idem: 391) : naust ‘boathouse’ (< *neh₂u-st-; idem: 384). 3× in AWb as against dozens in guom-, guam-; the interpretation of the hapax spelling gaom- is disputed, see EWAhd (iv: 562 with lit.), whose authors assume that goumo arose due to contamination with goumen ‘eat, feast’ (cf. also ALEW 400). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 vocalism 269 21; Hamp 1979: 163–166; Markey 1989: 599–600; Huld 1990: 398–400; Oettinger 2003; Kroonen 2013: 31) — The word for ‘apple’ occupies a curious position in the study of Indo-European. On the one hand, the word has often figured as a key example in the question of Indo-European l-stems (Fraenkel 1936b: 172– 176; Adams 1985; Olsen 2010: 76; Beekes 2011: 195; Stifter 2019: 204–207); on the other, it has not infrequently been regarded as a probable loanword from a nonIE source.77 The evidence for an ablauting l-stem comes from the following: Lt. obuolỹs and OIr. ubull < *abūl- (cf. in detail Stifter 2019) point to lengthened grade suffix *-ōl-, while Germanic *apla- and Slavic *ābl-uka- (trad. *ablъko/ъ) suggest a zero-grade *-l-. Words for ‘apple tree’ tend to show full-grade: Pr. E wobalne; MR яболонь (Шахматов 1915: 151), Pl. jabłoń, Sln. jáblana ‘apple tree’ and OIr. aball, MW avall ‘apple tree’ suggest *-al-n- (a-grade? cf. 7.6; or rather Slavic *-ol-n-, Celtic *-l-n-); ON apaldr also suggests *apal-(d)ra- as against West Germanic *apla-. Lt. obelìs points to *-el- which could potentially be secondary for *-ol-, as Lt. sẽser- ‘sister obl.’ ≪ *suesor- (cf. Skt. svásāram acc.sg.). Aside from the limited distribution, the argument for a non-IE origin essentially comes down to the presence of the phoneme *b. If one does not accept the existence of *b in PIE, the word must be interpreted as a borrowing; by contrast, if one does accept such a phoneme, the word is unproblematic (cf. NIL 264). Hamp (1979: 163) initially speculated that the long vowel in BaltoSlavic might be a reflection of some non-IE feature, but retracted this view in an addendum, preferring to evoke the recently discovered Winter’s law (cf. Winter 1978: 438). This has become the communis opinio (NIL 263), as the supposition of *b in principle explains both the Germanic and Balto-Slavic data. However, I do not consider there to be sufficient evidence for a phoneme *b in PIE (see Pedersen 1951: 10–12; Гамкрелидзе/Иванов 1984: 6–7; Lubotsky 2013; see also Olander 2020). Aside from Lt. trobà ‘peasant house’, where the vocalism is unexpected (see the following entry), all of the other examples of Winter’s law from *b are ambiguous at best. Thus, Winter’s only other example was Lt. grėb́ ti ‘rake, gather up; snatch’, where the secondary nature of the acute is shown by Lv. grebt ‘carve, hollow out’, R грести́ ‘rake up, gather together’,78 77 78 I will not get into the attempts to connect the families of Lat. mālum and Pashto maṇá (both ‘apple’) to this word, except to say that the supposed irregular change *b > *m (Blažek 1995: 17) or, conversely, *m > *b (Гамкрелидзе/Иванов 1984: 639–640; Cheung/Aydemir 2015: 85–86) are both completely ad hoc (cf. Kroonen 2016: 88). Kortlandt (1988: 393; followed by Derksen 1996: 321–322) has assumed the confusion of two roots, *gʰrebʰ- ‘to dig’ and *gʰreb- ‘to grab’, but the latter is based only on Lt. grėb́ ti (LIV 201; OCS грабити ‘steal, snatch’ belongs rather with Lt. gróbti; on the Germanic forms, see Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 270 chapter 7 and other cases do not inspire confidence, either.79 If we reject the phoneme *b, as I would recommend, then we should interpret the correspondences as irregular. The long, acute vowel in Balto-Slavic as opposed to the short vowel elsewhere can be viewed in this context (see in particular the following example). On the other hand, it cannot entirely be excluded that the word was borrowed into Balto-Slavic prior to Winter’s law. Note in this respect the fact that the word can be reconstructed to Proto-Balto-Slavic, and appears to show archaic ablaut, which would favour an early adoption. In this case, this alternation could not be characterized as one of length, and the only challenge to an Indo-European origin would be the necessity to reconstruct *b. ‘cottage’. Lt. trobà ‘peasant house; room’, HLv. tràba² ‘old, worn-out building; improvised hut’ (ME iv: 227; EH ii: 692); Oscan trííbúm acc.sg. ‘house’80 ~ OIr. treb, MW tref ‘residence, estate’; further cf. Go. þaurp ‘field’, OE þrop, þorp, OHG dorf ‘hamlet, estate’ (Hamp 1978: 187; Huld 1990: 398; de Vaan 2008: 626) — There is no necessity in including Lat. trabs ‘beam, tree’ in this etymology (see Ernout/Meillet 698; Untermann 2000: 766 contra Walde/Hofmann ii: 696–697; de Vaan 2008: 626, etc.). As with the word for ‘apple’, discussed immediately above, the analysis of this word is intrinsically linked to the status of the phoneme *b. In this case, both Oscan and Germanic provide independent evidence in favour of this reconstruction, and the long vowel in Baltic could be attributed to Winter’s law (Derksen 2015: 472). However, the Baltic *ā-vocalism also presents problems, as from an original *o lengthened by Winter’s law, we would anticipate *ô.81 ALEW (p. 1298) makes reference to “Neoablaut”, but this is difficult, as the word does not show any evidence of ablaut within Baltic (leaving aside the doubtful vien-trēb ‘alone’ beside -treĩb, cf. ME iv: 667). In a similar context, Derksen has referred to “the well-known East Baltic reshuffling of the ablaut relations” (2002: 9); however, to justify this position, we would need more concrete argumentation. As it stands, ‣ 79 80 81 Kroonen 2013: 187). For the secondary acute, as well as the variant grėb̃ ti (cf. 3pres. grẽbia in Alytus), compare also Lt. rėṕ ti (dial. rėp̃ ti) ‘take, rob’ beside rẽplės ‘tongs’, Alb. rjep, rrjep ‘skin, flay’ (cf. LIV 507, where the secondary nature of the acute is taken for granted; likewise ALEW 991; for further discussion of this kind of metatony, see Pronk 2012: 29–32). For Lt. drėb̃ ti (3pres. drẽbia), Derksen (2015: 138) is again content to assume contamination of two verbs, but the euphemistic sense ‘strike’ is hardly to be separated from senses such as ‘pour (e.g. porridge); make from clay; slouch’. On OCS слабъ ‘weak’, which is reconstructed *sleb- (LIV addenda s.v. *(s)leb-), see Kroonen (2011a: 258–259). The Umbrian hapax trebeit 3sg.pres. ‘lingers, dwells(?)’ is normally ascribed a short *ĕ (Buck 1904: 62; Untermann 2000: 759 with lit.). But as far as I can make out, the length is ambiguous. Buck (op. cit. 26) writes “oftenest there is no designation of the length” and on the spelling of *ē in Umbrian (p. 34), “i occurs frequently […] but e is far more common”. Lt. núogas ~ Skt. nagná- ‘naked’; púodas ‘pot’ ~ OE fæt ‘vessel, jar’, etc. (Winter 1978: 345). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 vocalism 271 it would appear that *a (see 7.6) and *b, both of which are of doubtful status, would have to be assumed if we are to attribute the length and acute accent to Winter’s law. On the other hand, Oscan also shows a long vowel in this word, suggesting that it may have some other origin. Either way, the relationship between Baltic *-ā̆-, Oscan *-ē- and Celtic *-eis impossible to adequately explain in Indo-European terms (Beekes 1969: 191), independently of whether one assumes an originally long vowel in Baltic or one that was secondarily lengthened by Winter’s law. For Baltic *a against *e elsewhere, see 7.2.2. On the other hand, the Germanic evidence is rather troubling, as at face value it seems to imply a syllabic *r̥ . This might be more consistent with an inherited origin (compare ‘furrow’, p. 208). It might be possible to view the Germanic *u vocalism in the context of the *e ∞ *u alternation discussed under 7.3.2, but this would require the additional assumption of the metathesis of *r (compare similarly ‘sturgeon’, pp. 236–237). ‘fresh’. Lt. prėś kas; R прѣ́сный, Sln. prẹ́sən ‘fresh, unleavened’ ~ OE fersc ‘fresh, unsalted’, OHG frisc ‘raw, fresh’ — It was previously assumed that BaltoSlavic reflected *proisk- and Germanic *prisk- (still Torp 1919: 135; Walde/ Pokorny ii: 89). However, since Būga (1922: 277) demonstrated that the supposed Lithuanian variant *príeskas does not exist, the etymology has largely been rejected, with the Germanic forms usually not even mentioned as possible comparanda (cf. Trautmann 1923: 231; LEW 652; REW ii: 429–430; Smoczyński 2018: 1018).82 It nevertheless seems difficult to imagine that the similarity is a mere coincidence in view of the precise agreement in meaning and correspondence of four consonants. The comparison can only be made by assuming parallel loanwords from another source. See also the discussion of Finnic *rē̮ska ‘fresh’, whose vocalism may also pose issues, in Chapter 3, fn. 159. [‘lynx’. Lt. lū́šis, Lv. lũsis, Pr. E luysis; R рысь, Sln. rȋs ‘lynx’ ~ OE lox, OHG luhs ‘lynx’ — See the detailed discussion on pp. 180–181.] ? ‘ash’. Lt. úosis, Lv. uôsis; Pr. E woasis; R я́ сень, Slk. jaseň, SCr. jȁsēn ‘ash tree’ ~ Lat. ornus; OIr. uinnius, MW onn (coll.) ‘ash tree’ (Machek 1954: 108, 1968: 217) — Bg. dial. о̀ сен need not imply a variant with *ŏ- in Balto-Slavic (pace БЕР iv: 936; Andersen 1996a: 142–143). Its distribution largely corresponds to that of осѝ ка ‘aspen’ (central Bulgaria, east of Sofia), which it was apparently influ- ‣ ‣ ‣ 82 The word is also omitted from Stang’s treatment of the Balto-Slavic–Germanic isoglosses. He does (1972: 40), however, adduce a pair which is remarkably similar, both belonging to a similar semantic field and showing the same correlation in vocalism: Pl. obrzazg, obrzask ‘tart flavour (of wine)’, R брѣ́згать ‘be squeamish, fussy’ ~ Nw. dial. brisk ‘sharp or bitter taste’. Note, however, that an ablauting *bʰroisg- : *bʰrisg- would indeed be possible here. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 272 chapter 7 enced by. The latter is a preserved archaism (cf. Cz. osika ‘aspen’), while dial. ясѝ ка ‘aspen’, on the contrary, shows the influence of я̀ сен ‘ash’ (thus already Zubatý 1892: 254 fn.).83 I prefer to keep ON askr, Arm. hacʿi ‘ash tree’ Alb. ah ‘beech’ (< *h₂esk-o-) apart. Kortlandt (1988: 391) has suggested to start from a paradigm *Heh₃-s- : *Hh₃-es-, a solution which has been followed by a number of Leiden-affiliated scholars (e.g. Schrijver 1991: 78; de Vaan 2008: 435; Kroonen 2013: 38). While such a paradigm seems possible on paper, it is difficult to imagine its survival into core PIE in a peripheral, non-basic vocabulary item.84 Furthermore, this is not the only possibility. It would be equally possible, both phonologically and morphologically, to assume a reduplicated *h₃e-h₃s- in Balto-Slavic as against *h₃e/os- elsewhere. Compare Lt. néndrė against Hitt. nāta- ‘reed’ (see p. 240, where other possible parallels are adduced). In this context, it is important to consider the similar correlation in the word for ‘elbow’. A long acute vowel is found in Lt. úolektis, Lv. uôlekts ‘ell (measure of length)’, and possibly Pr. E woaltis ‘ell, forearm’ matching Gr. ὠλένη ‘forearm’, while Lt. alkū́nė, Pr. E alkunis, OCS лакъть, R ло́ коть ‘elbow’ are not consistent with a laryngeal, and match Gr. ὀλέκρανον ‘point of the elbow’.85 Here, the reconstruction *Heh₃-l- : *Hh₃-el- (Kortlandt loc. cit.; Lubotsky 1990: 131–132) is even more uncomfortable, as it would have to have survived into the respective prehistories of Greek and Balto-Slavic, while *Heh₃l- : *Hh₃l- (Kroonen 2013: 22) is unlikely to work for the Balto-Slavic data. Thus, if we were to explain the long vowel of Lt. úosis (etc.) as the result of borrowing from a non-IE source, consistency would demand we use the same explanation for ‘elbow’. Yet considering that ‘elbow’ has plausible cognates at least in Indo-Iranian (Skt. aratní- ‘elbow, ell’) and more or less basic semantics, 83 84 85 Thus also SCr. jàsika beside jȁsēn, and Sln. jasíka and jásen reported in the same village by Erjavec (1883: 290). A variant with *e- perhaps underlies Sln. jesíka, whence jésen (and Kajkavian jȅsēn, cf. РСА viii: 741). Or does this variation have something to do with the frequent occurrence of ja- for je- in South and West Slavic (Andersen 1996a: 74–76)? On the association of ‘ash’ and ‘aspen’, see Normier 1981: 25–26 with lit. Note in this respect that the very similar word for ‘mouth’ *h₁eh₃-s, obl. *h₁h₃-s- still preserved its archaic paradigm in Hitt. ais (for *ās, cf. CLuw. āssa), obl. iss- (thus Kloekhorst 2008: 166–167), but was levelled in the rest of PIE, where it was probably reinterpreted as a root noun: Lat. ōs (gen.sg. ōris), Skt. ā́s-, OIr. (rare; cf. LEIA A-4) á ‘mouth’. (For alternative views on the Hittite form, see Melchert 2010 and NIL 388–389 with lit.). The short vowel is assured in all early attestations, while ὠλέκρανον is a corruption of later editors (Isépy/Primavesi 2014: 126–127). Lat. ulna ‘ell, elbow’, OIr. uilen ‘elbow’ and Go. aleina ‘ell’ reflect a short vowel, but are ambiguous as this may be the result of pretonic shortening (В. Дыбо 1961: 13, 25, 2008: 561; cf. Schrijver 1991: 352; Kroonen 2013: 22). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 vocalism 273 a non-IE etymology is not attractive. If the only thing separating úosis from úolektis is its semantics and geography, then its non-IE origin cannot be considered certain. [? ‘nettle’. Lt. notrė,̃ Lv. nâtre; Pr. E noatis ‘nettle’ ~ OSw. nätla, nätsla, OE netele, OHG nezzila ‘nettle’ — See the detailed discussion on p. 203.] † ‘harrow’. Lt. akėč́ ios, Lv. ecêšas, Pr. E aketes pl. ‘harrow’ ~ OE egeþe, OHG egida; MW oget ‘harrow’ (Oštir 1930: 15) — The alleged connection with Bel. dial. (cf. ДАБМ No. 233) асе́ць ‘a kind of drying barn’ (not a rack!) should be abandoned for semantic and phonological reasons. On the Slavic suffix, see now Pronk/Pronk-Tiethoff (2018: 285–286). Kroonen et al. (2022: 13) point out that Germanic can reflect *ageþjō-. However, there is still a mismatch with regard to the medial syllable, which is long (and acute) in Baltic. In the opinion of Pisani (1968: 19–20), the foreign origin of the Baltic harrow is supported by the similarity of harrows used in the Baltic to those used in Rome. Lat. occāre ‘till, harrow’ (and the late occa ‘harrow’ — a back formation?) has also usually been adduced here. It has been claimed that the Baltic vowel could be analogical after the verb seen in Lt. akėt́ i ‘to harrow’ (Топоров ПЯ i: 67–68). On the other hand, the verb, like OHG egen, eckan ‘to harrow’ has itself been seen as a potential back-formation, which is supported by the ja-present (Lt. akėj́ a, Lv. ecêju; cf. ALEW 13). One could alternatively assume secondary suffix replacement on the model of forms such as Lt. vežėč́ ios ‘one-horse cart’ (cf. Smoczyński 2018: 10). Although I think this word may well be of non-IE origin, the potential for analogy means that there is no certain evidence. See also Kroonen et al. 2022: 13. † ‘moss’. Lt. mūsaĩ ‘mould film (on beer, wine, etc.)’; Lat. mūscus ‘moss’86 ~ R мох, Cz. mech, SCr. dial. mȃh ‘moss’; OE mos, OHG mos ‘moss; swamp’ — Beside the Lithuanian evidence, Latin mūscus may also suggest an original long vowel, so we could suppose an alternation *ū ∞ *ŭ on this basis. On the other hand, the Latin form may equally reflect a full-grade *meus-, corresponding to OE mēos, Du. obs. mies ‘moss’. It is tempting to attribute the Baltic lengthened vowel to a secondary development (cf. Smoczyński 2018: 831), and indeed Būga (1914: 198–199; RR i: 585) has adduced Žemaitian evidence that would point to a short -u-,87 supported by the forms mọ̾ sᵃ acc.sg. and mọsuôtẹ ‘mould over’ ‣ ‣ ‣ 86 87 The Latin length is considered uncertain by Walde/Hofmann (ii: 134) and de Vaan (2008: 397) and the vowel is given as short by TLL. While metrical evidence is lacking, the long vowel is clearly demonstrated by the Romance reflexes (cf. Ernout/Meillet: 424), cf. Italian mùschio, Spanish musgo ‘moss’, etc. Musomìs aptraukė Salantai, musojaĩ Kvėdarna. As for ⟨Allus apmuſſójęs⟩ ‘Kahmicht bier’ Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 274 chapter 7 cited for Mosėdis by Vanagienė (2014: 455–456). Although the reason for this lengthened vowel is unclear,88 it is very unlikely that the two variants should be explained as parallel loanwords from a non-Indo-European source. Note in addition that the vowel here is circumflex, while other potential examples of length alternations show an acute. † ‘poplar’. Lt. túopa; Lat. pōpulus ‘poplar’ ~ R то́ поль, Slk. topoľ, SCr. dial. topòla ‘poplar’89 (Machek 1954: 132; Matasović forthc.) — The Lithuanian form, which is now part of the standard language, goes back to Būga (1908: 87; 1921: 433), where it is attributed to the East Aukštaitian dialect of Salakas. Gliwa (2008) is sceptical that this is an inherited word, and considers it more likely we are dealing with a clipping of the Slavic loanword tópelis. As for the -úo-, Gliwa’s assumption of an original Žemaitian form jars with the reported East Lithuanian distribution. The vocalism could, however, be explained as a dialectal adaptation of literary short /ò/.90 Therefore, despite disagreeing in the details, I would support Gliwa’s suggestion, and suspect that túopa may indeed be a dialectal neologism based on a Slavic loanword.91 ‣ 7.5.2 Baltic Short ∞ Slavic Long ‘iron’. Lt. geležìs, Lv. dzèlzs (dial. dzelezs), Pr. E gelso ‘iron’ ~ OCS желѣзо, Sln. želẹ́zọ ‘iron’ (Mikkola 1903: 41; Meillet 1909: 70; Machek 1968: 725; ALEW 351) — ‣ 88 89 90 91 (Mielcke ii: 291), it is tempting to assume an error; cf. the immediately preceding “Kahmicht muſótas”. In the LKŽ, all of Būga’s examples have been corrected to mūs- (cf. Mūsõms aptraukė cited for Salantai s.v. mūsà), but there is no reason to doubt their reliability. It would seem most promising to start from the verb (ap-)mūsóti -ja ‘become covered with a mould film’, where for the lengthened grade we could compare iterative formations like bylóti -ja ‘speak’, although the verb in question does not have an iterative meaning. The Slavic word is often considered a borrowing from Latin (e.g. Ernout/Meillet 924; Machek 1968: 647); however, finding a suitable source form presents difficulties (see REW iii: 121). Compare Salakas forms such as puoľka (= literary pòlka) ‘a dance’, kaľiduõras (= virtual *kolidòras, literary korìdorius) ‘corridor’ (Zinkevičius 1966: 69–70). In a similar area, we find tòpalas ‘poplar’ Kazitiškis (just 15km from Salakas), tòpolis Kupiškis (LKŽ). Some of these forms from the LKŽ might even be normalizations of dialectal /túop-/, but note that Vosylytė (2013: 377) only cites forms with short /ò/, e.g. tↄ̾.pↄ.l’ↄ. gen.sg. Kupiškis. On the other hand, Gliwa (2008: 241), is rather dismissive of the LBŽ’s further citations from the South Aukštaitian dialects of Seiniai and Alytus, stating that these may ultimately trace back to Būga, but without evidence. Another of Gliwa’s arguments is that the poplar (Populus alba and Populus nigra) is not found in Lithuania. This statement is consistent with the distribution maps on https://euforgen.org/, but not with those of the European Atlas of Forest Tree Species, where both species are marked as native to Lithuania. The only widespread dialectal term for ‘poplar’ quoted in the LKŽ which is not loaned from Slavic то́ поль is jõvaras — another Slavic loanword (Skardžius 1931: 90; LEW 195). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 vocalism 275 In Lithuanian, there is rather a lot of evidence for an original root noun; particularly note the nom.pl. gẽležes recorded in several dialects which otherwise only appear to have old consonant stems.92 However, an ablauting *ǵʰ-stem (Tremblay 2004) can hardly come into question here, first of all due to the lack of parallels, and secondly due to the acute intonation in Slavic (R желѣ́зо, cf. Зализняк 2019: 508; SCr. (Čak.) želȅzo; cf. Derksen 2015: 555). Thirdly, there is the obvious chronological issue of reconstructing an archaic Indo-European nominal paradigm for a designation for ‘iron’. The difference in vowel length would rather speak in favour of the word entering Baltic and Slavic independently. On the various unsuccessful external comparisons, particularly with Gr. χαλκός ‘copper, bronze’, see Thorsø et al. (2023: 113). ‘ruffe’. Lt. dial. (S Aukšt.) ežgė,̃ also rarely ežegỹs (cf. eʒʒ́égis, Ruhig ii: 220) ‘ruffe’; Pr. E assegis · persk93 ~ Kash. jôżdż (gen.sg. jażdża), Pl. jazgarz, Cz. ježdík ‘ruffe’ — The Slavic forms require a reconstruction *ēzg- or *āzg-. Although the word has a limited distribution within Slavic, the discrepancy in vowel length rules out the possibility of a Baltic loanword. Derksen (2008: 155; 2015: 159) states that the -g- in Lithuanian “may be the well-known Baltic intrusive velar”. This can clearly not be correct, first and foremost due to the trisyllabic form attested in Prussian and Lithuanian dialects.94 ALEW (p. 309) explains the Slavic vocalism as due to the influence of Pl. jaź, Cz. jesen ‘ide’, yet this is a very different kind of fish (cf. Sławski SEJP i: 533). While the traditional etymological comparison with ežỹs ‘hedgehog’ might be semantically acceptable (Trautmann 1910: 305; Derksen 2015: 159), it cannot be substantiated without ad hoc morphological assumptions. ‣ 92 93 94 The form is widespread in Uteniškiai dialects: Dusetos, Užpaliai, Debeikiai (Zinkevičius 1966: 264), Leliūnai (Papildymų kartoteka). From these dialects, Zinkevičius otherwise cites only nom.pl. dùres ‘door’, díeveres ‘brothers-in-law’, óbeles ‘apples’, vóveres ‘squirrels’ and (from Debeikiai) aũses ‘ears’. All of these are probably or possibly old consonant stems. The form gẽležes is also cited from the South Aukštaitian dialect Seinai, and is the only form cited by Zinkevičius from this dialect (from the LKŽ we can also add dùres nom.pl. ‘door’ and žuvès gen.sg. ‘fish’). I therefore do not think that the ALEW (351) is justified in calling the consonant stem inflection secondary here, despite the i-stem inflection in the earliest texts. See also Tremblay (2004: 239). Interpreted by Trautmann (1910: 305) as Perca fluviatilis, i.e. ‘perch’ (thus also Endzelīns 1943: 145; Топоров ПЯ i: 133; PKEŽ i: 104), but as correctly noted by ALEW (309), it can hardly be excluded that the actual meaning of the Prussian word was ‘ruffe’, which is considered a kind of perch in German folk taxonomy (“Kaulbarsch”). But even without these forms, the idea that ežgė ̃ should somehow be a back formation(?) from egžlỹs (attested lexicographically, cf. egʒ́lys, Ruhig ii: 220), which has “preserved the original constellation” is implausible. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 276 table 17 chapter 7 Possible examples of long ∞ short alternations Baltic Slavic Germanic Elsewhere ‘apple’ *âBōl- *âBl- *abl- Celt. *aBōl- ‘cottage’ *trâB- – *t(u)rb- ‘fresh’ *prêsk- *prêsk-n- *prisk- ‘lynx’ *lûḱ- *[l]ûḱ- *luk-s- ? ‘ash’ *âs- *ôs-en- – ? ‘nettle’ *nât- *nât- *nad- ‘iron’ *Geleǵʰ- *Gelêǵ⁽ʰ⁾-o- – ‘ruffe’ *eǵʰegʰ-i̯- *ēźgʰ- – It. *trēbCelt. *treB- Gr. *lunkIt. *os-VnCelt. *os-nCelt. *ninat- 7.5.2.1 Conclusion The certain and possible evidence long ∞ short vocalic alternations is collected in Table 17, above. As in previous tables, long vowels which turn up as acute in Balto-Slavic are written with the caret ⟨ ˆ ⟩ (see p. 185 for more help reading the table). Forms which do not provide relevant data are presented in light grey. Shaded cells indicate forms containing long vowels. Quite a large number of examples have been identified which show an unexpected long vowel in Balto-Slavic by contrast to other European comparanda. In every case, the vowel is acute, and remarkably, a Proto-Balto-Slavic reconstruction is possible, suggesting that we are dealing with a relatively significant time depth. Where we find an acute vowel, it is possible that something other than length is responsible, such as glottalization. 7.5.3 i/u ∞ ∅ ‘alder’. Lt. al̃ksnis, dial. (NE) alìksnis (cf. p. 33), Lv. àlksnis; R ольха́ , Pl. olcha; ON ǫlr, OHG elira ‘alder’ ~ Lat. alnus ‘alder’ (Machek 1954: 130; Polomé ‣ Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 vocalism 277 1990: 334; Huld 1990: 401–402; Derksen 2002: 6, 2008: 307; de Vaan 2008: 34–35; Kroonen 2013: 22; Pronk 2019a: 154; Matasović forthc.) — The Slavic vocalism presents difficulties. South Slavic in general suggests *elixā- (trad. *( j)elьxa): SCr. obs. jelha (> jóha; Skok i: 772), Sln. ję́łša, Bg. елша̀ , dial. елха̀ ‘alder’. Polish olcha and Cz. olše demand initial *a- (trad. *o-), as does apparently East Slavic, where one usually anticipates the preservation of *e- before *ĭ (trad. *ь) in the following syllable (Шахматов 1915: 140–141; REW i: 389, s.v. ёвня). It has been suggested that some forms may result from contamination with the word for ‘spruce’, cf. R dial. ёлха ‘alder’ beside ёлка ‘spruce’ (Kortlandt apud Schrijver 1991: 41), and Bg. елха̀ ‘conifer, fir tree’, dial. ‘alder’. While it is unlikely that contamination with the word for ‘spruce’ can explain all cases of *e- (Derksen 2008: 370),95 the evidence is difficult to evaluate in view of the more general problems with initial vowels in Slavic (cf. p. 244 and Andersen 1996a: 128–130). A more remarkable issue is posed by the Latin form. While all the remaining data points to *alis-, Latin is only consistent with a reconstruction *als-no(Walde/Hofmann i: 31; Schrijver 1991: 42). This can be considered clear evidence for an irregular alternation *-i- ∞ *-∅- and therefore offers some empirical support for the hypothesis that the word for ‘alder’ is of non-IE origin. ‘ground elder’. Lt. garšvà (dial. gáršva), Lv. gãrsa ‘ground elder’ ~ OHG gires · macedonicum; cf. gierisch 1604 ‘aegopodium’, giersig 1616 ‘wild angelica’, modern Giersch ‘ground elder’ (DWb vii: 7388–7389); MLG gers, gersele · grot petercilie (MoLG Heers with unclear anlaut, cf. Marzell i: 125) — Although the Germanic and Baltic forms are usually compared without question (e.g. IEW 445; EWAhd iv: 370–372), the almost consistently disyllabic form in OHG (AWb iv: 285), which can hardly be explained as svarabhakti (cf. Reutercrona 1920: 137, 169), as well as perhaps Early MoHG gierisch (see above) and Swiss dial. Gerrist (Schw. Id. ii: 404), seem to suggest a disyllabic preform, e.g. *gʰeru/is- or *gʰirVs (where *V is not *a), which cannot easily be compared with the Baltic words. Thus, if the comparison is correct, we are dealing both with an irregular loss of the second syllable vowel in Baltic, which can hardly be explained in Indo-European terms, as well as an alternation between a front vowel in Germanic and back vowel in Baltic (on which see 7.2.2). ‣ 95 The Russian dialectal evidence is in fact more complicated, as we also find forms like dial. ело́ха (РЭС xi: 325–326). However, even if we assume an original *eluxā- (trad. *( j)elъxa), this dialectal variant must in any case be analogical (after gen.pl. ело́х). Rather than multiplying entities, I suspect that this form is ultimately the result of a dialectal hardening of /l’/ in the sequence /l’x/, although more evidence would be desirable (Мирская apud Касаткин 1999: 177 mentions the dialectal forms О́ lга ‘Olga’, ско́ lко ‘how much’, but the distribution of this phenomenon is not clear to me). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 278 chapter 7 ‘aspen’. Lt. ẽpušė, dial. ãpušė ‘aspen’ ~ Lv. apse; Pr. E abse; R оси́ на; ON poet. ‣ǫsp?‘aspen’ , Arm. dial. opʿi ‘poplar’ (for refs. and further discussion, see p. 219) — In view of the Lithuanian data, the Latvian word has been derived by syncope from an earlier *apuse (e.g. Schulze 1913: 288; Smoczyński 1989: 40). However, syncope does not generally occur in Slavic, so here *aps- really is required, in line with Armenian. Būga (1922: 226) has suggested that the Lithuanian word arose through contamination with pušìs ‘pine’, an idea which has been taken over in other etymological works (Trautmann 1923: 11–12; LEW 14; Smoczyński 2018: 40; note also ALEW 45). However, a change *-ps- > *-puš- involves both a vowel epenthesis and a change in sibilant quality,96 and seems hardly imaginable, especially since the trees in question are not very similar. Endzelīns (1943: 136; cf. Būga 1908: 118), starts instead from *aps-ušē-, with loss of the first *s by dissimilation, but such a dissimilation would be unparalleled (cf. Zinkevičius 1966: 181–182). Already Hoops (1905: 123–124) drew attention to a group of similar forms in the Turkic languages. Chuvash ăvăs ‘alder’ seems to reflect an earlier *abus (Räsänen 1969: 3; Мудрак 1993: 29) or *abïs (ЭСТЯ i: 607–608). The *b is also supported by e.g. Siberian Tatar ausaq, dial. (Tomsk) apsaq (Тумашева 1992: 25, 32) ‘aspen’, which reflects a derivative *abs-ak with regular syncope (cf. А. Дыбо 2007: 130). The similarity of Turkic *abus and the possible Proto-Baltic form *apuš- is striking, but since the Turkic *-u- may be due to anaptyxis in a final cluster *-bs (cf. СИГТЯ vi: 65), it does not unambiguously support the reality of this Baltic reconstruction. In fact, the reconstruction *abs rather than *abus might better account for Khakas os and Tatar usaq ‘alder’, which appear, at face value, to reflect Turkic *os (the expected Khakas reflex of *abus would be *ōs, with a long vowel; А. Дыбо 2007: 19). Hoops (followed by СИГТЯ iv: 131) had considered these to be Slavic loanwords, but it would be rather remarkable if Tomsk apsaq were unrelated to Tatar usaq, with its identical suffixation; moreover, a suitable Slavic source is unattested.97 A very similar correlation is found between Chuv. avăt- and Old Turkic öt- ‘sing (of birds)’. which is reconstructed *ebt- by Мудрак (СИГТЯ vi: 96 97 Būga had previously expressed the view that *s became *š after labials (1911: 3). This would help his case, but as it is clearly contradicted by Lt. vapsvà ‘wasp’, he had apparently already rejected the development by the time of this proposal, where he states explicitly that the expected Lithuanian form would be *apsė. Hoops suggests a Slavic donor form *osa; however, such a form is only attested as a relic in West Slavic, and does not occur in East Slavic at all. Note that the Belarusian аса́ cited in ЭССЯ (xxxii: 93) does not exist. In the original source (Лемцюгова 1970: 7), it is only a reconstructed form (*АСА sic.) based on toponymic evidence. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 vocalism 279 166; in his notation *ĕvt-), implying a reduction *eb- > *ö- before a consonant. Therefore, one way to connect the Khakas/Tatar forms would be to assume a parallel development *abs > *ŏs. Perhaps more problematic, and something which does not seem to have been noted, is the fact that we find *-b- in Turkic rather than *-p-. This might in fact be better accounted for by assuming a disyllabic donor form in which *p had become lenited intervocalically. We can recall here Hoops’ (loc. cit.) suggestion of an Iranian source, but are faced again with the issue that no trace of the word has been found in Indo-Iranian.98 The main issue with deriving the Turkic words from a (para-)Baltic source is the word’s broad distribution in Siberia. This could theoretically be accounted for by assuming a loan already into ‘dialectal’ Proto-Turkic (but from where?) or by assuming a later horizontal spread through the Turkic dialects, which could perhaps provide an alternative account for the irregularity within Turkic, but is difficult to substantiate in any detail. Compare also the discussions of the words for ‘honeycomb’ (p. 133), ‘mink’ (p. 143) and ‘elm’ (pp. 235–236). All in all, the word for ‘alder’ raises a number of problems which preclude its reconstruction, and it is possible that this could be explained by assuming parallel loanwords from an unattested source language. However, this does not really help to resolve the word’s problematic distribution within Siberian Turkic. ? ‘beehive’. Lt. avilỹs, (Žem.) aulỹs ‘beehive’, Lv. (Kurzeme) aũlis, also avelis ‘wooden beehive’ ~ R у́лей, Pl. ul ‘beehive’, Sln. ȗlj ‘hollow tree; beehive’ — The word is generally connected to Lat. alvus ‘belly’, alveus ‘hollow vessel’, Gr. αὐλός ‘pipe, hollow tube’ (Trautmann 1923: 18; LEW 25–26; REW iii: 181; ALEW 77– 78). On semantic grounds, the comparison can hardly be faulted, as Latin alvus is also used in the sense ‘beehive’. On the other hand, the Baltic forms are far easier to explain starting from *avil-, with the aul-forms deriving by syncope, and indeed, the Lithuanian variants have led Zinkevičius (1966: 138) to doubt the IE etymology. ALEW speculates that avilỹs is due to reanalysis on the basis of an unspecified root *au-, but the main issue is that the suffix -ilỹs is not ana- ‣ 98 Leaving aside the supposed connection with the Indo-Iranian word for ‘shovel; shoulderblade’ (Friedrich 1970: 50–52; Гамкрелидзе/Иванов 1984: 627; Šorgo 2020: 434), which rests on pure speculation (cf. KEWA iii: 547 with older lit.; Normier 1981: 24, fn. 21). The *b is also problematic to Pedersen’s suggestion of a “pre-Armenian” source (1906: 462) and А. Дыбо’s suggestion of a Tocharian source (2007: 130). The discovery of Arm. opʿi causes additional problems for both proposals with regard to vocalism. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 280 chapter 7 lysable. Therefore, despite the attractiveness of the IE etymology, the unclear -i- might suggest an irregular correspondence with Slavic, which could point to a foreign origin. † ‘hazel (2)’. Lt. kaſsula ‘Schaft’ (Bretke apud Bezzenberger 1877: 293), dial. (S Aukšt.) kasulà ‘plough shaft’ (LKŽ), kaſſúlas ‘Jägerspieß’ (Ruhig i: 213) ~ Lat. corylus (< *kosVlo-), OHG hasal, OIr. coll (< *koslo-) ‘hazel’ (Huld 1990: 401; Matasović 2013: 84, forthc.) — As Latin can simply reflect *kose/olo-, with suffixal ablaut, the irregularity depends entirely on the Baltic data, whose appurtenance is uncertain on semantic grounds. I therefore exclude this example. ‣ 7.6 IE *a It has been suggested that many of the words traditionally reconstructed with *a are rather loans from non-IE sources (Kuryłowicz 1956a: 194–195; Kuiper 1995: 65–68; Pronk 2019a: 154). The argument is essentially that most words for which *a has been reconstructed are geographically restricted, have a technical meaning, and often involve other irregularities. The following have already been regarded as probable loanwords on other grounds in the previous sections: – *ǵʰuak- ‘torch’ (p. 167): illegal root structure99 – *dʰraK- ‘dregs’ (pp. 190–191): *gʰ ∞ *k, illegal root structure – *dʰalK- ‘scythe’ (p. 191): *gʰ ∞ *k, illegal root structure – *kanaP- ‘hemp’ (pp. 206–207): *b ∞ *p, *nn ∞ *n – *bʰar(s)dʰ- ‘beard’ (pp. 224–225): *sdʰ ∞ *dʰ – *bʰaB- ‘bean’ (pp. 228–229): *bʰ ∞ *w – *trā/ēb̆ - ‘cottage’ (pp. 270–271): *ā ∞ ē, IE *b Still, given the continued disagreement as to whether *a should be reconstructed, it might seem overly dismissive to label any word appearing to suggest the reconstruction *a (for which the most decisive evidence comes from ItaloCeltic and Greek) as being of non-IE origin, especially where no other evidence supports this hypothesis (compare the similar considerations with regard to the word for ‘apple’ on pp. 268–270). Here, I will briefly treat a few words which fall into this category: ? ‘post (2)’. Lt. stãbas ‘pillar; idol, statue’; ON stafr ‘staff, cane; post, support’ ~ OIr. sab ‘pole, stake’ (Beekes 2000b: 12) — The Irish word is unlikely to be bor- ‣ 99 Schrijver (1991: 465) has considered Lat. fax an example of the unrounding of *wo in an open syllable (pre-Latin *fwak- < *fwok-), but the development is clearly contradicted by Lat. forum ‘open space’ (< *fworo- < *dʰuoro-; cf. idem: 472), and can hardly be accepted. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 vocalism 281 rowed from Old Norse due to the unique substitution st- → s- (cf. Marstrander 1915: 97, 125). A reconstruction *stHbʰ- would be unproblematic for Germanic and Celtic, but whether it could account for the Baltic data is disputed. According to one widespread current view, in Balto-Slavic, “a laryngeal was never vocalized” (Beekes 1988: 23; cf. also Lubotsky 1981: 89; Smoczyński 2006: 187– 188). On the other hand, it seems the evidence is not exactly decisive. The “classic” view, at least, states that the Indo-European ‘schwa’ yielded Balto-Slavic *a (cf. Brugmann 1897: 177; Arumaa 1964: 80–81; Stang 1966: 22; Matasović 2008: 89), for which the Paradebeispiele — Lt. statùs ‘upright; steep’ (= Gr. στατός ‘standing’) and OR споръ ‘abundant’, Cz. sporý ‘stocky; abundant’ (= Skt. sphirá‘fat’) — still maintain much of their initial appeal. The reluctance in accepting this sound law seems mainly to be based on the small number of examples, but as long as no counter-evidence exists, it cannot be rejected out of hand.100 I therefore agree with Villanueva Svensson (2008: 12) that the issue is in need of “a full and unprejudiced study”.101 ? ‘mast’. OCS мостъ, SCr. mȏst ‘bridge’; OE mæst ‘mast’, OHG mast ‘mast, pole’ ~ OIr. maide ‘stick, staff; beam’; Lat. mālus ‘mast, pole’ (Kuryłowicz 1956a: 195; Pronk 2019a: 151) — Again, a reconstruction *mHsd- is possible at least for the extra-Balto-Slavic evidence (Schrijver 1991: 167). The Slavic *t might be more elegantly explained by positing a Germanic loanword (Stender-Petersen 1927: 281–283; Matasović 2008: 50; Kroonen 2013: 357), although this is uncertain for semantic and accentological reasons (REW ii: 163; Pronk-Tiethoff 2012: 180). ‣ 100 101 Alternative explanations for both have been suggested, but the question is, even if these explanations are in principle possible, are they an improvement on the traditional etymologies? Since Kortlandt (1982: 26), in my view rightly, rejects the notion that OR стръıи ‘father’s brother’ is derived from *ph₂tr-, it is unclear on what basis he is opposed to the development. He takes statùs as a derivative of statýti ‘put (upright)’, in turn from a reduplicated *ste-sth₂- (Kortlandt 1989b), but the opposite derivation seems more likely from a Baltic perspective (cf. Smoczyński 1999: 23); moreover, in Kortlandt’s scenario, the -a- still has to be explained as secondary. For споръ, the reconstruction *su-para- (i.e. *sъporъ; Kortlandt 1980: 352) appears to be in contradiction to the attested Old Russian evidence (cf. СРЯ 11–17 xxvii: 72). Other examples to consider here are Lt. mãtas ‘measure’ <? *mh₁-to- (Darden 1990: 63; Smoczyński 1999: 23), and OCS столъ ‘throne, bench’ (beside Go. stols ‘seat, throne’), cf. Smoczyński (1999: 20). For alternations such as CS носъ ‘nose’ : Lt. nósis and OCS соль ‘salt’ : Lv. sā̀ls, Matasović (1997: 135) has reconstructed *nh₂s-, *sh₂l-. At first sight, this indeed seems preferable to assuming the preservation of an extremely archaic paradigm *naʔs- : obl. nʔas- into Proto-Balto-Slavic (Kortlandt 1985: 119), but Sanskrit nás- ‘nose’ must also be accounted for (cf. Lubotsky 1981: 90). For ‘salt’, an equally possible reconstruction is *sh₂-ēl : *sh₂-el-, provided Eichner’s Law is rejected (see Pronk 2019b: 144–145). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 282 chapter 7 Admittedly, assuming a suffixed formation *mazd-to- for Slavic is not much of an improvement (Kiparsky 1934: 47; Derksen 2008: 326–327). ? ‘corner’. Lt. kam̃ pas ‘corner, angle’ ~ Gr. κάμπτω ‘bend (the knee), turn back (a chariot)’ (Kuryłowicz 1956a: 195; Beekes 2000a: 28) — Here, a reconstruction *kh₂mp- is improbable, if not impossible (Beekes 2010: 632). On the other hand, the possibility that Gr. κάμπ- is secondary for *κάπ- with the analogical restoration of -μ- from the full-grade (Pronk 2019a: 149) remains plausible, if somewhat convoluted (cf. χανδάνω ‘hold’ for *χαδάνω < *gʰnd-nH- beside perf. κεχόνδει). OIr. camm, MW cam ‘crooked, bent’, if they belong here, could possibly reflect a zero-grade *kmp- (with a development *-mp- > *-mb-; Thurneysen 1946: 117). ‣ ∵ In conclusion, none of these potential examples of *a are entirely watertight, and therefore the question as to whether the apparent presence of such a phoneme is sufficient to prove a non-Indo-European origin need not be answered here. However, the characterizations of authors such as Kuryłowicz do seem to be generally valid, and I consider this to be another potential criterion which could favour a non-Indo-European origin, at least where other evidence is available. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 chapter 8 Analysis In the above pages (including two examples in 5.3.1), I have discussed 92 word families which might plausibly be explained as loanwords from unattested non-Indo-European sources. In 16 cases, it was found that the evidence is too ambiguous or uncompelling, and these cases will not be fed into the further analysis. Of the remaining examples, I have considered 46 to be probable loanwords, and a further 30 have been accepted as possible, but uncertain. In this section, I will analyse the data from an extra-linguistic perspective, as well as attempt to draw some broader conclusions about the dataset as a whole. In this context, the certain cases will be used as my core data set, with uncertain examples only being incorporated where this could provide additional useful information. 8.1 Semantics The majority of the words treated here fall into the following broad semantic categories (uncertain cases are listed in square brackets): a. Wild animals (12+7): Mammals: bison, lynx, roe (deer) [+ badger, boar] Birds: bird of prey (see 5.3.1), grouse, oriole, pigeon [× 2], swan (× 2) Aquatic animals: ruffe, seal, sturgeon [+ cod, frogspawn, salmon, fishing trap] b. Wild plants (11 +8): Trees: alder, hornbeam, maple [+ ash, aspen, elm (× 2)] Tree parts: leaf, nut [+ bast] Edible plants: (wild) carrot, goosefoot, ground elder, ramsons [+ nettle] Other: (false) hellebore, reed [+ heather, sedge] c. Cultivated plants and agriculture (9+5): Crops: bean, hemp, oats, rye [+ lentil, millet] Fruits and vegetables: apple, garlic, pear, turnip Agriculture: scythe [+ furrow, ploughshare, aftermath] © Anthony Jakob, 2024 | doi:10.1163/9789004686472_010 Anthony Jakob - 978-90-04-68647-2 This is an open access chapter distributed under the terms of the cc by-nc-nd 4.0 license. Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 284 chapter 8 These three categories already account for 70% of the certain examples. Beyond this, three relatively clear semantic groups can be identified with at least two certain examples:1 d. Apiculture (3 +2): drone, honeycomb, wax [+ bee, beehive] e. Structures (2+2): cottage/estate, oven [+ mast, post] f. Metallurgy (2): iron, silver Each of these six categories will be discussed below in more detail, but first I would like to point out some absences. Most remarkable here is the absence of geographical terminology and terms for natural phenomena, especially considering that these semantic areas have received much attention from other researchers focusing on questions of substrate contact (cf. Kalima 1919: 257– 258; Bertoldi 1932: 94; Ariste 1971: 9–10; Polomé 1986: 662; Saarikivi 2004; Aikio 2009: 41, 2012a: 85; Beekes 2014: 47–51). The only word fitting into this category is the uncertain case ‘lightning’, but even here the precise semantics might point towards borrowing in a religious or cult context (see p. 202). While it is true that some suggestions of non-IE origin have been made in connection to Baltic geographical terms (e.g. ‘meadow’, p. 198), I have found none of these to be compelling.2 Another semantic category which is under-represented, although perhaps less surprisingly, is that of animal husbandry. Outside of words connected with apiculture (on which see 8.1.4, below), the only term in my corpus which falls into this category is the adjective ‘in calf’. As in many Indo-European languages, the Baltic lexicon for domestic livestock is conservative, with most important terms being directly inherited from the proto-language.3 We have also observed that several words connected to livestock breeding were loaned into ProtoFinnic (see 3.6.1), and the main foreign source for words in this semantic field appears to have been Germanic (see Chapter 2). This points towards a continuity in animal husbandry practices among Baltic-speaking populations since Proto-Indo-European times, and relatively advanced stockbreeding practices compared to their non-Indo-European neighbours. 1 With one certain example, we can also note body parts: beard [+ heel, palate]. Note in this context Ariste’s mention of “somatic words” as good candidates for substrate loans (1963: 17). 2 Lt. mãrios ‘sea; (Curonian) lagoon’ has often been considered to be of non-IE origin (Nehring 1959; Hamp 1979: 162–163; Sausverde 1996: 136), but since Latin mare ‘sea’ has been regarded as either a regular cognate (Schrijver 1991: 474–475) or analogical (Vine 2011), this word has not come into consideration here. See also Chapter 7, fn. 50 on Lt. pélkė ‘marsh’. 3 For instance, Lt. avìs ‘sheep’, ožỹs ‘goat’, Lv. gùovs ‘cow’ (= Skt. ávi- ‘sheep’, ajá- ‘goat’, gáv‘cow’), Lt. par̃šas ‘piglet, castrated boar’ (= Lat. porcus ‘pig’). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 analysis 285 The remaining certain cases are difficult to group together in any meaningful way, largely because their meanings are too general to be categorized, or because they cover multiple possible semantic fields. For instance, the words for ‘dregs’ (also ‘yeast’) and ‘fresh’ (also ‘unleavened’), as well as ‘oven’ (categorized here under Structures) could all be associated with breadmaking, but since the attested meanings for each term are not limited to this semantic domain, such a grouping is too optimistic.4 8.1.1 Wild Animals Terms for animals have often been mentioned as especially strong candidates for borrowing in substrate contact situations (e.g. Bertoldi 1932: 94; Philippa et al. i: 22; Matasović 2013: 76). In this context, it is notable that the words for wild animals show a more limited distribution in comparison to other semantic categories. Out of 12 probable cases, 9 are limited to Baltic, Slavic and Germanic. This suggests that we are dealing with localized terms rather than extensive horizontal transmission, which is consistent with a substrate mechanism. Table 18, overleaf, illustrates the distribution of the most certain cases in this semantic category.5 Forms for which a common proto-form can (theoretically) be reconstructed have been enclosed in dotted lines. Thus, the word for ‘lynx’ is potentially reconstructible for Balto-Slavic, and also for GraecoArmenian (note that the Balto-Slavic and Graeco-Armenian reconstructions are not reconcilable). As regards the motivation for the borrowing of animal names, it seems natural to assume that words for local species for which no term was previously available would be most prone to adoption (Haspelmath/Tadmor 2009: 51; Aikio 2012a: 85). However, this explanation can only apply in the minority of the cases in our corpus. One such case is ‘lynx’: the animal’s current range does not extend beyond the forest steppe, and as the animal’s preferred habitat is dense woodland (Nowak/Paradiso 1983: 1072),6 it is unlikely it would have frequented the grasslands further south. The animal is absent in the steppe, but is recorded in the Neolithic from Trypillia (Mallory 1982: 208), a culture with 4 Both Baltic and Slavic borrowed the Germanic word for ‘bread’ (see p. 38), although Lithuanian has also preserved an older inherited term, dúona ‘bread, loaf’ (LEW 111). Latvian màize ‘bread’ is derived from mìeži ‘barley’. Other uncategorized words are ‘thousand’, ‘torch’ (p. 167) and the uncertain cases ‘circle’, ‘corner’, ‘fast’ and ‘people’. 5 The abbreviations used are as follows: B = Baltic, S = Slavic, G = Germanic, C = Celtic, It = Italic, Gr = Greek. Under “+”, I have indicated all other comparanda (with the usual abbreviations). 6 To establish current distribution, in addition to the references cited, I have referred to the data on the IUCN Red List website (https://www.iucnredlist.org/); for birds, this data has been supplemented with the maps from Birds of the World (https://birdsoftheworld.org/). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 286 table 18 chapter 8 Distribution of borrowed animal names bison lynx roe bird of prey grouse oriole pigeon (1) swan (1) swan (2) ruffe seal sturgeon B S G ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ? ✓ ✓ ✓ ✓ ✓ ✓ C ✓ ✓ ✓ ✓ ✓ It ✓ ✓ Gr + ✓ Arm. Eg. ✓ which Indo-Europeans potentially came into contact during their early fragmentation (Kroonen et al. 2022: 33–34). Note that the wide distribution and possibility of reconstructing common proto-forms for multiple branches might suggest that this word was borrowed comparatively early. Lexical gaps could also account for the borrowing of a word for ‘seal’, an animal which is not found inland, and perhaps also ‘sturgeon’. The sturgeon is anadromous, meaning it migrates upriver to spawn. Migrations are usually relatively short, but as much as 1000 river kilometres may occasionally be travelled (Holčík et al. 1989: 376; Brevé et al. 2022: 1164–1165). In addition, the now endangered stellate sturgeon previously spawned in river basins across the Pontic-Caspian area (cf. Mallory 1983: 267, 275). Nevertheless, the adoption of a foreign term could have been motivated by differences in species, a geographical gap in the distribution of sturgeon species, or changing dietary habits among migrating populations, which might have caused the original term for the animal to have been lost.7 However, most of the animals discussed here must have been known to speakers of Indo-European. Among the mammals, the roe deer, as well as the badger and wild boar, are widely distributed throughout Europe and are 7 Similar considerations could apply to the word for ‘salmon’, which may have originally referred to the anadromous salmon trout (Diebold 1976). Note that Mallory stresses the paucity of salmonid remains in the Pontic-Caspian region (1983: 268). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 analysis 287 also present in the Pontic-Caspian steppe (cf. Mallory 1982: 206–207, 211, 212; Bellquist 1993: 336–337; Anthony 2007: 175). The same can be said of the bison, which still occurred east of the Dnieper into the Middle Ages (Benecke 2005), although admittedly does not appear to have been very frequent in the steppe since the Neolithic (Mallory 1982: 213). Among the birds, the golden oriole breeds throughout all of Europe including the steppe, and the same is true of the wood pigeon (cf. Mallory 1991: 231). The rock dove also occurs natively in the steppe; the exact vectors of spread of the domesticated and feral pigeon are difficult to trace, but it is now of course ubiquitous (Gilbert/Shapiro 2013). The mute swan breeds in many parts of the steppe, and was therefore presumably known to Indo-European speakers. It is common in Northern Europe, although its distribution admittedly becomes more patchy east of the Dniester. Among the fish, the ruffe has a very extensive distribution throughout Eurasia. The above facts make it is improbable that these terms were borrowed to fill lexical gaps within the Indo-European languages.8 It has been remarked that substrate loanwords tend to involve animals of low economic significance (Schrijver 1997: 295; cf. also Matasović forthc.). Perhaps this idea derives from an expectation that terms for economically important animals should rather be transmitted horizontally, for instance through trade.9 However, economic significance is rather a cultural and subjective phenomenon. Even migratory passerines, such as the golden oriole, may have economic value: in coastal Egypt, they are hunted for food and sold on as delicacies (Eason, Rabia & Attum 2016). We therefore should seek a cultural motivation for borrowing. In the case of wild animals, the most obvious cultural context is hunting. In the Eastern Baltic context, it has been suggested that the transition from a hunter-gatherer to a stockbreeding economy passed through a transitional stage where the reliance on hunting and gathering remained significant (Zvelebil/Dolukhanov 1991: 268 with lit.; Piličiauskas et al. 2017: 541), and cultural exchange might have been amplified by a later in-flow of hunter-gatherer-derived populations during the Bronze Age (Mittnik et al. 2018; Saag et al. 2019). In a context of language shift, 8 A word for ‘roe’ can probably be reconstructed for Proto-Indo-European on the basis of Gr. (Herodotus) ζορκάς (~ δορκάς) and MW iwrch ‘roe deer’ (< *iork-), cf. IEW 513. For the ‘swan’, only a common Italo-Celtic form can be given: Lat. olor, MW alarch (?< *h₁el-r-) (Schrijver 1995a: 76). 9 One does indeed find, for instance, a clustering of terms for insects and reptiles among the Finnic substrate terms in the Russian dialects (Мызников 2004: 113–116). Above, on p. 117, I have suggested that the common denominator between these animal terms might rather be their negative perception. This brings us back to the idea of low-status vocabulary (cf. Chapter 5, fn. 20). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 288 chapter 8 technical vocabulary associated with a particular economy may be transferred into the target language, and this is likely to be more robust where activities related to this economy continue to be practiced (Brenzinger 1992). The borrowing of the technical term ‘fishing trap’, if reliable, would also favour this interpretation. 8.1.2 Wild Plants Like animal names, plant names have often been viewed as central candidates for substrate borrowings (Bertoldi 1932: 94; Philippa et al. i: 22; Matasović 2013: 76; Soosaar 2021). The names for wild plants were a key focus of Václav Machek, and the irregularities observed in these can be seen as having given rise to his version of the substrate theory (Machek 1944–1946; 1950b; 1954; see p. 155). Although a small number of wild plant names show a narrow distribution comparable with that of the wild animals, they on the whole tend to exhibit comparanda outside of the Baltic region (see Table 19, overleaf). First of all, it is remarkable that several of the tree names which have come into question here have been assessed as uncertain, and have therefore been omitted from the table (thus ‘aspen’, ‘ash’ and two words for ‘elm’). Indeed, words for trees very often seem to show minor phonological issues, to the extent that irregularities have been viewed as a mere quirk typical of tree names (cf. E. Itkonen 1946: 306; Friedrich 1970: 108), an opinion which is perhaps justified by the perception of such terms as belonging to a ‘dialectal’ phase of IndoEuropean (cf. Hirt 1905: 189; Schrader/Nehring ii: 630; and e.g. Ernout/Meillet 23 s.v. alnus). A case could sometimes be made for such an interpretation, especially where identical forms are found in neighbouring branches (compare the example of ‘alder’, above). Aikio (2015a: 45–46) has argued that a number of West Uralic terms in this semantic field should in fact be explained as substrate words, noting that they show irregular sound correspondences: – F haapa, E haab; Sá. N suhpi, Ma. E šopke ‘aspen’ – F dial. vahtera, E vaher; Md. E ukštor; Ma. EW waštar ‘maple’ – F pähkinä, E pähkel; Md. M päšťä ‘nut, hazelnut’; Ma. EW pükš ‘nut’; Udm. paš-pu ‘hazel’ (pu ‘tree’) At the Sub-Indo-European workshop in Leiden, September 2021, Aikio has additionally adduced the word for ‘alder’ (F leppä, Sá. N leaibi, Md. E ľepe), which shows a clear resemblance to the Balto-Slavic word for ‘lime tree’ (see p. 89). What is remarkable is that these words also show a rather broad geographical distribution, with cognates found from Sámi to Mari or from Finnic to Permic. This probably implies that the words were adopted at a time when these branches were closer together, and it might be possible to talk of ‘dialectal Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 289 analysis table 19 Distribution of borrowed plant names alder hornbeam maple leaf nut carrot goosefoot ground elder ramsons hellebore reed B S G ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ C ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ It Gr ? + F Sá Ma Md ✓ ✓ ✓ ✓ Proto-Uralic’ here, too (the words for ‘maple’, for instance, can almost be treated as regular cognates). Of the uncertain cases, the words for ‘aspen’ and ‘elm (2)’ also have potential comparanda in Turkic.10 In this case, some kind of borrowing must certainly have taken place, but we are still left with the question as to where were these words adopted from, and what motivated their borrowing. As the hornbeam is not currently found in the steppe, and spread to southeastern Europe only during the Atlantic Period, being earlier restricted to Italy (Sauer 1988: 152–154), it is unlikely that the Indo-Europeans would have known this tree, and its borrowing might have been motivated by a lexical gap. The alder, on the other hand, is very widespread in Europe and should have been present in the steppe (cf. Friedrich 1970: 72–73 with lit.). The same can be said of the aspen, ash and elm. The motivation for borrowing must again have somehow been associated with differences in cultural practices. There are few reliably reconstructible words for specific trees, and it is quite possible that trees were of lesser importance to steppe pastoralists than to the European populations they replaced. This might be implied by the large-scale deforestation (or “steppification”) 10 Here we can note that Kroonen (2013: 39) has compared the European words for ‘aspen’ with F haapa, etc. I am not convinced, however, that there is sufficient similarity to warrant a comparison. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 290 chapter 8 of Northern Europe coinciding with the arrival of the Corded Ware Culture (Pelisiak 2016: 218–219; Haak et al. 2023: 71–72; Allentoft et al. forthc.). Although one can hardly conclude that the Indo-Europeans did not value wood, it may be suggested that distinguishing varieties of trees was not a top priority for steppederived pastoralists. I have divided the remaining plants into ‘edible’ and ‘other’.11 This division is somewhat arbitrary, as it is difficult to know what was interpreted as food by prehistoric populations. Cultural groups may differ in plant preferences, despite there being no significant difference in plant availability (Welcome/Van Wyk 2019). Reeds are fully edible, and may have been eaten, but since reeds also have numerous other uses (e.g. weaving mats, producing ropes), I have categorized them under ‘other’. On the other hand, nettles may also be twined into string and woven into textiles. My categorization as an ‘edible’ plant is partially influenced by the semantic shift to ‘vegetable tops’ observed for this word in Slavic (but this is not necessarily indicative of its earlier uses). Furthermore, the knowledge of which plants are poisonous (such as the hellebore) is obviously most vital to those gathering plants for consumption. Goosefoot, Chenopodium, is a plant whose remains are found in abundance at Yamnaya sites, with indications that it was eaten (Anthony 2007: 326, 439), so it is probable that Indo-European speakers had a word for the plant. One might assume that a decline in the consumption of this plant could have been associated with a shift towards cultivated cereals, although there is plenty of evidence of Chenopodium consumption even in Iron Age agricultural contexts (Kroll 1990; Behre 2008: 68–69; Ślusarska 2021: 189). Evidence for both wild garlic and nettles have been recovered in the Bronze Age Srubnaya Culture in the same region (Anthony et al. 2005: 408) as well as pollen belonging to Apiaceae (the family to which the carrot and ground elder belong). Plants from this family might have been consumed as vegetables in Western Russia already during the Neolithic (Kittel et al. 2020: 196). It appears that the borrowing of these plant names can in no case be confidently associated with a lexical gap; on the contrary, there is evidence that many of these species were actively consumed both in the steppe and in Europe. Notably, the wave of deforestation coinciding with the emergence of animal husbandry in Northern Europe actually coincides with an increase in evidence for both Chenopodiaceae and Urtica (nettles) in the palynological record (Pelisiak 2016: 218–219). 11 For information about plant uses, I have referred to the Plants for a Future database at https://pfaf.org/, where copious further references are provided. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 analysis 291 The borrowing of these plant names into dispersing Indo-European dialects, if not associated with a change in dietary preferences which may be the result of shifting subsistence practices, could indeed be connected to the principle of “low economic significance” signalled by Schrijver (1997: 295). Wild plants form a small but integral part of both pastoralist and agriculturalist diets (cf. Zanina et al. 2021; van Amerongen 2016: 215–226), but gathering of plants outside of a hunter-gatherer economy is presumably perceived as of secondary importance. On the other hand, we may consider a sex bias in the transmission of these terms. Ethnological evidence shows that plant gathering and preparation is cross-culturally most often the sole domain of women (Murdoch/Provost 1973: 207, 210).12 A male sex-bias in the migrations of steppe-derived populations, combined with female exogamy (Knipper et al. 2017; Saag et al. 2017; Mittnik et al. 2019), would provide a plausible context whereby indigenous terms, passed down from mother to daughter, could resist replacement during language shift. 8.1.3 Cultivated Plants and Agriculture First, it should be noted that cultivated plants and their wild equivalents are not always linguistically differentiated. Following from the discussion on pp. 229– 231, I have listed ‘carrot’ as a wild plant, even though the term normally refers to the cultivated variety in the modern languages. On the other hand, I have included ‘garlic’ as a cultivated plant in view of the fact that the word, wherever it occurs, is differentiated from the wild Allium ursinum (another probable loanword; see ‘ramsons’, pp. 246–247). Many of the crop terms have comparanda beyond Balto-Slavic, Germanic and Italo-Celtic, and many show an extremely broad distribution, suggestive of largely horizontal rather than vertical borrowing. On this basis, it can be hypothesized that the majority of terms for cultivated plants spread as Wanderwörter. In the case of founder crops, the spread of these words is unlikely to be directly associated with the spread of the crops themselves. Instead, it is more likely to be related to the spread of peoples and the adoption of an agricultural lifestyle. The trajectory of spread is usually difficult to establish on linguistic grounds. The material covered in this category is as follows:13 12 13 “A few men, especially those who hunt and fish, also gather some edible plants from time to time. However, it was not customary, and their knowledge of these plants was quite limited in comparison to the women’s” (Ertuğ 2000: 175 in a study of a community pursuing mixed hunter-gatherer/agricultural subsistence in Anatolia). Additional abbreviations used in this table: Ro = Romance, Pm = Permic, T = Turkic. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 292 table 20 bean hemp oats lentil rye apple garlic pear turnip scythe chapter 8 Distribution of borrowed agricultural terms B S G ? ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ? ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ C It ✓ Ro ✓ ✓ ✓ ? ✓ ✓ ✓ ✓ Gr + ✓ ? Berber Arm. Iran. F Md Ma ? Md Pm T ✓ It is notable that a disproportionate number of borrowed terms for crops are shared with Italic (including a word for ‘scythe’), which seems to bring the centre of gravity towards central and southern Europe. In addition, we often find historically identical forms in several branches. In certain cases, it is possible that a word spread within IE; as discussed above (p. 194), the word for ‘rye’ might well have entered Baltic, and possibly even Slavic, through Germanic mediation. Similarly, the Baltic word for ‘hemp’ may well have been borrowed from Slavic. Nevertheless, most of the forms cannot be explained as borrowings from any attested language. The process of Neolithization in the Eastern Baltic is extremely interesting and differs markedly from that in Central Europe. While the arrival of Corded Ware can be dated to the early 3ʳᵈ millennium bce (Piličiauskas 2018), the first individuals do not show evidence of admixture with Anatolian Farmer populations, suggesting an independent, direct migration from the steppe (Mittnik et al. 2018: 8). Although later individuals do show evidence of this ancestry, there remains no solid evidence for agriculture until the middle of the second millennium, where a few barley grains have been recovered from western Lithuania (Grikpėdis/Motuzaitė Matuzevičiūtė 2017). Here still, we also find abundant wild plant remains, suggesting a mixed subsistence involving only small-scale cultivation; moreover, it cannot be decided with certainty whether the aforementioned grains were cultivated locally or imported (Grikpėdis/Motuzaitė Matuzevičiūtė 2020: 162). A radical reassessment of early agriculture in the East Baltic has taken place in the past few years (cf. Piličiauskas et al. 2016; Girininkas 2019). Far from Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 analysis 293 earlier claims of cultivated grains already in the 3ʳᵈ millennium (Rimantienė 1992: 109–110), it has now become apparent that there is no solid evidence of agriculture prior to the Late and Final Bronze Age, i.e. the 1ˢᵗ millennium bce. This is, at least, not in contradiction with dietary data,14 where a shift to a diet incorporating cereals can only be demonstrated from the Late Bronze Age (Piličiauskas et al. 2017). If we examine the cereal terminology in the East Baltic languages, it becomes immediately apparent that the arrival of its speakers in the region cannot be equated with the first steppe migrations, as has sometimes been suggested (Mallory 1989: 108; Rimantienė 1992: 137–138; Parpola 2012: 133; Mittnik et al. 2018: 8). What we find is that the cereal terminology in Baltic is generally archaic, with some terms directly inherited from (core) Proto-Indo-European (see below). As a result, we must assume a continuity in agricultural practices among Balto-Slavic peoples during their migration from the Indo-European homeland. This points to a much later date for the arrival of Baltic-speaking populations in the Baltic Region, the most probable proxy being the Late Bronze Age hillfort phenomenon (Lang 2016: 18, 2018). At least the following Baltic cereal terms appear to be inherited: – Lt. javaĩ m.pl. ‘cereal’ (= Skt. yáva- ‘grain, crop’) ́ enys ‘seed, linseed’, Pr. E semen ‘seed’ (= Lat. sēmen) – Lt. sėmuõ, pl. sėm – Lt. dúona ‘bread’ (= Manichaean Sogdian δʾn ‘seed’) – Lt. žìrnis, Lv. zir̃nis ‘pea’, Pr. E syrne ‘grain’ (= Lat. grānum ‘grain’) – Lt. pẽlūs m.pl, Lv. pȩlus f.pl.; Pr. E pelwo ‘chaff’ (= Skt. palā́va- ‘chaff’) – Lt. árti, Lv. ar̂t ‘to plough’ (= Gr. ἀρόω, Lat. arō) – Pr. E wagnis ‘coulter’ (= Gr. ὀφνίς ‘ploughshare’)15 While it cannot be excluded that some of these terms originally referred to wild grains, their consistent agricultural meaning favours an early association with agriculture (see the survey in Kroonen et al. 2022). Particularly relevant are terms connected to ploughing, as archaeological evidence for ploughs and other agricultural tools in the East Baltic appears to be unreliable before the Late Bronze Age, coinciding with the archaeobotanical evidence (Lang 2007: 107; Luik/Maldre 2007: 33; Piličiauskas et al. 2016: 190–191; Girininkas 2019: 68– 72). 14 15 Inferred from the isotopic ratios of carbon and nitrogen in the bone collagen of ancient individuals. This word has been replaced in East Baltic, however, so is only indirectly relevant to this question (see the discussion of another word for ‘ploughshare’ on pp. 213–215). The Greek word is only known from a Hesychian gloss, but the formal correspondence with Prussian is ideal, and the word is also known from Germanic. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 294 chapter 8 Interestingly, most of the inherited terms are generic in character, suggesting that, while agriculture was certainly practiced, it remained rudimentary, with different crop types perhaps not being distinguished. One inherited term has become semantically specified in East Baltic — the term for ‘pea’, representing a semantic shift from a generic term for ‘grain’ (as preserved in Prussian syrne and OCS зрьно ‘grain’). As noted above (see 3.6.2), the semantic shift from ‘grain’ to ‘pea’ is rather surprising, as the pea, while one of the earliest crops to appear in the East Baltic, is recorded in small quantities (Pollmann 2014: 409), making its status as a staple crop improbable. Table 21, overleaf, shows the Balto-Slavic terms for various specific crops arranged in order of their appearance in the archaeobotanical record. The periodization is based on the useful survey of the archaeobotanical evidence by Grikpėdis and Motuzaitė Matuzevičiūtė (2020). Terms that are highlighted in bold have been suggested in this book to be borrowings from non-Indo-European sources. Shaded cells indicate that a common proto-form could theoretically be set up for multiple sub-branches. If we start from the hypothesis that the arrival of East Baltic-speaking populations in the Baltic region was associated with the emergence of diversified agriculture, then the crops should fall into two groups: the earliest crops, which might have been brought by the Balts themselves, and for which terms might already have been present in the Baltic languages prior to their arrival (i.e. possible “pre-migration terms”) and later crops, adopted already in situ, for which any terms must postdate such a migration. Pre-migration crops: The only term reconstructible for Proto-Balto-Slavic refers to a variety of wheat. This term is continued by Lt. (Žem.) pūraĩ, Lv. dial. pûŗi² ‘winter wheat’, corresponding to RCS пъıро (rendering Gr. ὀλύρα, ζέα, cf. СДРЯ 1759), SCr. dial. pȉr ‘spelt’ (Skok ii: 660), Sln. píra ‘spelt; (dial.) millet’ and further to Gr. πῡροί ‘wheat’. Due to the meaning and limited distribution, a non-IE origin has been suggested (Frisk ii: 631; Lubotsky 1988: 136); however, the comparison is impeccable on formal grounds, and we must reckon with the possibility of an inherited cereal term (Nieminen 1956: 170–172; Kroonen et al. 2022: 21). The semantic specialization in Baltic is explained by the word’s marginalization in favour of the loaned kviečiaĩ, probably associated with a transition to free-threshing wheats (see below).16 Beyond this, a shared word for ‘barley’ can be reconstructed for Proto-Baltic. It is possible that this could be connected to the role of barley as a pion- 16 Note that according to ME (iii: 449–450), pūŗi was used in some parts of Kurzeme as a general term for ‘wheat’. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 295 analysis eer crop in more northern latitudes (Motuzaitė Matuzevičiūtė 2018), although according to our workinghypothesis, the earliest barley finds in the Eastern Baltic should predate the arrival of the Balts. Nevertheless, the existence of a shared Baltic term might suggest barley was one of the first crops to have been cultivated by Baltic speakers. The origin of the term is unknown, however (Smoczyński 2018: 798; Kroonen et al. 2022: 15–16), and a post-Proto-Baltic diffusion cannot be ruled out. table 21 Chronology of cereal terms in Baltic East Baltic Prussian (E) Slavic LATE BRONZE AGE (1ˢᵗ millennium bce) — “pre-migration” barley Lt. miẽžiai Lv. mìeži hulled wheat (spelt) ? Lt. pūraĩ ‘winter – wheat’ RCS пъıро SCr. pȉr broomcorn millet Lt. sóros Lv. obs. sāre prassan [?← Pl.] Pl. proso SCr. prȍso pea Lt. žìrnis Lv. zir̃nis [keckers ← MLG] R горо́х SCr. grȁh (broad) bean Lt. pupà Lv. pupa babo [?← Pl.] R боб SCr. bȍb false flax, Camelina sativa Lt. jùdros Lv. idra [?← F] – ? Pl. rydz ? Sln. rȋdžək moasis RCS ꙗчьмъı Sln. jéčmen ROMAN IRON AGE (1ˢᵗ–8ᵗʰ c. ce) — “post-migration” free-threshing (bread) wheat [Lt. kviečiaĩ Lv. kvìeši ← Go.] gaydis OCS пьшеница rye [Lt. rugiaĩ Lv. rudzi ← Go.] rugis [?← G] R рожь Sln. ȓž [?← G] oats Lt. ãvižos Lv. àuzas wyse R овёс Sln. óvəs flax Lt. linaĩ Lv. lini [?← Sl.] Pr. G lino, lynno [?← Sl.] R лён Sln. lȃn Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 296 chapter 8 table 21 Chronology of cereal terms in Baltic (cont.) East Baltic Prussian (E) Slavic hemp Lt. kanãpės [?← Sl.] knapios [?← Sl.] R конопля́ Pl. konopie turnip Lt. rópė – R рѣ́ па SCr. rȅpa MIDDLE AGES (13ᵗʰ–14ᵗʰ c. ce) lentil Lt. lę̃šis [Lv. lȩ̃ca ← R] opium poppy [Lt. aguonà moke [?← Sl.] Lv. maguône ← G] [lituckekers] RCS лѧча SCr. léća R мак Sln. màk Although false flax (Camelina sativa) is normally interpreted as a weed in southern European Neolithic contexts (Zohary/Hopf 2012: 111), it appears that it was cultivated before flax in the Eastern Baltic, perhaps serving both as an oil plant and as animal fodder (Pollmann 2014: 412–413). No certainly old designation for false flax can be identified in Balto-Slavic.17 It is conceivable that the modern word for ‘flax’, which could theoretically be reconstructed for Proto-Balto-Slavic, was applied to this plant, or served as a general designation of oil plants. Pollmann notes that the same area where abundant remains of Camelina were identified archaeologically was later known for flax cultivation (2014: 413). However, it cannot be entirely excluded that the East Baltic terms were adopted from North Russian as late as the Middle Ages (cf. 1.1). The East Baltic designations for ‘millet’ and ‘bean’ are both possible borrowings from non-IE sources, although for ‘bean’, I have considered the inclusion of the Baltic data uncertain (see pp. 228–229). For millet, the main evidence is the existence of comparanda in Mordvin, which cannot be explained as direct borrowings. In principle, it is possible that the Balts picked up millet cultivation from Central Europe, where millet was well established from the 2ⁿᵈ millennium (Filipović et al. 2020). However recent investigations demonstrate 17 Perhaps the best candidate for a Proto-Slavic term is R ры́ жик, Pl. rydz (see atlas.roslin.pl/​ plant/6517, accessed 9 November 2023), Sln. rȋdžək (Pleteršnik ii: 426) ‘false flax’, which all derive from an adjective continued by R ры́ жий ‘red-haired’, Pl. dial. rydzy ‘copper-red’, Sln. (Pleteršnik) rȋdž ‘fuchsgelb’. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 analysis 297 another centre of spread in Central Asia (Motuzaitė Matezuvičiūtė et al. 2022). Widespread evidence of millet can be identified in the Pontic steppe region as well as in northwest Kazakhstan from the 1ˢᵗ millennium bce. It is possible that an eastern centre of spread could account for the linguistic facts more effectively, although more evidence is required to establish the archaeological plausibility of this scenario. If true, the word for ‘millet’ can be identified as a Wanderwort with its roots in an unidentified Central Asian language. Post-migration crops: In Chapter 2, I have argued that the East Baltic term for ‘wheat’ is a loan from East Germanic. Since the possible timeframe for contacts with Germanic coincides more or less with the first reliable evidence for free-threshing wheat, in particular bread wheat, Triticum aestivum (Grikpėdis/Motuzaitė Matezuvičiūtė 2020: 164), there is a plausible archaeological context for the adoption of this foreign term (note also that a term for ‘bread’ was borrowed from Germanic). Considering the similar chronology of rye cultivation in the region, it is probable that the word for ‘rye’ was taken from the same source. The Baltic word for ‘hemp’ is possibly a Slavic loanword, as is the word for ‘flax’ (see above). Interestingly, at least two “post-migration” crop names — ‘oats’ and ‘turnip’ — are clear borrowings from unknown sources.18 The comparanda for both of these point towards central or southern Europe. Both terms are shared with Italic, and are actually attested in literary sources in Latin several centuries before they emerge in the Baltic archaeological record,19 which strongly implies a trajectory from south to north. However, a proximate source of borrowing cannot be identified in any known language. Both words are also present in Slavic, but the reconstruction of a Balto-Slavic prototype is impossible, implying the Balts and Slavs must have been in contact with distinct Central European agricultural groups carrying related words. Evidence of the linguistic landscape in north-eastern Europe is practically non-existent until the Late Middle Ages, so that the existence of unrecorded languages during the first millennium ce which later went extinct need not surprise us. However, since we are clearly dealing with Wanderwörter, even if the terms are originally of non-Indo-European origin, it cannot be ruled out that they were transmitted into Balto-Slavic through unattested Indo-European languages. This is imaginable in cases such as ‘turnip’, where the Baltic term is historically identical to the equivalent in Germanic and Latin. On the other 18 19 For the Baltic word for ‘lentil’, see the discussion on pp. 201–202. A derivative of the word for ‘turnip’ is also found early in Greek, but in a secondary meaning. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 298 chapter 8 hand, little can be said with certainty; neither can it be established that the languages with which Baltic and Slavic were in contact were related with each other, despite possessing similar words for crops. I have also included the fruit trees apple and pear in this subsection, although they might be better described as wild. Both were first domesticated after the dispersal of the Indo-Europeans, as the cultivation of these plants must be done through grafting rather than from seed (Mallory/Adams 1997: 26; Zohary/Hopf 2012: 138, 140). The distribution of the crab apple and wild European pear is similar, encompassing most of Europe, and the western half of the Pontic-Caspian steppe (see Zohary/Hopf 2012: 137, 139). The pear is not found north of Latvia (cf. Schrader/Nehring i: 147), and as a consequence, there is no old word for ‘pear’ in Finnic. Both plants can be and are consumed in their wild form. A possible candidate for an inherited word for ‘apple’ is Gr. μῆλον, which has convincingly been argued to be cognate to Hittite samlu- ‘apple’ (Kroonen 2016). If this originally referred to the wild apple, then the spread of the Greek word into Lat. mālum and Alb. mollë (Schrader/Nehring i: 53) might be associated with the emergence of domesticated varieties in the early historical period. However, it is difficult to rule out a post-PIE loanword.20 For pear, we have no comparisons which go beyond two neighbouring branches, and no inherited term can be reconstructed with confidence, although it is theoretically possible that Gr. ἄπιον and Lat. pirum ‘pear’ could reflect an inherited *h₂pis-o-. In this case, the term would originally refer to a wild variety and only secondarily to the cultivated pear. 8.1.4 Apiculture As words for ‘honey’ and ‘mead’ can be reconstructed for Proto-Indo-European, it has been assumed that PIE speakers must have been involved in apiculture (Гамкрелидзе/Иванов 1984: 603); however, since wild honey hunting has been practiced since the Mesolithic, there is no necessity to believe the IndoEuropeans were familiar with domesticated honey bees (Schrader/Nehring i: 139–140; van Sluis 2022: 4, 26; cf. Crane 1999: 162). The complete absence of beeswax residues on pottery in the Neolithic Eurasian Steppe, despite good conditions for its preservation, probably speaks against any active apiculture (Roffet-Salque et al. 2015: 229). Three words have been classed as probable loanwords in this semantic field (see Table 22, overleaf). 20 If Kroonen’s comparison (2016: 88–89) with Georgian msxali ‘pear’ is valid, then the loanword would have to be very early, predating the loss of the laryngeals. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 299 analysis table 22 Distribution of borrowed apicultural terms drone honeycomb wax B S G ✓ ✓ ✓ ✓ ✓ ✓ ✓ C It Gr + ✓ ✓ Md Ma Tur The borrowing of terms for bees along with the technology for their domestication would be unsurprising, as bees may have been a mere pest to honey hunters, and therefore of less importance (Vennemann 1998: 477–478). Interestingly, however, in an actual case of language shift studied by Brenzinger (1992), we find the opposite situation: after shifting to speak Maasai, originally Yaaku beekeeping communities continued to use a substrate word for ‘honey’, while words for various kinds of ‘bee’ had recently fallen out of use (idem: 234– 235). This of course need not worry us too much, as we cannot expect all cases of language shift to be identical. Among the apicultural terms, only the words for ‘wax’ and ‘drone’ can potentially be reconstructed for Proto-Balto-Slavic, both of which show good evidence of foreign origin. The terms for ‘bee’, and also perhaps ‘beehive’ (see pp. 279–280), which have been considered uncertain pre-European loanwords, cannot be reliably reconstructed for Balto-Slavic. As a result, it is uncertain whether speakers of Proto-Balto-Slavic were engaged in apiculture. Tree beekeeping is already attested in Latvia in the Middle Ages, and was only completely superseded in the East Baltic region by (log) hive beekeeping in the 18ᵗʰ century (Crane 1999: 132–133, 233–234). A potentially Proto-Balto-Slavic term related to the use of tree hives is Lt. dial. genỹs, geinỹs, Lv. dzeĩnis, dial. dzenis ‘climbing rope (for accessing tree hives)’ (apparently → the Võro hapax kõno in the same sense; Vaba 1990b: 173) which corresponds regularly to R dial. жень (Nižnij Novgorod, Kostroma; СРНГ ix: 129), Bel./Uk. (Polesia) жэнь, же̑нь, жинь (ДАБМ No. 313; ЕСУМ ii: 193; Никончук apud ЭСБМ iii: 270) ‘climbing rope’ (Būga 1916: 156).21 21 ЕСУМ (ii: 193) suspect that the Slavic word is loaned from Baltic. The distribution would appear to favour this, even though the Russian word is attested rather far from the Baltic territories. It is uncertain whether a Baltic loanword can be expected to have undergone the first palatalization. Note, however, OR ижера pro *игера ‘Ingrians’ ← Ingr. Inkeroin cited on p. 18 and the hydronym Селижа́ ровка (beside OR Серегѣрь), also of presumed Finnic origin (REW ii: 605; Крысько 1994: 83). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 300 chapter 8 8.1.5 Structures Despite a perception of the Indo-Europeans as primarily nomadic (Kuhn 1862: 371; Anthony 2007: 321–322; Anthony/Ringe 2015: 211; see also the literature review in Häusler 2002: 3–48), there is evidence that at least the later stages of Indo-European unity were associated with a level of sedentarization (Kroonen et al. 2022: 32–36), and some clear inherited terminology exists relating to the erection of fixed or semi-fixed structures (Mallory/Adams 2006: 219–229), most notably the verbal root in HLuw. tama-, Gr. δέμω ‘to build’, which is the basis of the nominal derivatives in most branches, e.g. Lycian tãma ‘building’, Arm. tun, Lat. domus ‘house’, and probably Lt. nãmas ‘house’ (cf. IEW 198– 199). Insight into the technology of house-building among Indo-European speakers may be provided by Skt. dehī-,́ Osc. feíhúss acc.pl., Gr. τεῖχος, Sln. zȋd ‘(surrounding) wall’, apparently derived from the verbal root for ‘to mould (clay)’.22 Mallory/Adams (2006: 223) emphasize that this word does not generally mean ‘wall of a house’, although Pr. E seydis · want ‘wall’ and Gr. τοῖχος, derived from the same root, do appear to be generic terms.23 In any case, it is tempting to speculate that this might be a reflection of the construction of temporary wattle and daub huts, as known from ethnographical parallels of nomadic pastoralists (e.g. Evans/Pritchard 1940: 65). On the other hand, a word for some kind of fortification must be reconstructed on the basis of Lt. pilìs, Lv. pils, Skt. (RV) púr- ‘fortress, stronghold’, Gr. πόλις ‘city, citadel’. Considering the possible association of the arrival of the Balts in the region with the appearance of fortified settlements (Lang 2016: 18, 2018a) and the implication of continuity provided by the linguistic data, it seems attractive to assume the construction of hillforts already started in the Indo-European homeland. Note, for instance, the Early Bronze Age hillfort at Mykhailivka on the Lower Dnieper (Anthony 2007: 324). Whatever the details of Indo-European house construction, it is likely to have greatly differed from that of Neolithic Europe. According to Della Volpe (1996: 152), timber-framed longhouses, generally being devoid of any defensive structures, predominate in the pre-Indo-European context. 22 23 Compare Go. (ga-)digan* (rendering Gr. πλάσσω ‘mould, form’), Lat. fingō ‘mould, fashion (clay, wax, etc.); sculpt’, ToB tsaikaṃ ‘mould (pottery); build’, as well as (with apparent metathesis) Lt. žiẽsti ‘mould (pottery)’, OCS съзьдати ‘build, create’ (IEW 245). Likewise, Mac. ѕид is a generic term for ‘wall’. The usual word for ‘wall (of a house)’ in Balto-Slavic is Lt. síena, R стѣна́ (while in Mac., стена means ‘rock face’). As this word is possibly related to Go. stains ‘stone’, it might reflect a shift towards stone architecture (for a discussion of the relationship between these words, see 6.3.2). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 301 analysis table 23 Distribution of borrowed terms for structures B cottage oven ✓ ✓ S G C It ✓ ✓ ✓ ✓ Gr + ✓ In the centre of an Indo-European home, there was presumably a hearth (Hitt. hāssā- ‘hearth, fireplace’, Lat. āra ‘altar’; IEW 68–69). On the other hand, ovens are considered to have spread into Europe as part of the Neolithic package, emerging during the eighth millennium bce in Anatolia (Barbaro et al. 2021: 1161). Domed clay ovens are known from households in Neolithic sites immediately adjacent to Yamnaya (Anthony 2007: 143, 166), and it seems quite probable that a word for ‘oven’ would have been taken over from such farming populations. 8.1.6 Metallurgy The only metal term in Balto-Slavic with direct Indo-European cognates is the word for ‘gold’, Lt. áuksas, Pr. E ausis (iii ausin acc.sg.) = Lat. aurum ‘gold’. While the narrow distribution has led to speculations of a direct or indirect loan relationship (Kretschmer 1896: 150; Pisani 1968: 11), on formal grounds, a common inheritance cannot be excluded (see Driessen 2003).24 The following terms can theoretically be dated to Proto-Balto-Slavic, two implying ablaut (for a discussion of the further etymologies of the metal names, which remain uncertain, see Thorsø et al. 2023: 117): – Lt. švìnas, Lv. svins ‘lead’ (< *ḱuin-) ~ OR свиньць, Sln. svínəc ‘lead’ (< *ḱuein-) – Lt. obs. álvas (áłwu inst.sg. in Daukša), Lv. al̂va, dial. al̂vs ‘tin’ (cf. Endzelīns 1923: 157) (< *HolH-u̯ -) ~ OCS олово, SCr. ȍlovo ‘lead’, R о́лово ‘tin’ (< *HolH-eu̯ -) – ? Pr. E wutris ‘smith’ ~ CS вътрь ‘smith’ (see Miklosich 1865: 113; SJS i: 352) On the other hand, the terms for two other metals do not permit the reconstruction of a common Balto-Slavic preform, and these may be interpreted as loanwords from unknown sources (see Table 24, overleaf). 24 I am rather convinced by the interpretation of To. B yasa, A wäs ‘gold’ as a loanword from Samoyed *wäsa (> Ngan. basa ‘metal, iron’, Taz Selkup kē̮si̮ ‘iron’; cf. Kallio 2004: 132–133). In any case, connecting the Tocharian with the European terms raises serious morphological issues (see Thorsø et al. 2023: 105–106). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 302 chapter 8 table 24 iron silver Distribution of borrowed metallurgical terms B S G C ✓ ✓ ✓ ✓ ✓ ? It Gr + Basque The absence of a common Balto-Slavic term for ‘iron’ is hardly surprising, as the split of this branch undoubtedly predated the Iron Age. While some iron artefacts may have been imported into the East Baltic region from elsewhere already in the Final Bronze Age (Lang 2007: 121), local iron production probably began during the first centuries ce, where it was produced in smelting furnaces from bog ores (Stankus 2001; Rundberget et al. 2020: 96).25 The Slavic word was evidently adopted from a related source, pointing to the spread of a localized smelting practice. An Indo-European word for ‘silver’ can be reconstructed on the basis of YAv. ərəzata-, Lat. argentum, OIr. argat, and probably Arm. arcatʿ ‘silver’, but this word appears to have been replaced in the northern European branches. The word for ‘silver’ in Balto-Slavic and Germanic is a widespread Wanderwort, whose centre of spread might be located in Iberia (Thorsø et al. 2023: 118), an idea that would be supported by the comparanda in Basque and Celtiberian (the latter probably being adopted locally after the southward migration of Celtic speakers). Although the word seems to be reconstructible to Proto-Germanic, it cannot be reconstructed for Proto-Balto-Slavic, or even Proto-Baltic, again suggesting that the word was absorbed into already diffuse linguistic groups. 8.2 Stratification The main methodological novelty in this section has been an attempt to identify alternations which do not merely reoccur, but which show a particular geographical patterning. I reasoned that a geographical distribution would both support the validity of an alternation, and potentially provide us with some information on the dialectal makeup of the underlying substratum. In 25 Although A. Merkevičius apud Lang 2018b dates the appearance of iron metallurgy in Lithuania to 300 bce. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 303 analysis table 25 Alternations showing a geographical patterning B *-VNT- ∞ *-VT*g⁽ʰ⁾ ∞ *k *b⁽ʰ⁾ ∞ *p *T ∞ *Tʰ *ž ∞ *š *st ∞ *(t)s *sd ∞ *d *a-CC ∞ *CVC *ke- ∞ *ka/o*ā̆ ∞ *ē ̆ *V̄ ∞ *V *V ∞ *V̄ S G C It ? Gr Examples 5 7 2 [+ 2] 3 [+ 1] 1 [+ 3] 3 [+ 1] 2 1+2 3 3 [+ 2] 4 [+3] 2 total, I have identified seven consonantal and five vocalic alternations which can be said to show a geographical distribution on the basis of at least three (or two certain) examples. This is presented in Table 25, above. Dark shaded cells consistently show the rightmost variant, while light shaded cells indicate a hesitation between the two. The shading is based on both certain and uncertain examples (the number of the latter is indicated in brackets). In the case of *a-CC ∞ *CVC, I have also included Schrijver’s examples of ‘blackbird’ and ‘ore’ (see 7.1.1), which seem to show a related phenomenon, even though they have not fallen under the scope of this work. Previous studies have often tended to treat the palaeo-European contact languages as a monolithic layer, whereby the irregularities present in the IndoEuropean reflexes are reflections of synchronic features of a single substrate language (Kuiper 1968; Schrijver 1997; Beekes 2014; see p. 166). The presence of geographical patterns contradicts this assumption, as such distributions are more easily explained as the result of dialectal or diachronic differences in the source language. In any case, it seems highly improbable that the linguistic landscape was homogenous among sedentary Neolithic farming populations prior to the expansion of Indo-European (see the discussion in Anthony 2007: 80–81). A deeper analysis of the stratification based on distribution alone is very difficult, as none of the alternations obviously correlate with each other. An Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 304 chapter 8 exception is *g⁽ʰ⁾ ∞ *k and *b⁽ʰ⁾ ∞ *p, but this actually results from the fact that the two alternations co-occur in two of the relevant word families, and we thereore cannot speak of the coherence of two independent sets. The fact that no clear patterns emerge on this higher level need not dishearten us. On the one hand, the number of examples of each alternation is small, and there is perhaps simply insufficient evidence for meaningful patterns to emerge. On the other hand, these alternations represent manifestations of complex contact situations which may have taken place in different locations and at different times, and therefore a complex picture is exactly what we should expect. It is perhaps more instructive to examine which kinds of alternations cooccur (cf. Šorgo 2020: 461–462). The word for ‘pigeon (1)’, for instance, shows both *g⁽ʰ⁾ ∞ *k and *-VNT- ∞ *-VT-. This might well suggest that the two alternations are somehow related. Indeed, on p. 185 (and in Jakob forthc. a.), I have noted that the word for ‘pigeon’ shows a similar structure to several other bird names, including another word for ‘pigeon (2)’ which potentially shows the alternation *b⁽ʰ⁾ ∞ *p. The full set of words (including one plant name) is as follows: ‘pigeon (1)’ ‘grouse’ OCS голѫбь, OE culufre, Lat. columba Lt. jerubė ̃ ‘hazel grouse’, Sln. jerę̑b, ?OHG reba-huon ‘partridge’ ‘oriole’ Lt. volungė,̃ Pl. wilga, MHG wite-wal ‘swan (1)’ Pl. łabędź, R ле́бедь, ON ǫlpt ‘goosefoot’ Lt. balánda, R лобода́ ‘goosefoot’, OHG melta ‘orache’ + ‘swallow’ Lat. hirundō, Gr. χελῑδών, Alb. dallëndyshe ? ‘pigeon (2)’ Lt. balañdis, Lat. palumbēs This collection of terms is perhaps the strongest evidence for a particular stratum: as well as clustering in a particular semantic field, they show similar kinds of alternations, in particular, a semi-regular correlation between voiced stops in the north (always Baltic, usually Slavic) and voiceless in the south (i.e. Italic), and a second syllable of the shape *VND, whereby the nasal is sometimes absent (although always present in Italic). In addition, I have noted ‘swan (1)’ as a plausible example of the alternation *a-CC ∞ *CVC. This might encourage us to view this alternation as yet another feature of this stratum. Indeed, the classic example of this alternation is another bird name (cf. Lat. merula, OHG amsla ‘blackbird’). Aside from this, potential ‘prefixal’ elements have been identified in the words for ‘grouse’ and ‘oriole’. Finally, it is tempting to adduce the word for ‘bison’ here, as OHG wisunt shows a similar disyllabic root structure with a second syllable in *VND, although here the initial syllable appears itself to be a ‘prefixal’ element. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 305 analysis table 26 The *VND substrate C G pigeon (1) grouse oriole swan (1) ? pigeon (2) + swallow + blackbird goosefoot ? bison (total). B S It Gr + Eg Core 1 6 5 Alb 6 4 1 Viewing the stratum as a whole (see Table 26, above), we can see a Central European ‘core’, consisting of Balto-Slavic, Germanic and (slightly less so) Italic, and a periphery. Notably, all of the words attested in the peripheral languages are also attested in Italic, and indeed Italy can be seen as a sort of interface between Central Europe and the Mediterranean on the one hand, and with Celtic on the other. It is certainly not the case, however, that the words were borrowed into the ‘peripheral’ branches directly from Latin or an Italic language. If words belonging to this stratum are not originally Mediterranean, they must have been carried into the region by speakers of unattested, presumably nonIndo-European languages. This implies a significant antiquity, which is already suggested by the attestation of ‘pigeon (1)’ in Egyptian in the 12ᵗʰ c. bce (see p. 189). On the other hand, the words ‘swan (1)’ and ‘oriole’ show irregular variation even within Slavic, suggesting that at least one variant was adopted after the dialectal fragmentation of this branch. This places us in a very broad timeframe stretching some two millennia, and raises serious doubts as to the internal coherence of the stratum. One suggestion, borrowing the analytical tools of botany, would be to interpret Slavic as a “centre of diversity”, and suggest that Slavic was geographically closest to the ‘core’. This is potentially supported by the fact that Slavic takes an intermediate position in the voicing alternations, implying contacts with multiple source languages or dialects. Considering the limitation to animal and plant names, and in particular bird names, we are most probably dealing with a largely vertical borrowing context; in other words, a linguistic substrate. By contrast, a number of techAnthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 306 chapter 8 nological borrowings, particularly those showing a broad distribution, were probably borrowed horizontally through trade, and can be characterized as Wanderwörter. Here we may include most of the terms for crops and cultivated plants (see 8.1.3), as well as the word for ‘silver’ and probably ‘thousand’ (see 3.5.4). It is probable that other (sub)strata existed. If we consider the alternations which do not occur in any of the words in the ‘bird name’ stratum, it is curious that *ke- ∞ *ka- typically involves a “non-core” distribution: of the three examples, two involve Celtic, and two involve Greek. The word for ‘ramsons’ shows a particularly broad distribution. Somewhat comparable is the alternation *T ∞ *Tʰ, which always (by necessity) involves Greek. The words in these categories seem to cluster semantically in the domain of wild and cultivated plants. At least ‘drone’ can be reconstructed for Proto-Balto-Slavic, suggesting a certain antiquity.26 The nasal alternation in the word for ‘lynx’ is also unlike that attested in the ‘bird names’. Above, I have briefly mentioned that this word could be a particularly old loanword. There are multiple possible indications of this: – The word shows an unusually large distribution, being present in five IndoEuropean branches. – It can possibly be reconstructed for Proto-Balto-Slavic,27 as well as for ProtoGraeco-Armenian. – Importantly, the sibilant in Balto-Slavic implies that these borrowings predated satemization. Aside from the word for ‘lynx’ only half a dozen words can be securely reconstructed for Proto-Balto-Slavic. The following may be mentioned: – In three branches: ‘apple’, ‘lynx’, ‘nettle’ – East Baltic + Slavic: ‘alder’, ‘drone’, ‘fresh’28 – Prussian + Slavic: ?‘bean’ (if not a Slavic loan in Prussian) While the possibility of reconstructing a word for Proto-Balto-Slavic may be seen as implying its relative antiquity, it is not a watertight indication: ‘rye’ and ‘hemp’, which would theoretically be reconstructible for Proto-Balto-Slavic, must have entered Baltic recently in view of the cultivation history of these 26 27 28 A shared reconstruction might also be attempted for ‘ramsons’, although this requires that Baltic *š results from the RUKI law, which I consider dubious (Chapter 1, fn. 14). A similar obstacle exists to the reconstruction of the words for ‘wax’ and ‘reed’ for Proto-Balto-Slavic (on the latter, see the note under ‘furrow’ on p. 224). The only obstacle to this is the Slavic *r-, on which see Chapter 6, fn. 18. More dubiously, we may be able to reconstruct words for ‘aspen’ (provided Baltic *u is not old; cf. pp. 277–278), ‘hornbeam’ (irregular Baltic s-) and ‘lightning’ (the dental in Slavic and East Baltic is ambiguous). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 analysis 307 plants. Nevertheless, the six certain examples cited above appear to be good candidates for Proto-Balto-Slavic loanwords. Support for the antiquity of the word for ‘apple’ may be seen in its adoption into a fairly unproductive noun class (the l-stems). Depending on one’s analysis (see pp. 268–270 for a discussion), this word — like ‘cottage’ — may additionally be interpreted as predating Winter’s law, which would certainly imply a Proto-Balto-Slavic antiquity.29 If the East Baltic comparanda for ‘bean’ are accepted (see p. 228), then the word can no longer be reliably reconstructed for Proto-Balto-Slavic. In fact, it would point to the opposite extreme: a word which is even irregular between East and West Baltic. A few other such words can be cited which exhibit similar behaviour: among the certain cases, we can mention ‘bison’ and ‘thousand’, both of which exhibit the alternation *st ∞ *(t)s, and also ‘silver’, a widespread Wanderwort. Less certain examples are ‘badger’, ‘salmon’, and also ‘oats’, depending on the analysis of the Prussian data (pp. 239–240). To summarize, it is clear, at least, that we are not dealing here with a chronologically or geographically localized borrowing event; however, due to the number of variables and small number of examples, it is difficult to comprehensively stratify the material. Nevertheless, there are indications of at least three chronological layers — one early layer, exemplified by the word for ‘lynx’, which may represent a borrowing event close to the steppe chronologically aligned with the disintegration of the proto-language, a late layer, apparently postdating the split of East and West Baltic (providing a Proto-Baltic stage ought to be reconstructed at all), and an intermediate layer. In addition, one group of words, primarily comprising bird names, seems to form a robust cluster and perhaps represents a set of loanwords from related source languages. 29 Other candidates for loanwords predating satemization, and therefore potentially contemporaneous with the word for ‘lynx’, are the uncertain cases ‘elm (1)’ and ‘furrow’. It must be noted, however, that the reconstruction of the word for ‘elm’ to Proto-Balto-Slavic is not entirely straightforward, as much of the Slavic evidence speaks in favour of accentual mobility (cf. В. Дыбо 2002: 469), and the word for ‘furrow’ is irregular between Baltic and Slavic (see p. 224). Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 Conclusion One of the first goals set out in this monograph was to scrutinize the stereotype of the Baltic languages as ‘pure’ Indo-European languages which have experienced minimal external influences. Through a detailed study of all possible contact events, both with attested and unattested languages, I have attempted to demonstrate that a significant part of the Baltic lexicon cannot be explained on internal grounds. However, most interactions which have contributed to the Baltic lexicon appear to have been with unknown languages, by contrast, for instance, to Finnic, which is known to have been in contact with several IndoEuropean languages throughout its history. No Indo-European loanword layers can be identified with certainty in Baltic prior to those with Gothic at the start of the Common Era. Contacts with Slavic, as far as we can make out, only started after the northern migration of (pre-)North Russian speakers. In addition, a notable layer of Baltic loanwords can be identified in Finnic, suggesting a significant contact event. However, even though the source of these loans seems to be more closely affiliated to East Baltic than West Baltic, and there is evidence that the source language has undergone some specifically East Baltic semantic and formal developments, it remains improbable that this was the direct ancestor of the attested East Baltic languages. There does not appear to be any old Finnic contribution to the Baltic lexicon, and the evidence seems to support the notion of an East Baltic substrate, most probably spoken to the east of the current Baltic territories, which was absorbed by Finnic some time before the Common Era. In addition, we see small layers of loanwords in both Sámi and Mordvin, suggesting some peripheral contact with this or a closely related Baltic language. We have approached the question of non-Indo-European components in the Baltic lexicon from multiple angles. First, we have attempted to find words common to both Baltic and Finnic which are unlikely on phonological grounds to have been adopted directly from one to the other. Although there are few relatively clear cases, there are a number of convincing examples which allow for a hypothesis that Baltic and Finnic were independently in contact with similar, and probably also distinct, ‘autochthonous’ populations upon their arrival in the Baltic Sea region. As we are able to operate in the context of regular sound correspondences (or more specifically, their absence), it is somewhat easier to make a case for a non-Indo-European element in Baltic lexical items with Indo-European comparanda. Almost fifty relatively clear cases were identified. Some initial efforts can be made to stratify this material, and at least one relatively coherent and © Anthony Jakob, 2024 | doi:10.1163/9789004686472_011 Anthony Jakob - 978-90-04-68647-2 This is an open access chapter distributed under the terms of the cc by-nc-nd 4.0 license. Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 conclusion 309 distinct stratum can be identified, consisting primarily of bird names with a second syllable of the shape *V(N)D. In addition to this, we can distinguish a number of widespread Wanderwörter, most significantly in the domain of cultivated plants, whose proximate source in the Baltic languages cannot be identified in any known language, and which may be reasoned at least partly to have originated among Neolithic farming populations. Contact with unattested languages is an area of study which has long been marginalized, partly for the reason that it is considered impossible to study, a priori unscientific, or inevitably circular. As a result of these prejudices, this subfield remains in its infancy. One of the goals of this work has been to demonstrate the feasibility of constraining the study of such contacts within scientific principles. It is important that a ‘substrate’ hypothesis is treated as strictly and objectively as a hypothesis of cognancy, and built on the basis of positive evidence. In this context, a suggestion of non-Indo-European origin can be viewed as a reasoned scientific solution to the problem of irregular sound correspondence, and not as a last resort or throwaway suggestion. It is certainly not true that the Baltic languages have developed in a vacuum, void of contact with other languages. Instead, most of the language contact has taken place in a preliterary context, with languages which never came to be written down, or of which no written trace has yet been uncovered. This is actually precisely what we should expect, since the area where the Balts have come to reside has been populated since the end of the last Ice Age by numerous genetically distinct populations, undoubtedly bringing with them different languages, while writing has only reached the region in the Middle Ages. In this context, traces of foreign languages preserved in the modern Baltic languages can be seen as a valuable resource and a potential key to unlocking the population history of the region. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 Bibliography Note on alphabetization: titles and names written in the Cyrillic script are presented in the original spelling, but alphabetized according to the usual academic transcription into the Latin alphabet: а б в г д е ж з и й к л м н о п р с т у ф х ц ч ш щ ъ ы ь э ю я ѣ a b v g d e ž z i j k l m n o p r s t u f x c č š šč ” y ’ e ju ja ě For the purposes of alphabetization, diacritics are ignored, so that Çabej precedes Цаболов (= Cabolov). The Cyrillic letter x is alphabetized under x (rather than h, as sometimes found). For ease of navigation, Ukrainian/Belarusian г have been alphabetized under g. Web links are valid as of 9 November 2023. Abbreviated Titles ALEW ALEW² АУМ AWb БДА БЭР CDIAL ČJA ClG i ДАБМ Wolfgang Hock (2019, ed.), Altlitauisches etymologisches Wörterbuch [Version 1.1]. Berlin: Humboldt-Universität. Wolfgang Hock (2020, ed.), Altlitauisches etymologisches Wörterbuch [Version 2.0]. Accessed online at https://alew.hu‑berlin.de/. Тетяна В. Назарова et al. (1984–2001, eds.), Атлас української мови. 3 vols. Київ: Наукова думка. Elisabeth Karg-Gasterstädt & Theodor Frings (1952– , eds.) Althochdeutsches Wörterbuch. 7 vols. [A–R]. Leipzig: Sächsische Akademie der Wissenschaften. Иван Ю. Кочев (2001, ed.), Български диалектен атлас. Обобщаващ том. София: Труд. Владимир И. Георгиев, Иван Дуриданов et al. (1971– , eds.), Български етимологичен речник. 8 vols. [а–фя̀ калка]. София: Българската академия на науките/М. Дринов. Turner, Ralph L. (1973 [1966]). A comparative dictionary of Indo-Aryan languages. London: Oxford University Press. Jan Balhar & Pavel Jančák (1992–2011, eds.), Český jazykový atlas [2ⁿᵈ edition]. 5 vols. Praha: Academia. Accessed online at https://cja.ujc.cas​ .cz/. Adolfas Ivaškevičius et al. (1995–1997, eds.), Clavis Germanico-Lithvana. Vols. 1–2. Vilnius: Mokslo ir enciklopedijų leidykla. Рубэн І. Аванесаў, Кандрат Крапіва & Юзэфа Ф. Мацкевіч (1963, Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 312 ДАРЯ i DOE DWb DWb² eDIL EES EH EMS ERHJ ЭСБМ ESJS ЭССЯ ЭСТЯ ЕСУМ EWA EWAhd bibliography eds.), Дыялекталагічны атлас беларускай мовы. Мінск: Акадэмія навук БССР. Рубен И. Аванесов & Софья В. Бромлей (1986, eds.), Диалектологический атлас русского языка i. Фонетика. Москва: Наука. Angus Cameron, Ashley Crandell Amos, Antonette diPaolo Healey et al. (2018, eds.), Dictionary of Old English: A to I. Accessed online at tapor.library.utoronto.ca/doe (subscription resource). Moriz Heyne et al. (1854–1971, comp.), Deutsches Wörterbuch von Jacob Grimm und Wilhelm Grimm. 33 vols. Leipzig: Hirzel. Deutsches Wörterbuch von Jacob Grimm und Wilhelm Grimm. Neubearbeitung (A–F). Accessed online at https://www.woerterbuchnetz.de/​ ?sigle=DWB2&lemid=A00001. Gregory Toner, Maire Ní Nhaonaigh et al. (2019, eds.), Electronic Dictionary of the Irish Language. Accessed online at dil.ie. Iris Metsmägi, Meeli Sedrik & Sven-Erik Soosaar (2002, eds.), Eesti etümoloogiasõnaraamat. Tallinn: Eesti Keele Sihtasutus. Janis Endzelīns & Edita Hausenberg (1956). Ergänzungen und Berichtigungen zu K. Mühlenbachs Lettisch-deutschem Wörterbuch. 2 vols. Chicago. Anu Haak, Evi Juhkam, Mari Kendla et al. (1994– , eds.), Eesti murrete sõnaraamat. 7 vols. [a–päevätuss]. Tallinn: Eesti Teaduste Akadeemia. Ranko Matasović, Dubravka Ivšić Majić & Tijmen Pronk (2016–2021, eds.), Etimološki rječnik hrvatskoga jezika. 2 vols. Zagreb: Institut za hrvatski jezik i jezikoslovlje. Віктар У. Мартынаў & Генадзь А. Цыхун (1978– , eds.), Этымалагічны слоўнік беларускай мовы. 14 vols. [а–тэчка]. Мінск: Навука і тэхніка. Eva Havlová et al. (1989–2018, eds.), Etymologický slovník jazyka staroslověnského. 19 vols. Praha: Academia; Brno: Tribun EU. Олег С. Трубачев & Андрей А. Журавлев (eds., 1971– ), Этимологический словарь славянских языков. 42 vols. [*a–*perzъ]. Москва: Наука. Эрванд В. Севортян, Лия С. Левитская et al. (1974–2003, eds.), Этимологический словарь тюркских языков. 7 vols. Москва: Наука/ Восточная литература РАН. Олександр С. Мельничук et al. (1982–2012, eds.), Eтимологiчний словник української мови. 6 vols. Київ: Наукова думка. Mayrhofer, Manfred (1992–2001). Etymologisches Wörterbuch des Altindoarischen. 3 vols. Heidelberg: Winter. Albert L. Lloyd, Otto Springer & Rosemarie Lühr (1988– , eds.), Etymologisches Wörterbuch des Althochdeutschen. 7 vols. [a–skazzôn]. Göttingen–Zürich: Vandenhoeck & Ruprecht. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 bibliography 313 GPC Gareth A. Bevan & Patrick J. Donovan (1967–2002, eds.), Geiriadur Prifysgol Cymru. 4 vols. Cardiff: Gwasg Prifysgol Cymru. Also 2nd online edition, accessed at geiriadur.ac.uk/gpc/gpc.html. ГСБМ Аляксандр М. Булыка (1982–2007, ed.), Гістарычны слоўнік беларускай мовы. 37 vols. Мінск: Навука i тэхніка/Беларуская навука. IEW Pokorny, Julius (1959). Indogermanisches etymologisches Wörterbuch. Bern–München: Francke. KEWA Mayrhofer, Manfred (1956–1980). Kurzgefaßtes etymologisches Wörterbuch des Altindischen. 4 vols. Heidelberg: Winter. KKS Pertti Virtaranta & Raija Koponen (1968–2005, eds.), Karjalan kielen sanakirja. 6 vols. Helsinki: Suomalais-Ugrilainen Seura. Accessed online at kaino.kotus.fi/kks. LÄGLOS Andries D. Kylstra, Sirkka-Liisa Hahmo, Tette Hofstra & Osmo Nikkilä (1991–2012, eds.), Lexikon der Alteren Germanischen Lehnworter in den Ostseefinnischen Sprachen. 3 vols. Amsterdam: Rodopi. LEIA Vendryes, Joseph, Édouard Bachallery & Pierre-Yves Lambert (1959– 1996). Lexique étymologique de l’irlandais ancien. 7 vols. Dublin: Institute for Advanced Studies. LEW Fraenkel, Ernst (1955–1966). Litauisches etymologisches Worterbuch. 2 vols. Heidelberg: Winter/Göttingen: Vandenhoeck & Ruprecht. LIV Helmut Rix et al. (2001, eds.), Lexicon der indogermanischen Verben [2nd edition]. Wiesbaden: Reichert. LIV addenda Kümmel, Martin (2015). Addenda und Corrigenda zu LIV². Accessed online at martinkuemmel.de/liv2add.pdf. LKA Elena Grinaveckienė, Kazys Morkūnas et al. (1977–1991, eds.), Lietuvių kalbos atlasas. 3 vols. Vilinus: Mokslas. LKŽ Juozas Balčikonis & Vytautas Vitkauskas (1941–2002, eds.), Lietuvių kalbos žodynas. 20 vols. Vilnius: Mokslas, Lietuvių kalbos instituto leidykla. LSJ Henry J. Liddel, Robert Scott, Henry S. Jones & Roderick McKenzie (1940, eds.), A Greek–English Lexicon [9ᵗʰ edition]. Oxford: Clarendon Press. LVPPV Laimdots Ceplītis, Aina Miķelsone, Tamāra Porīte & Silvija Raģe (1995, eds.), Latviešu valodas pareizrakstības un pareizrunas vārdnīca. Riga: Avots. MdWb Kaino Heikkilä (1990–1996, comp.), H. Paasonens Mordvinisches Wörterbuch. 6 vols. Helsinki: Suomalais-Ugrilainen Seura. ME Janis Endzelīns (1923–1925, ed.), K. Mühlenbachs Lettisch-deutsches Wörterbuch. 4 vols. Rīga: Latvju grāmata. MWb Mittelhochdeutsches Wörterbuch (online edition) [a–kochlêhen]. Accessed online at http://mhdwb‑online.de/. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 314 NdsWb NIL ODS OED OstWb PKEŽ PrWb РЭС REW RJA РСА Schw. Id. SD СДРЯ СДРЯ 11–14 SEK bibliography Hans Janßen et al. (1965– ). Niedersächsiches Wörterbuch. 10 vols. [a– skrofulös]. Neumünster–Kiel–Hamburg: Wachholtz. Dagmar S. Wodtko, Britta Irslinger & Carolin Schneider (2008). Nomina im Indogermanischen Lexikon. Heidelberg: Winter. Ordbok over det danske Sprog (online edition). Accessed online at https://ordnet.dk/ods. Oxford English Dictionary (online edition). Oxford University Press. Accessed online at https://www.oed.com/ (subscription resource). [N.B. entries updated for the OED Third Edition are marked OED³]. Steinitz, Wolfgang, Liselotte Böhnke, Gert Sauer & Brigitte Schulze (1966–1993). Dialektologisches und etymologisches Wörterbuch der ostjakischen Sprache. 15 issues. Berlin: Akademie. Mažiulis, Vytautas (1988–1997). Prūsų kalbos etimologijos žodynas. 4 vols. Vilnius: Mokslas. Erhard Riemann & Ulrich Tolksdorfs (1974–2005, eds.), Preußisches Wörterbuch. Deutsche Mundarten Ost- und Westpreußens. 6 vols. Neumünster: Wachholtz. Аникин, Александр Е. (2007– ). Русский этимологический словарь. 16 issues [а–житьё]. Москва: Рукописные памятники Древней Руси. Vasmer, Max (1953–1958). Russisches etymologisches Wörterbuch. 3 vols. Heidelberg: Winter. Đuro Daničić, Pero Budmani et al. (1880–1976, eds.), Rječnik hrvatskoga ili srpskoga jezika. 23 vols. Zagreb: Jugoslavenska akademija znanosti i umjetnosti. Речник српскохрватског књижевног и народног језика (1959– ). 21 vols. [а–покупити]. Београд: Српска академиjа наука. Friedrich Staub et al. (1881– , eds.), Schweizerisches Idiotikon. 17 vols. [a– zih]. Frauenfeld: Huber/Basel: Schwabe. Constantinus Szyrwid [= Konstantinas Sirvydas] (1677). Dictionarium trium lingvarum. 4ᵗʰ edition. Vilnæ: Societas Iesu. [SD¹ = first edition, digitized online at seniejirastai.lki.lt/db.php?source=20]. Измаил И. Срезневский (1893–1912). Матерiалы для словаря древне-русскаго языка по письменнымъ памятникамъ. Санктпетербургъ. Рубен И. Аванесов (vols. 1–5), Игорь С. Улуханов (vol. 6), Вадим Б. Крысько (7– ) (1988– , eds.), Словарь древнерусского языка (xi–xiv вв.). 12 vols. [а–съотъходьнъ]. Москва: Русский язык. Boryś, Wiesław & Popowska-Taborska, Hanna (1994–2006). Słownik etymologiczny Kaszubszczyny. 6 vols. Warszawa: Slawistyczny Ośrodek Wydawniczy. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 bibliography СИГТЯ iv СИГТЯ v СИГТЯ vi SJS SKES Сл. Акад. Sł. Warsz. SMS СРГК СРГС СРЯ 11–17 СРНГ SSA SSKJ² SSN SSP TLL 315 Эдхям Р. Тенишев (2001, ed.), Сравнительно-историческая грамматика тюркских языков: Лексика. Москва: Наука. Эдхям Р. Тенишев (2002, ed.), Сравнительно-историческая грамматика тюркских языков: Региональные реконструции. Москва: Наука. Эдхям Р. Тенишев & Анна В. Дыбо (2006, ed.), Сравнительно-историческая грамматика тюркских языков: Пратюркский язык-основа. Москва: Наука. Joseph Kurz (1966–1997, ed.), Slovník jazyka staroslověnského. 4 vols. Praha: Academia, Euroslavica. Also corrected online edition at http://​ gorazd.org/gulliver/. Yrjö H. Toivonen, Erkki Itkonen & Aulis J. Joki (1955–1981, eds.), Suomen kielen etymologinen sanakirja. 7 vols. Helsinki: Suomalais-Ugrilainen Seura. Словарь Академіи Россійской (1789–1794). 6 vols. Санктпетербургъ: Императорская Академія Наукъ. Jan Karłowicz et al. (1898–1927, eds.), Słownik jęzika polskiego. Warszawa. Suomen murteiden sanakirja (online edition). Accessed at kaino.kotus​ .fi/sms. Александр С. Герд. (1994–2002, ed.), Словарь русских говоров Карелии и сопредельных областей. 6 vols. Санкт-Петербург: СанктПетербургский государственный университет. Александр И. Федоров (1999–2006, ed.), Словарь русских говоров Сибири. Новосибирск: Наука. Степан Г. Бархударов, Галина А. Богатова et al. (1975– , eds.), Словарь русского языка xi–xvii вв. 31 vols. [а–улокъ]. Москва: Наука. Федот П. Филин (vols. 1–23), Федор П. Сороколетов (vols. 24–45), Сергей А. Мызников (46– ) (1966– , eds.), Словарь русских народных говоров. 52 vols. [а–цванки]. Ленинград–Санкт-Петербург: Наука. Erkki Itkonen, Ulla-Maija Kulonen et al. (1992–2001, eds.), Suomen sanojen alkuperä. 3 vols. Helsinki: Suomalaisen Kirjallisuuden Seura. Slovar slovenskega knjižnega jezika [2ⁿᵈ edition] (2014). 2 vols. Ljubljana: Inštitut za slovenski jezik. Ivor Ripka & Adriana Ferenčíková (1994– , eds.), Slovník slovenských nárečí. 3 vols [a–r]. Bratislava: Veda. Stanisław Urbańczyk et al. (1953–1995, eds.), Słownik staropolski. 11 vols. Warszawa–Wrocław: Ossolineum. Thesaurus linguae latinae (1900– ). Leipzig: Teubner/Berlin: De Gruyter. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 316 ТРС UEW VKS VMS bibliography Татарско-русский словарь (2007). 2 vols. Казань: Маргариф. Rédei, Károly (1986–1992). Uralisches etymologisches Wörterbuch. 3 vols. Wiesbaden: Harrassowitz. Elna Adler, Silja Grünberg & Merle Leppik (2012, eds.), Vadja keele sõnaraamat. Tallinn: Eesti Keele Sihtasutus. Valdek Pall (1989, ed.), Väike murdesõnastik (online edition). Accessed online at https://eki.ee/dict/vms/. List by Author Абаев, Василий И. (1949). Осетинский язык и фольклор i. Москва: Наука. Абаев, Василий И. (1958). Историко-этимологический словарь осетинского языка i. A-K’. Москва–Ленинград: Наука. Абаев, Василий И. (1965). Скифо-европейские изоглоссы: на стыке Востока и Запада. Москва: Главная редакция восточной литературы. Абаев, Василий И. (1973). Историко-этимологический словарь осетинского языка ii. L-R. Ленинград: Наука. Абаев, Василий И. [= V.I. Abaev] (1996). Corrections and additions to the Ossetic etymological dictionary. In: Hans H. Hock (ed.), Historical, Indo-European, and Lexicographical Studies [FS Zgusta], Berlin–New York: De Gruyter, 197–220. Adams, Douglas Q. (1985). The Indo-European Word for ‘apple’ Again. Indogermanische Forschungen 90, 79–82. Adams, Douglas Q. (2013). A Dictionary of Tocharian B: Revised and Greatly Enlarged. Amsterdam–New York: Rodopi. Ahlqvist, August (1871). De vestfinska språkens kulturord. Helsinki: Finska litteratursällskapet. Aikio, Ante (2002). New and Old Samoyed Etymologies. Finnisch-Ugrische Forschungen 57, 9–57. Aikio, Ante (2004). An essay on substrate studies and the origin of Saami. In: Irma Hyvärinen, Petri Kallio & Jarmo Korhonen (eds.), Etymologie, Entlehnungen und Entwicklungen [FS Koivulehto]. Helsinki: Société Néophilologique, 5–34. Aikio, Ante (2006a). On Germanic-Saami contacts and Saami prehistory. SuomalaisUgrilaisen Seuran Aikakauskirja 91, 9–55. Aikio, Ante (2006b). Etymological Nativization of Loanwords: a Case Study of Saami and Finnish. In: Ida Toivonen & Diane Nelson (eds.), Saami Linguistics. Amsterdam– Philadelphia: John Benjamins, 17–52. Aikio, Ante (2009). The Saami Loanwords in Finnish and Karelian. Diss. University of Oulu. Aikio, Ante (2012a). An essay on Saami ethnolinguistic prehistory. In: Riho Grünthal Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 bibliography 317 & Petri Kallio (eds.), A Linguistic Map of Prehistoric Northern Europe. Helsinki: Suomalais-Ugrilainen Seura, 63–117. Aikio, Ante (2012b). On Finnic long vowels, Samoyed vowel sequences, and ProtoUralic *x. In: Tiina Hyytiäinen, Lotta Jalava, Janne Saarikivi & Erika Sandman (eds.), Per urales ad Orientem: iter polyphonicum multilingue [FS Janhunen]. Helsinki: Suomalais-Ugrilainen Seura, 227–250. Aikio, Ante (2014a) On the reconstruction of Proto-Mari vocalism. Journal of Language Relationship 11, 125–157. Aikio, Ante (2014b). Studies in Uralic Etymology ii: Finnic Etymologies. Linguistica Uralica 50, 1–19. Aikio, Ante (2014c). Studies in Uralic Etymology iii: Mari Etymologies. Linguistica Uralica 50, 81–93. Aikio, Ante (2015a). The Finnic ‘secondary e-stems’ and Proto-Uralic vocalism. Suomalais-Ugrilaisen Seuran Aikakauskirja 95, 25–66. Aikio, Ante (2015b). Studies in Uralic Etymology iv: Ob-Ugric Etymologies. Linguistica Uralica 51, 1–20. Aikio, Ante (2021). Studies in Uralic Etymology v: Permic Etymologies. Linguistica Uralica 57, 161–179. Aikio, Ante (2022). Proto-Uralic. In: Marianne Bakró-Nagy, Johanna Laakso & Elena Skribnik (eds.), The Oxford Guide to the Uralic Languages. Oxford: University Press, 3–27. Aikio, Ante (in prep.). Uralic Etymological Dictionary (draft version of entries A–Ć). Draft paper published on academia.edu (17 January 2020). Aitzetmüller, Rudolf (1958). Das Hexaemeron des Exarchen Johannes i. Graz: Akademie. Alatalo, Jarmo [= Ярмо Алатало] (1998). Сӱ̄ссыӷӯй э̄җипсан. Қэ̄ тқый қӯланни. Максимкин Яр–Helsinki: Национальный клуб максимоярочка. Alatalo, Jarmo (2004, ed.). Sölkupisches Wörterbuch. Helsinki: Finnisch-Ugrische Gesellschaft. Alessio, Giovanni (1946). Un’ oasi linguistica preindoeuropea nella regione baltica? Studi Etruschi 19, 141–176. Allen, James (2020). Ancient Egyptian Phonology. Cambridge: University Press. Allentoft, Morten E. et al. (forthc.). Population genomics of Stone Age Eurasia. Preprint version posted on bioRxiv, 5 May 2022. doi.org/10.1101/2022.05.04.490594. Аллмяэ, Райли, Маре Аун, & Лиина Малдре (2018). Предварительные результаты изучения остеологического материала курганных могильников Рысна-Сааре i и ii в Северной Сетумаа (Юго-Восточная Эстония). Археология и история Пскова и Псковской земли. Материалы lii заседания. Псков, 298–310. Alminauskis, Kazimieras (1934). Die Germanismen des Litauischen i. Die deutschen Lehnwörter im Litauischen. Diss. Universität Leipzig. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 318 bibliography Ambrazas, Saulius (1993). On the development of diminutives in the Baltic languages. Linguistica Baltica 2, 47–67. Ambrazas, Saulius (2000). Daiktavardžių darybos raida ii. Lietuvių kalbos vardažodiniai vediniai. Vilnius: Mokslo ir enciklopedijų leidybos institutas. Ambrosiani, Björn (2005). Birka and Scandinavia’s trade with the East. Russian History 32, 287–296. van Amerongen, Yvonne (2016). Wild West Frisia. Diss. Universiteit Leiden. Andersen, Henning (1996a). Reconstructing Prehistorical Dialects: Initial Vowels in Slavic and Baltic. Berlin: Mouton de Gruyter. Andersen, Henning (1996b). Взгляд на славянскую прародину: доисторические изменения в экологии и культуре. Вопросы языкознания 1996/5, 65–106. Andersen, Henning (2003). Slavic and the Indo-European migrations. In: Henning Andersen (ed.), Language Contacts in Prehistory: Studies in Stratigraphy. Amsterdam–Philadelphia: Benjamins, 45–76. Anderson, Walter (1938). Zum Lehnwort im Baltendeutsch. Zeitschrift für Mundartforschung 14, 146–149. Аникин, Александр (1992). [Review of] G. Holzer. Entlehnungen aus einer bisher unbekannten indogermanischen Sprache im Urslavischen und Urbaltischen. Вопросы языкознания 1992/5, 169–173. Аникин, Александр (1998). Этимология и балто-славянское лексическое сравнение в праславянской лексикографии. Материалы для балто-славянского словаря i. [*a–*go-]. Новосибирск: Сибирский хронограф. Аникин, Александр (2000). Этимологический словарь русских диалектов Сибири [2ⁿᵈ edition]. Москва–Новосибирск: Наука. Аникин, Александр (2003). Этимологический словарь русских заимствований в языках Сибири. Новосибирск: Наука. Аникин, Александр (2004). [Review of] Słownik prasłowiański. Tom xiii. Goda — Gyža. Балто-славянские исследования 16, 369–386. Аникин, Александр (2005). Опыт словаря лексических балтизмов в русском языке. Новосибирск: Наука. Аникин, Александр (2014). Проблемы изучения балтизмов в русском языке. Siberian Journal of Philology 2014/4, 189–195. [Аникин РЭС — see РЭС under Abbreviated Titles.] Anthony, David W. (2007). The Horse, the Wheel and Language. Princeton–Oxford: Princeton University Press. Anthony, David W. & Ringe, Don (2015). The Indo-European Homeland from Linguistic and Archaeological Perspectives. Annual Review of Linguistics 1, 199–219. Anthony, David W. et al. (2005). The Samara Valley Project: Late Bronze Age Economy and Ritual in the Russian Steppes. Eurasia Antiqua 11, 395–417. Ariste, Paul (1962). Mõnda substraadist. Keel ja Kirjandus 5, 13–17. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 bibliography 319 Ariste, Paul (1971). The Earliest Substrata in the Baltic-Finnic Languages. Néprajz és Nyelvtudomány 15, 5–11. Ariste, Paul (1975). Substraatseid kalanimetusi. Keel ja Kirjandus 18, 467–472. Arumaa, Peeter (1930). Litauische mundartliche Texte aus der Wilnaer Gegend. Dorpat: Universitas Tartuensis. Arumaa, Peeter (1964). Urslavische Grammatik i. Einleitung. Lautlehre. Heidelberg: Winter. Arumaa, Peeter (1976). Urslavische Grammatik ii. Konsonantismus. Heidelberg: Winter. Ахметьянов, Рифкат Г. (1981). Общая лексика духовной культуры народов Среднего Поволжья. Москва: Наука. Ахметьянов, Рифкат Г. (2015). Этимологический словарь татарского языка. 2 vols. Казань: Магариф–Вакыт. Babel, Anna M. (2016). Affective motivations for borrowing: Performing local identity through loan phonology. Language & Communication 49, 70–83. Bacevičiūtė et al. (2004) = Rima Bacevičiūtė, Audra Ivanauskienė, Asta Leskauskaitė & Edmundas Trumpa (eds.), Lietuvių kalbos tarmių chrestomatija. Vilnius: Lietuvių kalbos institutas. Bailey, Harold W. (1979). Dictionary of Khotan Saka. Cambridge–New York–Melbourne: Cambridge University Press. Bakkum, Gabriël C.L. M. (2009). The Latin Dialect of the Ager Faliscus. Amsterdam: University Press. Balaišis, Vytautas (1994). Das Problem der gotischen Diphthonge ai, au und die litauischen Lehnwörter kvietỹs „Weizen“, kliẽpas „Laibbrot“. Baltistica 4 Priedas, 5– 12. Bańkowski, Andrzej (2000). Etymologiczny słownik języka polskiego. 2 vols. Warszawa: Wydawnictwo naukowe PWN. Barbaro, Cecilia Conati et al. (2021). A Multidisciplinary Approach to the Study of Early Neolithic Pyrotechnological Structures. The Case Study of Portonovo (Marche, Italy). Open Archaeology 7, 1160–1175. Barber, Elizabeth J.W. (1991). Prehistoric Textiles. Princeton: University Press. Bartens, Raija (1999). Mordvalaiskielten rakenne ja kehitys. Helsinki: Suomalais-Ugrilainen Seura. Batisti, Roberto (2021). Birds of a feather? Latin columba ‘pigeon, dove’ and Greek κόλυμβος ‘grebe’. Indogermanische Forschungen 126, 205–228. Becker, Johann P. (1904). Kurische sprache in Perwelk. Beiträge zur Kunde der indogermanischen Sprachen 28, 257–281. Bednarczuk, Leszek (1976). Zapożyczenia ugrofińskie w językach bałtosłowiańskich. Acta Baltico-Slavica 9, 39–64. van Beek, Lucien (2011). The “Saussure effect” in Greek: a reinterpretation of the evidence. Journal of Indo-European Studies 39, 129–175. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 320 bibliography van Beek, Lucien (2013). ‘Struck (with Blood)’. The Meaning and Etymology of παλάσσω and Its Middle Perfect πεπάλακτο. Mnemosyne 66, 541–565. van Beek, Lucien (2018). Greek πέδιλον ‘sandal’ and the Origin of the e-Grade in PIE ‘foot’. In: Lucien van Beek et al. (eds.), Farnah [FS Lubotsky]. Ann Arbor–New York: Beech Stave Press. van Beek, Lucien (2022). The Reflexes of Syllabic Liquids in Ancient Greek. Leiden– Boston: Brill. Beekes, Robert S.P. (1969). The Development of the Proto-Indo-European Laryngeals in Greek. Den Haag–Paris: Mouton. Beekes, Robert S.P. (1971). A European substratum word. Orbis 20, 132–137. Beekes, Robert S.P. (1975). [Review of] E.J. Furnée, Die wichtigsten konsonantischen Erscheinungen des Vorgriechischen. Lingua 36, 69–82. Beekes, Robert S.P. (1982). GAv. må the PIE word for ‘moon, month,’ and the perfect participle. Journal of Indo-European Studies 10, 53–64. Beekes, Robert S.P. (1987). The PIE Words for ‘Name’ and ‘Me’. Die Sprache 33, 1–12. Beekes, Robert S.P. (1988). PIE. RHC- in Greek and Other Languages. Indogermanische Forschungen 93, 22–45. Beekes, Robert S.P. (1996). Ancient European Loanwords. Historische Sprachforschung 109, 215–236. Beekes, Robert S.P. (1998). The origin of Lat. aqua, and of *teutā ‘people’. Journal of Indo-European Studies 26, 459–466. Beekes, Robert S.P. (2000a). European substratum words in Greek. In: Michaela Ofitsch & Christian Zinko (eds.), 125 Jahre Indogermanistik in Graz. Graz: Leykam, 21–31. Beekes, Robert S.P. (2000b). Roots with nasal infix in Pokorny. Amsterdamer Beiträge zur älteren Germanistik 54, 3–23. Beekes, Robert S.P. (2010). Etymological Dictionary of Greek. Leiden–Boston: Brill. Beekes, Robert S.P. (2011) = Michiel de Vaan (ed.), Comparative Indo-European Linguistics [2ⁿᵈ edition]. Amsterdam–Philadelphia: Benjamins. Beekes, Robert S.P. (2014) = Stefan Norbruis (ed.), Pre-Greek: Phonology, Morphology, Lexicon. Leiden–Boston: Brill. Behre, Karl-Ernst (1992). The history of rye cultivation in Europe. Vegetation History and Archaeobotany 1, 141–156. Behre, Karl-Ernst (2008). Collected seeds and fruits from herbs as prehistoric food. Vegetation History and Archaeobotany 17, 65–73. Беляева, Ольга П. (1973). Словарь говоров Соликамского района Пермской области. Пермь: Пермский государственный педагогический институт. Bellquist, Julie B. (1993). “Badger” in Indo-European. Journal of Indo-European Studies 21, 331–346. Benecke, Norbert (2005). The Holocene distribution of European bison — the archaeozoological record. Antropologia-Arkeologia 57, 421–428. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 bibliography 321 Bennett, William H. (1949). The Monophthongization of Gothic ái áu. Language 25, 15– 21. Bentlin, Mikko (2008). Niederdeutsch-finnische Sprachkontakte. Helsinki: SuomalaisUgrilainen Seura. Bereczki, Gábor (1994). Grundzüge der tscheremissischen Sprachgeschichte i. Szeged. Bereczki, Gábor (2013). Etymologisches Wörterbuch des Tscheremissischen. Wiesbaden: Harrassowitz. de Bernardo Stempel, Patrizia (1987), Die Vertretung der indogermanischen Liquiden und nasalen Sonanten im Keltischen. Innsbruck: Institut für Sprachwissenschaft. de Bernardo Stempel, Patrizia (1999). Nominale Wortbildung des älteren Irischen. Tübingen: Niemeyer. Berneker, Erich (1899). Von der Vertretung des idg. ēŭ im baltisch-slavischen Sprachzweig. Indogermanische Forschungen 10, 145–167. Berneker = Berneker, Erich (1908–1914). Slavisches etymologisches Wörterbuch. 2 vols. Heidelberg: Winter. Bertoldi, Vittorio (1932). Problèmes de substrat. Essai de méthodologie dans le domaine préhistorique de la toponymie et du vocabulaire. Bulletin de la Société de linguistique de Paris 32, 93–184. Bezlaj = Bezlaj, France (1977–2007). Etimološki slovar slovenskega jezika. 5 vols. Ljubljana: Mladinka knjiga; Založba ZRC. Bezzenberger, Adalbert (1877). Beiträge zur Geschichte der litauischen Sprache. Göttingen: Peppmüller. Bezzenberger, Adalbert (1882). Ein lettisches lautgesetz. Beiträge zur Kunde der indogermanischen Sprachen 7, 273–277. Bezzenberger, Adalbert (1885). Lettische Dialekt-Studien. Göttingen: Vandenhoeck & Ruprecht. Bezzenberger, Adalbert (1887). Mundartliche Texte. Mitteilungen der Litauischen litterarischen Gesellschaft 2, 29–48. Bezzenberger, Adalbert (1911). [Review of] Reinhold Trautmann, Die altpreußischen Sprachdenkmäler. Zeitschrift für vergleichende Sprachforschung 44, 285–331. Bichlmeier, Harald (2016). Zur Frage des nichtindogermanischen Substrats im Germanischen: Ein kurzer Überblick. Slovo a slovesnost 77, 316–336. Bjørnflaten, Jan I. (2006). Chronologies of the Slavicization of Northern Russia Mirrored by Slavic Loanwords in Finnic and Baltic. In: Juhani Nuorluoto (ed.), The Slavicization of the Russian North: Mechanisms and Chronology. Helsinki: Yliopisto, 50–77. Blažek, Václav (1995). Indo-European ‘Apple(s)’. Sborník prací Filozofické fakulty brněnské univerzity A 43, 15–20. Blažek, Václav (2003). Semitic *ṯáwar- “bull” and its relatives. In: Lionel Bender, David Appleyard & Gábor Takács (eds.), Afrasian [FS Diakonoff]. München: Lincom, 7–13. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 322 bibliography Blažek, Václav (2018). Indo-European Dendronyms in the Perspective of External Comparison. Journal of Indo-European Studies 46, 1–45. Blažek, Václav (2019). Tocharian B lasto in Balto-Slavic Perspective. Journal of IndoEuropean Studies 47, 213–222. Blažek, Václav & Janyšková, Ilona (2015). Slavic *tisъ — an evergreen problem of Slavic etymology? In: Ilona Janyšková & Helena Karlíková (eds.), Etymological Research into Old Church Slavonic, 69–98. Bliujienė, Audronė & Curta, Florin (2011). Exotic Lands, Quixotic Friends: Eastern Lithuania and the Carpathian Basin in Late Antiquity and the Early Middle Ages (ad c. 380 to c. 620). Medieval Archaeology 55, 29–64. Blokland, Rogier (2005). The Russian Loanwords in Literary Estonian. Diss. Universiteit Groningen. Boček/Malčík = Vít Boček & Petr Malčík (2011, eds.), Václav Machek: Korespondence. 2 vols. Praha: Lidové noviny. Boisacq, Émile (1916). Dictionnaire Étymologique de la Langue Grecque. Heidelberg: Winter. Bopp, Franz (1833). Vergleichende Grammatik des Sanskrit, Zend, Griechischen, Lateinischen, Litthauischen, Gothischen und Deutschen. Berlin: Akademie der Wissenschaften. Boryś, Wiesław (1977). Czakawskie brńuse. Rocznik Slawistyczny 38, 93–95. Boryś, Wiesław (2008). Słownik etymologiczny języka polskiego. Kraków: Wydawnictwo Literackie. Boutkan, Dirk (1998). On the form of North European substratum words in Germanic. Historische Sprachforschung 111, 102–133. Boutkan, Dirk (2003a). On Gothic magaþ ~ Old Frisian megith and the form of some North European substratum words in Germanic. Amsterdamer Beiträge zur älteren Germanistik 58, 11–28. Boutkan, Dirk (2003b). Lithuanian šlãkas, Norse slag. In: Alfred Bammesberger & Theo Venneman (eds.), Languages in Prehistoric Europe. Heidelberg: Winter, 245– 252. Boutkan, Dirk & Kossmann, Maarten (2001). On the etymology of ‘silver’. North-Western European Language Evolution 38, 3–15. Boutkan, Dirk & Siebinga, Sjoerd M. (2005). Old Frisian Etymological Dictionary. Leiden–Boston: Brill. Braune, Wilhelm (1876). Ueber den altpreussischen diphthong ai. Beiträge zur vergleichenden Sprachforschung 8, 91–101. Breidaks, Antons B. [= А.Б. Брейдак] (1983). Некоторые данные балтизмов финноугорских языков. Baltistica 19, 47–51. Brenzinger, Matthias (1992). Lexical retention in language shift: Yaaku/MukogodoMaasai and Elmolo/Elmolo-Samburu. In: Matthias Brenzinger (ed.), Language Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 bibliography 323 Death: Factual and Theoretical Explorations with Special Reference to East Africa. Berlin: Mouton de Gruyter, 213–254. Brevé, Niels W.P. et al. (2022). Historical reconstruction of sturgeon (Acipenser spp.) spatiotemporal distribution and causes for their decline in North-Western Europe. Biodiversity and Conservation 31, 1149–1173. Brodowski = Jacob Brodowskÿ. Lexicon Germanico-Lithuanicum et Lithuanico-Germanicum [Manuscript]. Accessed online at epaveldas.lt/preview?id=MAB04‑0000708 16. Brückner, Aleksander (1877). Litu-slavische Studien i. Die slavischen Fremdwörter im Litauischen. Weimar: Böhlau. Brückner, Aleksander (1927). Słownik etymologiczny języka polskiego. Kraków: Spółka Wydawnicza. Brückner, Aleksander (1929). Die germanischen Elemente im Gemeinslavischen. Archiv für Slavische Philologie 42, 125–146. Brugmann, Karl (1897). Grundriss der vergleichenden Grammatik der indogermanischen Sprachen. Zweite Bearbeitung i. Einleitung und Lautlehre. Strassburg: Trübner. Brugmann, Karl (1906). Grundriss der vergleichenden Grammatik der indogermanischen Sprachen. Zweite Bearbeitung ii. Vergleichende Laut-, Stammbildungs- und Flexionslehre. Strassburg: Trübner. Buck, Carl Darling (1904). A Grammar of Oscan and Umbrian. Boston: Ginn. Būga, Kazimieras (1908). Aistiški studijai: Tyrinėjimai lygintijo prūsų, latvjų ir liėtuvjų kalbômoksljo srityjẹ i. Peterburgas: Imperatoriškoji Mokslų Akademija. Būga, Kazimieras (1911). Apie lietuvių asmens vardus. Vilnius: Kukta [= RR i: 201–269]. Būga, Kazimieras [= К.К. Буга] (1912). Lituanica. Извѣстiя отдѣленiя русскаго языка и словесности императорской академiи наукъ 17/1, 1–49 [= RR i: 339–383]. Būga, Kazimieras (1913). Kann man Keltenspuren auf baltischem Gebiet nachweisen? Rocznik sławistyczny 6, 1–38 [= RR i: 496–530]. Būga, Kazimieras [= Казимиръ Буга] (1914). Славяно-балтійскія этимологіи 135–160. Русскiй филологическiй вѣстникъ 72, 187–202 [= RR i: 460–472]. Būga, Kazimieras [= Казимиръ Буга] (1915). Славяно-балтійскія этимологіи 161–185. Русскiй филологическiй вѣстникъ 73, 335–343 [= RR i: 473–479]. Būga, Kazimieras [= Казимиръ Буга] (1916). Славяно-балтійскія этимологіи 186–261. Русскiй филологическiй вѣстникъ 75, 141–156 [= RR i: 479–495]. Būga, Kazimieras (1922). Kalba ir senovė. Kaunas: Svietimo Ministerija [= RR ii: 7–330]. Būga, Kazimieras (1923a). Upių vardų studijos ir aisčių bei slavėnų senovė. Tauta ir žodis 1, 1–44 [= RR iii: 493–550]. Būga, Kazimieras (1923b). [Review of] H. H. Bender. A Lithuanian etymological index. Tauta ir žodis 1, 397–405 [= RR iii: 648–657]. Būga, Kazimieras [= Kasimir Būga] (1923–1924). Die Metatonie im Litauischen und Lettischen. Zeitschrift für vergleichende Sprachforschung 51, 109–142; 52, 91–98, 250–302 [= RR ii: 386–483]. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 324 bibliography Būga, Kazimieras (1924a). Lietuvių kalbos žodynas. Kaunas: Svietimo Ministerija [= RR iii: 9–483]. Būga, Kazimieras (1924b). Šis-tas iš lietuvių ir indoeuropiečių senovės. Tauta ir žodis 2, 98–110 [= RR iii: 584–599]. Būga, Kazimieras (1925). Die litauisch-weißrussischen Beziehungen und ihr Alter. Zeitschrift für slavische Philologie 1, 26–55 [= RR iii: 749–778]. Būga RR = Zigmas Zinkevičius (1958–1962, comp.), K. Būga. Rinktiniai Raštai. 4 vols. Vilnius: Valstybinė politinės ir mokslinės literatūros leidykla. Bugge, Sophus (1888). Etymologische Studien ueber germanische Lautverschiebung iii. Beiträge zur Geschichte der deutschen Sprache und Literatur 13, 311–339. Булаховский, Леонид А. (1948). Общеславянские названия птиц. Известия академии наук СССР. Отделение литературы и языка 7/2, 97–124. Булаховский, Леонид А. (1958). Исторический комментарий к русскому литературному языку [5ᵗʰ edition]. Киев: Радянська школа. Булаховский, Леонид А. (1968). Морфологическая проблематика русских наименований птиц. Вопросы языкознания 1968/4, 100–106. Byrd, Andrew M. (2010). Reconstructing Indo-European Syllabification. Diss. University of California. Çabej, Eqrem (1976). Studime gjuhësore: Studime etimologjike në fushë të shqipes. 2 vols. Prishtinë: Rilindja. Цаболов, Руслан Л. (2001). Этимологический словарь курдского языка i. A–M. Москва: Восточная литература. Цаболов, Руслан Л. (2010). Этимологический словарь курдского языка ii. N–Ž. Москва: Восточная литература. Campbell, Alistair (1959). Old English Grammar. London: Oxford University Press. Черепанова, Евгения А. (1973). К географии балтизмов в говорах полесья. Baltistica 9, 71–74. Chantraine, Pierre (1933). La formation des noms en grec ancien. Paris: Klincksieck. Chantraine DELG = Chantraine, Pierre (1968–1980). Dictionnaire étymologique de la langue grecque. 4 vols. Paris: Klincksieck. Charpentier, Jarl (1930). Der Name des Roggens. Arkiv för nordisk filologi 46, 63–73. Cheung, Johnny (2002). Studies in the Historical Development of the Ossetic Vocalism. Wiesbaden: Reichert. Cheung, Johnny (2007). Etymological Dictionary of the Iranian Verb. Leiden–Boston: Brill. Cheung, Johnny & Aydemir, Hakan (2002). Turco-Afghanica: On East Iranian *amarnā and Turkic alma, alïmla, almïla ‘apple’. In: Михаил С. Пелевин (ed.), «На пастбище Мысли Благой» [FS Стеблин-Каменский]. Санкт-Петербург: Контраст, 73– 94. Cigale, Matej (1860). Deutsch-slovenisches Wörterbuch. 2 vols. Ljubljana: Blasnik. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 bibliography 325 Clackson, James (1994). The Linguistic Relationship between Armenian and Greek. Oxford: Wiley–Blackwell. Clauson, Gerard (1972). An Etymological Dictionary of Pre-Thirteenth-Century Turkish. Oxford: University Press. Collinder, Björn (1955). Fenno-Ugric Vocabulary: An Etymological Dictionary of the Uralic Languages. Stockholm: Almqvist & Wiksell. Collinder, Björn (1960). Comparative Grammar of the Uralic Languages. Stockholm: Almqvist & Wiksell. Čop, Bojan (1973). Oštirs sprachwissenschaftliche Ideenwelt. Linguistica 13, 13–96. Cordes, Gerhard (1973). Altniederdeutsches Elementarbuch. Heidelberg: Winter. Craddock, Jerry R. (1969). Latin Legacy Versus Substratum Residue. Berkeley–Los Angeles: University of California Press. Crane, E. Eva (1999). The World History of Beekeeping and Honey Hunting. New York– London: Routledge. Даль² = Даль, Владиміръ И. (1880–1882). Толковый словарь живаго великорусскаго языка [2ⁿᵈ edition]. 4 vols. С.-Петербургъ–Москва: Вольфъ. Даль³ = Иван А. Бодуэн-де-Куртенэ (1903–1909, ed.), Толковый словарь живого великорусскаго языка Владимiра Даля [3ʳᵈ edition]. 4 vols. С.-Петербургъ–Москва: Вольфъ. Darden, Bill J. (1990). Laryngeals and Syllabicity in Balto-Slavic and Indo-European. Papers from the 26ᵗʰ Annual Meeting of the Chicago Linguistics Society ii. The Parasession on the Syllable in Phonetics & Phonology, 61–70. Daukantas, Simonas (1846). Dajnes Źiamajtiû (pagał Zôdiû Dajnînikû îszraszytas). Petropìlie: Kray. Decaux, Étienne (1966). La neutralisation n’est pas un changement structurel. Bulletin de la Société de linguistique de Paris 61, 57–81. Delamarre, Xavier (2003). Dictionnaire de la langue gauloise [2ⁿᵈ edition]. Paris: Editions Errance. Della Volpe, Angela (1996). Indo-European Architectural Terms and the Pre-Indo-Europeans. In: Karlene Jones-Bley & Martin E. Huld (eds.), The Indo-Europeanization of Northern Europe. Washington: Institute for the Study of Man, 148–165. Demiraj, Bardhyl (1997). Albanische Etymologien. Untersuchungen zum albanischen Erbwortschatz. Amsterdam–Atlanta: Rodopi. Derksen, Rick (1996). Metatony in Baltic. Amsterdam–Atlanta: Rodopi. Derksen, Rick (1999). North European *KulP- ‘beak, snout’. Baltistica 34, 69–76. Derksen, Rick (2000). Old Icelandic jarpi ‘hazel-grouse’, rjúpa ‘ptarmigan’ and their Germanic and Balto-Slavic cognates. In: Dirk Boutkan & Arend Quak (eds.), Language Contact. Substratum, superstratum, adstratum in Germanic Languages. Amsterdam–Atlanta: Rodopi, 75–87. Derksen, Rick (2002). “Rozwadowski’s change” in Baltic. Baltu filoloģija 9, 5–12. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 326 bibliography Derksen, Rick (2003). Slavic *jь-. In: Peter Houtzagers et al. (eds.), Dutch Contributions to the Thirteenth International Congress of Slavists, Ljubljana: Linguistics. Amsterdam–Atlanta: Rodopi, 97–105. Derksen, Rick (2007). Balto-Slavic Etymological Studies and Winter’s Law: A Concise Review of Dybo 2002. In: Mate Kapović & Ranko Matasović (eds.), Tones and Theories: Proceedings of the International Workshop on Balto-Slavic Accentology. Zagreb: Institut za hrvatski jezik i jezikoslovlje, 39–46. Derksen, Rick (2008). Etymological Dictionary of the Slavic Inherited Lexicon. Leiden– Boston: Brill. Derksen, Rick (2010). The development of PIE *eu in Baltic and Slavic. In: Elena Stadnik-Holzer & Georg Holzer (eds.), Sprache und Leben der frühmittelalterlichen Slaven [FS Katičić]. Frankfurt am Main: Peter Lang, 37–41. Derksen, Rick (2015). Etymological Dictionary of the Baltic Inherited Lexicon. Leiden– Boston: Brill. Derksen, Rick (2019). ‘Henbane’ in Slavic. In: Uwe Bläsing, Jasmine Dum-Tragut & Theo M. van Lint (eds.), Armenian, Hittite, and Indo-European Studies [Gedenkschrift Weitenberg]. Leuven: Peeters, 113–119. Derksen, Rick (2020). On the earliest Slavic borrowings in East Baltic. Baltu filoloģija 29, 33–53. Deroy, Louis (1956). La valeur du suffixe préhellénique -nth- d’après quelques noms grees en -νϑος. Glotta 35, 171–195. Diebold, A. Richard (1976). Contributions to the Indo-European Salmon Problem. In: William M. Christie, Jr. (ed.), Current Progress in Historical Linguistics. Amsterdam– New York: North-Holland, 341–387. Dini, Pietro U. (2014) = Milda B. Richardson & Robert E. Richardson (2014, trans.), Foundations of Baltic Languages. Vilnius. Doerfer, Gerhard (1965). Türkische und Mongolische Elemente im Neupersischen ii. Wiesbaden: Steiner. Doerfer, Gerhard (1971). Bemerkungen zur Methodik der türkischen Lautlehre. Orientalistische Literaturzeitung 66, 325–344. Donner, Otto (1884). Über den Einfluss des Litauischen auf die finnischen Sprachen. Internationale Zeitschrift für allgemeine Sprachwissenschaft i, 257–271. Donner/Joki = Aulis J. Joki (1944, ed.), Kai Donners kamassisches Wörterbuch. Helsinki: Suomalais-Ugrilainen Seura. Dressler, Wolfgang (1974). [Review of] Die wichtigsten konsonantischen Erscheinungen des Vorgriechischen […] by Edzard J. Furnée. Language 50, 736–738. Driessen, Michiel (2003). *h₂é-h₂us-o-, the Proto-Indo-European term for ‘gold’. Journal of Indo-European Studies 31, 347–362. Dunkel, George E. (2014). Lexikon der indogermanischen Partikeln und Pronominalstämme ii. Lexikon. Heidelberg: Winter. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 bibliography 327 Durkin-Meisterernst, Desmond (2004). Dictionary of Manichaean Texts iii/1: Dictionary of Manichaean Middle Persian. Turnhout: Brepols. Дыбо, Анна В. (2007). Лингвистические контакты ранних тюрков. Лексический фонд. Пратюркский период. Москва: Восточная литература. Дыбо, Анна В. (2008). Материальный быт ранних тюрок. Жилище. In: Анна В. Дыбо (ed.), Природное окружение и материальная культура пратюркских народов. Москва: Восточная литература, 219–272. Дыбо, Владимир А. (1961). Сокращение долгот в кельто-италийских языках и его значение для балто-славянской и индоевропейской акцентологии. Вопросы славянского языкознания 5, 9–34. Дыбо, Владимир А. (2000). Морфонологизованные парадигматические акцентные системы i. Москва: Языки русской культуры. Дыбо, Владимир А. (2002). Balto-Slavic Accentology and Winter’s Law. Studia Linguarum 3, 295–515. Дыбо, Владимир А. (2008). Германское сокращение индоевропейских долгот, германский «Verschärfung» (закон Хольцмана) и балто-славянская акцентология. Аспекты компаративистики 3, 537–608. Eason, Perri, Basem Rabia & Omar Attum (2016). Hunting of migratory birds in North Sinai, Egypt. Bird Conservation International 26, 39–51. Eckert, Rainer (1983). Die Nominalstämme auf -i im Baltischen. Berlin: Akademie der Wissenschaften. Ehrlich, Freydis et al. (2021). Application of morphometric and stable isotope analyses for distinguishing domestic and wild geese. International Journal of Osteoarchaeology 32, 457–466. Elger, Georg (1683). Dictionarium Polono-Latino-Lottauicum. Vilnius: Societas Iesu. Emmerick, Ronald E. (1968). Saka Grammatical Studies. London: Oxford University Press. Emmerick, Ronald E. (1989). Khotanese and Tumshuqese. In: Rüdiger Schmitt (ed.), Compendium Linguarum Iranicarum. Wiesbaden: Reichert, 204–229. Endzelīns, Jānis [= І. Эндзелинъ] (1899). Латышскія заимствованія изъ славянскихъ языковъ. Живая старина 9/3, 285–312 [= DI i: 80–112]. Endzelīns, Jānis [= J. Endzelin] (1905). Zur erweichung der gutturale im Lettischen. Beiträge zur Kunde der indogermanischen Sprachen 29, 178–195 [= DI i: 290–306]. Endzelīns, Jānis [= И. Эндзелинъ] (1907). О происхожденіи литовско-латышскаго ë. Извѣстія Отдѣленія русскаго языка и словесности Императорской Академіи наукъ 12/1, 40–66 [= DI ii: 9–30]. Endzelīns, Jānis [= J. Endzelin] (1909). Lettische Miscellen. Zeitschrift für vergleichende Sprachforschung 42, 375–380 [= DI ii: 84–89]. Endzelīns, Jānis [= И. Эндзелинъ] (1911). Славяно-балтийскiе этюды. Харьковъ: Зильбербергъ [= DI ii: 167–354]. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 328 bibliography Endzelīns, Jānis [= J. Endzelin] (1914a). Miszellen. Indogermanische Forschungen 33, 119–127 [= DI ii: 528–537]. Endzelīns, Jānis [= J. Endzelin] (1914b). Zu den kurischen Bestandteilen des Lettischen. Indogermanische Forschungen 33, 96–104 [= DI ii: 504–512]. Endzelīns, Jānis [= J. Endzelin] (1923). Lettische Grammatik. Heidelberg: Winter. Endzelīns, Jānis [= J. Endzelin] (1924). Germanisch-baltische Miszellen. Zeitschrift für vergleichende Sprachforschung 52, 110–128 [= DI iii/1: 395–412]. Endzelīns, Jānis [= J. Endzelin] (1932). [Review of] E. Blese, Latviešu personu vārdu studijas i. Zeitschrift für Slavische Philologie 9, 245–262 [= DI iii/2: 498–517]. Endzelīns, Jānis [= J. Endzelin] (1933). Zur Herkunft der baltischen Lehnwörter des Finnischen. Mémoires de la Société finno-ougrienne 67, 76–83 [= DI iii/2: 128–134]. Endzelīns, Jānis (1937). Sīkumi. Filologu biedrības raksti 17, 162–166 [= DI iii/2: 286– 292]. Endzelīns, Jānis (1943). Senprūšu valoda. Riga: University Press [= DI iv/2: 9–351]. Endzelīns DI = Daina Zemzare, Hermanis Bendiks et al. (eds., 1974–1982). Jānis Endzelīns. Darbu izlase. 4 vols. Rīga: Zinātne Endzelīns/Schmalstieg 1971 = William R. Schmalstieg & Benjamiņš Jēgers (1971, trans.), Jānis Endzelīns’ Comparative Phonology and Morphology of the Baltic Languages. Den Haag–Paris: Mouton. [Endzelīns ME — see ME under Abbreviated Titles.] Erhart, Adolf (1995). Archaisch oder konservativ? Das Anatolische und das Baltische. Linguistica Baltica 4, 1–12. Erjavec, Fran (1883). Is pótne torbe. Letopis Matice Slovenske 1882/1883, 195–351. Erman/Grapow = Erman, Adolf & Grapow, Hermann (1926–1961). Wörterbuch der aegyptische Sprache. 5 vols. Leipzig: Hinrichs. Ernout/Meillet = Ernout, Alfred & Meillet, Antoine (1951). Dictionnaire étymologique de la langue latine [3ʳᵈ edition]. Paris: Klincksieck. Ertuğ, Füsun (2000). An Ethnobotanical Study in Central Anatolia (Turkey). Economic Botany 54/2, 155–182. Falileyev, Alexander & Isaac, Graham R. (2003). Leeks and garlic: the Germanic ethnonym Cannenefates, Celtic *kasn- and Slavic *kesn-. North-Western European Language Evolution 42, 3–12. Falk/Torp = Falk, Hjalmar S. & Torp, Alf (1910–1911). Norwegisch-dänisches etymologisches wörterbuch. 2 vols. Heidelberg: Winter. Федотов = Федотов, Михаил Р. (1996). Этимологический словарь чувашского языка. 2 vols. Чебоксары: Чувашский государственный университет гуманитарных наук. Feist, Sigmund (1910). Die germanische und die hochdeutsche lautverschiebung sprachlich und ethnographisch betrachtet. Beiträge zur Geschichte der deutschen Sprache und Literatur 37, 307–354. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 bibliography 329 Feist, Sigmund (1913). Kultur, Ausbreitung und Herkunft der Indogermanen. Berlin: Weidmann. Feist, Sigmund (1932). The Origin of the Germanic Languages and the Indo-Europeanising of North Europe. Language 8, 245–254. Fennel, Trevor G. (1982). The First Latvian Grammar: J.G. Rehehusen’s “Manductio ad linguam lettonicam …”. Melbourne: Latvian Tertiary Committee. Fiesel, Eva (1928). Namen des griechischen Mythos im Etruskischen. Göttingen: Vandenhoeck & Ruprecht. Filipović, Dragana et al. (2020). New AMS ¹⁴C dates track the arrival and spread of broomcorn millet cultivation and agricultural change in prehistoric Europe. Scientific Reports 10:13698. Foley, James (1977). Foundations of theoretical phonology. Cambridge: University Press. Forssman, Berthold (2001). Lettische Grammatik. Dettelbach: Röll. Fortescue, Michael & Vajda, Edward (2022). Mid-Holocene Language Connections between Asia and North America. Leiden–Boston: Brill. Fraenkel, Ernst (1936a). [Review of] E. Dickenmann, Untersuchungen über die Nominalkomposition im Russischen. Zeitschrift für Slavische Philologie 13, 205–236. Fraenkel, Ernst (1936b). Die indogermanischen -l-Stämme. Zeitschrift für vergleichende Sprachforschung 63, 168–201. Fraenkel, Ernst (1936c). Zur Suffixsubstitution im Litauischen. Zeitschrift für vergleichende Sprachforschung 63, 201–206. Fraenkel, Ernst (1937). Zur eu-Erweiterung indogermanischer Wurzeln unter besonderer Berüksichtigung [sic.] von Baltisch, Slavisch und Griechisch. Mélanges Émile Boisacq i, 355–381. Fraenkel, Ernst (1950a). Die baltischen Sprachen. Heidelberg: Winter. Fraenkel, Ernst (1950b). [Review of] J.B. Hofmann: Etymologisches Wörterbuch des Griechischen. Gnomon 22, 235–239. Fraenkel, Ernst (1951). Zur indoeuropäischen Wortbildung und Etymologie. Lingua Posnaniensis 3, 113–134. [Fraenkel LEW — See LEW under Abbreviated Titles.] Freudenthal, Axel O. & Vendell, Hermann (1886). Ordbok öfver estländsk-svenska dialekterna. Helsingfors: Tidnings- & Tryckeri-Aktibolagets tryckeri. Friedrich, Paul (1970). Proto-Indo-European Trees. Chicago–London: University of Chicago. Frisk = Frisk, Hjalmar (1954–1972). Griechisches Etymologisches Wörterbuch. 3 vols. Heidelberg: Winter. Frischbier = Frischbier, Hermann (1882–1883). Preussisches Wörterbuch: Ost- und westpreussische Provinzialismen in alphabetischer Folge. 2 vols. Berlin: Enslin. Furlan, Metka (2013). Novi etimološki slovar slovenskega jezika. Poskusni zvezek. Ljubljana: Znanstvenoraziskovalni center. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 330 bibliography Furnée, Edzard J. (1972). Die wichtigsten konsonantischen Erscheinungen des Vorgriechischen. Den Haag–Paris: Mouton. Галинская, Елена А. (1993). О хронологии некоторых изменений в системе вокализма праславянского языка. In: Борис А. Успенский & Мария Н. Шевелёва (eds.), Исследования по славянскому историческому языкознанию [Gedenkschrift Хабургаев]. Москва: Московский государственный университет, 35–46. Гальковскiй, Николай М. (1913). Борьба христiанства съ остатками язычества въ древней Руси ii. Древне-русскiя слова и поученiя, направленныя противъ остатковъ язычества въ народѣ. Москва: А. Снегирева. Гамкрелидзе, Тамаз В. & Иванов, Вячеслав В. (1984). Индоевропейский язык и индоевропейцы. Тбилиси: Университет. Garde, Paul [= П. Гард] (1974). К истории восточнославянских гласных среднего подъема. Вопросы языкознания 1974/3, 106–115. Garnier, Romain (2014). Nouvelles considérations sur l’effet Kortlandt. Glotta 90, 139– 159. Garnier, Romain (2015). [Review of] Robert Stephen Paul Beekes, Pre-Greek: Phonology Morphology, Lexicon. Journal asiatique 303, 329–332. Gāters, Alfred (1953). Der baltische Name für das Grummet. Zeitschrift für vergleichende Sprachforschung 71, 113–117. Gebauer = Gebauer, Jan (1903–1916). Slovník staročeský. 2 vols. [A–N]. Praha: Česká akademie věd a umění/Unie. Gelumbeckaitė, Jolanta (2017). The evolution of Baltic. In: Jared Klein, Brian Joseph & Matthias Fritz (eds.), Handbook of Comparative and Historical Indo-European Linguistics. Berlin: De Gruyter, 1712–1715. Георгиев, Владимир И. (1957). Тракийският език. София: Българска академия на науките. Георгиев, Владимир И. (1973). [Review of] Furnée, Edzard J.: Die wichtigsten konsonantischen Erscheinungen des Vorgriechischen. Kratylos 17, 164–167. Герд, Александр С. (1970). Из истории прибалтийско-финских названий рыб (sintti, tintti, tint). Советское финно-угроведение 6/2, 87–91. Герд, Александр С. (1980). Лексические заимствования или слова-автохтоны. iv Всесоюзная конференция балтистов. Тезисы докладов. Рига, 125. Герд, Александр С. (1981). Из истории прибалтийско-финских названий рыб (siika, siiga; muhju). In: Георгий М. Керт & Мария И. Зайцева (eds.), Прибалтийскофинское языкознание, 50–54. Геров = Геровъ, Найденъ (1895–1904). Рѣчник на Блъгарскый языкъ. 5 vols. Пловдивъ: Съгласие. Gerullis, Georg (1922). Die altpreußischen Ortsnamen. Berlin–Leipzig: De Gruyter. Gharib, Badrozzaman [= Qarīb, Badr al-Zamān] (1995). Sogdian Dictionary. Tehran: Farhangan. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 bibliography 331 Giannelli, Ciro & Vaillant, André (1958). Un lexique macédonien du xviᵉ siècle. Paris: Institut d’études slaves. Gilbert, M. Thomas P. & Shapiro, Michael D. (2013). Pigeons, Domestication of. In: Claire Smith (ed.), Encyclopedia of Global Archaeology. New York: Springer. Gippert, Jost (2017). Armeno-Albanica ii: Exchanging doves. In: Bjarne S.S. Hansen, Adam Hyllested, Anders R. Jørgensen & Guus Kroonen (eds.), Usque ad radices [FS Olsen]. København: Museum Tusculanum, 179–192. Girininkas, Algirdas (2019). Hoes and the First Farmers on Lithuanian Territory: From Hoe to Ard. Archaeologia Baltica 26, 66–79. Gliwa, Bernd (2008). Ist lit. túopa ‘Pappel’ wirklich ein Erbwort? Baltistica 43, 239–244. Gliwa, Bernd (2009). Pflanzennamen mit Präfix pa- und angrenzendes im Litauischen und Lettischen. Baltistica 44, 77–90. Gołąb, Zbigniew (1982). Kiedy nastąpilo rozszczepienie językowe Baltów i Słowian. Acta Baltico-Slavica 14, 121–134. Gołąb, Zbigniew (1992). The Origins of the Slavs. Columbus: Slavica. Gorbachov, Slava [Yaroslav] (2006). “North Indo-European” *tuhₓsn̥ tih₂ ‘1000’: a new proposal [Conference handout]. AATSEEL Annual Meeting, Philadelphia (29 December 2006). Grabowski, Radosław (2011). Changes in cereal cultivation during the Iron Age in southern Sweden. Vegetation History and Archaeobotany 20, 479–494. Griffin, David A. (1959). Los mozarabismos del «Vocabulista» atribuído a Ramón Martí (Continuación). al-Andalus 24, 333–380. Grikpėdis, Mindaugas & Motuzaitė Matuzevičiūtė, Giedrė (2016). The beginnings of rye (Secale cereale) cultivation in the East Baltics. Vegetation History and Archaeobotany 25, 601–610. Grikpėdis, Mindaugas & Motuzaitė Matuzevičiūtė, Giedrė (2017). A review of the earliest evidence of agriculture in Lithuania and the earliest direct AMS date on cereal. European Journal of Archaeology 21/2, 1–16. Grikpėdis, Mindaugas & Motuzaitė Matuzevičiūtė, Giedrė (2020). From barley to buckwheat: Plants cultivated in the Eastern Baltic region until the 13th–14th century ad. In: Santeri Vanhanen & Per Lagerås (eds.), Archaeobotanical studies of past plant cultivation in northern Europe. Groningen: Barkhuis. Grinaveckienė, Elena & Mackevič, Juzefa F. (1989). Lie. (i)au vokalizmo atliepai baltarusių lituanizmuose. Baltistica 3 Priedas, 70–80. Grinaveckis, Vladas (1972). Žem. jeĩnis, jeĩnė ir mum̃ siai. Baltistica 8, 73–77. Grünthal, Riho (2012). Baltic loanwords in Mordvin. In: Riho Grünthal & Petri Kallio (eds.), A Linguistic Map of Prehistoric Northern Europe. Helsinki: SuomalaisUgrilainen Seura, 297–343. Haak, Wolfgang et al. (2023). The Corded Ware Complex in Europe in Light of Current Archeogenetic and Environmental Evidence. In: Kristian Kristiansen, Guus Kroonen Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 332 bibliography & Eske Willerslev (eds.), The Indo-European Puzzle Revisited. Cambridge: University Press, 63–80. Haas, Otto (1960). Das frühitalische Element: Versuch über die Sprache der ersten Indogermanen Italiens. Wien: Notring. Hackstein, Olav (1992). Eine weitere griechisch-tocharische Gleichung: Griechisch πτῆξαι und tocharisch B pyāktsi. Glotta 70, 136–165. Häkkinen, Jaakko (2009). Kantauralin ajoitus ja paikannus: perustelut puntarissa. Suomalais-Ugrilaisen Seuran Aikakauskirja 92, 9–56. Häkkinen, Kaisa (2004). Nykysuomen etymologinen sanakirja. Juva: WSOY. Häkkinen, Kaisa & Lempiäinen, Terttu (1996). Die ältesten Getreidepflanzen der Finnen und ihre Namen. Finnisch-ugrische Forschungen 53, 115–182. Hamann, Silke R. (2003). The Phonetics and Phonology of Retroflexes. Diss. Universiteit Utrecht. Hamp, Eric P. (1965). Evidence in Keltic. In: Werner Winter (ed.), Evidence for Laryngeals. The Hague: Mouton, 224–235. Hamp, Eric P. (1970a). Lithuanian ugnìs, Slavic ognь. In: Thomas F. Magner & William R. Schmalstieg (eds.), Baltic Linguistics. University Park: Pennsylvania State University, 75–79. Hamp, Eric P. (1970b). Productive suffix ablaut in Baltic. Baltistica 6, 27–32. Hamp, Eric P. (1971). Varia iii. Ériu 22, 181–187. Hamp, Eric P. (1978). Sound Change and the Etymological Lexicon. In: Donka Farkas, Wesley M. Jacobsen & Karol W. Todrys (eds.), Papers from the Parasession on the Lexicon. Chicago: Linguistic Society, 184–195. Hamp, Eric P. (1979). The North European word for ‘apple’. Zeitschrift für Celtische Philologie 37, 158–166. Hamp, Eric P. (1981). Refining Indo-European Lexical Entries. Zeitschrift für vergleichende Sprachforschung 95, 81–83. Hamp, Eric P. (1990). The Pre-Indo-European language of Northern (Central) Europe. In: Thomas L. Markey & John A.C. Greppin (eds.), When Worlds Collide. Ann Arbor: Karoma, 291–305. Hamp, Eric P. (2003). On *ri in Latin and Albanian krip. Glotta 59, 230–231. Hao, Chen (2019). Bark: A Study on the Spiritual World of the Early Türks. Studia UraloAltaica 52, 15–20. Hasiuk, Michał (1970). Die Ferndissimilation des k, g in den litauischen Dialekten. In: Velta Rūķe-Draviņa (ed.), Donum Balticum [FS Stang]. Stockholm: Almquist & Wiksell, 352–360. Haspelmath, Martin & Tadmor, Uri (2009, eds.), Loanwords in the World’s Languages: A Comparative Handbook. Berlin: De Gruyter Mouton. Häusler, Alexander (2002). Nomaden, Indogermanen, Invasion. Zur Entstehung eines Mythos. Halle: Orientwissenschaftliches Zentrum. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 bibliography 333 Havlová, Eva (1966). Slav. čelověkъ. Scando-Slavica 12, 80–86. Havlová, Eva (1994). Etymologické metody Václava Machka. Slavia 63, 389–394. [Havlová ESJS — see ESJS under Abbreviated Titles.] Hehn, Victor (1870). Kulturpflanzen und Hausthiere in ihrem Übergang aus Asien nach Griechenland und Italien sowie in das übrige Europa. Berlin: Gebrüder Borntraeger. Hehn, Victor (1911) = [8ᵗʰ edition of Hehn 1870, ed. Otto Schrader]. Heidermanns, Frank (1993). Etymologisches Wörterbuch der germanischen Primäradjektive. Berlin–New York: De Gruyter. Heikkilä, Mikko (2013). Itämerensuomen ehdollisesta äänteenmuutoksesta /ai/ > /ei/. Virittäjä 117, 583–592. Heikkilä, Mikko (2014). Bidrag till Fennoskandiens språkliga förhistoria i tid och rum. Helsinki: Unigrafia. Heikkinen, Marja E. et al. (2015). Relationship between wild greylag and European domestic geese based on mitochondrial DNA. Animal Genetics 46, 485–497. Helimski, Eugen (1997a). Die matorische Sprache. Szeged: JATE. Helimski, Eugen (1997b). The southern neighbours of Finno-Ugrians: Iranians or an extinct branch of Aryans (“Andronovo Aryans”)? In: Tette Hofstra et al. (eds.), Finnisch-ugrische Sprachen in Kontakt, 117–125. Helimski, Eugen [= Хелимский, Евгений] (2001). Уральцы и их предшественники: Белые пятна на этноисторической карте Северной Евразии и уральские языки. In: Congressus Nonus Internationalis Fenno-Ugristarum iv. Dissertationes sectionum: Linguistica i. Tartu: Paar, 332–336. Hendriks, Pepijn (2014). Innovation in Tradition: Tönnies Fonne’s Russian-German Phrasebook (Pskov, 1607). Amsterdam–New York: Rodopi. Henning, Walter B. (1946). The Sogdian Texts of Paris. Bulletin of the School of Oriental and African Studies 11, 713–740. Henning, Walter B. (1963a). The Kurdish Elm. Asia Major 10, 68–72. Henning, Walter B. (1963b). Coriander. Asia Major 10, 195–199. Henning, Walter B. (1965). A Grain of Mustard. 29–47. AIΩN-Linguistica 6, 29–47. Herne, Gunnar (1954). Die slavischen Farbenbennenungen: Eine semasiologisch-etymologische Untersuchung. Uppsala: Almqvist & Wiksell. Hertzen, Erik von (1973). Itämerensuomen lainakerrostumien ikäämisestä. SuomalaisUgrilaisen Seuran Aikakauskirja 72, 77–105. Hill, Eugen (2016). Phonological evidence for a Proto-Baltic stage in the evolution of East and West Baltic. International Journal of Diachronic Linguistics and Linguistic Reconstruction 13, 205–232. Hillman, Gordon (1978). On the Origins of Domestic rye. Anatolian Studies 28, 157– 174. Hintze, Almut (1994). Der Zamyād-Yašt: Edition, Übersetzung, Kommentar. Wiesbaden: Reichert. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 334 bibliography Hirt, Hermann (1892). Vom schleifenden und gestossenen Ton in den indogermanischen Sprachen i. Indogermanische Forschungen 1, 1–42. Hirt, Hermann (1896). Akzentstudien i: Germ. got. þūsundi. Indogermanische Forschungen 6, 344–349. Hirt, Hermann (1898). Grammatisches und etymologisches. Beiträge zur Geschichte der deutschen Sprache und Literatur 23, 288–357. Hirt, Hermann (1905). Die Indogermanen: Ihre Verbreitung, ihre Urheimat und ihre Kultur i. Strassburg: Trübner. Hirt, Hermann (1907). Die Indogermanen: Ihre Verbreitung, ihre Urheimat und ihre Kultur ii. Strassburg: Trübner. Hirt, Hermann (1909). Etymologie der neuhochdeutschen Sprache. München: Beck. Hirt, Hermann (1927). Indogermanische Grammatik i. Heidelberg: Winter. Hogg, Richard M. (1992). A Grammar of Old English i. Phonology. Oxford: Blackwell. Holčík, Juraj et al. (1989). Acipenser sturio Linnaeus, 1758. In: Juraj Holčík (ed.), The Freshwater Fishes of Europe. Vol. i/2. Wiesbaden: AULA, 367–393. Holopainen, Sampsa (2019). Indo-Iranian borrowings in Uralic. Diss. University of Helsinki. Holopainen, Sampsa (2021). On the question of substitution of palatovelars in IndoEuropean loanwords into Uralic. Suomalais-Ugrilaisen Seuran Aikakauskirja 98, 197– 233. Holopainen, Sampsa (2022). Uralilaisen lingvistisen paleontologian ongelmia — mitä sanasto voi kertoa kulttuurista? In: Kaisla Kaheinen et al. (eds.). Hämeenmaalta Jamalille [FS Salminen]. Helsinki: Helda Open Books, 101–112. Holopainen, Sampsa & Junttila, Santeri (2022). Die alten arischen und baltischen Lehnverben der uralischen Sprachen. Dettelbach: Röll. Holopainen, Sampsa, Juha Kuokkala & Santeri Junttila (2017). Indoeurópai jövevényszavak és a második szótagi labiális magánhangzók fejlődése az uráli nyelvekben. A nyelvtörténeti kutatások újabb eredményei 9, 109–136. Holthausen, Ferdinand (1899). Engl. culver — russ. gólubĭ ‘Taube’. Indogermanische Forschungen 10, 112. Holthausen, Ferdinand (1934). Altenglisches etymologisches Wörterbuch. Heidelberg: Winter. Holton, David et al. (2020). The Cambridge Grammar of Medieval and Early Modern Greek. 4 vols. Cambridge: University Press. Holub, Josef & Kopečný, František (1952). Etymologický slovník jazyka českého. Praha: Státní nakladatelství učebnic. Holzer, Georg (1989). Entlehnungen aus einer bisher unbekannten indogermanischen Sprache im Urslavischen und Urbaltischen. Wien: Akademie der Wissenschaften. Holzer, Georg (2001). Zur Lautgeschichte des baltisch-slavischen Areals. Wiener Slavistisches Jahrbuch 47, 33–50. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 bibliography 335 Honti, László (1982). Geschichte des obugrischen Vokalismus der ersten Silbe. Budapest: Akadémiai Kiadó. Hoops, Johannes (1905). Waldbäume und Kulturpflanzen im germanischen Altertum. Straßburg: Trübner. Hoops, Johannes (1915–1916, ed.). Reallexikon der Germanischen Altertumskunde iii. Straßburg: Trübner. Huld, Martin E. (1981). Albanian corrigenda to Friedrich’s Proto-Indo-European trees. Zeitschrift für vergleichende Sprachforschung 95, 302–308. Huld, Martin E. (1984). Basic Albanian Etymologies. Columbus: Slavica. Huld, Martin E. (1990). The linguistic typology of the Old European substrata in North Central Europe. In: Karlene Jones-Bley & Martin E. Huld (eds.), The Indo-Europeanization of Northern Europe. Washington, DC: Institute for the Study of Man, 389– 423. Hupel, August W. (1818). Ehstnische Sprachlehre für die beyden Hauptdialekte, den revalschen und dörptschen, nebst einem vollständigen ehstnischen Wörterbuche. 2 vols. Mitau: Steffenhagen und Sohn. Hyllested, Adam (2009). PIE *-bʰ- in Nouns and Verbs: Distribution, Function, Origin. In: Rosemarie Lühr & Sabine Ziegler (eds.), Protolanguage and Prehistory. Wiesbaden: Reichert, 202–214. Ильинскiй, Григорій А. (1910). Славянскія этимологіи xxi–xxv. Русскiй филологическiй вѣстникъ 63, 322–341. Ильинскiй, Григорій А. (1916). Звукъ ch въ славянскихъ языкахъ ii. Извѣстія Отдѣленія русскаго языка и словесности 20/4, 135–186. Ильинский, Григорий А. (1926). К этимологии балтийско-славянских названий „зубра“. Tauta ir žodis 4, 51–56. Иллич-Свитыч, Владислав М. (1960). К этимологии слов морковь и тыква. Этимологические исследования по русскому языку 1, 16–26. Иллич-Свитыч, Владислав М. (1963). Именная акцентуация в балтийском и славянском. Москва: Академия наук СССР. Ильминский, Николай И. (1860–1861). Матеріалы къ изученію киргизскаго нарѣчія. Ученыя записки, изданныя императорскимъ казанскимъ университетомъ 1860/3–4, 1861/1. Imberciadori, Giulio (2022). The Etymology of Toch. B tapre, A tpär ‘high’ [Conference handout]. International Etymological Symposium (Greifswald, 16–18 May 2022). Ipsen, Gunther (1924). Der Alte Orient und die Indogermanen. In: Johannes Friedrich et al. (eds.), Stand und Aufgaben der Sprachwissenschaft [FS Streitberg]. Heidelberg: Winter, 200–237. Isépy, Peter & Primavesi, Oliver (2014). Helladios und Hesychios. Neues zum Text der Bibliotheke des Photios (Cod. 279). Zeitschrift für Papyrologie und Epigraphik 192, 121– 142. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 336 bibliography Ишкильдина, Линара К. (2018). Этимология слов с анлаутным «ш» в башкирском языке. Томский журнал ЛИНГ и АНТР 22, 31–38. Itkonen, Erkki (1946). Zur Frage nach der Entwicklung des Vokalismus der ersten Silbe in den finnisch-ugrischen Sprachen, insbesondere im Mordwinischen. FinnischUgrische Forschungen 29, 232–337. Itkonen, Erkki (1948). Vokaalikombinaatiot ja vartalotyypit. Virittäjä 52, 124–144. Itkonen, Erkki (1953). Zur Geschichte des Vokalismus der ersten Silbe im Tscheremissischen und in den permischen Sprachen. Finnisch-Ugrische Forschungen 31/3, 149– 345. Itkonen, Erkki (1977). Die Umwandlung einiger a- und ä-Stämme zu e-Stämmen im Urfinnischen. Suomalais-Ugrilaisen Seuran Aikakauskirja 75, 5–12. Itkonen, Terho (1982). Laaja, lavea, lakea ja laakea: lisiä suomen kj-sanojen vaiheisiin. Virittäjä 86, 121–139. Itkonen, Terho (1987). Erään vokaalivyyhden selvittelyä. Pitkän vokaalin lyhenemistä vai lyhyen pitenemistä h-loppuisessa ensi tavussa? Virittäjä 91, 164–208. Itkonen, Terho (1997). Reflections on Pre-Uralic and the “Saami-Finnic protolanguage”. Finnisch-Ugrische Forschungen 54, 229–266. Itkonen, Toivo I. (1958). Koltan- ja kuolanlapin sanakirja i. Helsinki: SuomalaisUgrilainen Seura. Иванов, Вячеслав В. (2002). Ранние коптские заимствования в славянском. In: Татьяна М. Николаева (ed.), Славянская языковая и этноязыковая системы в контакте с неславянским окружением. Москва: Языки славянской культуры, 55–60. Иванов, Вячеслав В. & Топоров, Владимир Н. (1958). К постановке вопроса о древнейших отношениях балтийских и славянских языков (iv Международный съезд славистов: Доклады). Москва: Академия наук СССР. Iversen, Rune & Kroonen, Guus (2017). Talking Neolithic. Linguistic and Archaeological Perspectives on How Indo-European Was Implemented in Southern Scandinavia. American Journal of Archaeology 121, 511–525. Jagersma, Abraham H. (2010). A descriptive grammar of Sumerian. Diss. Universiteit Leiden. Jakob, Anthony (forthc. a.). Three pre-Balto-Slavic bird names, or: A more austere take on Oštir. To appear in: Guus Kroonen (ed.), Sub-Indo-European Europe. De Gruyter Mouton. Jakob, Anthony (forthc. b.). On some potential ghost words in Baltic. To appear in Baltistica 58(1). Jakob, Anthony (forthc. c.). The palaeo-Baltic substrate: a methodological exploration. To appear in the proceedings of Hansakansan kanssa. Internationales etymologisches Symposium, Greifswald. Jakob, Anthony (forthc. d.). Symmetry, sound change and the Baltic loanwords in Finnic. To appear in: Sampsa Holopainen & Juho Pystynen (eds.), Areal effects Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 bibliography 337 in prehistoric contacts between Uralic and Indo-European [working title]. Helsinki: Suomalais-Ugrilaisen Seura. Jamison Commentary = Jamison, Stephanie. Rigveda Translation: Commentary. Accessed online at rigvedacommentary.alc.ucla.edu. Janhunen, Juha (1977). Samojedischer Wortschatz: Gemeinsamojedische Etymologien. Helsinki. Janko, Josef (1912). O pravěku slovanském. Praha: Otto. Jasanoff, Jay. H. (2002). The nom. sg. of Germanic n-stems. In: Alfred Wedel & Hans-Jörg Busch (eds.), Verba et Litterae: Explorations in Germanic Languages and German Literature. Newark: Linguatext, 31–46. Jēgers, Benjamiņš (1969). Einige baltische und slavische Verwandte der Sippe von lit. dir̃ti. Studi Baltici 10, 63–112. Jóhannesson, Alexander (1942). Isländische Beiträge zum indogermanischen Wörterbuch. Zeitschrift für vergleichende Sprachforschung 67, 220–223. Joki, Aulis J. (1973). Uralier und Indogermanen. Helsinki: Suomalais-Ugrilainen Seura. Jokl, Norbert (1923). Linguistisch-kulturhistorische Untersuchungen aus dem Bereiche des Albanischen. Berlin–Leipzig: De Gruyter. Jokl, Norbert (1924). Albaner. In: Max Ebert (ed.), Reallexikon der Vorgeschichte i. Berlin: De Gruyter, 84–94. Junttila, Santeri (2011). A collection of forgotten etymologies: revisiting the most improbable Baltic loanwords in Finnic. In: Cornelius Hasselblatt et al. (eds.), Language Contact in Times of Globalization. Amsterdam–New York: Rodopi, 77–132. Junttila, Santeri (2012). The prehistoric context of the oldest contacts between Baltic and Finnic languages. In: Riho Grünthal & Petri Kallio (eds.), A Linguistic Map of Prehistoric Northern Europe. Helsinki: Suomalais-Ugrilainen Seura, 261–296. Junttila, Santeri (2014). Die alten baltischen Lehnwörter im neuen estnischen etymologischen Wörterbuch. Baltu filoloģija 23, 126–135. Junttila, Santeri (2015a). Proto-Finnic loanwords in the Baltic languages? An old hypothesis revisited. In: Santeri Junttila (ed.), Contacts between the Baltic and Finnic languages. Helsinki: Suomalais-Ugrilainen Seura, 12–38. Junttila, Santeri (2015b). Revisiting the ‘separate Proto-Balto-Slavic borrowings’ of Saami. In: Harry Mantila et al. (eds.), Congressus duodecimus internationalis FennoUgristarum, Oulu. Book of abstracts, 476–478. Junttila, Santeri (2016a). Tiedon kumuloituminen ja trendit lainasanatutkimuksessa. Helsinki: Yliopisto. Junttila, Santeri (2016b). Die baltisch-slawische Frage im Lichte der alten Baltischen Lehnwörter des ostseefinnischen. Baltistica 51/2, 217–238. Junttila, Santeri (2017). Lähtökielen sanansisäisten soinnittomien klusiilien edustus kantasuomen balttilaislainoissa i. Suomalais-Ugrilaisen Seuran Aikakauskirja 96, 127–148. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 338 bibliography Junttila, Santeri (2018). Altlettgallische Lehnwörter in den mordwinischen Sprachen? Finnisch-Ugrische Forschungen 64, 72–91. Junttila, Santeri (2019). Lähtökielen sanansisäisten soinnittomien klusiilien edustus kantasuomen balttilaislainoissa ii. Suomalais-Ugrilaisen Seuran Aikakauskirja 97, 35–73. Junttila, Santeri (in prep.). Kantasuomen balttilaislainat. Sanakirjan alustava versio keskustelun pohjaksi. [2ⁿᵈ draft, dated 17 April 2020]. Accessed online at: blogs​ .helsinki.fi/santerijunttila/luonnokset/ Juška = Juška, Antanas [= А. Юшкевичъ] (1898–1904). Литовскiй словарь съ толкованiемъ словъ на русскомъ и польскомъ языкахъ. 3 issues [A–K]. Санктпетербургъ: Императорская Академія Наукъ. Kalima, Jalo (1919). Die ostseefinnischen Lehnwörter im Russischen. Hesinki: Société Finno-Ougrienne. Kalima, Jalo (1924). Zur karelisch-olonetzischen lautgeschichte. Finnisch-Ugrische Forschungen 24, 163–176. Kalima, Jalo (1928). Eräiden sanojen alkuperästä. Virittäjä 32, 102–111. Kalima, Jalo (1929). Zur Herkunft der slavischen Lehnwörter im Ostseefinnischen. Zeitschrift für Slavische Philologie 6, 154–172. Kalima, Jalo (1934). Russ. стержень und slav. stěna, d. Stein. In: Germanisch-romanische Studien [FS Suolahti]. Helsinki: Suomalainen tiedeakatemia. Kalima, Jalo (1936). Itämerensuomalaisten kielten balttilaiset lainasanat. Helsinki: Suomalaisen kirjallisuuden seura. Kalima, Jalo (1939–1940). Fi. ativo. Finnisch-Ugrische Forschungen 26, 211–214. Kalima, Jalo (1950). Zur slavischen Etymologie. Zeitschrift für Slavische Philologie 20, 406–417. Kalima, Jalo (1956). Die slavischen Lehnwörter im Ostseefinnischen. Berlin–Wiesbaden: Harrassowitz. Kallio, Petri (1998). Vanhojen balttilaisten lainasanojen ajoittamisesta. In: Riho Grünthal & Johanna Laakso (eds.), Oekeeta asijoo [FS Suhonen]. Helsinki: SuomalaisUgrilainen Seura, 209–217. Kallio, Petri (2006). On the Earliest Slavic Loanwords in Finnic. In: Juhani Nuorluoto (ed.), The Slavicization of the Russian North: Mechanisms and Chronology. Helsinki: Yliopisto, 154–166. Kallio, Petri (2007). Kantasuomen konsonanttihistoriaa. In: Jussi Ylikoski & Ante Aikio (eds.), Sámit, sánit, sátnehámit [FS Sammallahti]. Helsinki: Suomalais-Ugrilainen Seura, 229–249. Kallio, Petri (2008a). On the “early Baltic” loanwords in common Finnic. In: Alexander Lubotsky et al. (eds.), Evidence and counter-evidence: Essays in honour of Frederik Kortlandt i. Balto-Slavic and Indo-European linguistics. Amsterdam–New York: Brill, 265–277. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 bibliography 339 Kallio, Petri (2008b). The Etymology of Finnish sauna ‘Sauna’. In: Kees Dekker, Alasdair MacDonald & Hermann Niebaum (eds.), Northern Voices [FS Hofstra]. Leuven– Paris–Dudley: Peeters, 314–319. Kallio, Petri (2009). Stratigraphy of Indo-European loanwords in Saami. In: Tiina Äikäs (ed.), Máttut–máddagat: the Roots of Saami ethnicities, societies and spaces / places. Oulu: Giellagas Institute, 30–45. Kallio, Petri (2012). Jälkitavujen diftongit kantasuomessa. Fenno-Ugrica Suecana Nova Series 14, 31–40. Kallio, Petri (2014). The Diversification of Proto-Finnic. In: Joonas Ahola & Frog (eds.), Fibula, Fabula, Fact: The Viking Age in Finland. Helsinki: Suomen Kirjallisuuden Seura, 155–168. Kallio, Petri (2015). The language contact situation in prehistoric Northeastern Europe. In: Robert Mailhammer, Theo Vennemann & Birgit A. Olsen (eds.), The Linguistic Roots of Europe: Origin and Development of European Languages. København: Museum Tusculanum, 77–102. Kallio, Petri (2016). Historical phonology from Proto-Finnic to Proto-Livonian. Journal of Estonian and Finno-Ugric Linguistics 7/1, 39–65. Kallio, Petri (2018). Ensitavun diftongit kantasuomessa. In: Sampsa Holopainen & Janne Saarikivi (eds.), Περὶ ὀρθότητος ἐτύμων: Uusiutuva uralilainen etymologia. Helsinki: Suomalais-Ugrilainen Seura, 251–268. Kallio, Petri (2021). The Position of Leivu. In: Uldis Balodis & Karl Pajusalu (eds.), Studies on the South Estonian language islands (Leivu, Lutsi, Kraasna). Tartu: Ölikool, 123–143. Kallio, Petri (forthc.). Substrates in Finnic. To appear in: Ante Aikio & Santeri Palviainen (eds.), Substrate Languages in Northern Europe. De Gruyter Mouton. Kapović, Mate (2006). Reconstruction of Balto-Slavic personal pronouns. Diss. University of Zagreb. Karaliūnas, Simas (1993). Some remarks on the names for pigeon in Lithuanian. Linguistica Baltica 2, 109–113. Karaliūnas, Simas (2004). Baltų praeitis istoriniuose šaltiniuose i. Vilnius: Lietuvių kalbos institutas. Karg, Sabine & Robinson, David (2000). Secondary food plants from medieval sites in Denmark: fruits, nuts, vegetables, herbs and spices. In: Karin Viklund (ed.), Nordic Archaeobotany–NAG 2000 in Umeå (= Archaeology and Environment 15). University of Umeå, 133–142. Karničar, Ludwig (1990). Der Obir-Dialekt in Kärnten: die Mundart von Ebriach/Obirsko im Vergleich mit den Nachbarmundarten von Zell/Sele und Trögern/Korte. Wien: Akademie der Wissenschaften. Karsten, Torsten E. (1915). Germanisch-finnische Lehnwortstudien. Helsinki: Finnische Literaturgesellschaft. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 340 bibliography Karulis, Konstantīns (1992). Latviešu etimoloģijas vārdnīca. 2 vols. Rīga: Avots. Касаткин, Леонид Л. (1999). Современная русская диалектная и литературная фонетика как источник для истории русского языка. Москва: Наука. Касьян, Алексей С. (2012). Отражение праславянской фонемы *y в старопсковском диалекте разговорника Т. Фенне как архаизм. Вопросы языкознания 2012/4, 73–100. Kaukienė, Audronė & Jakulytė, Dalia (2015). Stonais po leipen zaidiantẽ: kitokia xv a. baltiško teksto interpretacija. Res humanitariae 17, 40–52. Kazlauskas, Jonas (1968). Lietuvių kalbos istorinė gramatika. Vilnius: Mintis. Kendla, Mari & Viikberg, Jüri (2015). Protoeurooplaste keelepärandist. Emakeele Seltsi aastaraamat 61, 135–154. Keresztes, László (1986). Geschichte des modwinischen Konsonantismus ii. Etymologisches Belegmaterial. Szeged. Keresztes, László (1987). Geschichte des mordwinischen Konsonantismus i. Szeged. Kettunen, Lauri (1930). Vatjan kielen äännehistoria. Helsinki: Suomalaisen Kirjallisuuden Seura. Kettunen, Lauri (1938). Livisches Wörterbuch mit grammatischer Einleitung. Helsinki: Suomalais-Ugrilainen Seura. Kießling, Roland & Mous, Maarten (2003). The Lexical Reconstruction of West-Rift Southern Cushitic. Köln: Rüdiger Köppe. Kilian, Lothar (1986). Zur Frage eines westfinnischen Substrats in Litauen aus der Sicht von Archäologie und Hydronymie. Zeitschrift für Ostforschung 35, 481–502. Kim, Ronald I. (2018). The phonology of Balto-Slavic. In: Jared Klein, Brian Joseph & Matthias Fritz (eds.), Handbook of Comparative and Historical Indo-European Linguistics. Berlin: De Gruyter, 1974–1985. Kiparsky, Valentin (1934). Die gemeinslavischen Lehnwörter aus dem Germanischen. Helsinki: Suomalainen Tiedeakatemia. Kiparsky, Valentin (1936). Fremdes im Baltendeutsch. Helsinki: Société néophilologique. Kiparsky, Valentin (1948). Chronologie des relations slavobaltiques et slavofinnoises. Revue des Études Slaves 24, 29–47. Kiparsky, Valentin (1949). Tuhkuri ‘vesikko’. Virittäjä 53, 58–68. Kiparsky, Valentin (1952). The Earliest Contacts of the Russians with the Finns and Balts. Oxford Slavonic Papers 3, 67–79 Kiparsky, Valentin (1959a). Kielitieteen peruskysymyksiä. Virittäjä 63, 421–424. Kiparsky, Valentin (1959b). Über etymologische Wörterbücher. Neuphilologische Mitteilungen 60, 209–230. Kiparsky, Valentin (1963). Russische historische Grammatik i. Entwicklung des Lautsystems. Heidelberg: Winter. Kiparsky, Valentin (1968). Slavische und baltische b/p-Fälle. Scando-Slavica 14, 73–97. Kiparsky, Valentin (1972). Zur Etymologie des lit. šeškas, lett. sesks ‘Iltis’. Baltistica 8, 157–160. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 bibliography 341 Kiparsky, Valentin (1973). О балтизмах русского литературного языка. Baltistica 9, 67–70. Kiparsky, Valentin (1975). Russische historische Grammatik iii. Entwicklung des Wortschatzes. Heidelberg: Winter. Kitson, Peter R. (1997). Old English bird-names. English Studies 78, 481–505. Kittel, Piotr et al. (2020). On the border between land and water: The environmental conditions of the Neolithic occupation from 4.3 until 1.6 ka bc at Serteya, Western Russia. Geoarchaeology 36, 173–202. Kleyner, Svetlana D. (2005). From Yellow to Blue — or Not? In: Dafydd Johnston, Elena Parina & Maxim Fomin (eds.), ‘Yn llawen iawn, yn llawn iaith’. Aberystwyth: Centre for Advanced Welsh and Celtic Studies, 47–52. Klimas, Antanas (1957). Lithuanian and Indo-European. Lituanus 1957/4, 14–16. Klimas, Antanas (2002). Lithuanian in the 21st Century: Linguistic Snippets and Tidbits. Lituanus 48/3, 52–80. Klingenschmitt, Gert (1982). Das altarmenische Verbum. Wiesbaden: Reichert. Kloekhorst, Alwin (2008). Etymological Dictionary of the Hittite Inherited Lexicon. Leiden–Boston: Brill. Kloekhorst, Alwin (2014). The Proto-Indo-European Acrostatic Inflection Reconsidered. In: Norbert Oettinger & Thomas Steer (eds.), Das Nomen im Indogermanischen. Wiesbaden: Reichert, 140–163. Kluge, Friedrich (1907). Zum altpreußischen Wortschatz. Indogermanische Forschungen 21, 358–361. Kluge/Götze = Alfred Götze (1948, ed.), Friedrich Kluge. Etymologisches Wörterbuch der deutschen Sprache [14ᵗʰ edition]. Berlin: De Gruyter. Kluge/Seebold = Elmar Seebold (2011, ed.). Kluge: Etymologisches Wörterbuch der deutschen Sprache [25ᵗʰ edition]. Berlin–New York: De Gruyter. Knipper, Corina et al. (2017). Female exogamy and gene pool diversification at the transition from the Final Neolithic to the Early Bronze Age in central Europe. Proceedings of the National Academy of Sciences 114/38. Koivulehto, Jorma (1970). Suomen laiva-sanasta. Virittäjä 74, 178–183. Koivulehto, Jorma (1971). Germanisch-finnische Lehnbeziehungen i. Finn. Vordervokal für germ. Hintervokal. Neuphilologische Mitteilungen 72, 577–607. Koivulehto, Jorma (1972). Germanisch-finnische Lehnbeziehungen ii. Wirtschaftliches und Landwirtschaftliches. Neuphilologische Mitteilungen 73, 575–628. Koivulehto, Jorma (1979a). Lainoja ja lainakerrostumia. Virittäjä 83, 267–301. Koivulehto, Jorma (1979b). Baltisches und Germanisches im Finnischen: die finn. Stämme auf -rte und die finn. Sequenz VrtV. In: Erhard F. Schiefer (ed.), Explanationes et tractationes Fenno-Ugricae in honorem Hans Fromm. München: Fink, 129– 164. Koivulehto, Jorma (1981). Reflexe des germ. /ē¹/ im Finnischen und die Datierung der Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 342 bibliography germanisch-finnischen Lehnbeziehungen. Beiträge zur Geschichte der deutschen Sprache und Literatur 103, 167–203. Koivulehto, Jorma (1983). Suomalaisten maahanmuutto indoeurooppalaisten lainasanojen valossa. Suomalais-Ugrilaisen Seuran Aikakauskirja 78, 107–132. Koivulehto, Jorma (1984). Itämerensuomalais-germaaniset kosketukset. In: Sven-Erik Åström (ed.), Suomen väestön esihistorialliset juuret. Helsinki: Societas Scientiarum Fennica, 191–204. Koivulehto, Jorma (1986). Die Sieverssche Regel im Lichte der germanisch-finnischen Lehnbeziehungen. In: Bela Brogyanyi & Thomas Krömmelbein (eds.), Germanic Dialects: Linguistic and Philological Investigations. Amsterdam–Philadelphia: John Benjamins, 249–294. Koivulehto, Jorma (1990). Zu den ältesten Kontakten zwischen Ostseefinnisch und Balto-Slavisch. In: Sari Vaula (ed.), Itämerensuomalaiset kielikontaktit. Helsinki: Valtion painatuskeskus, 148–153. Koivulehto, Jorma (1992a). Direkte Kontakte des Lappischen mit dem Baltischen. In: Pál Deréky et al. (eds.), Festschrift für Károly Rédei zum 60. Geburtstag. Wien–Budapest: MTA Nyelvtudományi Intézet, 299–304. Koivulehto, Jorma (1992b). Der Typus palje ‘(Blase)balg’, turve ‘Torf’ unter den Lehnwörtern des Ostseefinnischen. Suomalais-Ugrilaisen Seuran Aikakauskirja 84, 163– 190. Koivulehto, Jorma (1993a). Zur Etymologie von fi. susi, tosi — und kesi: eine Entgegnung. Linguistica Uralica 29/1, 21–36. Koivulehto, Jorma (1993b). Suomi. Virittäjä 97, 400–406. Koivulehto, Jorma (1998). Kiire ‘päälaki’ ja muuta etymologista rajankäyntiä. In: Riho Grünthal & Johanna Laakso (eds.), Oekeeta asijoo [FS Suhonen]. Helsinki: Suomalais-Ugrilainen Seura, 235–246. Koivulehto, Jorma (1999). Verba mutuata. Helsinki: Suomalais-Ugrilainen Seura. Koivulehto, Jorma (2000). Reflexe des urbaltischen *ā in baltischen Lehnwörtern des Ostseefinnischen. Linguistica Baltica 8, 103–124. Koivulehto, Jorma (2001). The earliest contacts between Indo-European and Uralic speakers in the light of lexical loans. In: Christian Carpelan, Asko Parpola & Petteri Koskikallio (eds.), Early Contacts between Uralic and Indo-European: Linguistic and Archaeological Considerations. Helsinki: Suomalais-Ugrilainen Seura, 235–263. Koivulehto, Jorma (2002). Contact with non-Germanic languages ii: Relations to the East. In: Oskar Bandle et al. (eds.), The Nordic Languages i. Berlin: De Gruyter Mouton, 583–594. Колесов, Владимир В. (1980). Историческая фонетика русского языка. Москва: Высшая школа. Kortlandt, Frederik (1977). Historical laws of Baltic accentuation. Baltistica 13, 319– 330. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 bibliography 343 Kortlandt, Frederik (1978). I.-E. palatovelars before resonants in Balto-Slavic. In: Jacek Fisiak (ed.), Recent developments in historical phonology. Den Haag: Mouton, 237– 243. Kortlandt, Frederik (1979a). Three problems of Balto-Slavic phonology. Zbornik za Filologiju i Lingvistiku 22/2, 57–63. Kortlandt, Frederik (1979b). On the history of the Slavic nasal vowels. Indogermanische Forschungen 84, 259–272. Kortlandt, Frederik (1980). [Review of] V.M. Illich-Svitych, Nominal accentuation in Baltic and Slavic. Lingua 51, 346–354. Kortlandt, Frederik (1982). IE *pt in Slavic. Folia Linguistica Historica 3, 25–28. Kortlandt, Frederik (1983). On final syllables in Slavic. Journal of Indo-European Studies 11, 167–185. Kortlandt, Frederik (1984–1985). On reduced vowels in Slavic. Zbornik Matice srpske za filologiju i lingvistiku 27–28, 367–368. Kortlandt, Frederik (1985). Long vowels in Balto-Slavic. Baltistica 21, 112–124. Kortlandt, Frederik (1988). Remarks on Winter’s law. In: Adriaan A. Barentsen, Ben M. Groen & Rob Sprenger (eds.), Dutch Contributions to the Tenth International Congress of Slavists: Linguistics. Amsterdam: Rodopi, 387–396. Kortlandt, Frederik (1989a). Od praindoevropskog jezika do slovenskog (fonološki razvoj). Zbornik za filologiju i lingvistiku 32/2, 41–58. Kortlandt, Frederik (1989b). Lithuanian statýti and related formations. Baltistica 25/2, 104–112. Kortlandt, Frederik (1993). Tokie šalti rytai. Baltistica 28, 45–48. Kortlandt, Frederik (1995). Lithuanian verbs in -auti and -uoti. Linguistica Baltica 4, 141– 143. Kortlandt, Frederik (1998). The Language of the Old Prussian Catechisms. Res Balticae, 117–129. Kortlandt, Frederik (2000). Initial a- and e- in Old Prussian. Linguistica Baltica 8, 125– 127. Kortlandt, Frederik (2007). The development of the Indo-European syllabic resonants in Balto-Slavic. Baltistica 42, 7–12. Kortlandt, Frederik (2018). Proto-Baltic? Baltistica 53, 175–185. Kossmann, Maarten (1999). Essai sur la phonologie du proto-berbère. Köln: Köppe. Kossmann, Maarten (2020). Proto-Berber phonological reconstruction: An update. Linguistique et Langues Africaines 6, 11–42. Kott = Kott, František Št. (1878–1893). Česko-německý slovník zvláště gramaticko-frazeologický. 7 vols. Praha: Kolář/Simáček. Krahe, Hans (1962). Die Struktur der alteuropäischen Hydronomie. Abhandlungen der geistes- und sozialwissenschaftlichen Klasse 1962/5. Krajčovič, Rudolf (1975). A Historical Phonology of the Slovak Language. Heidelberg: Winter. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 344 bibliography Kregždys, Rolandas (2004). Liet. barzdà ‘barba’: samplaikos -zd- kilmės klausimu. Acta Linguistica Lithuanica 51, 15–27. Kregždys, Rolandas (2012). Baltų mitologemų etimologijos žodynas i. Kristburgo sutartis. Vilnius: Lietuvos kultūros tyrimų institutas. Кренке, Николай А. (2019). Древности бассейна москвы-реки от неолита до средневековья. Москва–Смоленск: Свиток. Kretschmer, Paul (1896). Einleitung in die Geschichte der griechischen Sprache. Göttingen: Vandenhoeck & Ruprecht. Kretschmer, Paul (1921). Der Götterbeiname Grabovius auf den Tafeln von Iguvium. In: Festschrift Adalbert Bezzenberger. Göttingen: Vandenhoeck & Ruprecht, 89–96. Kretschmer, Paul (1940). Die vorgriechischen Sprach- und Volksschichten. Glotta 28, 231–278. Kroll, Helmut (1990). Melde von Feudvar, Vojvodina: Ein Massenfund bestätigt Chenopodium als Nutzpflanze in der Vorgeschichte. Prähistorische Zeitschrift 65, 46– 48. Kroonen, Guus J. (2011a). False exceptions to Winter’s law. In: Rick Derksen & Tijmen Pronk (eds.), Accent Matters: Papers on Balto-Slavic Accentology. Amsterdam–New York: Rodopi–Brill, 251–261. Kroonen, Guus J. (2011b). The Proto-Germanic n-stems: a study in diachronic morphology. Amsterdam: Rodopi. Kroonen, Guus J. (2012). Non-Indo-European root nouns in Germanic: evidence in support of the Agricultural Substrate Hypothesis. In: Riho Grünthal & Petri Kallio (eds.), A Linguistic Map of Prehistoric Northern Europe. Helsinki: Suomalais-Ugrilainen Seura, 239–260. Kroonen, Guus J. (2013). Etymological Dictionary of Proto-Germanic. Leiden–Boston: Brill. Kroonen, Guus J. (2016). On the Origin of Greek μῆλον, Latin mālum, Albanian mollë and Hittite šam(a)lu- ‘apple’. Journal of Indo-European Studies 44, 85–91. Kroonen, Guus J. (2017). The development of the Proto-Indo-European instrumental suffix in Germanic. Indogermanische Forschungen 122, 105–110. Kroonen, Guus J. (forthc.). For the nth time: The Pre-Greek νϑ-suffix revisited. To appear in: Guus Kroonen (ed.), Sub-Indo-European Europe. De Gruyter Mouton. Kroonen, Guus J. & Lubotsky, Alexander (2009). Proto-Indo-European *tsel- ‘to sneak’ and Germanic *stelan- ‘to steal, approach stealthily’. Studia Etymologica Cracoviensia 14, 237–241. Kroonen, Guus J., Anthony Jakob et al. (2022). Indo-European cereal terminology suggests a Northwest Pontic homeland for the core Indo-European languages. PLoS ONE 17/10. Крысько, Вадим Б. (1994). Древний новгородско-псковский диалект на общеславянском фоне. Вопросы языкознания 1994/6, 16–93. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 bibliography 345 Крысько, Вадим Б. (2014). Маргиналии к «Этимологическому словарю славянских языков» (вып. 34–38). Вопросы языкознания. 2014/1, 100–119. Kuhn, Adalbert (1862). [Review of] Adolphe Pictet. Les origines Indo-Europennes ou les Aryas primitifs i. Beiträge zur vergleichenden Sprachforschung 2, 369–382. Kuiper, Frans B.J. (1956). The Etymology of ἄνθρωπος. In: Heinz Kronasser (ed.), ΜΝΗΜΗΣ ΧΑΡΙΝ: Gedenkschrift Paul Kretschmer i. Wien: Sprachgesellschaft, 211– 226. Kuiper, Frans B.J. (1968). Pre-Hellenic Labio-Velars? Lingua 21, 269–277. Kuiper, Frans B.J. (1995). Gothic bagms and Old Icelandic ylgr. North-Western European Language Evolution 25, 63–88. Кулешов, Вячеслав С. (2010). Вокализм ранних славянских заимствований в прибалтийско-финских языках и реконструкция вокалической системы северно-словенского диалекта. Finnisch-Ugrische Mitteilungen 32/33, 345–354. Куликовскiй, Германъ И. (1898). Словарь областного олонецкаго нарѣчiя. СанктъПетербургъ: Императорская академія наукъ. Kulonen, Ulla-Maija (1988). Ovatko reuna ja leuka sittenkin balttilaisia lainasanoja? Virittäjä 92, 84–87. Kümmel, Martin (2000). Das Perfekt im Indoiranischen. Wiesbaden: Reichert. Kümmel, Martin (2017a). Even more traces of the accent in Armenian? In: Bjarne S.S. Hansen, Adam Hyllested, Anders R. Jørgensen & Guus Kroonen (eds.), Usque ad radices [FS Olsen]. København: Museum Tusculanum. Kümmel, Martin (2017b). Agricultural terms in Indo-Iranian. In: Martine Robbeets & Alexander Savelyev (eds.), Language Dispersal Beyond Farming. Amsterdam– Philadelphia: Benjamins. Kuokkala, Juha (2012). Jälkitavun labiaalivokaalit saamessa ja itämerensuomessa. MA Thesis Helsingin yliopisto. Kurila, Laurynas (2021). Lithuania in the Migration Period — Quiet Refuge or Battlefield? Lithuanian Heritage January/February 2021, 17–21. Kurschat, Friedrich (1883). Wörterbuch der littauischen Sprache ii. Littauisch-deutsches Wörterbuch. Halle: Buchandlung des Waisenhauses. Kuryłowicz, Jerzy (1956a). L’apophonie en indo-européen. Wrocław: Ossolineum. Kuryłowicz, Jerzy (1956b). Die Doppelvertretung von idg. ei, oi im Lettolitauischen. In: Heinz Kronasser (ed.), ΜΝΗΜΗΣ ΧΑΡΙΝ: Gedenkschrift Paul Kretschmer i. Wien: Sprachgesellschaft, 227–236. Кузьменко, Юрий К. (2013). К вопросу о неиндоевропейских субстратах в германском. Индоевропейское языкознание и классическая филология 17, 511–535. Lacey, Mohamed Eric R. (2013). Birds and Bird-lore in the Literature of Anglo-Saxon England. Diss. University College London. Lamprecht, Arnošt, Dušan Šlosar & Jaroslav Bauer (1977). Historický vývoj češtiny. Praha: Státní pedagogické nakladatelství. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 346 bibliography Lang, Valter (2007). The Bronze and Early Iron Ages in Estonia. Tartu: Ölikool. Lang, Valter (2015). Formation of Proto-Finnic — an archaeological scenario from the Bronze Age / Early Iron Age. In: Harri Mantila et al. (eds.), Congressus Duodecimus Internationalis Fenno-Ugristarum. Plenary Papers. Oulu: Yliopisto, 63–84. Lang, Valter (2016). Early Finnic-Baltic contacts as evidenced by archaeological and linguistic data. Journal of Estonian and Finno-Ugric Linguistics 7, 11–38. Lang, Valter (2018a). Fortified Settlements in the Eastern Baltic: From Earlier Research to New Interpretations. Achaeologia Lituana 19, 13–33. Lang, Valter (2018b). The Eastern Baltic. In: Colin Haselgrove et al. (eds.), The Oxford Handbook of the European Iron Age. Oxford: University Press, 173–192. Lange, Jacob (1772–1773). Vollständiges deutschlettisches und lettischdeutsches Lexicon. 2 vols. Mitau: Steffenhagen. Lanszweert, René (1984). Die Rekonstruktion des baltischen Grundwortschatzes. Frankfurt am Main–Bern–New York: Peter Lang. Ларин, Борис А. (1959). Русско-английский словарь-дневник Ричарда Джемса (1618– 1619 гг.). Ленинград: Университет Larsson, Jenny H. (2018). The lexicon of Baltic. In: Jared Klein, Brian Joseph & Matthias Fritz (eds.), Handbook of Comparative and Historical Indo-European Linguistics. Berlin: De Gruyter, 1681–1698. Лаучюте, Юрате А. (1982). Словарь балтизмов в славянских языках. Ленинград: Наука. Laumane, Benita (1973). Zivju nosaukumi latviešu valodā. Rīga: Zinātne. Лаумане, Бенита [= Laumane, Benita] & Непокупный, Анатолий П. (1968). К постановке вопроса об атласе морской и рыболовецкой лексики балтийского моря. Baltistica 4, 69–78. Legzdiņa, Dardega et al. (2020). Re-evaluating the Bronze and Earliest Iron Age in Latvia: changes in burial traditions in the light of ¹⁴C dates. Radiocarbon 62, 1845– 1868. Lehmann, Winfred P. (1951). The Distribution of Proto-Indo-European /r/. Language 27, 13–17. Lehmann, Winfred P. (1986). A Gothic Etymological Dictionary. Leiden: Brill. Lehtinen, Meri (1967). On the Origin of the Balto-Finnic Long Vowels. Ural-altaische Jahrbücher 39, 147–152. Lehtiranta, Juhani (2001). Yhteissaamelainen sanasto. Helsinki: Suomalais-Ugrilainen Seura. Lejeune, Michel (1974). Manuel de la langue vénète. Heidelberg: Winter. Лемцюгова, Валянціна П. (1970). Беларуская айканімія: лінгвістычны аналіз назваў населеных пунктаў Мінскай вобласці. Мінск: Навука і тэхніка. Lemeškin, Ilja (2014). Petrus Wickerau vs. Petrus Turnau. Kretos pėdsako autorystės klausimu. Baltistica 49, 139–161. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 bibliography 347 Leskien, August (1891). Die Bildung der Nomina im Litauischen. Abhandlungen der philologisch-historischen Classe der königlich sächsischen Gesellschaft der Wissenschaften 12, 151–618. Leskien, August (1922). Handbuch der altbulgarischen Sprache [6ᵗʰ edition]. Heidelberg: Winter. Leumann, Manu (1942). Idg. sk̑ im Altindischen und im Litauischen (Schluß.). Indogermanische Forschungen 58, 113–130. Leumann, Manu (1977). Lateinische Grammatik i. Lateinische Laut- und Formenlehre. München: Beck. Левичкин, Александр Н. & Мызников, Сергей А. (2014). Словарь пинежских говоров: Проект. Пробные словарные статьи. Санкт-Петерург: Нестор-История. Levin, Jules F. (1973). -jā Stems and -ē Stems in the Elbing Vocabulary. In: Arvids Ziedonis, Jr. (ed.), Baltic Literature and Linguistics. Toronto: University, 189–196. Levin, Jules F. (1974). The Slavic Element in the Old Prussian Elbing Vocabulary. Berkeley: University of California Press. Levin, Jules F. (1992). Pigeons, cows and April in Lithuania. Linguistica Baltica 1, 85–91. Levin, Jules F. (2003). The North Slavic-Lithuanian contact area: mutual influence and resistance. In: Robert A. Maguire & Alan Timberlake (eds.), American Contributions to the 13th International Congress of Slavists i. Linguistics. Bloomington: Slavica, 139– 148. Lewis/Short = Lewis, Charlton T. & Short, Charles (1879, eds.), A Latin Dictionary. Oxford: Clarendon Press. Librado, Pablo et al. (2021). The origins and spread of domestic horses from the Western Eurasian steppes. Nature 598, 634–640. Lidén, Evald (1897). Studien zur altindischen und vergleichenden Sprachgeschichte. Upsala: Almqvist & Wiksell. Lidén, Evald (1906). Ein gotisches Lehnwort im Altpreussischen. Beiträge zur Geschichte der deutschen Sprache und Literatur 31, 600–602. Liewehr, Ferdinand (1954). Zur Ausdrucksverstärkung im Slawischen. Zeitschrift für slavische Philologie 23, 89–115. Liewehr, Ferdinand (1956). Über expressive Sprachmittel im Slawischen. Zeitschrift für slavische Philologie 25, 11–27. Лигорио, Орсат (2012). С.-х. рода, лат. ardea. Зборник Матице српске за филологију и лингвистику 55, 21–28. van Linde, Paul (2001). Borrowed or inherited? Congressus Nonus Internationalis FennoUgristarum v. Dissertationes sectionum: Linguistica ii. Tartu, 288–293. van Linde, Paul (2007). The Finnic vocabulary against the background of interference. Diss. Universiteit Groningen. Liukkonen, Kari (1973). Lisiä balttilais-suomalaisten lainasanasuhteiden tutkimukseen. Virittäja 77, 17–32. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 348 bibliography Liukkonen, Kari [= Лиукконен, Кари] (1987). Восточнославянские отглагольные существительные на -m-. Helsinki. Liukkonen, Kari (1999). Baltisches im Finnischen. Helsinki: Finnisch-Ugrische Gesellschaft. Lockwood, William B. (1990). Latin columba, palumbēs, Greek κόλυμβος. Historische Sprachforschung 103, 261–263. Loewenthal, Wilhelm (1901). Die slavischen Farbenbezeichnungen. Leipzig: Preis. Loma, Aleksandar (2012). Vom Indogermanischen zum Slavischen und zurück. Probleme und Perspektiven einer vertiefenden Wortforschung im Bereich der slavischen Sprachen. In: Bohumil Vykypěl & Vít Boček (eds.), Methods of Etymological Practice. Praha: Lidové noviny. Lõo, Jaan (1911). Meie koduloomade nimetused. Eesti kirjandus 6, 49–61, 81–89. Łoś, Jan (1922). Gramatyka polska i. Głosownia historyczna. Lwów–Warszawa–Kraków: Ossolineum. Lõugas, Lembi, Aivar Kriiska & Lina Maldre (2007). New dates for the Late Neolithic Corded Ware Culture burials and early animal husbandry in the East Baltic region. Archaeofauna 16, 21–31. Lubotsky, Alexander (1981). On the reduced grade -a- in Sanskrit. Lingua 55, 75–95. Lubotsky, Alexander (1988). The system of nominal accentuation in Sanskrit and ProtoIndo-European. Leiden–New York–København–Köln: Brill. Lubotsky, Alexander (1990). La loi de Brugmann et *H₃e-. In: Jean Kellens (ed.), La reconstruction des laryngeals. Liège–Paris: Université de Liège, 129–136. Lubotsky, Alexander (1994). Avestan ϑβōrəštar- and the Indo-European root √turḱ-. Die Sprache 36, 94–102. Lubotsky, Alexander (1995). Sanskrit h < *dh, bh. In: Ярослав В. Васильков & Никита В. Гуров (eds.), Стхапакашраддха. Сборник статей памяти Г. А. Зографа. СанктПетербург: Петербургское востоковедение, 124–144. Lubotsky, Alexander (1997). The Indo-Iranian reflexes of PIE *CRHUV. In: Alexander Lubotsky (ed.), Sound Law and Analogy [FS Beekes]. Amsterdam–Atlanta: Rodopi. Lubotsky, Alexander (2001). The Indo-Iranian substratum. In: Christian Carpelan, Asko Parpola & Petteri Koskikallio (eds.), Early Contacts between Uralic and Indo-European: Linguistic and Archaeological Considerations. Helsinki: Suomalais-Ugrilainen Seura, 301–317. Lubotsky, Alexander (2004). Avestan siiazd-, Sanskrit sedh-, Latin cēdere. In: Adam Hyllested et al. (eds.), Per aspera ad asteriscos [FS Rasmussen]. Innsbruck: Innsbrucker Beiträge zur Sprachwissenschaft, 322–332. Lubotsky, Alexander (2011). The origin of Sanskrit roots of the type sīv- ‘to sew’, dīv- ‘to play dice’, with an Appendix on Vedic i-Perfects. In: Stephanie W. Jamison, H. Craig Melchert & Brent Vine (eds.), Proceedings of the 22nd UCLA Annual Indo-European Conference. Bremen: Hempen, 105–126. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 bibliography 349 Lubotsky, Alexander (2013). False Labels in Indo-European Reconstruction: Laryngeal Loss in Compounds and Marginal Phonemes [Conference handout]. ComparativeHistorical Linguistics of the 21st century: Issues and Perspectives (Moscow, March 20– 22, 2013). Lühr, Rosemarie (1993). Zur Semantifizierung von Zahlwörtern: das Wort ‘Tausend’ — eine germanisch-baltoslavische Isoglosse? Linguistica 33, 117–136. Luik, Heidi & Maldre, Lina (2007). Bronze Age bone artefacts from Narkūnai, Nevieriškė and Kereliai fortified settlements. Raw materials and manufacturing technology. Archaeologia Lituana 8, 5–39. Lunt, Horace G. (2001). Old Church Slavonic Grammar [7ᵗʰ edition]. Berlin–New York: Mouton de Gruyter. Лыткин, Василий И. (1952). Древнепермский язык. Москва: Академия наук СССР. Лыткин, Василий И. (1964). Исторический вокализм пермских языков. Москва: Наука. Лыткин, Василий И. & Гуляев, Евгений С. (1970). Краткий этимологический словарь коми языка. Москва: Наука. Mac an Bhaird, Alan (1980). Tadhg mac Cein and the Badgers. Ériu 31, 150–155. Machek, Václav (1924). Slov. klъkъ. Listy Filologické 51, 125–131. Machek, Václav (1930). Slovanské jméno piva olъ. Slavia 8, 209–217. Machek, Václav (1934). Recherches dans le domaine du lexique Balto-Slave. Brno: Filosofická fakulta. Machek, Václav (1944–1946). Drobné výklady o jménech rostlin. Naše řeč 28, 106–109, 175–180; 29, 107–114; 30, 61–68. Machek, Václav (1947). Einige slavische Fischnamen. Zeitschrift für Slavische Philologie 19, 53–67. Machek, Václav (1950a). Einige slavische Vogelnamen. Zeitschrift für Slavische Philologie 20, 29–51. Machek, Václav (1950b). Quelques noms slaves de plantes. Lingua Posnaniensis 2, 145– 161. Machek, Václav (1951a). Quelques mots slavo-germaniques. Slavia 20, 200–218. Machek, Václav (1951b). Trois noms slaves de couleurs. Lingua Posnaniensis 3, 96–111. Machek, Václav (1954). Česká a slovenská jména rostlin. Praha: Akademie věd. Machek, Václav (1961). Zum Wortschatz des Litauischen: Teil 2. Zeitschrift für Slavische Philologie 29, 345–356. Machek, Václav (1968). Etymologický slovník jazyka českého. Praha: Academia. Mägiste, Julius (1923). Etümologiseerimiskatseid. Eesti Keel 2, 33–39. Mägiste, Julius (1939). Läänemere-sm. ja volga-sm. keelte võrdluste alalt. Eesti Keel 18, 60–81. Mägiste, Julius (1959). Gibt es im Tscheremissischen baltische Lehnwörter? Uralaltaische Jahrbücher 31, 169–176. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 350 bibliography Mägiste, Julius (1962). Sm. lunka, lunki ‘tuohi’ ja mord. ľeńǵe ‘niini’. Virittäjä 66, 244– 247. Mägiste, Julius (1970). Estn. (h)õis ‘Blüte’ und seine Sippe — ein balt. Lehnwort? In: Velta Rūķe-Draviņa (ed.), Donum Balticum [FS Stang]. Stockholm: Almquist & Wiksell, 352–360. Mailhammer, Robert (2007). The Germanic Strong Verbs. Berlin–New York: Mouton de Gruyter. Mallory, James P. (1982). Indo-European and Kurgan Fauna i. Wild Mammals. Journal of Indo-European Studies 10, 193–222. Mallory, James P. (1983). Proto-Indo-European and Kurgan Fauna ii. Fish. Journal of Indo-European Studies 11, 263–279. Mallory, James P. (1989). In Search of the Indo-Europeans. London: Thames & Hudson. Mallory, James P. (1991). Kurgan and Indo-European Fauna iii. Birds. Journal of IndoEuropean Studies 19, 223–234. Mallory, James P. & Adams, Douglas Q. (1997). Encyclopedia of Indo-European Culture. London–Chicago: Fitzroy Dearborn. Mallory, James P. & Adams, Douglas Q. (2006). The Oxford Introduction to Proto-IndoEuropean and the Proto-Indo-European World. Oxford: University Press. Marchand, James W. (1973). The Sounds and Phonemes of Wulfila’s Gothic. Den Haag– Paris: Mouton. Markey, Thomas L. (1989). The spread of agriculture in western Europe: Indo-European and (non-) pre-Indo-European linguistic evidence. In: David R. Harris & Gordon C. Hillman (eds.), Foraging and Farming. London–New York: Routledge, 585–619. Marstrander, Carl J.S. (1915). Bidrag til det norske sprogs historie i Irland. Kristiania: Dybwad. Martirosyan, Hrach K. (2008). Etymological dictionary of the Armenian inherited lexicon. Leiden–Boston: Brill. Мартынов, Виктор В. (1985). Прусско-славянские эксклюзивные изолексы. Этимология 1982, 3–13. Marzell = Marzell, Heinrich (1943–1979). Wörterbuch der deutschen Pflanzennamen. 4 vols. Leipzig: Hirzel/Stuttgart–Wiesbaden: Steiner. Masing, Oskar (1926). Niederdeutsche Elemente in der Umgangssprache der baltischen Deutschen. Riga: Löffler. Matasović, Maja (2011). Analiza najstarijih latinskih posuđenica u hrvatskom. Diss. Sveučilište u Zagrebu. Matasović, Ranko (1997). Odrazi indoeuropskih laringala u slavenskim jezicima. Croatica 45/46, 129–146. Matasović, Ranko (2004). The Proto-Indo-European syllabic resonants in Balto-Slavic. Indogermanische Forschungen 109, 337–354. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 bibliography 351 Matasović, Ranko (2005). The Centum Elements in Balto-Slavic. In: Gerhard Meiser & Olav Hackstein (eds.), Sprachkontakt und Sprachwandel. Wiesbaden: Reichert, 363– 374. Matasović, Ranko (2008). Poredbeno-povijesna gramatika hrvatskoga jezika. Zagreb: Matica Hrvatska. Matasović, Ranko (2009). Etymological Dictionary of Proto-Celtic. Leiden–Boston: Brill. Matasović, Ranko (2013). Substratum words in Balto-Slavic. Filologija 60, 75–102. Matasović, Ranko (2020). Language of the bird names and the pre-Indo-European substratum. In: Romain Garnier (ed.), Loanwords and Substrata, 331–344. Matasović, Ranko (forthc.). Proto-Slavic Forest Tree Names: Substratum or PIE origin? To appear in: Guus Kroonen (ed.), Sub-Indo-European Europe. De Gruyter Mouton. Mathiassen, Terje (1995). Nochmals zum ie-Komplex im Ostbaltischen. Linguistica Baltica 4, 41–53. Maticsák, Sándor (2014). The Erzya-Mordvin continuation of nominal derivational suffixes in the proto-language composed of a single sibilant or affricate. FinnischUgrische Mitteilungen 38, 115–141. Matzinger, Joachim (2006). Der Altalbanische Text Mbsuame e Krështerë (Dottrina cristiana) des Lekë Matrënga von 1592. Dettelbach: Röll. Mayer, Harvey E. (1978). Die Divergenz des Baltischen und des Slavischen. Zeitschrift für Slavische Philologie 40/1, 52–62. Mayer, Harvey E. (1990). Baltic via Glottalic Indo-European. Baltistica 26, 100–104. Mayrhofer, Manfred (1986). Indogermanische Grammatik i. Heidelberg: Winter. [Mayrhofer EWA, Mayrhofer KEWA — See EWA, KEWA under Abbreviated Titles.] Mažiulis, Vytautas (1979). [Review of] В. Н. Топоров. Прусский язык. Словарь (A–D). Baltistica 15, 146–149. [Mažiulis PKEŽ — See PKEŽ under Abbreviated Titles.] McKenzie, Roderick (1918). Notes sur l’histoire des diphtongues ie et uo dans les langues baltiques. Bulletin de la Société de linguistique de Paris 21/2, 156–174. Meillet, Antoine (1894). De quelques difficultés de la théorie des gutturales indoeuropéennes. Mémoires de la Société de linguistique de Paris 8, 277–304. Meillet, Antoine (1905). Études sur l’étymologie et le vocabulaire du vieux slave ii. Paris: Champion. Meillet, Antoine (1907). Les alternances vocalique en vieux slave. Mémoires de la Société de linguistique de Paris 14, 332–390. Meillet, Antoine (1908–1909). De quelques emprunts probables en grec et en latin. Mémoires de la Société de linguistique de Paris 15, 161–164. Meillet, Antoine (1909). [Review of] Slavisches etymologisches Wörterbuch von Dr Erich Berneker. Rocznik Slawistyczny 2, 57–71. Meillet, Antoine (1912). À propos de Avestique zrazdā-. Mémoires de la Société de linguistique de Paris 18, 60–64. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 352 bibliography Meillet, Antoine (1913). La langue Lituanienne et la langue Lette. Annales des nationalités 2/5–6, 204–205. Meillet, Antoine (1920). Les noms du “feu” et de l’ “eau” et la question du genre. Mémoires de la Société de linguistique de Paris 21, 249–256. Meillet, Antoine & Vaillant, André (1933). Le slave commun *jarębĭ et le suffixe en -b-. Revue des études slaves 13, 101–102. Meineke, Eckhard (1998). Glas. § 1. Sprachliches. In: Rosemarie Müller et al. (ed.), Reallexikon der germanischen Altertumskunde [2ⁿᵈ edition] xii. Getränke–Greiftierstil. Berlin–New York: De Gruyter, 139–142. Meiser, Gerhard (1998). Historische Laut- und Formenlehre der lateinischen Sprache [2ⁿᵈ edition]. Darmstadt: Wissenschaftliche Buchgesellschaft. Meissner, Torsten (2013). Griechische Etymologie [Review of Beekes 2010]. Kratylos 58, 1–31. Melchert, H. Craig (2010). The Word for ‘mouth’ in Hittite and Proto-Indo-European. International Journal of Diachronic Linguistics and Linguistic Reconstruction 7, 55– 63. Meringer, Rudolf (1909). Sprachlich-sachliche Probleme. Wörter und Sachen i, 164– 210. Merkevičius, Algimantas & Muradian, Lijana (2016). Ankstyviausi žirgų palaikai laidojimo objektuose Lietuvoje. Archaeologia Lituana 16, 28–39. Metsäranta, Niklas (2020). Periytyminen ja lainautuminen: Marin ja permiläisten kielten sanastontutkimusta. Diss. Helsingin yliopisto. Meyer, Gustav (1891). Etymologisches Wörterbuch der albanesischen Sprache. Strassburg: Trübner. Meyer, Gustav (1892). Etymologisches. Indogermanische Forschungen 1, 319–329. Mielcke = Mielcke, Christian (1800). Littauisch-Deutsches und Deutsch-Littauisches Wörter-Buch. 2 vols. Königsberg. Miežinis, Mykolas (1894). Lietuviszkai-latviszkai-lenkiszkai-rusiszkas žodynas. Tilžė: Noveskis. Mikkola, Jooseppi J. (1894). Berührungen zwischen den westfinnischen und slavischen Sprachen. Helsinki: Suomalais-Ugrilainen Seura. Mikkola, Jooseppi J. (1896a). Litauische lehnwörter im Slavischen. Beiträge zur Kunde der indogermanischen Sprachen 21, 118–121. Mikkola, Jooseppi J. (1896b). Etymologische beiträge. Beiträge zur Kunde der indogermanischen Sprachen 21, 218–225. Mikkola, Jooseppi J. (1897). Baltische etymologien. Beiträge zur Kunde der indogermanischen Sprachen 22, 239–255. Mikkola, Jooseppi J. (1898). Slavica ii. Indogermanische Forschungen 8, 302–304. Mikkola, Jooseppi J. (1899). Baltische etymologien ii. Beiträge zur Kunde der indogermanischen Sprachen 25, 73–76. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 bibliography 353 Mikkola, Jooseppi J. (1903). Baltisches und Slavisches. Öfversigt af Finska VetenskapsSocietens Förhandlingar 45/4, 2–47. Mikkola, Jooseppi J. (1906). Eräs baltilainen lainasana. Virittäjä 10, 78. Mikkola, Jooseppi J. (1908). [Review of] Vergleichende slavische Grammatik von Dr Wencel Vondrák. Rocznik Slawistyczny 1, 3–19. Mikkola, Jooseppi J. (1930). Vecākie sakari somu un baltu valodu starpā. Izglītības Ministrijas Mēnesraksts 1930, 436–446. Mikkola, Jooseppi J. (1938). Die älteren Berührungen zwischen Ostseefinnisch und Russisch. Helsinki: Suomalais-Ugrilainen Seura. Miklosich, Franz von (1865). Lexicon palaeoslovenico–graeco–latinum. Wien: Braumüller. Miklosich, Franz von (1878). Altslovenische Lautlehre [3ʳᵈ edition]. Wien: Braumüller. Miklosich, Franz von (1886). Etymologisches Wörterbuch der slavischen Sprachen. Wien: Braumüller. Militarev, Alexander & Kogan, Leonid (2005). Semitic Etymological Dictionary ii. Animal Names. Münster: Ugarit-Verlag. Militarev, Alexander & Nikolaev, Sergei (2020). Proto-Afrasian names of ungulates in light of the Proto-Afrasian homeland issue. Journal of Language Relationship 18/3, 199–226. Minkevičius, Karolis et al. (2020). New evidence on the southeast Baltic Late Bronze Age agrarian intensification and the earliest AMS dates of Lens culinaris and Vicia faba. Vegetation History and Archaeobotany 29, 327–338. Mittnik, Alissa et al. (2018). The genetic prehistory of the Baltic Sea region. Nature Communications 9/442. Mittnik, Alissa et al. (2019). Kinship-based social inequality in Bronze Age Europe. Science 366, 731–734. Младеновъ, Стефанъ (1909). Старитѣ германски елементи въ славянскитѣ езици. Сборникъ за народни умотворения, наука и книжнина 25. Motuzaitė Matuzevičiūtė, Giedrė (2018). The possible geographic margin effect on the delay of agriculture introduction into the East Baltic. Estonian Journal of Archaeology 22, 149–162. Motuzaitė Matuzevičiūtė, Giedrė et al. (2022). The integration of millet into the diet of Central Asian populations in the third millennium bc. Antiquity 96 (387), 560– 574. Mous, Maarten (1996). Was There Ever A Southern Cushitic Language (Pre-) Ma’a? In: Catherine Griefenow-Mewis & Rainer Voigt (eds.), Cushitic and Omotic Languages. Köln: Köppe, 201–211. Мудрак, Олег А. (1993). Исторические соответствия чувашских и тюркских гласных: опыт реконструкции и интерпретации. Москва: Институт востоковедения. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 354 bibliography Мудрак, Олег А. (2007). Аварская надпись на сосуде из клада Надь-Сент-Миклош. In: Михаил Е. Алексеев (ed.), Аспекты алтайского языкознания: материалы Тенишевских чтений. Москва: Советский писатель, 81–103. Mueller-Bieniek, Aldona (2010). Carrot (Daucus carota L.) in Medieval Kraków (S. Poland): a cultivated form? Journal of Archaeological Science 37, 1725–1730. Munkácsi, Bernat & Kálman, Béla (1986). Wogulisches Wörterbuch. Budapest: Akadémiai kiadó. Murdock, George P. & Provost, Caterina (1973). Factors in the Division of Labor by Sex: A Cross-Cultural Analysis. Ethnology 12/2, 203–225. Мызников, Сергей А. (2004). Лексика финно-угорского происхождения в русских говорах Северо-Запада. Санкт-Петербург: Наука. Мызников, Сергей А. (2019). Русский диалектный этимологический словарь. Москва–Санкт-Петербург: Нестор-История. Надькин, Дмитрий Т. (1988). Проблема редуцированного гласного и особенности развития ударения в мордовских языках. Актуальные вопросы мордовского языкознания 94, 3–11. Naktinienė, Gertrūda, Aldona Paulauskienė & Vytautas Vitkauskas (1988). Druskininkų tarmės žodynas. Vilnius: Mokslas. Напольских, Владимир В. (1990). Палеоевропейский субстрат в составе западных финно-угров. Uralo-Indogermanica: Балто-славянские языки и проблема уралоиндоевропейских связей ii. Москва: Академия наук СССР, 128–134. Напольских, Владимир В. (1997). Происхождение субстратных палеоевропейских компонентов в составе западных финно-угров. Балто-славянские исследования 1988–1996, 198–208. Напольских, Владимир В. [= Napoľskich, Wladimir] (2002). Zu den ältesten Beziehungen zwischen Finno-Ugriern und zentraleuropäischen Indogermanen. In: Rogier Blokland & Cornelius Hasselblatt (eds.), Finno-Ugrians and Indo-Europeans: linguistic and literary contacts. Maastricht, 265–271. Напольских, Владимир В. (2006). Балто-славянский языковой компонент в Нижнем Прикамье в середине i тыс. н. э. Славоведение 2006/2, 3–19. Напольских, Владимир В. (2010). Названия спорыни в удмуртском языке. Acta Linguistica Petropicani 6/1, 51–60. Напольских, Владимир В. (2015). К происхождению названий соли в финнопермских языках. Linguistica Uralica 51, 161–176. Напольских, Владимир В. & Энговатова, Ася В. (2000). [Review of] Симпозиум “Контакты между носителями нидоевропейских и уральских языков в неолите, энеолите и бронзовом веке в свете лингвистических и археологических данных”. Российская археология 2000/4, 224–232. Nehring, Alfons (1959). Idg. *mari, *mori. In: Wolfdietrich Rasch (ed.), Festschrift für Franz Rolf Schröder. Heidelberg: Winter. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 bibliography 355 Nesselmann, Georg H.F. (1851). Wörterbuch der Littauischen Sprache. Königsberg: Gebrüder Borntraeger. Nichols, Johanna (1998). The Eurasian spread zone and the Indo-European dispersal. In: Roger Blench & Matthew Spriggs (eds.), Archaeology and Language ii. Archaeological Data and Linguistic Hypotheses. London: Routledge, 220–266. Niedermann, Max (1918). Essais d’étymologie et de critique verbale latines. Neuchâtel: Attinger Frères. Nieminen, Eino (1934). Der stammauslaut der ins urfinnische entlehnten baltischen āfeminina und die herkunftsfrage. Finnisch-Ugrische Forschungen 22, 5–66. Nieminen, Eino (1940). Itämerensuomalaisten ja balttilaisten kielten välisten kosketusten tutkimuksesta [Review of Kalima 1936]. Virittäjä 44, 376–384. Nieminen, Eino (1944). Eräistä balttilaisista lainasanoistamme ja niiden vastineista. Virittäjä 48, 238–250. Nieminen, Eino (1945). Reuna ja leuka. Virittäjä 49, 40–55. Nieminen, Eino (1953). Ein Beitrag zu der ostslavischen und ostseefinnischen Badeterminologie. Lingua Posnaniensis 4, 211–228. Nieminen, Eino (1955). Die baltischen und ostseefinnischen Ausdrücke für Segel. Zeitschrift für vergleichende Sprachforschung 72, 129–160. Nieminen, Eino (1956). Die Benennungen der Hirse in den baltischen Sprachen. Zeitschrift für vergleichende Sprachforschung 74, 162–176. Nieminen, Eino (1957). Über einige Eigenschaften der baltischen Sprache, die sich in den ältesten baltischen Lehnwörtern der ostsefinnischen Sprachen abspiegelt. Sitzungsberichte der Finnischen Akademie der Wissenschaften 1956, 185–206. Nieminen, Eino (1963). Ausdrücke für Bau-, Arbeits- und Nutzhölzer fremden Ursprungs im Finnischen. Zeitschrift für Slavische Philologie 31, 233–241. Nikkilä, Osmo (1982). Wörter für ‘Körper’ germanischen Ursprungs im Ostseefinnischen. Советское финно-угроведение 18, 251–260. Nikkilä, Osmo (1999). Sporadischer Konsonantenwechsel im Ostseefinnischen. Finnisch-Ugrische Mitteilungen 21/22, 129–160. Nikkilä, Osmo (2002). Suomen sanojen alkuperät päätökseen [Review of SSA iii]. Virittäjä 106, 124–136. Nikolaev, Alexander (2018). Greek Etymology in the 21ˢᵗ century [Conference handout]. Society for Classical Studies 149ᵗʰ Annual Meeting, Boston, Massachusetts. Николаев, Сергей Л. (1988). Следы особеностей восточнославянских племенных диалектов в современных великорусских говорах i. Кривичи. Балто-славянские исследования 1986, 115–154. Николаев, Сергей Л. (1989). Следы особеностей восточнославянских племенных диалектов в современных великорусских говорах i. Кривичи (окончание). Балто-славянские исследования 1987, 187–225. Николаев, Сергей Л. (1990). К истории племенного диалекта кривичей. Советское славяноведение 4, 54–63. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 356 bibliography Николаев, Сергей Л. (1995). Вокализм карпато-украинских говоров i. Покутскобуковинско-гуцульский ареал. Славяноведение 1995/3, 105–122. Николаев, Сергей Л. (2001). Из исторической фонетики и просодии северозападных говоров. Вопросы русского языкознания 9, 86–121. Николаев, Сергей Л. (2017). К этимологии и сравинтельно-исторической фонетике имен северогерманского (скандинавского) происхождения в «Повести временных лет». Вопросы ономастики 14/2, 7–54. Николаев, Сергей Л. (2020). «Слово о полку Игореве»: Реконструкция стихотворного текста. Москва–Санкт-Петербург: Нестор-История. Николаев, Сергей Л. & Хелимский, Евгений А. (1990). Славянские (новгородопсковские) заимствования в прибалтийско-финских языках: -а и -i в рефлексах имен мужского рода. Uralo-Indogermanica: Балто-славянские языки и проблема урало-индоевропейских связей ii. Москва: Академия наук СССР, 41–43. Nikulin, Andrey (2016). Six Uralic Etymologies [Unpublished article]. Accessed online at academia.edu/28497819. Nilsson, Torbjörn K. (1996). Is Balto-Finnic kypärä ‘helm, casque’ really a loanword from Baltic, or is it indigenous? Word 47, 185–191. Nilsson, Torbjörn K. (2001). A new treatise on Baltic loanwords in Finnish [Review of Liukkonen 1999]. Linguistica Baltica 9, 177–194. Nirvi, Ruben E. (1964). Sanoja ja käyttäytymistä: Sanahistoriallisia tutkimuksia ii. Helsinki: Suomalaisen Kirjallisuuden Seura. Noreen, Adolf (1894). Etymologisches. Indogermanische Forschungen 4, 320–326. Noreen, Adolf (1923). Altnordische Grammatik i. Altisländische und altnorwegische Grammatik [4ᵗʰ edition]. Halle: Niemeyer. Normier, Rudolf (1981). Zu Esche und Espe. Die Sprache 27, 22–29. Nowak, Ronald M. & Paradiso, John L. (1983). Walker’s mammals of the world [4ᵗʰ edition]. Baltimore–London: Johns Hopkins University Press. Nussbaum, Alan (1986). Head and Horn in Indo-European. Berlin: De Gruyter. Nuutinen, Olli (1987a). Malo — balttilainen lainasanarelikti. Virittäjä 91, 514–530. Nuutinen, Olli (1987b). Kalannimi seipi. In: Mauno Koski, Eeva Lähdemäki & Kaisa Häkkinen (eds.), Fennistica festiva in honorem Göran Karlsson septuagernarii. Åbo Akademis förlag, 107–115. Nuutinen, Olli (1987c). Balttilaisten lainojen substituutiotapauksia. Virittäjä 91, 52–63. Nuutinen, Olli (1989). Järvi — balttilainen laina. Virittäjä 93, 497–503. Nuutinen, Olli (1992). Pälvi — balttilainen laina. Virittäjä 96, 403–406. Oettinger, Norbert (2003). Neuerungen in Lexikon und Wortbildung des NordwestIndogermanischen. In: Alfred Bammesberger & Theo Vennemann (eds.), Languages in Prehistoric Europe. Heidelberg: Winter, 183–193. Ohala, John J. (1978). Southern Bantu vs. the World: The Case of Palatalization of Labials. Proceedings of the 4th Annual Meeting of the Berkeley Linguistics Society, 370–386. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 bibliography 357 Ojansuu, Heikki (1908). Hehko, hehvo, hieho. Virittäjä 12, 33–35. Ojansuu, Heikki (1916). Suomen kielen tutkimuksen työmaalta. Sarja esitelmiä i. Jyväskylä: Gummerus. Ojansuu, Heikki (1921). Lisiä suomalais-balttilaisiin kosketuksiin. Helsinki: Suomalaisen Kirjallisuuden Seura. Ojansuu, Heikki (1922). Eesti Etümologiad. Eesti Keel 1, 137–139. Olander, Thomas (2015). Proto-Slavic Inflectional Morphology. Leiden–Boston: Brill. Olander, Thomas (2020). To *b or not to *b: Proto-Indo-European *b in a phylogenetic perspective. Historische Sprachforschung 133, 182–208. Olsen, Birgit (1999). The Noun in Biblical Armenian. Berlin–New York: De Gruyter. Olsen, Birgit (2010). Derivation and Composition. Innsbruck: Institut für Sprachen und Literaturen. Olsen, Birgit & Thorsø, Rasmus (2022). Armenian. In: Thomas Olander (ed.), The IndoEuropean Language Family: A Phylogenetic Perspective. Cambridge: University Press. Orel, Vladimir (2000). A Concise Historical Grammar of the Albanian Language. Leiden–Boston–Köln: Brill. Orel, Vladimir (2003). A Handbook of Germanic Etymology. Leiden–Boston: Brill. Osthoff, Hermann (1898). Griechische und lateinische Wortdeutungen iii. Indogermanische Forschungen 8, 1–68. Osthoff, Hermann (1901). Etymologische Parerga i. Leipzig: Hirzel. Oštir, Karel (1921). Beiträge zur alarodischen Sprachwissenschaft i. Wien–Leipzig: Beyers. Oštir, Karel (1929). Japodi. Etnolog 3, 87–113. Oštir, Karel (1930). Drei vorslavisch-etruskische Vogelnamen. Ljubljana: Znanstveno Drustvo. Otrębski, Jan (1939). Indogermanische Forschungen. Vilnius: Towarzystwo przyjaciół nauk. Otrębski, Jan (1955). Aus der Geschichte der litauischen Sprache. Lingua Posnaniensis 5, 23–40. Otrębski, Jan (1966). Die ältesten germanischen Lehnwörter im Baltischen und Slavischen. Die Sprache 12, 50–64. Ozoliņš, Kaspars (2015). Revisiting Proto-Indo-European Schwebeablaut. Diss. University of California. Paasonen, Heikki (1896). Kielellisiä lisiä suomalaisten sivistyshistoriaan. Suomi 3/13. Paasonen, Heikki (1897). Die türkischen Lehnwörter im Mordwinischen. Helsingfors: Finnische Litteratur-Gesellschaft. Paasonen, Heikki (1903). Mordvinische Lautlehre. Helsinki: Société Finno-Ugrienne. Paasonen, Heikki (1906). Über die benennung des roggens im syrjänisch-wotjakischen und im mordwinischen. Journal de la Société Finno-Ougrienne 23, 1–7. Paasonen, Heikki (1909a). Mordwinische Chrestomathie [2ⁿᵈ edition]. Helsinki: Suomalais-Ugrilainen Seura. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 358 bibliography Paasonen, Heikki (1909b). Eräs liettualaisperäinen sivistyssana. Suomalainen Tiedeakatemia. Esitelmät ja pöytäkirjat 1, 16–17. [Paasonen MdWb — See MdWb under Abbreviated Titles.] Pabrėža, Jurgis A. (1834). Waardaa tayslyynee Augimiu atsyrądąntiu Augmyniićźioo źemaytyyniee [Manuscript]. Accessed online at https://www.epaveldas.lt/preview​ ?id=C10000689960. Pajusalu, Karl, Arvo Krikmann & Eberhard Winkler (2009). Lexical relations between Salaca Livonian and Estonian dialects. Linguistica Uralica 45, 283–298. Pakerys, Antanas (2003). Lietuvių bendrinės kalbos fonetika. Vilnius: Enciklopedija. Palionis, Jonas (1995). Lietuvių literatūrinės kalbos istorija [2ⁿᵈ edition]. Vilnius: Mokslas. Palmér, Axel I., Anthony Jakob et al. (2021). Proto-Indo-European ‘fox’ and the reconstruction of an athematic ḱ-stem. Indo-European Linguistics 9, 234–263. van Pareren, Remco (2006). Baltische leenwoorden in het Mordwiens. MA Thesis Universiteit Groningen. van Pareren, Remco (2008). Die direkten baltischen Lehnwörter im Mordwinischen. Finnisch-Ugrische Mitteilungen 30/31, 69–147. Parpola, Asko (2012). Formation of the Indo-European and Uralic language families in the light of archaeology: Revised and integrated ‘total’ correlations. In: Riho Grünthal & Petri Kallio (eds.), A Linguistic Map of Prehistoric Northern Europe. Helsinki: Suomalais-Ugrilainen Seura, 119–184. Parry-Williams, Thomas H. (1923). The English element in Welsh. London: Cymmrodorion. Павловъ, Алексѣй С. (1890). Неизданный памятникъ русскаго церковнаго права xii вѣка. С.-Петербургъ. Pedersen, Holger (1895). Das indogermanische s im Slavischen. Indogermanische Forschungen 5, 33–87. Pedersen, Holger (1905). Die nasalpräsentia und der Slavische akzent. Zeitschrift für vergleichende Sprachforschung 38, 297–421. Pedersen, Holger (1906). Armenisch und die nachbarsprachen. Zeitschrift für vergleichende Sprachforschung 39, 334–485. Pedersen, Holger (1909). Vergleichende Grammatik der keltischen Sprachen i. Einleitung und Lautlehre. Göttingen: Vandenhoeck & Ruprecht. Pedersen, Holger (1913). Vergleichende Grammatik der keltischen Sprachen ii. Bedeutungslehre (Wortlehre). Göttingen: Vandenhoeck & Ruprecht. Pedersen, Holger (1935). Lit. iau. Studi Baltici 4, 150–154. Pedersen, Holger (1951). Die gemeinindoeuropäischen und die vorindoeuropäischen Verschlusslaute. København: Munksgaard. Пекарский = Пекарский, Эдуард К. (1907–1930). Словарь якутского языка. 13 issues. Санкт-Петербург–Ленинград: Академия наук. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 bibliography 359 Pelisiak, Andrzej (2016). The beginnings of mobile husbandry in the mountain periphery of Southeastern Poland. In: Martin Furholt, Ralf Grossmann & Marzena Szmyt (eds.), Transitional landscapes? The 3rd millennium bc in Europe. Bonn: Habelt, 210– 227. Перевощиков, Петр Н. (1962, ed.). Грамматика современного удмуртского языка. Ижевск: Удмуртское книжное издательство. Persson, Per (1912). Beiträge zur indogermanischen Wortforschung. Uppsala: Akademiska bokhandeln. Petersson, Herbert (1909). Etymologien. Indogermanische Forschungen 23, 384–404. Petersson, Herbert (1921). Studien über die indogermanische Heteroklisie. Lund: Gleerup. Petit, Daniel (2000). Lituanien taũsti, čiùtnas, et le nom du «peuple» en indo-européen. Bulletin de la Société de linguistique de Paris 45, 119–146. Petit, Daniel (2003). Note sur le lituanien kraĩtis. In: Norbert Ostrowski & Ona Vaičiulytė-Romančuk (eds.), Prace bałtystyczne: Język, literatura, kultura. Warszawa: Uniwersytet Warszawski, 92–100. Petit, Daniel (2004). Apophonie et catégories grammaticales dans les langues baltiques. Leuven–Paris: Peeters. Petit, Daniel (2005). Das litauische Adverb ankstì ‘früh’. Studia Etymologica Cracoviensia 10, 141–156. Petit, Daniel (2010). Untersuchungen zu den baltschen Sprachen. Leiden–Boston: Brill. Peust, Carsten (1999). Egyptian Phonology: An Introduction to the Phonology of a Dead Language. Göttingen: Peust & Gutschmidt. Peyrot, Michaël (2018). Tocharian Agricultural Terminology: Between Inheritance and Language Contact. In: Guus Kroonen, James P. Mallory & Bernard Comrie (eds.), Talking Neolithic. Washington, DC: Institute for the Study of Man, 242–277. Philippa et al. = Marlies Philippa, Frans Debrabandere & Arend Quak (2003–2009, eds.). Etymologisch woordenboek van het Nederlands. Amsterdam: University Press. Пичхадзе, Анна А. (2002). О происхождении славянского перевода Хроники Георгия Амартола. Лингвистическое источниковедение и история русского языка 2001, 232–249. Piggot, Stuart (1969). The Earliest Wheeled Vehicles and the Caucasian Evidence. Proceedings of the Prehistoric Society 34, 266–318. Pijnenburg, Wilhelmus J.J. (1989). Eine germanisch-baltoslawische Isoglosse. Historische Sprachforschung 102, 99–106. Piličiauskas, Gytis (2018). Virvelinės keramikos kultūra Lietuvoje 2800–2400 cal bc. Vilnius: Lietuvos istorijos institutas. Piličiauskas, Gytis, Dalia Kisielienė & Giedrė Piličiauskienė (2016). Deconstructing the concept of Subneolithic farming in the southeastern Baltic. Vegetation History and Archaeobotany 26, 183–193. Piličiauskas, Gytis et al. (2017). The transition from foraging to farming (7000–500 cal bc) in the SE Baltic. Journal of Archaeological Science: Reports 14, 530–542. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 360 bibliography Pinault, Georges-Jean (2008). Chrestomathie tokharienne. Leuven–Paris: Peeters. Pinault, Georges-Jean (2009). On the formation of the Tocharian demonstratives. In: Elisabeth Rieken & Paul Widmer (eds.), Pragmatische Kategorien: Form, Funktion und Diachronie [Proceedings]. Wiesbaden: Reichert, 221–245. Пирейко, Лия А. (1976). Талышско-русский словарь. Москва: Русский язык. Pisani, Vittore (1968). Rom und die Balten. Baltistica 4, 7–21. Pleteršnik = Pleteršnik, Maks (1894–1895). Slovensko-nemški slovar. 2 vols. Ljubljana: Knezoškofijstvo. Plöger, Angela (1973). Die russischen Lehnwörter der finnischen Schriftsprache. Wiesbaden: Harrassowitz. Plöger, Angela (1982). Über die Entstehung des finnischen Stammtyps CV̄ C(C)a/ä. Finnisch-Ugrische Forschungen 44, 66–98. Подвысоцкiй, Александръ (1885). Словарь областного архангельскаго нарѣчiя. С.Петербургъ: Академія наукъ. Pogatscher, Alois (1898). Altenglisch ƀr aus mr. In: Jakob Schipper (ed.), Festschrift zum viii. allgemeinen deutschen Neuphilologentage in Wien Pfingsten 1898. Wien– Leipzig: Braumüller, 97–107. Погодинъ, Александръ Л. (1903). Слѣды корней-основъ въ славянскихъ языкахъ. Варшава: Варшавскій учебный округъ. Polák, Václav (1955). Trois notules étymologiques concernant les mots de substrat. Časopis pro moderní filologii 37, 54–56. Pollmann, Britta (2014). Environment and agriculture of the transitional period from the Late Bronze to early Iron Age in the eastern Baltic: an archaeobotanical case study of the lakeshore settlement Luokesa 1, Lithuania. Vegetation History and Archaeobotany 23, 403–418. Polomé, Edgar C. (1986). The Non-Indo-European Component of the Germanic Lexicon. In: Annemarie Etter (ed.), o-o-pe-ro-si [FS Ernst Risch]. Berlin–New York: De Gruyter, 661–672. Polomé, Edgar C. (1989). Substrate Lexicon in Germanic. North-Western European Language Evolution 14, 53–73. Polomé, Edgar C. (1990). The Indo-Europeanization of Northern Europe: the Linguistic Evidence. In: Karlene Jones-Bley & Martin E. Huld (eds.), The Indo-Europeanization of Northern Europe. Washington, DC: Institute for the Study of Man, 331–338. Polomé, Edgar C. (1992). Indo-European and Substrate Languages in the West. Archivio Glottologico Italiano 67, 66–85. Popowska-Taborska, Hanna & Boryś, Wiesław (1996). Leksyka kaszubska na tle słowianskim. Warszawa: Akademia Nauk. Porzig, Walter (1954). Die Gliederung des indogermanischen Sprachgebiets. Heidelberg: Winter. Posti, Lauri (1946). Etymologisia huomioita. Virittäjä 50, 385–387. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 bibliography 361 Posti, Lauri (1953). From Pre-Finnic to Late Proto-Finnic. Finnisch-Ugrische Forschungen 31/1, 1–91. Posti, Lauri (1962). Silakka sanan alkuperästä. Virittäjä 66, 282–287. Posti, Lauri (1977). Some new contributions to the stock of Baltic loanwords in Finnic languages. Baltistica 13, 263–270. Потебня, Александръ А. (1881). К исторiи звуков русскаго языка iii. Этимологическiя и другiя замѣтки. Варшава: Земкевичъ и Нолковский. Prellwitz, Walther (1891). Die deutschen Lehnwörter im Preussischen und Lautlehre der deutschen Lehnwörter im Litauischen. Göttingen: Vandenhoeck & Ruprecht. Prellwitz, Walther (1897). Studien zur indogermanischen etymologie und wortbildung. Beiträge zur Kunde der indogermanischen Sprachen 22, 76–114. Prellwitz, Walther (1909). Etymologische Miszellen. Zeitschrift für vergleichende Sprachforschung 42, 385–387. Преображенский = Преображенскій, Александръ Г. (1910–1916). Этимологическiй словарь русскаго языка. 2 vols. Москва: Лисснеръ & Совко. Press, J. Ian (1986). Aspects of the Phonology of the Slavonic Languages: The Vowel y and the Consonantal Correlation of Palatalization. Amsterdam: Rodopi. Priestly, Tom M.S. (1997). Vandals, Veneti, Windischer: The Pitfalls of Amateur Historical Linguistics. Slovene Studies 19, 3–41. Pronk, Tijmen (2012). Proto-Indo-European long vowels and Balto-Slavic accentuation. Baltistica 47, 205–247. Pronk, Tijmen (2013). On the Development of *in, *im, *un and *um in Slavic. In: Marko Jesenšek (ed.), Miklošičeva monografija. Ob dvestoletnici rojstva Franca Miklošiča. Ljutomer: Gimnazija Franca Miklošiča, 117–140. Pronk, Tijmen (2015). Singulative n-stems in Indo-European. Transactions of the Philological Society 113/3, 327–348. Pronk, Tijmen (2018). Old Church Slavonic ( j)utro, Vedic uṣár- ‘daybreak, morning’. In: Lucien van Beek et al. (eds.), Farnah [FS Lubotsky]. Ann Arbor: Beech Stave Press, 298–306. Pronk, Tijmen (2019a). Proto-Indo-European *a. Indo-European Linguistics 7, 122–163. Pronk, Tijmen (2019b). Eichner’s law: a critical survey of the evidence. Münchener Studien zur Sprachwissenschaft 73, 121–155. Pronk, Tijmen (2022). Balto-Slavic. In: Thomas Olander (ed.), The Indo-European Language Family: A Phylogenetic Perspective. Cambridge: University Press. Pronk, Tijmen & Pronk-Tiethoff, Saskia (2018). Balto-Slavic Agricultural Terminology. In: Guus Kroonen, James P. Mallory & Bernard Comrie (eds.), Talking Neolithic. Washington, DC: Institute for the Study of Man, 278–314. Pronk-Tiethoff, Saskia (2012). The Germanic loanwords in Proto-Slavic. Amsterdam– New York: Rodopi. Puhvel, Jaan (1990). On Terms for ‘nit’ in Baltic and Balto-Finnic. In: Baiba Metuzāle- Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 362 bibliography Kangere & Helge Rinholm (eds.), Symposium Balticum [FS Rūk̦e-Dravin̦ a]. Hamburg: Buske. Puhvel, Jaan (1997). Hittite Etymological Dictionary iv. Words beginning with K. Berlin– New York: Mouton De Gruyter. Pystynen, Juho (2014a). Palatal Unpacking in Finnic [Conference talk]. 19th Conference of the Finno-Ugric Studies Association of Canada. Pystynen, Juho (2014b). Interplay of minor soundlaws: Samic glide clusters. [Blog post, dated 19 July 2014, hosted at protouralic.wordpress.com]. Pystynen, Juho (2016). Another Phonological Relict in South Estonian. [Blog post, dated 23 July 2016, hosted at protouralic.wordpress.com]. Pystynen, Juho (2017). *ü > *i, *ü in Samoyedic. [Blog post, dated 7 October 2017, hosted at protouralic.wordpress.com]. Pystynen, Juho (2018). Itämerensuomen pitkien vokaalien alkuperä. Diss. Helsingin yliopisto. Pystynen, Juho (2019). Seistä. In: Santeri Junttila & Juha Kuokkala (eds.), Petri Kallio Rocks: Liber semisaecularis. Helsinki, 153–160. Pystynen, Juho (2020a). On the development of *i in Permic. Finnisch-ugrische Forschungen 65, 62–97. Pystynen, Juho (2020b). Secondary apocope in Mordvin. [Blog post, dated 14 February 2020, hosted at protouralic.wordpress.com]. Pystynen, Juho (2020c). Notes on Proto-Mansi word-final vocalism. In: Sampsa Holopainen et al. (eds.), Ёмас сымыӈ нэ̄ кве во̄ ртур э̄ тпост самын патум [FS UllaMaija Forsberg]. Helsinki: Suomalais-Ugrilaisen Seuran Toimituksia, 249–261. Pystynen, Juho (2023). Sorsa ja sotka. Äännesuhteita sukelluksissa. In: Sampsa Holopainen, Jeongdo Kim & Niklas Metsäranta (eds.), Elämä ja etymologia [FS Saarikivi]. Helsinki, 336–364. Qvigstad, Just (1893). Nordische Lehnwörter im Lappischen. Christiania: VidenskabsSelskabs forhandlinger. Rannamäe, Eve (2016). Development of sheep populations in Estonia as indicated by archaeofaunal evidence and ancient mitochondrial DNA lineages from the Bronze Age to the Modern Period on sheep. Diss. University of Tartu. Räsänen, Martti (1920). Die tschuwassischen Lehnwörter im Tscheremissischen. Helsinki: Suomalais-Ugrilainen Seura. Räsänen, Martti (1969). Versuch eines etymologischen Wörterbuchs der Türksprachen. Helsinki: Suomalais-Ugrilainen Seura. Rasmussen, Jens E. (1999). Selected Papers on Indo-European Linguistics i. København: Museum Tusculanum. Ravila, Paavo (1935). Die stellung des lappischen innerhalb der finnisch-ugrischen sprachfamilie. Finnisch-Ugrische Forschungen 23, 20–65. Rédei, Károly (1968). Szófejtések. Nyelvtudományi közlemények 70, 155–161. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 bibliography 363 Rédei, Károly (1986). Zu den indogermanisch-uralische Sprachkontakten. Wien: Akademie der Wissenschaften. Rédei, Károly (2000). [Review of] Kari Liukkonen, Baltisches im Finnischen. Linguistica Uralica 36, 226–229. [Rédei UEW — see UEW under Abbreviated Titles.] Reutercrona, Hans G. (1920). Svarabhakti und erleichterungsvokal im altdeutschen bis ca. 1250. Heidelberg: Winter. Rhea. Samuel A. (1872). Brief Grammar and Vocabulary of the Kurdish Language of the Hakari District. Journal of the American Oriental Society 10, 118–155. Rießler, Michael (2022). Kildin Saami. In: Marianne Bakró-Nagy, Johanna Laakso & Elena Skribnik (eds.), The Oxford Guide to the Uralic Languages. Oxford: University Press, 219–239. Rietz = Rietz, Johann E. (1862–1867). Svenskt dialektlexikon: ordbok öfver svenska allmogespråket [1962 reprint]. Lund: Gleerup. Rimantienė, Rimutė (1992). The Neolithic of the Eastern Baltic. Journal of World Prehistory 6, 97–143. Ritter, Ralph-Peter (1977). Zur Frage der finnischen Evidenz für die Sievers’sche Regel im Germanischen. Die Sprache 23, 171–179. Ritter, Ralph-Peter (1993). Studien zu den ältesten germanischen Entlehnungen im Ostseefinnischen. Frankfurt am Main: Lang. Ritter, Ralph-Peter (1998). Zu den Reflexen der baltischen Liquiddiphthonge im Ostseefinnischen. Res Balticae 1998, 113–116. Rix, Helmut (1976). Historische Grammatik des Griechischen: Laut- und Formenlehre. Darmstadt: Wissenschaftliche Buchgesellschaft. Rix, Helmut (1999). Schwach charakterisierte lateinische Präsensstämme zu Seṭ-Wurzeln mit Vollstufe i. In: Heiner Eichner & Hans Christian Luschützky (eds.), Compositiones indo-germanicae: in memoriam Jochem Schindler. Praha–Wien: Enigma, 515–535. Robbeets, Martine (2004). Does Doerfer’s Zufall mean ‘cognate’? The case of the initial velar correspondence in Altaic. Turkic Languages 8, 147–179. Roffet-Salque, Mélanie et al. (2015). Widespread exploitation of the honeybee by early Neolithic farmers. Nature 527, 226–231. Rosenschon, Ursula (1993). Sechs Seiten medizinischer Rezepte im glagolitischen Psalter 3/N des Sinaiklosters. Sudhoff’s Archiv 77, 129–159. Rozwadowski, Jan M. (1915). Przyczynki do historycznej fonetyki języków słowiańskich. Rocznik slawistyczny 7, 9–21. Ruhig = Ruhig, Philipp (1747). Littauisch-Deutsches und Deutsch-Littauisches Lexicon. 2 vols. Königsberg: Hartung. Rūķe-Draviņa, Velta (1955). Die Benennungen des Wacholders im Baltischen. ORBIS: Bulletin International de Documentation Linguistique 4/1, 390–409. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 364 bibliography Rūķe-Draviņa, Velta (1964). Die Expansion der russischen Buchweizenbenennungen nach Westen. Scando-Slavica 10, 107–131. Rundberget, Bernt et al. (2020). Bloomery ironmaking in Latvia — a comparative study of Iron Age and medieval technologies. Historical Metallurgy 52/2, 96–109. Saag, Lehti et al. (2017). Extensive Farming in Estonia Started through a Sex-Biased Migration from the Steppe. Current Biology 27, 2185–2193. Saag, Lehti et al. (2019). The Arrival of Siberian Ancestry Connecting the Eastern Baltic to Uralic Speakers further East. Current Biology 29, 1701–1711. Saarikivi, Janne (2004). Is there Palaeo-European substratum interference in western branches of Uralic? Suomalais-Ugrilaisen Seuran Aikakauskirja 90, 187–214. Saarikivi, Janne (2011). Saamelaiskielet — nykypäivää ja historiaa. In: Irja SeurujärviKari, Petri Halinen & Risto Pulkkinen (eds.), Saamentutkimus tänään. Helsinki: Suomalaisen Kirjallisuuden Seura, 77–119. Saarikivi, Janne (2018). Finnic and other western Uralic borrowings in Permian. In: Sampsa Holopainen & Janne Saarikivi (eds.), Περὶ ὀρθότητος ἐτύμων: Uusiutuva uralilainen etymologia. Helsinki, 269–355. Saarikivi, Janne (2022). The divergence of Proto-Uralic and its offspring: A descendent reconstruction. In: Marianne Bakró-Nagy, Johanna Laakso & Elena Skribnik (eds.), The Oxford Guide to the Uralic Languages. Oxford: University Press, 28–58. Sabaliauskas, Algirdas (1959). Относительно происхождения названия растений в балтийских языках. In: Ēvalds Sokols (ed.), Rakstu krājums [Gedenkschrift Endzelīns]. Rīga: Zinātņu akadēmija, 219–242. Sabaliauskas, Algirdas (1960a). О происхождении названия мака в балтийских языках. Kраткие сообщения Института славяноведения 28, 70–73. Sabaliauskas, Algirdas (1960b). Dėl kai kurių baltų kalbų žemės ūkio augalų pavadinimų kilmės. Lietuvių kalbotyros klausimai 3, 257–267. Sabaliauskas, Algirdas (1968). Baltų kalbų naminių gyvulių pavadinimai. Lietuvių kalbotyros klausimai 10, 101–190. Sabaliauskas, Algirdas (1974). Zur Herkunft von litauisch pal̃vė. Zeitschrift für Slawistik 19, 200–211. Sabaliauskas, Algirdas (1990). Lietuvių kalbos leksika. Vilnius: Mokslas. Sadziński, Witold & Witczak, Krzysztof T. (2016). Rozważania o nazwach bałtyckich fok. Pomorania Antiqua 25, 49–66. Salminen, Tapani (1999). Euroopan kielet muinoin ja nykyisin. In: Paul Fogelberg (ed.), Pohjan poluilla. Helsinki: Societas Scientiarum Fennica, 13–26. Salminen, Tapani (2002). Problems in the taxonomy of the Uralic languages in the light of modern comparative studies. In: Александр Е. Кибрик (ed.), Лингвистический беспредел. Москва: Московский государственный университет, 44–55. Salmons, Joseph [Joe] (1992). Northwest Indo-European Vocabulary and Substrate Phonology In: Perspectives on Indo-European Language, Culture and Religion [FS Polomé]. McLean, Virginia: Institute for the Study of Man, 265–279. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 bibliography 365 Salmons, Joseph (2004). How (non-)Indo-European is the Germanic lexicon? … And what does that mean? In: Irma Hyvärinen, Petri Kallio & Jarmo Korhonen (eds.), Etymologie, Entlehnungen und Entwicklungen [FS Koivulehto]. Helsinki: Société Néophilologique, 311–321. Sammallahti, Pekka (1977). Suomalaisten esihistorian kysymyksiä. Virittäjä 81/2, 119– 136. Sammallahti, Pekka (1988). Historical phonology of the Uralic languages with special reference to Samoyed, Ugric and Permic. In: Denis Sinor (ed.), The Uralic languages: description, history and foreign influences. Leiden: Brill, 478–554. Sammallahti, Pekka (1998). The Saami Languages: an Introduction. Kárášjohka: Davvi Girji. Sammallahti, Pekka (1999). The Indo-European loanwords in Saami. In: Christian Carpelan, Asko Parpola & Petteri Koskikallio (eds.), Early contacts between Uralic and Indo-European: linguistic and archaeological considerations. Helsinki: Suomalais-Ugrilainen Seura, 397–415. Sauer, Jonathan D. (1988). Plant Migration: The Dynamics of Geographic Patterning in Seed Plant Species. Berkeley: University of California Press. Saukkonen, Pauli (1962). Eräs etymologiointiperiaate. Virittäjä 66, 342–344. Sausverde, Ērika (1996). Seewörter and Substratum in Germanic, Baltic and Baltic Finno-Ugric Languages. In: Karlene Jones-Bley & Martin E. Huld (eds.), The IndoEuropeanization of Northern Europe. Washington, DC: Institute for the Study of Man, 133–147. Шахматовъ, Алексѣй А. (1893). Изслѣдованiя в области русской фонетики. Варшава: Варшавскій учебный округъ. Шахматовъ, Алексѣй А. (1915). Очеркъ древнѣйшаго перiода исторiи русскаго языка. Петроградъ: Императорская академия наукъ. Шахматовъ, Алексѣй А. (1916a). Повѣсть временныхъ лѣтъ i. Вводная часть. Текстъ. Примѣчанія. Петроградъ: Археографическая комиссия. Шахматовъ, Алексѣй А. (1916b). Введенiе въ курсъ исторiи русскаго языка i. Историческiй процессъ образованiя русскихъ племенъ и нарѣчiй. Петроградъ: Научное дѣло. Schaarschmidt, Gunter (1997). A Historical Phonology of the Upper and Lower Sorbian Languages. Heidelberg: Winter. Schaffner, Stefan (2001). Das Vernersche Gesetz und der innerparadigmatische grammatische Wechsel des Urgermanischen im Nominalbereich. Innsbruck: Institut für Sprachen und Literaturen. Schiller, Karl & Lübben, August (1875–1881). Mittelniederdeutsches Wörterbuch. 6 vols. Bremen: Kühtmann/Fischer. Schindler, Jochem (1970). [Review of] Anttila, Raimo: Proto-Indo-European Schwebeablaut. Kratylos 15, 146–152. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 366 bibliography Schmalstieg, William R. (1983). Slavic kamy and mati: a structural view. Journal of IndoEuropean Studies 11, 151–165. Schmeller/Bergmann = Joseph Bergmann (1855, ed.), Johann Andreas Schmeller’s sogenanntes Cimbrisches Wörterbuch. Wien: Kaiserlich-königliche Hof- und Staatsdruckerei. Schmid, Wolfgang P. (1968). Alteuropäisch und Indogermanisch. Abhandlungen der geistes- und sozialwissenschaftlichen Klasse 1968/6. Schmid, Wolfgang P. (1977). Sprachwissenschaftliche Bemerkungen zu den Wörtern für Bauer und Dorf im Slavischen. In: Herbert Jankuhn, Rudolf Schützeichel & Fred Schwind (eds.), Das Dorf der Eisenzeit und des frühen Mittelalters. Göttingen: Vandenhoeck & Ruprecht, 41–61. Schmidt, Johannes (1875). Zur Geschichte des indogermanischen Vocalismus. Zweite Abteilung. Weimar: Böhlau. Schmidt, Johannes (1891). Die urheimath der Indogermanen und das europäische zahlsystem. Abhandlungen der Königlichen Akademie der Wissenschaften zu Berlin aus dem Jahre 1890. Schmidt, Johannes (1895). Kritik der Sonantentheorie. Weimar: Böhlau. Schmidt, Karl H. (1977). Der Beitrag der keltiberischen Inschrift von Botorrita zur Rekonstruktion der protokeltischen Syntax. Word 28, 51–62. Schmidt, Peter [= Šmits, Pēteris] (1930). Sprachliche Zeugnisse über die Urheimat der Balten und ihre Kultur. In: Fritzis Adamovičs (ed.), Die Letten. Riga: Walters & Rapa, 71–88. Schmidt-Brandt, Robert (1967). Die Entwicklung des indogermanischen Vokalsystems. Heidelberg: Groos. Schrader, Otto (1883). Sprachvergleichung und Urgeschichte. Jena: Costenoble. Schrader, Otto (1901). Reallexikon der indo-germanischen Altertumskunde. Strassburg: Trübner. Schrader/Nehring = Schrader, Otto & Nehring, Alfons (1917–1929). Reallexikon der indogermanischen Altertumskunde [22ⁿᵈ edition]. 2 vols. Berlin–Leipzig: De Gruyter. Schrijver, Peter (1991). The reflexes of the Proto-Indo-European laryngeals in Latin. Amsterdam: Rodopi. Schrijver, Peter (1995a). Studies in British Celtic Historical Phonology. Amsterdam– Atlanta: Rodopi. Schrijver, Peter (1995b). De etymologie van de naam van de Cannenefaten. Amsterdamer Beiträge zur älteren Germanistik 41, 13–22. Schrijver, Peter (1997). Animal, vegetable and mineral: some Western European substratum words. In: Alexander Lubotsky (ed.), Sound Law and Analogy [FS Beekes]. Amsterdam–Atlanta: Rodopi, 293–316. Schrijver, Peter (1999). On Henbane and Early European Narcotics. Zeitschrift für celtische Philologie 51, 17–45. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 bibliography 367 Schrijver, Peter (2001). Lost Languages in Northern Europe. In: Christian Carpelan, Asko Parpola & Petteri Koskikallio (eds.), Early Contacts between Uralic and IndoEuropean: Linguistic and Archaeological Considerations. Helsinki: SuomalaisUgrilainen Seura, 417–425. Schrijver, Peter (2018). Talking Neolithic: The Case for Hatto-Minoan and its Relationship to Sumerian. In: Guus Kroonen, James P. Mallory & Bernard Comrie (eds.), Talking Neolithic. Washington, DC: Institute for the Study of Man, 336–374. Schuchardt, Hugo (1922). Heimisches und fremdes Sprachgut. Revista Internacional de los Estudios Vascos 13, 70–82. Schuhmann, Roland (2016). Where is the Substrate in the Germanic Lexicon? In: Bjarne S.S. Hansen et al. (eds.), Etymology and the European Lexicon. Wiesbaden: Reichert, 377–384. Schulze, Wilhelm (1910). Etymologisches. Sitzungsberichte der königliche preussische Akademie der Wissenschaften 1910, 787–808. Schulze, Wilhelm (1913). Der Fuchs. Zeitschrift für vergleichende Sprachforschung 45, 287–288. Schumacher, Stefan (2004). Die keltischen Primärverben. Innsbruck: Innsbrucker Beiträge zur Sprachwissenschaft. Schumacher, Stefan (2013). Historische Phonologie. In: Schumacher, Stefan & Matzinger, Joachim. Die Verben des Altalbanischen. Wiesbaden: Harrassowitz, 205– 276. Schuster-Šewc = Schuster-Šewc, Heinz (1978–1996). Historisch-etymologisches Wörterbuch der ober- und niedersorbischen Sprache. 24 issues. Bautzen: Domowina. Schützeichel, Rudolf (2004). Althochdeutscher und altsächsischer Glossenwortschatz. 12 vols. Niemeyer: Tübingen. Seewald, P. (1865). Einige lettische Ausdrucksformen. Magazin, herausgegeben von der Lettisch-Literärischen Gesellschaft 13/2, 38–74. Sehwers, Johann (1936). Sprachlich-kulturhistorische Untersuchungen: vornehmlich über den deutschen Einfluss im Lettischen. Wiesbaden: Harrassowitz. Senn, Alfred (1925). Germanische Lehnwortstudien. Heidelberg: Winter. Senn, Alfred (1933). Nachträge zu Kluge’s Wörterbuch. The Journal of English and Germanic Philology 32, 504–529. Senn, Alfred (1943). The Historical Development of the Lithuanian Vocabulary. Bulletin of the Polish Institute of Arts and Sciences in America 1, 946–969. Seržant, Ilja A. (2006). Die Vermittlungsrolle des Hochlettischen bei den altrussischen und litauischen Entlehnungen im Lettischen. Acta Linguistica Lituanica 55, 89– 106. Seržant, Ilja A. [= Сержант, И.А.] (2008). Относительная хронология процессов палатализации прабалтийских *k и *g в латышском. Вопросы языкознания 2008/1, 121–129. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 368 bibliography Setälä, Eemil, N. (1902). Zur etymologie von Sampo. Finnisch-Ugrische Forschungen 2, 141–164. Setälä, Eemil, N. (1909). Beiträge zur finnisch-ugrischen wortkunde 9. Fi. kataja ‘wachholder’. Finnisch-Ugrische Forschungen 9, 126–128. Setälä, Eemil, N. (1929). Itämerensuomalaisten kielten kosketukset slaavilaisten kanssa. Suomalais-Ugrilaisen Seuran Aikakauskirja 43, 30–39. Shevelov, George Y. (1964). A Prehistory of Slavic: The Historical Phonology of Common Slavic. Heidelberg: Winter. Shevelov, George Y. (1979). A Historical Phonology of the Ukrainian Language. Heidelberg: Winter. Sillasoo, Ülle & Hiie, Sirje (2007). An archaeobotanical approach to investigating food of the Hanseatic period in Estonia. In: Sabine Karg (ed.), Medieval Food Traditions in Northern Europe. Copenhagen: National Museum of Denmark, 73–96. Simon, Zsolt (2008). How to find the Proto-Indo-European homeland? A methodological essay. Acta Antiqua Academiae Scientiarum Hungaricae 48, 289–303. Simon, Zsolt (2022). The Hurro-Urartian loan contacts of Armenian: A revision. Hungarian Assyriological Review 3, 63–89. Simon, Zsolt (forthc.). Indo-European Substrates: The Problem of the Anatolian Evidence. To appear in: Guus Kroonen (ed.), Sub-Indo-European Europe. De Gruyter Mouton. Skach, Elisabeth (2010). Die slavische Monophthongierung *ej > ī im Lichte der Lehnbeziehungen zum Baltischen und Ostseefinnischen. Wiener slavistisches Jahrbuch 56, 137–140. Skardžius, Pranas (1931). Die slavischen Lehnwörter im Altlitauischen. Kaunas: Spindulys. Skardžius, Pranas (1941). Lietuvių kalbos žodžių daryba. Vilnius: Lietuvos mokslų akademija. Skeat, Walter W. (1882). An Etymological Dictionary of the English Language. Oxford: Clarendon Press. Skjærvø, P. Oktor (2002). Khotanese manuscripts from Chinese Turkestan in the British Library: a complete catalogue with texts and translations. London: British Library. Skok = Skok, Petar (1971–1974). Etimologijski rječnik hrvatskoga ili srpskoga jezika. 3 vols. Zagreb: Akademija znanosti i umjetnosti. Скворцов = Михаил И. Скворцов (1985, ed.), Чувашско-русский словарь. Москва: Русский язык. Sławski, Franciszek (1960). Dubrownicka nazwa łabędzia. Rocznik sławistyczny 21, 37– 40. Sławski SEJP = Franciszek Sławski (1952–1982). Słownik etymologiczny języka polskiego. 5 vols. [*a–*łżywy]. Kraków: Towarzystwo miłośników jęzika polskiego. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 bibliography 369 Sławski SP = Franciszek Sławski (1974–2001, ed.), Słownik prasłowiański. 8 vols. [*a– *gyža]. Wrocław: Ossolineum. van Sluis, Paulus (2022). Beekeeping in Celtic and Indo-European. Studia Celtica 56, 1– 28. van Sluis, Paulus (forthc.). Substrate alternations in Celtic. To appear in: Guus Kroonen (ed.), Sub-Indo-European Europe. De Gruyter Mouton. van Sluis, Paulus, Anders R. Jørgensen, & Guus Kroonen (2023). European Prehistory between Celtic and Germanic: The Celto-Germanic isoglosses revisited. In: Kristian Kristiansen, Guus Kroonen & Eske Willerslev (eds.), The Indo-European Puzzle Revisited. Cambridge: University Press, 193–243. Ślusarska, Katarzyna (2021). Wild Resources in the Economy of Bronze and Early Iron Ages Between Oder and Bug Rivers — Source Overview. Open Archaeology 7, 177– 210. Smalinskienė, Gražina (1994). Lie. salà, la. sala „kaimas“. Lietuvių kalbotyros klausimai 31, 177–180. Сметанина, Зоя В. & Иванова, Галина А. (2018). Фонематические варианты в «Областном словаре вятских говоров». In: Людмила В. Калинина (ed.), Семантика. Функционирование. Текст [FS Черновая]. Киров: Радуга, 203–210. Śmieszek, Antoni (1909). Etymologica. Materyały i prace Komisyi Językowej Akademii Umiejętności w Krakowie 4, 391–408. Smith, Emil (1910). Om opindelsen til fiskenavene sild og sil. In: Magnus Olsen (ed.), Maal og minne. Kristiania: Bymaals-Laget, 139–142. Smoczyński, Wojciech (1983). Staropruskie lekcje i etymologie i. Biuletyn Polskiego Towarzystwa Językoznawczego 40, 167–183. Smoczyński, Wojciech (1989). Studia bałto-słowiańskie i. Wrocław: Ossolineum. Smoczyński, Wojciech (1998). Etimologijos pastabos v. Nomina postverbalia indas, pergas, išeiga ir kt. Baltistica 33, 251–260. Smoczyński, Wojciech (1999). Заметки о балто-славянских рефлексах и.-е. ларингальных. Kalbotyra 48/2, 7–26. Smoczyński, Wojciech (2000). Untersuchungen zum deutschen Lehngut im Altpreussischen. Kraków: Uniwersytet Jagielloński. Smoczyński, Wojciech (2006). Laringalų teorija ir lietuvių kalba. Vilnius: Lietuvių kalbos institutas. Smoczyński, Wojciech (2018). Lithuanian Etymological Dictionary. Berlin: Peter Lang. Smoczyński SEJL² = Smoczyński, Wojciech. Słownik etymologiczny języka litewskiego (online version). Accessed at https://rromanes.org/pub/alii/. Snoj, Marko (2003). Slovenski etimološki slovar. Ljubljana: Modrijan. Соболевскiй, Алексѣй И. (1910). Мелкія замѣтки по славянской и русской фонетикѣ. Русскiй филологическiй вѣстникъ 64, 178–179. Соболевскiй, Алексѣй И. (1914). Нѣсколько замѣтокъ по славянскимъ вокализму и лексикѣ. Русскiй филологическiй вѣстникъ 71, 431–452. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 370 bibliography Solmsen, Felix (1904). Slavische etymologien. Zeitschrift für vergleichende Sprachforschung 37, 575–601. Sommer, Ferdinand (1905). Griechische Lautstudien. Strassburg: Trübner. Sommer, Ferdinand (1914). Die indogermanischen iā- und io-Stämme im Baltischen. Leipzig: Teubner. Soosaar, Sven-Erik (2021). Võimalikke protoeuroopa substraadi jälgi eesti keele dendronüümides. Emakeele Seltsi aastaraamat 66, 295–309. Šorgo, Aljoša (2020). Characteristics of lexemes of a substratum origin in Proto-Germanic. In: Romain Garnier (ed.), Loanwords and substrata. Innsbruck: Innsbrucker Beiträge zur Sprachwissenschaft, 427–472. Сорвачева, Валентина А. & Безносикова, Люция М. (1990). Удорский диалект коми языка. Москва: Наука. Specht, Franz (1932). Die Flexion der n-Stämme im Baltisch-Slavischen und Verwandtes. Zeitschrift für vergleichende Sprachforschung 59, 213–298. Specht, Franz (1935). Zur baltisch-slavischen Spracheinheit. Zeitschrift für vergleichende Sprachforschung 62, 248–258. Specht, Franz (1947). Der Ursprung der Indogermanischen Deklination. Göttingen: Vandenhoeck & Ruprecht. van der Staaij, Robert J. (1975). A Reconstruction of Proto-Italic. Leiden: University Press. Stang, Christian S. (1957). Slavonic Accentuation. Oslo: Universitetsforlaget. Stang, Christian S. (1966). Vergleichende Grammatik der Baltischen Sprachen. Oslo: Universitetsforlaget. Stang, Christian S. (1972). Lexikalische Sonderübereinstimmungen zwischen dem Slavischen, Baltischen und Germanischen. Oslo: Universitetsforlaget. Stankus, Jonas (2001). Geležies gamybos lietuvoje apžvalga. Lietuvos archeologija 21, 171–182. Starostin, Sergei, Anna Dybo & Oleg Mudrak (2003). An Etymological Dictionary of Altaic Languages. Leiden: Brill. Steingass, Francis J. (1892). A Comprehensive Persian-English Dictionary. London: Routledge & K. Paul. Steinitz, Wolfgang (1964). Zur Periodisierung der alten baltischen und germanischen Lehnwörter im Finnischen. Wissenschaftliche Zeitschrift der Humbolt-Universität zu Berlin. Gesellschafts- und sprachwissenschaftliche Reihe 13, 335–338. Stender-Petersen, Adolf (1927). Slavisch-Germanische Lehnwortkunde. Göteborg: Elander. Stifter, David (1998). Old Irish ²fén ‘bog’. Die Sprache 40, 226–228. Stifter, David (2019). An apple a day … Indogermanische Forschungen 124, 171–218. Stifter, David (2023). With the Back to the Ocean. The Celtic Maritime Vocabulary. In: Kristian Kristiansen, Guus Kroonen & Eske Willerslev (eds.), The Indo-European Puzzle Revisited. Cambridge: University Press, 172–192. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 bibliography 371 Stifter, David (forthc.). Prehistoric layers of loanwords in Old Irish. To appear in: Guus Kroonen (ed.), Sub-Indo-European Europe. De Gruyter Mouton. Stika, Hans-Peter & Heiss, Andreas G. (2013). Bronzezeitliche Landwirtschaft in Europa — Der Versuch einer Gesamtdarstellung des Forschungsstandes. In: Karl-Heinz Willroth (ed.), Siedlungen der älteren Bronzezeit. Neumünster: Wachholtz, 183–216. Stockfleth, Nils Vibe (1852). Norsk-lappisk Ordbog. Christiania: Cappelen. Stokes, Whitley (1894). Urkeltischer Sprachschatz (= Vergleichendes Wörterbuch der indogermanischen Sprachen [4ᵗʰ edition], Vol. 2). Göttingen: Vandenhoeck & Ruprecht. Strömberg, Reinhold (1944). Griechische Wortstudien. Göteborg: Elander. Suhonen, Seppo (1973). Die jungen lettischen Lehnwörter im Livischen. Helsinki: Société Finno-Ougrienne. Suhonen, Seppo (1988). Die baltischen Lehnwörter der finnisch-ugrischen Sprachen. In: Denis Sinor (ed.), The Uralic Languages: Description, History and Foreign Influences. Leiden–New York–Copenhagen: Brill, 596–615. Suolahti, Hugo (1899). Die althochdeutschen Tiernamen. Darmstadt: Otto. Suolahti, Hugo (1909). Die deutschen Vogelnamen. Straßburg: Trübner. Szemerényi, Oswald (1967). Славянская этимология на индоевропейском фоне. Вопросы языкознания 1967/4, 3–24. Takács, Gábor (2001). Etymological Dictionary of Etyptian ii. b-, p-, f-. Leiden–Boston– Köln: Brill. Tamai, Tatsushi (2020). Tocharian Mss. of Weber and Macartney [Unpublished]. Accessed online at https://academia.edu/43390019. Tamm, Fredrik (1881). Svenska ord belysta genom slaviska och baltiska språken. Upsala Universitets Årsskrift 1881. Терентьев, Владимир А. (1990). Corrections to the “Suomen kielen etymologinen sanakirja” concerning Germanic, Baltic and Slavic loanwords. In: Uralo-Indogermanica: Балто-славянские языки и проблема урало-индоевропейских связей ii. Москва: Академия наук СССР, 30–32. Thomsen, Vilhelm (1870). Über den Einfluss der germanischen Sprachen auf der Finnisch-Lappischen. Halle: Buchhandlung des Waisenhauses. Thomsen, Vilhelm (1877). The Relations between Ancient Russia and Scandinavia and the Origin of the Russian State. Oxford: Parker. Thomsen, Vilhelm (1890). Berøringer mellem de finske og de baltiske (litauisk-lettiske) Sprog. København: Blanco Lunos. Thomson, Alexander I. (1927). Beiträge zur Geschichte des Diphthongs ы in den slavischen Sprachen. Zeitschrift für Slavische Philologie 4, 342–356. Thörnqvist, Clara (1948). Studien über die nordischen Lehnwörter im Russischen. Uppsala: Almqvist & Wiksell. Thorsø, Rasmus (forthc.). Alternation of diphthong and monophthong in Armenian Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 372 bibliography words of substrate origin. To appear in: Guus Kroonen (ed.), Sub-Indo-European Europe. De Gruyter Mouton. Thorsø, Rasmus, Andrew Wigman, Anthony Jakob et al. (2023). Word mining: metal names and the Indo-European dispersal. In: Kristian Kristiansen, Guus Kroonen & Eske Willerslev (eds.), The Indo-European Puzzle Revisited. Cambridge: University Press, 105–126. Threatte, Leslie (1980). The Grammar of Attic Inscriptions i. Phonology. Berlin–New York: De Gruyter. Thurneysen, Rudolf (1946). A Grammar of Old Irish [Revised and Enlarged Edition]. Dublin: Institute for Advanced Studies. Toivonen, Yrjö H. (1917). Huomioita lainasanatutkimuksemme alalta. SuomalaisUgrilaisen Seuran Aikakauskirja 34, 1–47. Toivonen, Yrjö H. (1928). Über alter und entwicklung des ackerbaus bei den finnischugrischen völkern. Mémoires de la Société Finno-ougrienne 58, 229–240. Толстой, Никита И. (1969). Славянская географическая терминология. Москва: Наука. Tomaschek, Wilhelm (1883). Ethnologisch-linguistische Forschungen über den Osten Europa’s. Das Ausland 36, 701–706. Tomaschek, Wilhelm (1889). Kritik der ältesten Nachrichten über den skythischen Norden. Sitzungsberichte der philosophisch-historischen Classe der kaiserlichen Akademie der Wissenschaften 107, 1–70. Топоров, Владимир Н. (1972). О балтийском элементе в подмосковье. Baltistica 1 priedas, 185–224. Топоров, Владимир Н. (1988–1997). Балтийский элемент в гидронимии поочья. Балто-славянские исследования 1986, 154–177; 1987, 47–69; 1988–1996, 276–310. Топоров ПЯ = Топоров, Владимир Н. (1975–1990). Прусский язык. 5 vols. Москва: Наука. Топоров, Владимир Н. & Трубачев, Олег Н. (1962). Лингвистический анализ гидронимов верхнего поднепровья. Москва: Академия наук. Töppen, Max (1867). Einige Reste der altpreussischen Sprache nebst antiquarischen Bemerkungen. Altpreußische Monatsschrift 4, 136–156. Torp, Alf (1919). Nynorsk etymologisk ordbok. Kristiania: Aschehoug. Trautmann, Reinhold (1910). Die altpreussischen Sprachdenkmäler. Göttingen: Vandenhoeck & Ruprecht. Trautmann, Reinhold (1923). Baltisch-Slavisches Wörterbuch. Göttingen: Vandenhoeck & Ruprecht. Treimer, Karl (1954). Ethnogenese der Slawen. Wien: Gerold. Tremblay, Xavier (2004). Sur χαλκός, lit. geležìs et turc qoruyžin. Historische Sprachforschung 117, 238–248. Трубачев, Олег Н. (1960). Происхождение названий домашних животных в славянских языках. Москва: Академия наук СССР. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 bibliography 373 Трубачев, Олег Н. (1971). Заметки по этимологии и сравнительной грамматике. Этимология 1968, 24–67. Трубачев, Олег Н. (1994). Праславянское лексическое наследие и древнерусская лексика дописьменного периода. Этимология 1991–1993, 3–23. [Трубачев ЭССЯ — see ЭССЯ under Abbreviated Titles.] Тумашева, Диляра Г. (1992). Словарь диалектов сибирских татар. Казань: Издательство Казанского университета. Uhlenbeck, Christianus C. (1900). Kurzgefasstes etymologisches Wörterbuch der gotischen Sprache [2ⁿᵈ edition]. Amsterdam: Müller. Ulmann, Carl Christian (1872). Lettisches Wörterbuch. Erster Theil: Lettisch-deutsches Wörterbuch. Riga: Brutzer. Untermann, Jürgen (2000). Wörterbuch des Oskisch-Umbrischen. Heidelberg: Winter. Uotila, Eeva (1983). Baltofennica iii: Finnish viekas, veikeä, vaikku ~ Baltic *veika-. Euroasiatica 4/6. Napoli: Istituto Universitario Orientale, 5–12. Uotila, Eeva (1986). Baltic impetus on the Baltic Finnic diphthongs. Finnisch-Ugrische Forschungen 47, 206–222. Uotila, Eeva (1990). More Baltic loans from ancient everyday contexts: Fi. suhta and hinta. In: Uralo-Indogermanica: Балто-славянские языки и проблема уралоиндоевропейских связей ii. Москва: Академия наук СССР, 134–140. de Vaan, Michiel (1999). The PIE root structure *Te(R)Dʰ-. Historische Sprachforschung 112, 1–25. de Vaan, Michiel (2003). The Avestan Vowels. Amsterdam–New York: Rodopi. de Vaan, Michiel (2008). Etymological Dictionary of Latin and the other Italic Languages. Leiden–Boston: Brill. de Vaan, Michiel (2018). The phonology of Albanian. In: Jared Klein, Brian Joseph & Matthias Fritz (eds.), Handbook of Comparative and Historical Indo-European Linguistics. Berlin: De Gruyter, 1732–1749. Vaba, Lembit (1990a). Die baltischen Sonderentlehnungen in den ostseefinnischen Sprachen. In: Sari Vaula (ed.), Itämerensuomalaiset kielikontaktit. Helsinki: Valtion painatuskeskus, 125–139. Vaba, Lembit (1990b). Gibt es ein baltisches Fragment im estnischen Wortschatz der Waldimkerei? Linguistica Uralica 26/3, 173–179. Vaba, Lembit (1992). [Review of] A.D. Kylstra, Sirkka-Liisa Hahmo, Tette Hostra, Osmo Nikkilä, Lexikon der älteren germanischen Lehnwörter in den ostseefinnischen Sprachen, Bd. i. A–J. Linguistica Uralica 28, 219–223. Vaba, Lembit (1997). Ergänzungen zur Substitution des baltischen wortanlautenden *st. Linguistica Uralica 33/3, 177–180. Vaba, Lembit (1998). Die Rolle des altpreussischen Sprachmaterials in etymologischen Untersuchungen ostseefinnischer Baltismen. In: Wojciech Smoczyński (ed.), Colloquium Pruthenicum Secundum. Kraków: Universitas, 177–175, Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 374 bibliography Vaba, Lembit (2011). Balti laenude uurimine avab meie kauget minevikku. Keel ja Kirjandus 54, 734–763. Vaillant, André (1929). Deux emprunts slaves au roman de Dalmatie. Revue des études slaves 9, 270–272. Vaillant, André (1950). Grammaire comparée des langues slaves i. Phonétique. Lyon: IAC. Vaillant, André (1958). Grammaire comparée des langues slaves ii. Morphologie. Première Partie: Flexion nominale. Lyon: IAC. Vaillant, André (1974). Grammaire comparée des langues slaves iv. La formation des noms. Paris: Klincksieck. Vanagienė, Birutė (2014). Šiaurės vakarų žemaičių žodynas i. A–O. Lietuvių kalbos institutas. Варбот, Жанна Ж. (1984). Праславянская морфонология словообразование и этимология. Москва: Наука. Васильев, Василий Л. (2012). Славянские топонимические древности Новгородской земли. Москва: Языки славянской культуры. Васильев, Василий Л. (2015). Проблематика изучения гидронимии балтийского происхождения на територии России. Linguistica 55, 173–186. Васильев, Леонид Л. (1929). О значении каморы в некоторых древнерусских памятниках xvi–xvii веков. Ленинград: Академия наук. Vasmer, Max [= М. Р. Фасмеръ] (1909). Греко-славянскiе этюды iii. Греческiе заимствованiя въ русскомъ языкѣ. Санктпетербургъ: Академія наукъ. Vasmer, Max (1913). Kritisches und Antikritisches zur neueren slavischen Etymologie v. Rocznik slawistyczny 6, 172–214. Vasmer, Max (1922). Prie germanybių baltų kalbose. Švietimo darbas 1922/3–6, 272–274. Vasmer, Max (1932). Beiträge zur historischen Völkerkunde Osteuropas i. Die Ostgrenze der baltischen Stämme. Sitzungsberichte der Preussischen Akademie der Wissenschaften. Philosophisch-historische Klasse 1932, 637–666. Vasmer, Max (1938). [Review of] J. Mikkola, Die älteren Beziehungen zwischen Ostseefinnischen und Russisch. Zeitschrift für Slavische Philologie 15, 448–455. [Vasmer REW — see REW under Abbreviated Titles.] Вахрос, Игорь С. (1963). К истории бани в свете ее наименований в славянских языках. Scando-Slavica 9, 153–168. Vennemann, Theo (1995). Etymologische Beziehungen im Alten Europa. Der GinkgoBaum: Germanistisches Jahrbuch für Nordeuropa 13, 39–115. Vennemann, Theo (1998). Germania Semitica. Biene und Imme. Sprachwissenschaft 23, 471–487. Vennemann, Theo (2003). Zur Frage der vorindogermanischen Substrate in Mittel- und Westeuropa. In: Theo Vennemann. Europa Vasconica — Europa Semitica. Berlin– New York: Mouton de Gruyter. Vennemann, Theo (2011). English as a Contact Language: Typology and Comparison. Anglia 129, 217–257. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 bibliography 375 Вершинин = Вершинин, Валерий И. (2004–2011). Этимологический словарь мордовских (эрзянского и мокшанского) языков. 5 vols. Йошкар-Ола. Viitso, Tiit-Rein (1978). The history of Finnic õ in the first syllable. Советское финноугроведение 14, 86–106. Viitso, Tiit-Rein (1981). Finnic Gradation: types and genesis. Советское финноугроведение 17/3, 176–185. Viitso, Tiit-Rein (1990). On the earliest Finnic and Balto-Slavic contacts. In: Sari Vaula (ed.), Itämerensuomalaiset kielikontaktit. Helsinki: Valtion Painatuskeskus, 140–147. Viitso, Tiit-Rein (1992). Finnic and its prehistoric Indo-European neighbors. In: László Honti et al. (eds.), Finnisch-ugrische Sprachen zwischen dem germanischen und dem slavischen Sprachraum. Amsterdam–Atlanta: Rodopi, 185–196. Viitso, Tiit-Rein (1998). Possible prehistoric contacts of Livonian. Linguistica Uralica 34, 8–22. Viitso, Tiit-Rein & Ernštreits, Valts (2012). Līvõkīel-ēstikīel-lețkīel sõnārōntõz = LivonianEstonian-Latvian Dictionary. Tartu–Rīga: Tartu Ülikool. Villanueva Svensson, Miguel (2008). [Reivew of] Wojciech Smoczyński, Laringalų teorija ir lietuvių kalba. Baltistica 43, 115–123. Villanueva Svensson, Miguel (2009). Indo-European *sk̑ in Balto-Slavic. Baltistica 44, 5–24. Villanueva Svensson, Miguel (2010). Baltic sta-presents and the Indo-European desiderative. Indogermanische Forschungen 115, 204–233. Villanueva Svensson, Miguel (2011). Indo-European long vowels in Balto-Slavic. Baltistica 46, 5–38. Villanueva Svensson, Miguel (2014). On the relationship between West Baltic and East Baltic. In: Tatjana Civjan, Marija Zavjalova & Artūras Judžentis (eds.), Baltai ir slavai: Dvasinių kultūrų sankirtos. Vilnius: Versmė, 162–176. Villanueva Svensson, Miguel (2015). The development of *ōu in Baltic. Indogermanische Forschungen 120/1, 313–328. Vine, Brent (1999). Greek ῥίζα ‘root’ and “Schwa secundum”. In: Vyacheslav V. Ivanov & Brent Vine (eds.), UCLA Indo-European Studies i, 5–30. Vine, Brent (2011). Initial *mo- in Latin and Italic. Münchener Studien zur Sprachwissenschaft 65, 261–286. Viredaz, Rémy (2020). Notes d’étymologie slave. In: Harald Bichlmeier, Ondřej Šefčík & Roman Sukač (eds.), Etymologus. Festschrift for Václav Blažek. Hamburg: Baar, 403– 422. Vosylytė, Klementina (2013). Kupiškėnų žodynas iv. Š–Ž. Vilnius: Lietuvių kalbos institutas. Vovin, Alexander (2005). The End of the Altaic Controversy. Central Asiatic Journal 49, 71–132. de Vries, Jan W. (1962). Altnordisches etymologisches Wörterbuch. Leiden: Brill. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 376 bibliography de Vries, Jan W. (1971). Nederlands Etymologisch Woordenboek. Leiden: Brill. Vycichl, Werner (1983). Dictionnaire étymologique de la langue copte. Leuven: Peeters. Vycichl, Werner (1990). La vocalisation de la langue égyptienne i. La phonétique. Cairo: Insitut Français d’archéologie orientale. Wagner, Heinrich (1981). Origins of Pagan Irish Religion. Zeitschrift für celtische Philologie 38, 1–28. Wälchli, Bernhard (1997). Mordwinisch und ältere baltische Wortsemantik. FinnischUgrische Forschungen 54, 305–324. Walde, Alois (1910). Lateinisches etymologisches Wörterbuch. Heidelberg: Winter. Walde/Hofmann = Johann B. Hofmann (1938–1954, ed.), Lateinisches etymologisches Wörterbuch von A. Walde [3ʳᵈ edition]. 2 vols. Heidelberg: Winter. Walde/Pokorny = Julius Pokorny (1927–1934, ed.), Alois Walde. Vergleichendes Wörterbuch der Indogermanischen Sprachen. 3 vols. Berlin–Leipzig: De Gruyter. Weber, Dieter (1997). Osset. bælon, D bælæu ‘Taube’ und lit. balañdis. Res Balticae 1997, 119–128. Weiss, Michael (1994). Life Everlasting: Latin iūgis “everflowing”, Greek ὑγιής “healthy”, Gothic ajukdūþs “eternity” and Avestan yauuaēǰī- “living forever”. Münchener Studien zur Sprachwissenschaft 55, 131–156. Weiss, Michael (2020). Outline of the Historical and Comparative Grammar of Latin [2ⁿᵈ edition]. Ann Arbor: Beech Stave Press. Welcome, Ashton K. & van Wyk, Ben-Erik (2019). Spatial patterns, availability and cultural preferences for edible plants in southern Africa. Journal of Biogeography 47, 584–599. Wexler, Paul (1977). A Historical Phonology of the Belorussian Language. Heidelberg: Winter. Wichmann, Yrjö (1902). Samojedisches lehngut im syrjänischen. Finnisch-Ugrische Forschungen 2, 165–183. Wichmann, Yrjö (1906). Zur geschichte der finnisch-ugrischen anlautenden s- und člaute im tscheremissischen. Finnisch-Ugrische Forschungen 6, 17–39. Wichmann, Yrjö (1911). Zur geschichte der finnisch-ugrischen anlautenden affrikaten nebst einem exkurs über die finnisch-ugrishen anlautenden klusile. FinnischUgrische Forschungen 11, 173–290. Wiik, Kalevi (1992). Suomen kielen ja murteiden syntykysymyksiä. In: Hannu Remes (ed.), Lähivertailuja 6. Joensuu: Joensuun yliopisto, 85–110. Wiklund, Karl B. (1896). Entwurf einer urlappischen Lautlehre i. Einleitung, Quantitätsgesetze, Accent, Geschichte der hauptbetonten Vokale. Helsinki: Suomalais-Ugrilainen Seura. Wilkin, Shevan et al. (2021). Dairying enabled Early Bronze Age Yamnaya steppe expansions. Nature 598, 629–633. Winkler, Eberhard & Pajusalu, Karl (2009). Salis-livisches Wörterbuch. Tallinn: Teaduste Akadeemia. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 bibliography 377 Winter, Werner (1950). Studien zum „Prothetischen Vokal“ im Griechischen. Hamburg: Hansischer Gildenverlag. Winter, Werner (1962). Die Vertretung indogermanischer Dentale im Tocharischen. Indogermanische Forschungen 67, 16–35. Winter, Werner (1978). The distribution of short and long vowels in stems of the type Lith. ėś ti : vèsti : mèsti and OCS jasti : vesti : mesti in Baltic and Slavic languages. In: Jacek Fisiak (ed.), Recent developments in historical phonology. The Hague: Mouton, 431–446. Winter, Werner (1982). Indo-European words for ‘tongue’ and ‘fish’: A reappraisal. Journal of Indo-European Studies 10, 167–186. Witczak, Krzysztof T. (1991). Armenian opʿi ‘white poplar, Populus alba L.’ and the development of PIE *ps in Armenian. Annual of Armenian Linguistics 12, 65–75. Witczak, Krzysztof T. (1996). The Pre-Germanic Substrata and Germanic Maritime Vocabulary. In: Karlene Jones-Bley & Martin E. Huld (eds.), The Indo-Europeanization of Northern Europe. Washington, DC: Institute for the Study of Man, 166–180. Witczak, Krzysztof T. (1997). Millet in Lithuanian and other Indo-European languages. Baltistica 32, 25–39. Witczak, Krzysztof T. (2001). Lith. atúodogiai, atúoriečiai ‘summer rye’, atólas (m.) ‘aftermath, after-grass’ and the Indo-European name for ‘grass’. Baltistica 36, 43–47. Witczak, Krzysztof T. (2003). Indoeuropejskie nazwy zbóż. Łódź: Uniwersytet. Witczak, Krzysztof T. (2020). Are There Traces of a Finno-Ugric Substratum in ProtoSlavic? Slavistična revija 68/1, 73–89. Witsen, Nicolaas (1785). Noord en Oost Tartarye [3ʳᵈ edition]. Amsterdam. Worrel, William H. (1934). Coptic Sounds. Ann Arbor: University of Michigan. Wrede, Ferdinand (1891). Über die Sprache der Ostgoten in Italien. Strassburg: Trübner. [Хелимский — see Helimski] Yakubovich, Ilya (2009). Two Armenian Etymologies. Caucasian and Near Eastern Studies 13, 266–272. Young, Steven (1998). Old Prussian bovids revisited. In: Wojciech Smoczyński (ed.), Colloquium Pruthenicum secundum. Kraków: Universitas, 201–216. Young, Steven (2009). Tone in Latvian borrowings from Old Russian. In: Tijmen Pronk & Rick Derksen (eds.), Stressing the past: Papers on Baltic and Slavic accentology. Amsterdam–New York: Rodopi, 177–192. Zair, Nicholas (2012). The reflexes of the Proto-Indo-European laryngeals in Celtic. Leiden: Brill. Зайцева, Мария И. & Муллонен, Мария И. (1972). Словарь вепсского языка. Ленинград: Наука. Зайцева, Нина Г. (2008). Отражение вепсско-северно-русских контактов в вепсском лексическом тезаурусе. Ученые записки петрозаводского государственного университета 2008/4, 79–83. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 378 bibliography Зайцева, Нина Г. (2010). Новый вепсско-русский словарь. Петрозаводск: Периодика. Зализняк, Андрей А. (1985). От праславянской акцентуации к русской. Москва: Наука. Зализняк, Андрей А. (2004). Древненовгородский диалект [2ⁿᵈ edition]. Москва: Языки славянской культуры. Зализняк, Андрей А. (2019). Древнерусское ударение: Общие сведения и словарь [2ⁿᵈ edition]. Москва: Языки славянских культур. Zanina, Oksana G. et al. (2021). Plant food in the diet of the Early Iron Age pastoralists of Altai. Journal of Archaeological Science: Reports 35. Желеховский = Желеховский, Євгений (1886). Малоруско-нїмецкий словар. 2 vols. Львів: Шевченко. Zeps, Valdis J. (1962). Latvian and Finnic Linguistic Convergences. Bloomington: Indiana University. Zinkevičius, Zigmas (1966). Lietuvių dialektologija. Vilnius: Mintis. Zinkevičius, Zigmas (1980–1981). Lietuvių kalbos istorinė gramatika i–ii. Vilnius: Mokslas. Zinkevičius, Zigmas (1984). Lietuvių kalbos istorija i. Lietuvių kalbos kilmė. Vilnius: Mokslas. Zinkevičius, Zigmas (1985). Lenkų–jotvingių žodynėlis? Baltistica 21/1, 61–82. Zinkevičius, Zigmas (1987). Lietuvių kalbos istorija ii. Iki pirmųjų raštų. Vilnius: Mokslas. Живлов, Михаил А. (2008). Этимология прапермского *våž ‘старый’. In: Языковые контакты в аспекте истории. Москва. Живлов, Михаил А. [= Zhivlov, Mikhail] (2014). Studies in Uralic vocalism iii. Journal of Language Relationship 12, 113–148. Живлов, Михаил А. (2015). Неиндоевропейский субстрат в финно-волжских языках [Conference handout]. X Чтения памяти С.А. Старостина (27 March 2015). Живлов, Михаил А. [= Zhivlov, Mikhail] (2023). Reconstruction of Proto-Uralic. In: Daniel Abondolo & Riitta-Liisa Valijärvi (eds.), The Uralic Languages [2ⁿᵈ edition]. London–New York: Routledge, 117–175. Zohary, Daniel, Maria Hopf & Ehud Weiss (2012). Domestication of Plants in the Old World [4ᵗʰ edition]. Oxford: University Press. Жолобов, Олег Ф. (2007). Корпус древнерусских списков Паренесиса Ефрема Сирина. i: РГАДА, Син. 38. Russian Linguistics 31, 31–59. Zubatý, Josef (1892). Zum baltischen ů. Beiträge zur Kunde der indogermanischen Sprachen 18, 241–266. Zubatý, Josef (1894). Slavische Etymologien. Archiv für slavische Philologie 16, 385–425. Zubatý, Josef (1898). [Review of] Recherches sur l’emploi du génitif-accusatif en VieuxSlave, par A. Meillet. Archiv für slavische Philologie 20, 392–397. Žulys, Vladas (1966). Keleto retų žodžių istorija. Baltistica 1, 151–161. Zupitza, Ernst (1896). Die germanischen Gutturale. Berlin: Weidmann. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 bibliography 379 Zupitza, Ernst (1900). Über doppelkonsonanz im Irischen. Zeitschrift für vergleichende Sprachforschung 36, 202–245. Zupitza, Ernst (1904). Miscellen. Zeitschrift für vergleichende Sprachforschung 37, 387– 406. Журавлев, Анатолий Ф. (1980). К этимологии слав. *vorb- ‘птица Passer, воробей’. Этимология 1978, 52–58. Zvelebil, Marek & Dolukhanov, Paul (1991). The Transition to Farming in Eastern and Northern Europe. Journal of World Prehistory 5/3, 233–278. Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 Word Index Baltic Languages Lithuanian aguonà 41, 296 ái 175n5 aĩrė, áirė 90 akėč́ ios 50, 273 akėt́ i 273 ãknos 175n5 ãkstinas 105n160 alìksnis 12, 33, 276 álksna 33 al̃ksnis 12, 276 alkū́nė 272 alùs 37, 41, 49 álvas 301 angìs 93n125 ankstì 83n108 añkštas 52, 83n108, 182n24 antstóti 83n108 apstùs, apstóti 83n108 ãpušė 278 árti 293 arumbė ̃ 175 ąsà 57, 124 asiaĩ 90 ãsilas 10, 11 ãšaka 240, 241 ãšutas 34, 238n10 ašva 55 ateĩvis, atéiva 51 atólas 104, 232 atstùs, atstóti 83n108 atvérti 59n47 áugti 170, 182n24 áuksas 12, 301 áukštas 182n24 aulỹs 279 ausìs 275n92 avilỹs 279 ãvinas 56 avìs 284n3 ãvižos 212, 239, 295 bagotỹrius 19n33 balánda 177, 180, 231, 304 balañdis 186, 209, 304 bãlas 209n84 bal̃nas 257n52 bálti 209 bam̃ bti 165 barzdà 224 baslỹs 57 baũbas 29 bãžmas 131n29 bažnìnčia, bažnýčia 184 Bebrunga 179 ber̃gždžias 225 béržas 169 bèsti 57 bíesas 27 bylóti 274n88 bìrkavas, bìrkuva 13, 19, 26, 27 birzdžiai 216, 223, 229n120 bir̃žė 208, 213, 224 bìržis 169 bìtė, bitìs 227 blusà 184n29 botãgas 165 božmas 131n29 briaunà 58, 70 bruĩšė 12 brunìs, brùnė 58n45 bruñšė 12 bruvìs 58n45 bruzdùs, bruzgùs 222, 222n111 bùbas 29 buĩlis 20 bulvõnas 9 buožmas See božmas bùrė 90 bur̃konas, burkūnas 229n121 burkuntai 229 burnà 165 burzdùs, burzgùs 222n111 bùsilas 11 bùžmas 130, 131 cìrkva 15n20 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 381 word index čérpė 17, 27 čerpyčia, čerepyčia 16, 16n22 dabà 60 dagỹs 54 dal̃gis, dal̃gė 191 dalià 120 daubà 210, 210n89 dègti 34, 54 degùtas 34, 238n10 dervà 49 dienì Žem. 68 dieverìs 275n92 dirvà, dir̃vinis 31 dóbilas 100 drãgės 190n40 drėb̃ ti 126n21, 270n79 dùbti, dubùs 210n89, 211 duktė ̃ 52, 82, 139 dū̃lė 13n14 dū́lis 58 dum̃ blas 210, 210n89, 221n106 dumbrà 210n89 dumlas 210 dùmti 182 dúona 285n4, 293 durà Žem. 60n57 dùrys 275n92 dùrti 60 egžlỹs 275n94 eĩti 40, 51 eldijà 80n105 ẽpušė 278 erẽlis 16, 244 erškẽtas 218, 236 erškėt̃ is, eršketỹs 218 eršketris 218n99 ešerỹs 73, 218n101 ešketras 218 ešva 55, 55n32 ežgė,̃ ežegỹs 275, 275n94 ežỹs 275 gaižùs 48 galvà 196 gar̃das 134 garnỹs 137n39 garšvà, gáršva 248, 277 gãtvė 37 gauraĩ 55, 251n38 gegužė,̃ gegė ̃ 120 geinỹs 299 geležìs 274, 275, 275n92 geltà, gel̃tas 53n22 gélti 55 geltónas 52 genỹs 299 gẽras 132 giẽžti 48 gylỹs 55 giminė ̃ 132 gìrti 132 glìnda 183, 184 glúosnis 194n49 gnýda 184 gniutìs 48 gomurỹs 267 grandìs 121 grėb́ ti, grėb̃ ti 269, 269n78, 270n78 grę̃žti 249 griẽkas 23, 23n41 grìkai, grìkiai 15 gróbti 269n78 grùmada, grumodas 9n6 grúodas 54, 128 grūšià 13n14 gùdas 42 gū́das 43n11 gùdė, gùdinti 43n11 gùdobelė 42, 43n11 gudrùs, gùdrinti 43n11 gul̃bė 189, 195, 206 gulbìs 189 gum̃ bas 48 guñtas 9 gúoba 194n49, 194n50 gùsti 42, 43n11 gvõltas 7 įgùsti See gùsti ýla 38, 39n4 ìlgas 169 -ininkas, -inykas 184 iñkaras 18n28 ìnkstas 181 ìrbė, irbenis 175n3, 180n17 iškūlýti 108n168 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 382 ìšperša, išperšėt́ i 208 įžymùs, įžymėt́ i 130 jaũ 58 jaukìnti 70n84 jáura 86 jáuti 169n25 javaĩ 32, 293 jéi 33n55, 175n5 jẽknos 175n5 jérbė 174n3 Jer̃biškiai 174n3 jerubė ̃ 33n55, 174, 174n3, 241, 259, 304 jerū̃bė 175n4 Jerùbiškiai 174n3 jerumbė ̃ 174, 180n17 jėŕ umbė 175 jõvaras 274n91 judėt́ i 121 jùdros 100, 295 jùnkti 181 jū́ra 87 jū̃s, jū́sų 2n5 jū́šė Žem. 12 kadagỹs 84 káimas, kaimýnas 40, 40n6, 50 káivinti See nukáivinti kãklas 122 kalavìjas 229n120 kalmaškà 17 kálpas 17, 26 kam̃ pas 282 kanãpės 207, 296 kãpas 76 karavõjus 16 kar̃bas 17, 27 kar̃bija, karbijà 18 kãrias 80 kárti 130 kárvė 55n30, 222n109 karvõjus 16 ką́ sti 213n92 kasulà, kasulas 280 kãtilas 10, 42 káušas 30 kazilaĩ 11 kẽlias 80 kel̃tvartis 246n24 word index kẽmeras, kemerai 245n22 keñkras 35n61 kẽpalas 38 kepùrė 122 ker̃džius 41 kermenaĩ 63 kermušė,̃ kérmušas 246n24 kermušis 246, 246n24, 247n28 keterà 46 kẽvalas 169n25 kiaũlė 21n35 kiaũltvartis 246n24 kiáunė 256n50 kiáurai, kiáurmušis 246n24 kiẽmas 40, 50 kiemerai 245n22 kirmìs 55, 203 kir̃sti 57n38 kir̃vis 35n61, 57 kìšti 98 klampà 48 klėt́ is 29n50 klẽvas 260 klíenas 66n68 kliẽpas 37, 38, 40, 65 klóti 29n50 kningà, knygà 184 kolidòras, korìdorius 274n90 korỹs 133, 155, 248 kortà 7 krañtas, kránta 50n13 kraũjas 80 kràusêt 60 kraušýti 60 kréisva 137 kreĩvas 137 krèkti 261 krekùčiai, *krekulai 261 kriáušė 13, 13n14, 197 krikščiónis 12 krìkštas 11, 12 krìkštyti 12n13 krùšti 60 kùbilas 11 kùila 20, 21, 21n35, 26 kuilỹs 20n35, 79n103 kū́lymas 108 kulkšnìs 257 kul̃nas, kulnìs 257n52 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 383 word index kùlšis, kùlšė 257 kulšnis 257 kūlýti See iškūlýti kumẽlė 55n32 kumelỹs 79n103 kùmetis 11, 28 kuodẽlis 22 kur̃bas 9 kur̃čias, kur̃tas 53 kur̃kti 261 kurkulaĩ 261 kùrpė 259 Kur̃šas 29 kùrtas 8, 10n8, 26 kur̃vis 9 kvãpas 126 kvėp̃ ti 126 kviečiaĩ 37, 38, 40, 194, 294, 295 kviẽsti 120 láibas 107n164 laišis 202n66 láiškas 265 laĩvas 90 lãkštas 265 lankà 131 lañktis 121 lãpas 76n95, 101n147 lasaša, lašaša 259 lãšis ‘lentil’ 202n66 lãšis ‘salmon’, lãšė, lašìs 98, 258, 258n59, 259 lašišà 98, 258 laũkas 76, 250n35 lazdà 266, 266n72 leĩšis 202n66 lẽmežis 213, 214 léngė 35n61 lénkas 19, 26 lénkė 35n61, 198 leñkti 121, 131 lestė 175n3 lę̃šis 202, 296 liaukà 70n83, 250 liáutis 102n150 liẽkas 58, 68, 68n77, 119, 136 líeknas 198 líepa 2, 89, 102n153, 135, 171 líesas 107 Lietuvà 29, 155 lýgmalis 122 lìkti 136 linaĩ 295 lýnas 93n126 lìnzė 202 lìpti 89 liū̃gas, liū́gas 198n61 liū́nas 198n61 lomà 130n26 lópeta 101 lopšỹs, lopišỹs 259 lóva 101 luišỹs 181n21 lùnkas 125, 135, 171, 181 lū́šis 180, 181n20, 271 mãrios 284n2 martì 51, 82 mãtas 281n101 maũmas 29 mẽdiga 101n148 mẽdis 80n104, 85 mẽdžiaga 85, 101n148 mẽdžias 54, 80, 80n104 meñtė 57 mẽs 2n5 midùs, mìdus 37, 42 mieras 24n42 miẽšti 170 miẽžiai 295 minià 196 mìnti 48, 196 morkvà, morkà 229 muĩlas 20, 26 muĩtas 20 mū̃mas 29 munkà, mūkà 175n4 mūsaĩ 273 mùsos, mùsas, musóti, musojaĩ 273, 273n87 mūsóti 274n88 mū́sų 2, 2n5 mùšti 246n24 nãmas 300 naũjas 80n104 néndrė 240, 272 nepuotis 52 nýtys, nýtis 57 nósis 281n101 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 384 notrė,̃ noterė ̃ 203, 240n17, 273 notūrà 7 nukáivinti 48 núogas 270n81 núoperša 208 obelìs 42, 268, 269, 275n92 obuolỹs 268, 269 óda 49n9 opščius 226 opšrùs 221, 226 ožẽlis 62 ožỹs 56, 284n3 ožkà 83 pãkulos 29 paliovà 102n150 pal̃vė 120 pam̃ pti 165 pántis 135 papar̃tis, papártis 240n17 par̃šas 73, 78, 284n3 pãslas 11n12 patì 51n17, 82 paũtas 130 peĩlis 135 pẽkus 37, 41, 231n127 pelenaĩ 16 pélkė 257n50, 284n2 pẽlūs 56, 293 penìmis 120n2 perkū́nas 43n12, 136 peršėt́ i 208 per̃ti 30 pielà 135 piemuõ 56, 70n81, 82n107 pilìs 300 pìnti 135 pipìras 14, 14n17 pyrãgas 8 pirmuonìs 82 pirtìs 30, 30n51, 76, 140 pó- 232 pòlka 274n90 Póvilas 14, 26, 28 pragà 60n53 praparšas 60n53, 208n82 praperšà 208 prėś kas 105n159, 271 word index prijùnkti See jùnkti pró- 232 próga, progas 60n53 próperša 60n53, 208 pū́das, pū̃das 19, 27 pul̃kas 8, 25, 28n48 pùndas 18, 19 púodas 270n81 puõkas (Žem. pọ̃ uks) 23 pupà 228, 295 pūraĩ 294, 295 pùsti 60n52 pušìs 278 putmuõ 132 pùtos 60n52 putrà, pùtelis 60 rabėti 241n18 rãgas 104 rãgės 104 raškà, rãškės, raškýti 131 rãtas, rãtai 57 raũdas 250n35 regėt́ i 132 rėk̃ šti 131 rẽplės 270n78 rėṕ ti, rėp̃ ti 270n78 ridìkas 14 riẽkti, riekė ̃ 66n68 riešas 238n10 ríešutas 34, 171, 238, 238n10 ríetas 66, 122, 127n23 rìšti 121 rópė 102n151, 207, 237, 266, 296 rubėti 241n18 rùdas 131 rudìkas 14 rugiaĩ 44, 191, 295 rùimas 21n37 rúonis 112, 266 saĩkas 40 salà 123, 123n13, 123n14 salavà 123n12 sãmanos 110 sárgas 220 selėt́ i 220 sėmuõ 56, 61, 293 sesuõ 52, 82, 269 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 385 word index sidãbras 221, 225, 245 síena 66, 220, 300n23 sỹkas 94, 95 sil̃kė 95 sývas 46n3 skárda 90n121 skavardà 17 sker̃džius 41 sketerà 46 skiẽtas 68n77 skìltis 64n65 skir̃pstas 195n51 sklandùs 48 skrõblas, skrúoblas 194 slãstai 107 slíekas 171 smirdas 14 smõgti 165 smuĩkas, smùikas 20 sniẽgas 2 sopãgas 8 sóros 136, 261, 295 spenỹs 262 spiñgis 130 spragà 60 sprógti 60 stãbas 280 stãgaras 46 stámbas 60n55, 171n31 stambùs 46 statýti 281n100 statùs 281, 281n100 stel̃bti 211n91 stíebas 60, 66, 74, 171n31 stìklas 10, 11n11, 42 stìrna 221 strãzdas 49, 194n49, 204 stul̃bti 211 stul̃pas 211 stum̃ bras 220, 242, 257 suitus 21, 21n36 sūnùs 3 súodžiai, súodys 136 súolas 220 svẽčias, svetỹs 80, 80n104 svíetas 27 šakà 53 šálmas 17, 26 šalnà 54, 123 šal̃nis 54n26 šálti 54 šãmas 93n126 šarmà 59, 127 šárvas 13, 37, 42, 44 šeimà 40, 51 šelmuõ 128 šerdìs 121 šer̃kšnas 59n50 šermùkšnis 247 šermuõ, šermuonėl̃ is 142, 143 šẽškas 85, 143 šiẽnas 65, 126 šikšnà 87, 135 šìlas 229n120 šil̃kas 14, 17, 26 šil̃tis, šil̃tas 54 šim̃ tas 106n162, 130 šìrmas 55n30, 59, 143 širmuonėl̃ is 142 šìršė, šir̃šilas 55, 55n29 širšuonis 82 šìrvas 55n30 šiùilė 21n37 šývas 46, 55n30 šỹvas, šývis 46n3 šlãpias 80n104 šliùižė 21n37 šlúota, šlúoti 59, 108 šmagóti 165 šókti 100 šónas 46 šùkos 127 šuõ 3, 82n107, 128 šùsti 76n95 šuvà 128 švìnas 301 takišỹs, tãkišas 97, 97n136, 259 tánkus 53, 53n24, 119, 171n31 tašýti 49n8, 126n21 táu 2 tauras 58, 250 taũras 250n36 taurė ̃ 58 tautà 249, 250 tavè, tãvi 2 tekėt́ i 97 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 386 tẽlias 80 tem̃ pti 58 tetervà 146 tę́vas 260 tìlės 57, 116 til̃pti 9n5 tìltas 57, 119 tỹnas 19n33 tiñklas 57 tìrdinti, tìrti 131 tolì, toliẽ, tõl 133 tópelis 274 tòpolis, tòpalas 274n90 tóšis 48 trãnas 200, 201 triáušėti 201 triaũšiai 201 triùšis 201, 224 trobà 248, 269, 270 trùšis 201 tū́bas 39, 42 tuĩnas 19n33, 21, 21n37, 25 tū́kstantis 105, 135, 222 tùlkas 8, 10n8, 27 tulkočius 9n6 tùlpinti 9n5 túopa 274 tur̃gus 8, 10, 26 tur̃tas 132 tùščias 53 ùblas 231 ugnìs 3, 161 ulangėlė 179 ulbėt́ i 169n26 *ungnis 160 Unguras* 19n30 ungurỹs 54, 79, 79n103, 83, 93 uodegà, úodas 125 úoksas 12, 47 uolà 130 úolektis 272, 273 úosis 194n49, 271–273 uõstai (Žem. ọ̃ ustâ) 23 úosvauti 47n6 ùrupė 101n145 vãgis 50 vãkaras 245 vanduõ 82n107 word index vánta 33n57 vapsvà, vãpsas 55, 278n96 vãrias 80 várža 213 vãsara 245 vãškas 106n163, 217 vãžis, važỹs 204n73 vẽčas Žem. 259 vedegà, vedigà 101n148, 125 vėgėlė ̃ 99 vėl̃ , vėl̃ ei, vėlùs 61 veñgras 19n30 vénteris 33, 33n57 ver̃pti 61 vérti 59 vėr̃ ūbė 175n4 ver̃žti 213, 216 vešėt́ i 170n29 vetušas 145, 259 vežėč́ ios 273 vėžỹs 99 vèžti 204n73 víenas 2 vierumbė ̃ 175 viẽšnė (Žem. vẹ̃išnẹ) 23 vìlbinti, vilbėt́ i 169n26 vìlkė, vil̃kas 79 vìlna 56 vìngis 121 vìnkšna 182, 182n24 výras 3 vìrbė 174n3 vìrdis 143 vìržis 215, 216, 223 vyšnė ̃ ‘cross beam’ 182n24 vyšnià, Žem. vỹšnė 23 Výtautas 30 výti 169n25 vìzgė 239 výža, vyžà 183 vókas 102 volungė ̃ 178, 179, 195, 257, 304 vorà 59 voras 101 vorùšis, vórupė 101n145 voverìs 275n92 žãgas 53n25 žáidas 40, 59, 70n81 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 387 word index žãlias 52, 79, 80n104 žam̃ bas 52 žãras 59 žargà 64n66 žąsìs 56, 138 žẽmė 129 žėrėt́ i 59, 128 žer̃gti 59 žiáunos 250 žydėt́ i 59 žíedas 40, 59 žiẽsti 300n22 žiótis 267n74 žìrgas 59 žìrnis 56, 79, 293, 295 žukmistras 13, 13n15 žuvìs 93, 275n92 žvãkė 167 Latvian Note: ⟨uo⟩ is alphabetized under o âbele 268 âbuõliņš 100n141 âbuõls ‘apple’ 100, 100n141, 268 âbuõls ‘clover’ See dâbuõls âda 49n9 aĩris, aĩre 90 akstins 105n160 àlksna 33 àlksnis 276 alus 37, 49 al̂va, al̂vs 301 ãmulis, amuols 100 apse 219, 278 âpsis, âpša 226 ar̂t 293 aši, ašķi 90 atãlêtiês 232 atãls 104, 232 atvirzīt See vir̃zît aũlis 279 àuns 56 àuzas 212, 239, 295 avelis 279 âzis 56, 80n105 baluôda 177, 231 baluôdis 50, 209 bā̀rda 224n113 bā̀rzda 224 best 57 bir̃kavs 13, 26 bìrze 208, 224 bite 227 braũna 71 bruņas 10, 42 buca 15n20 bul̃vāns 9, 26 bulens 9 bur̃kãns 31, 229 burkan̂ ts 229n122 burtenis 9 buŗa 90 caûna 257n50 ceļš 80 cȩpure 122 cȩ̄r̀ maûksis 247n28 cḕrme 63 cìems 40, 50 cìlvȩ̃ks, cilȩ̄ks 14 cìmds 49n9 cìrmenis, cirmis 55 cìrst 57n38 cìrvis 35n61, 57 cũka 73n88 daba 60 dâbuõls 100 dadzis 54 dagla, daglis 54, 124 daļa 120 dar̂va 49 daũba 210n89 degt 34, 54 dȩguts 34 deĩbiņa 165 duõmât (HLv. dùmuôť) 22, 23 dradži 190 drêbt 126n21 dubļi, dubra 211n89 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 388 dubt 210n87 dubums 210n89 duburs 211n89 dukurs 15n20 dũlis 58, 108 dum̃ bêris 210n89 dum̃ bla, dum̃ brs 210, 210n89 dumûksnis 210, 221n106 dùre 60, 82, 108 dur̃t 60 dzȩguze 120 dzeinis 299 dzelezs 274 dzȩl̂ta 54 dzȩltãns, dzȩltains 52, 53n22, 54 dzȩlts 53n22 dzèlzs 274 dzenis 299 dzil̂na 50 dzîvuôt 132 ecêšas 50, 273 ecêt 273 ȩñkurs 18n28 ẽrcis 244 ḕrglis 16 gàizs² 48 gãju 60n53 gãmurs 267 gãrds 134 gãrsa 277 gate 37 gatuve 38 gatve, gatva 37 gauri 55 glīsis, glīse 38, 39n3 gnĩda 184n27 gņuta 48 gùovs 284n3 grebt 269 grȩ̀ks 23 griķi 15 grumada 9n6 guds 42 gùlbis 189 gum̃ ba 48 word index idra 100, 295 ielukši 174 ìerube² 174 iêt 40 ĩlȩns 38 il̃kss 174 ir̃be 174, 241 ir̃bene 174n3 jà 33n55 jàu 58 jũdît 121 judras 100 jũra 87 kadags, kadȩgs 84 kaĩmiņš 50 kakls 122 kal̃ps 16, 17, 26 kàlst 108n168 kal̃va 50 kaņepes 207 kàpuôsts 22 kaps 76 kā̀rba 17, 27 kāre 133, 248, 248n30 kārms (HLv. kùorms) 16, 17 karuõgs 22 kaŗš 80 katls 10, 11n12, 42 kàuns 40n7 kaza 83 klàips 37, 38n1, 40 kļava, kļavs 260 klẽts 29n50 kliêns², kliẽns 66n68 kraûsis² 197 kreĩlis 137 krèiss 137 krieviņš 43n12 krìevs 19n32, 24, 43n12 krišķāns 12 krusts, krists 13, 13n16 kubls 11n12 kuĩlis 20n35 kūkuļi (HLv. kìukûļi) 23n40 kũla 108 kul̃da 9 kulksnis 257 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 389 word index kulna 257n52 kulša 257 kumeļš 79n103 kùrkt 261 kur̂kuļi 261 kurls, kur̃ns 53 kur̃pe 259 Kur̂sa² 29 kur̃ts 8, 26 kvȩ̂pi, kvêpji 125, 126 kvêpt 126 kvìeši 37, 40, 295 kvitêt 40n8 ķeĩris 137 ķève 55n32 ķìlis² (HLv. kèiļs) 21n35 laĩbs 107n164 laiska 265 laĩva 90 laksti 265 lañka 131, 198 lasis 98, 258 làuks 76, 250n35 làupît 162 lâva 101 lazda 266 lȩ̃ca 202, 296 lemesis 213 lemešs 214 leste 175n3 lìcis 121 liẽkna 198 lìeks 58, 68, 68n77, 136 liẽpa 2, 89, 102n153, 135, 171 liẽss 107 Lìetava 29 lini 295 lĩnis 93n126 luba 162 lûks 125, 135, 171, 181 lũsis 180, 271 maguône 41, 253, 296 màize 285n4 mala 122, 135 mā̀rša 51 mȩdaga 85 mežs 54, 80 miêrs 24n42 mìeži 285n4, 295 milna 202 mĩt 48 muõkas 22 muca 15n20 muĩta 20 mũsu 2 naba 52, 119 nâtre 203, 273 naujš 80n104 -niẽks (HLv. -inìks) 24 nĩt, nìt², nĩtis 57 uôdze 93n125 uõla 130 uôlekts 272 osa ⟨ôssa⟩ 55 ùosa 57, 124 uôsis 271 pàlis 57n39 pàtaga 165 pàuts 130 Pãvils 14, 26 pȩ̀lni 16 pȩlus 56, 293 pȩ̄r̀ kuôns 136 pḕrt 30 pils 300 pipars 14n17 pìrãgs 8 pìrts 30, 76 plẽšas 105n159 Pliskava 14, 28 pluts 14 puõsms 34 prùods² (HLv. prùds) 23n40 puĩka 9 pùka² (HLv. pỳuka) 23n40 pùlks 8, 25 pulna 9 pupa 228, 295 pur̂ns 165 pûŗi 294, 294n16 putra 60 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 390 rads 132 rags 104 ragus, ragavas 104 rats, rati 57 redzêt 132 rȩgavas 105 riẽksts 34, 171, 238 riẽta 122 risamais 121n6 rist 121 ruônis 112, 266 rubenis 175n6, 241, 259 rubinât 241 ruds 13, 131 rudzi 191, 295 rutks 13n16, 14 sàime 40, 51 saka 53 sâkt 100 sala 123, 123n13 sal̂na 54, 123 sā̀ls 59, 137, 281n101 sā̀lis, sā̀lis² 59 sams 93n126 sāms 129n25 sãns 46 sāre 136, 261, 295 sar̂gs 220 sar̂ma 59, 127 ser̂de 121 sȩr̂ma 64n64 sȩ̄r̀ maûksis, sȩ̄r̀ mûkslis 247 sȩr̂mulis 142 sȩsks 85 sidrabs 225 sìeks 40 siêna 220 sìens 126 siẽrs (HLv. sìrs) 24 sĩga 95 siksna 87, 135 silce 95 sìlts 54 sil̃ķe ̧ 95 sir̃ms 55n30, 59 sirna 221 sìrpis 193 sir̂senis 55 word index siruobs 15n20 sìts (HLv. sèits) 19n33 skābardis 195 skãrds 90n121 skàusts 46 skavardnīca 17 slagzds 107n166 slaka 165 slapjš 80n104 slasts, slazds 107 sliêka 171 sluôta 59, 108 smirds 14 snìegs 2 sùodît 22 suods, suôdri² 136 suôls 220 suõma 22 spraga 60 sprâgt 60 stars 59n49 stebere 74 stìebrs 60 stiepats 74, 97, 165 stikls 10, 42 stipt 97 stir̃na 221 strazds 49, 204 stùlps 211 stum̃ brs 220, 257 sũbrs 221 sudrabs 225 sùitâk² 26, 27, 139 suits, suišs 21, 26, 27 suĩtums 26 suka, sukas 127 sumbrs 221 sûra² 136, 261 svešs 80 svins 301 šķaunats 97n135 šķìets 68n77 šļaka 165, 165n17 tacis 97, 259 tãss 48 tâst², tastît 48n8 tàure 58 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 391 word index tàuta 249 teĩba 165 teļš 80 tèst, têst 48n8, 126n21 teteris 146 tev, tevi 2 tìept 58 tiêvs 260 tìkls 57 til̃ts 57 timnica, timenîca 15n20 tìmsa² 15n20 tir̂dît 131 tìrgus 8, 10, 13n16, 26 tirt 131 tît 57n40 tràba² HLv. 270 traušļi 201 trusis 201 tũba 39 tũkstuôtis 105, 135, 222 tukšs 53 tul̃ks 8, 27 tul̃pîties 9n5 tûna, tũna 21n37 turba 9 vâcele 102 vadzis 50 vakars 245 vâks 102 vãluôdze 178, 195, 257 vapsene 55 vãrde 143 var̂za² ‘weir’, varzi 213 var̂za, var̃za ‘tangle’, var̂zât 213 vaŗš 80 vasks 106n163, 217 vȩcs 145, 259 vȩdga 101n148, 125 vêdzele 99 vêl, vȩ̂ls 61 vepris 243 vȩr̂ga 125 vẽris 137 vḕrpt 61 ver̃t 59 vêzis 99 viêns 2 vientrēb, vientreib 270 vietēt 33n57 vîksna 182 vìksne² 213 vil̃na 56 virbe, vìrba² 174n3 vìrši 216 vir̃zît 213 vir̂ži² 216 vîtuõls 169n25 vùikls² 20 zàbaks 8 zaļš 52, 79, 80n104 zars 59, 123, 144 zȩga 53n25 ziêds 59 zir̂gs 59 zir̃nis 56, 79, 293, 295 zivs 93 zizlis, zizls 14 zùobs 52 zùoss 56, 138 žagas 53n25 žeberis 16 žer̃biņš, žerebiņš 16 žĩguris 79n103 žìvats² 8 Prussian Unmarked = Elbing Vocabulary. Note: ⟨c⟩, ⟨y⟩ and ⟨z⟩ are alphabetized under k, i and s, respectively. abse 219, 278 ains iii 2 aketes 50, 273 alkunis 272 alu 42, 49 angstainai iii 83n108 angurgis 54, 79, 83, 93 ansis 57 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 392 asilis 11 assegis 275 aswinan 113n173 attolis 79n102, 104, 232 auckstimiskan iii 83n108 ausis, iii ausin 12, 301 awins 56 babo 228, 295 balgnan 257n52 balsinis 257n52 bitte, TC bete 227 bordus 225n113 broakay 40n5 brunse 12 brunyos 11, 42 buccareisis 238 burwalkan iii 229n120 doalgis 191 dragios 190 driāudai, draudieiti iii 70n84 duckti iii 52 ebangelion iii 229n120 esketres 218, 236 etwiērpt, etwerpeis iii 70n84 gaydis, G gaide 42, 295 geytye, iii geits 42 gelatynan 52, 95n130 gelso 274 girmis 84, 203 girnoywis 181n21 grandis 121 gudde 42 gulbis 189 iau iii 58 iaukint iii 70n84, 71n84 ilmis 43 ylo 38 ioūs, ioūson iii 2n5 isrankīt 78n100 iuriay* See luriay iuse 12, 169n25 kaāubri iii 168 kadegis 79n102, 84 word index kāimaluke 40n6 kaimīnan iii 50 caymis 40 kalabian, iii kalbīan 229n120 kalopeilis 135 kampnit, kamnet G 231 carbio* See tarbio kargis* See kragis catils 11, 42 caune 257n50 keckers 295 kelmis 43n13 keuto 169n25 knapios 206, 207, 296 kragis 79n103, 80 crays, G kraise 97n136, 239n13 craysewisse (gloss) 97n136, 239 crausy, crausios 13, 197 krawia iii 80 crixti lāiskas iii 11, 12n13 crixtia iii 12n13 crixtianai iii 12 cuylis* See tuylis culczi 257 kulnis 257n52 kumetis 11 kurpe 259 kurwis 222n109 lalasso 98, 258 lapinis 43 laxde 266 leipen TC 2, 89 linis 93n126 lino, lynno G 295 lituckeckers 296 lopto 101, 101n148 luysis 180, 181, 271 lunkan 125, 171, 181 luriay 87 lasasso* See lalasso malunis 196n53 mārtin iii 51 mealde 202 median 54, 80 mes iii 2n5 moasis 295 moke 296 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 393 word index nabis 52 noatis 203, 273 noūson iii 2, 2n5 sturdis 239n12 suckis, iii suckans 13 zuit iii 20, 27, 78n100, 139 panto 135 paute G 130 pecku i–iii 41 pelky 257n50 pelwo 56, 293 pepelis 204n73 percunis 136 pievſſen TC 71n84 pippalins iii 204n73 pirsten, iii pīrstans 204n73 piuclan 70n84, 78 poalis 209 powiērpt 61n58 prassan 295 takes 97 talus 57n41 tarbio 18n25 tauris 250 tauto 249 tebbei iii 2 tien iii 2 tiēnstwei, tenseiti iii 70n84 tresde 203 trinie iii 201 tubo 39 tuylis 20n35, 79n103 tūsimtons iii 106n162 rugis 191, 295 zaidiantẽ TC 59 saligan 52, 79n103 salowis 95n130 sansy 56, 82 sarwis 13 sasintinklo 2n4, 57 seydis 300 seimīns iii 51 semen 56, 293 sylecke 95 sylo 229n120 singuris 79n103 sirablan iii 225, 245 sirgis 59, 78 syrne 84, 293, 294 sirsilis 55 sirwis 55, 83, 123, 124, 221 sywan 55n30 skerptus 195n51 scoberwis 195 scurdis 239n12 slayx 171 snaygis 2 spurglis 259n60 sticlo 11, 11n11, 42 stoberwis 195 umne (gloss) 231 umnode 231 unds iii 2n4 urs iii 101 uumpȋs 231, 254 wagnis 56, 56n36, 83, 293 waldwico 20 wargien 79n103, 80 wedigo 101n148, 125 wessis 204n73 wyse 213, 239, 295 wisge G 239 wissa 78n100 wissambs’ 220, 221, 242, 257 woaltis 272 woasis 271 wobalne 269 wobilis 100 woble 100, 268 wobsdis (gloss) 226 wobsdus 221, 226 wobse 55 wosee 56, 79, 83 wosux 226n115 wuysis 20 wundan 2n4 wutris 301 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 394 word index Slavic Languages Russian Note: ⟨ѣ⟩ is alphabetized under е. алёс, а́ лес 29, 33, 35 Асмоудъ OR 18 бараба́ н 230 белена́ 163 берёжая 225 берёза 169 берескле́т 216 бе́рест 215 бе́рковец, OR бьрковьскъ 13, 26, 28 бѣсъ OR 27 блоха́ 184n29 боб 228, 295 бобы́ ль 20n34 болва́ н 9 борзы́ й MR 222 борка́ н 31, 32n52, 36 борода́ 224 борозда́ 224 бо́ ртень 9 бо́чка 15n20 брат 194n50 брѣ́згать 271n82 брънѧ OR 10, 42 бры́ згать 165 бу́ дто 20 бурка́ н 31n52 бъчи* OR 15n20 бы́ дто 20 бык 20 быль MR 20 бьрковьскъ OR See бе́рковец вага́ н 102n152 ва́рягъ MR, варѧгъ OR 18n27 вѣ́ко 102 Велья 36n62 ве́нтерь 33 вервь, верёвка 31 Веремоудъ OR 18 ве́рес, ве́рест 215 ве́реск 215, 216 верея́ 24n43 ве́рша 213 вѣ́тер 25 ве́хорь 24 Витовтъ MR 30 ви́ шня 23, 214 воробии OR 164 воротъ, воротитисѧ OR 122n10 вы́ клый 20 вы́ мя 170 выть 20 вьрпати* OR 61n58 *вьсь OR 19n32 вяз 182 вя́ терь, вя́ тель 33 гни́ да 183 гнус 184 голубо́ й 176n7, 188n32 го́лубь 176n7, 178 голѧдь OR 19, 29 горб, горбу́ н 21n35 горо́ х 295 граб 194, 206 грамота OR 18n29 гремѣждь RCS 258 грести́ 269 гре́ча, гре́чка, гречи́ ха 15, 16 громѣждь RCS 258 гру́ да 69n78 гру́ ша 197 груши́ ть 198 дёготь 29, 34 дере́вня 29, 31 де́рево 31 дрозд 194n49, 203 ду́мать 22, 23 *дъхорь OR 25n46 еба́ ть 33n55 ёлха, ёлка, ело́ха 277, 277n95 ерѧбь RCS 174 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 395 word index жегъзоулѧ OR 120 желѣ́зо 275 жень 299 живо́ т, животы́ 8 жир 242 жито OR 15 жни́ во 88 жьзлъ OR 14 за́ быль 242 зепь 242 зубр 220, 242 зуфь, зуфрь MR 242 и́ волга 178, 195, 242 И́ горь 18 ижера OR 18, 299n21 иза́ боль 242 изарбавъ MR 242n21 изубрина 242n20 изумрутъ MR 242n21 изу́ фрь 242 изю́ брь 179, 242 и́ лим 235n5 ильмъ OR 234 капу́ ста 22 карма́ н 230 ки́ гачи 117 ки́ ла MR 20, 21n35, 26 киля́ к, килу́ н 21n35 клён 260 клѣть 29n50 кобе́з 168n23 ко́ бчик 168 ковш 29 коза́ 83 ко́злы 11 колк 257n53 ко́лпица, колпь 189 конопля́ 206, 296 ко́ поть 126n21 копы́ то 20n34 ко́ роб, OR коробъ 17, 27 коробьѧ ONovg. 18 коро́ ва 222n109 корсь OR 29 корчма́ 32n54 котьлъ OR 10, 42 кривичи OR 24 криво́ й 137 крупа́ 133, 256 кроуша OR 197n57 крушити MR 198 крьсти́ ти OR 12n13 крьстъ OR 13 кряк, кря́ кать 261 куде́ль 22 ку́ жель 236 ку́ коль 23n40 куна́ 257n50 къбьлъ* RCS 11 кяп 23 ла́ ва 102 лён 295 лебеда́ 177 ле́бедь 176, 234, 257n51, 304 лебя́ жий, MR лебежий 176n10 ле́мех, леме́ш 213 лечь 198 лим 235n5 линь 93n126 ли́ па 2, 89, 171 лист 265 Литва́ , OR литва 19n32, 29 лито́ ка, лито́ нья 175 лобода́ 177, 304 лог 198 ло́ коть 272 лопь 43n12 лосо́ сь 98, 258 луг 198 лука́ 131, 198 лупи́ ть 161 лы́ ко 125, 171, 181 ля́ га 35n61 ля́ жа 35n61 лях MR, лѧхъ OR 19, 26 лѧча RCS 201, 202n67, 296 ля́ ща 202n67 мак 253, 296 медвя́ дь, медвѣ́дь 23 межа́ 54 мѣръ 25 морко́ вь 229, 230n123 мостъ OR 57 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 396 мох му́ ки му́ма мы́ ло word index 273 22 29 20, 20n34, 26 нати́ на 203n70 недѣ́ля 24n43 не́рпа 159n13 -ник, OR -ьникъ 24 овёс, OR овьсъ 15, 212, 239, 295 ови́ н 32 овьсъ OR See овёс оже OR 33n55 О́ лга 277n95 оли OR 33n55 о́лово 301 Ольга OR 235n4 ольха́ 12, 276 о́мех, оме́ш 214 орѣ́х 25n46, 171, 238 осёл, OR осьлъ 10 осётр 218, 236 оси́ на 219, 278 осо́ ка 240 осо́ т 34 острога́ 160n14 о́ стрый 160n14 осьлъ OR See осёл ота́ ва 232 о́чередь 59n47 Павьлъ OR 14 па́ кля, па́ ккула ‘chaga’ 30 па́ кля, па́ кула ‘tow’ 29, 30 па́ перть, MR папортъ 30, 31 па́ рма 117 па́ смо 35 пѣнѧзь OR 10n10 переть, пере́дка MR 30 перт 140n45 перть 29, 30 Пероунъ OR 137 пила́ 135 пиро́г, OR пирогъ 8, 160n14 пиръ OR 160n14 плот 14 Пльсковъ OR 14 подчефре́ниться 35n61 полк, OR пълкъ 8, 25 по́лно 9 прѣ́сный 271 при́ переток 30 пруд 23n40 пры́ скать 165 поудъ OR 18 пу́ сма 29, 34, 35 пух 23, 23n40 пълкъ OR See полк пъıро RCS 294, 295 пьпьрь OR 14 *пьрть ONovg 30, 35, 142 пьшеница OR 15 ре́дька, OR рьдькъı* 14 рѣ́па 23, 102n151, 237, 266, 296 решето́ 57n40 родъ OR 132 рож 193n46 рожь, OR ръжь 15, 191, 295 руда́ , рудо́ й 131 ру́ дос 131n30 руно́ 267 роусь OR 19n32 ръжь OR → See рожь 15 ры́ жий 180n18, 296n17 ры́ жик 296n17 рысь 180, 271 рьдькъı* OR See ре́дька ряб 175 рябо́ й 174n2, 176n7 рябь 175, 176n7 ря́ па 23 са́ жа 136 сапо́ гъ OR 8 свѣтъ OR 27 свиньць OR 301 селёдка 31, 95n131 Селижа́ ровка 299n21 сельдь 31, 96 Серегѣрь OR 299n21 серп 193 сиг 95 си́ зый, сизя́ к 209n85 ско́лко 277n95 скора́ 120 смьрдъ OR 14 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 397 word index со́лод 193n47 сом 93n126 споръ OR 281, 281n100 сруб 15n20 сте́бель 60 стѣна́ 300n23 столб 211 сто́ рож 220 стръıи OR 281n100 стькло OR 10, 42 суди́ ть 22 сума́ 22 сыр 20n34 сыт, OR съıтъ 20, 21, 26 сьрна OR 221 темни́ ца, OR тьмьница 15n20 тенёта 57n40 тис 180, 265 товка́ ч 117 толк 8, 27 толма́ чь 236 толпа́ , толпи́ ться 9n5 то́ поль 274, 274n91 топта́ ть 165 то́ рба 9 торг, OR търгъ 8, 26, 28 треска́ ‘trout’ 255, 255n49 трѣска́ ‘splinter’, RCS трѣска 255, 255n49 тресть 201, 255 тро́ ска, MR троска 255 трость 201, 255 труба́ 137 тру́ тень 200 тур 253n45 търгъ OR See торг тъıти OR 232 тыни́ ть 25n47 тъıнъ OR, MR 21, 25 тьло OR 57n41 тьмьница OR See темни́ ца у́ горь 93 у́лей 279 усы́ 23 уха́ 169n25 холо́ п, OR холопъ 17, 26 хоромъ OR 17 хорт, хо́ ртица 8, 26 хору́ гвь 22 хотѣ́ть, хотя́ 196n53 хроуша RCS 197n57 цѣ́лый 23 цѣна́ 130n28 цѣп 23 цылье́къ 15n19 ця́ лой 23 ча́ шка 143 чек 15n19 человѣкъ OR 14, 15 чемери́ ца 245 чемерь RCS 245 чепе́ц 122 черв, червя́ к 35n61 черёмуха, черёма, черёмка 247n27 черемша́ 246 че́реп 27 черепо́ к 17 черп 35n61 чеса́ ть 248 чесно́ к 247 чилэк 15n19 шёлк 14, 17, 17n23 шеломъ OR 17 шест 143 ши́ жлик 117 шолкъ* OR See шёлк шоломъ OR 17 -ьникъ OR See -ник я́ блоко 268 яболонь MR 269 язы́ к 20 ꙗкорь OR 18 я́ сень 271 ꙗтвѧгъı OR 19, 29 ̣ꙗчьмъı RCS 295 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 398 word index Ukrainian вереск 215 вѣхорь, вѣтеръ MUk. 25n45 в’я́тір 33 горобе́ць 164 дере́вня 31 дьо́готь 34 же̑нь, жинь 299 ки́ ла 26 ко́ буз 168n23 ле́бідь 176, 176n10 ́ ліска, ліщи́ на 266 оря́ бок, о́ рябка 33n55, 174 осока́ 240 ре́дька 14 то́ впитися 9n5 трiска́ 255 хорт 8 чере́мха 247 чоловіќ 15 шовк 14, 17 щеть 127 нати́ ня 203n70 нать 203n70 Belarusian авін́ алёс *аса́ асёл асе́ць 32 33, 35 278n97 11 273 ́ бажніца 184 батогъ MBel. 165 борзды MBel. 222 бу́ сел 11 быллё 20 малання́ , маладня́ мо́ рква 229 мы́ ла 20 мыто MBel. 20 203n68 ́ націна 203n70 па́ кулле 30 ра́ бчык, рабо́ к 175 ́ габ, габіна 194n50 грама́ да 9n6 скаварада́ 17 смык, MBel. смыкъ 20 сяло́ 123n13 жэнь 299 жэ́ рабя 16 таўка́ ч 9n6 траска́ 255 калама́ жка 17 карава́ й 16 ́ кніга 184 хлѣбъ MBel. 38n2 лём 235, 235n5 луг 198n61 чалаве́к 15 ́ чарапіца 16 чэ́мер 245 чэ́ рап 17 шворобе́й, шурабе́й 164 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 399 word index Lechitic (unmarked = Polish) barzo OPl., barzdo 222, 222n110 bób 228 chart 8 chcieć 196n53 chociaż 196n53 chrzcić 12, 201n63 chybać 220n104 ciemierzyca 245 człowiek 15 dąć, dmę 182 deptać 165 drozd 203 drożdże 190 dupa 210 dziegieć 34 gab, gabina 194n50 gbeł OPl. 11 gołąb 178 gont 9 grece, grice OPl. 15 gromada 9n6 gruda 69n78 grusza 197 gwałt 7 ilem* OPl. 234 jabłoń 269 jarząbek 174, 175, 241 jazgarz 275 jaź 275 jaźwiec 226n114 jelito 175 jesiory 218n101 jôlsc Kash. 226n114 jôpsc Kash. 226 jôżdż Kash. 275 karta 7 kiełb 190n39 klon 260 kmieć OPl. 11 kobuz 168 kobz* (kobez*) 168n23 kocioł 11n12 kolimaga 17 kôłp Kash. 189, 195, 206 konopie 206, 296 krëszka Kash. 197 krobia 18n25 krzest OPl. 11 kubeł 11, 11n12 lebioda 177 lemiąż OPl., Kash. 215 lemiąże 214 lim 235n5 Litwa 29 łabędź 176, 187, 234, 234n2, 304 łąg 198 łąka 198 łosoś 98, 258 łupić 162 męka 175n4 młyn 196n53 nać 203 natura 7 obrzazg, obrzask 271n82 olcha 276, 277 olesie 33 oset 34 osioł 11 oskard 239n12 pakuły 30 paseł 11n12 pierz, pieprz OPl. 14 piorun 137 płet* OPl. 14 proso 295 rydz 295, 296n17 rydzy 180n18, 296n17 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 400 word index skrzek 261 smard OPl. 14 smardz, OPl. smarsz 222n110 sum 93n126 syrzisko OPl. 20 syty 20 szłom OPl. 17 śledź 96 trąd 200 treść 201 trzcina 201n63 trzemka 247n27 trzemucha 246 trzop 17 tur 250 ul 279 wąż 93n125 więcierz, Kash. wiącel 33 wilga 178, 179, 195, 257, 304 włodyka OPl. 20 wróbel 164, 168 wyżeł 20 ząbr 220, 242 Czech-Slovak Unmarked = Czech blín 163 bratr 194n50 brázda 224 brikáda 199 brzý OCz. 222 břest, břečťán 215 březa 216 březek 215, 216 krstiti OCz. 201n63 křepel, křepeliti 241 křtíti 201n63 kuláš 199 kuna 257n50 kužel 236 kýla, Slk. kyla 26 labuť 171, 176, 187 lava 102 lebeda 177 lemez Slk. 215 lemiez OCz. 214 lipa Slk. 89 list 265 lín 93n126 líska 266 loboda Slk. 177 lúpiť Slk. 162 čemer Slk. 245 česnek 247 dehet Cz., OCz. 34 drúh, drúk Slk. 187 dúpä Slk. 210 habr, hrab 194 hruška 197 jablko 268 jarabý, jarabica Slk. jaseň Slk. 271 jesen 275 jeseter 218, 236 ježdík 275 kelka 257n53 kobec Slk. 168 krb 17n24 krpě 259 176n7 mech 273 mlýn 196n53 mše 265n71 mydlo Slk. 26 nať Slk. 203n70 oheň 181n23 olše 277 osika 272 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 401 word index otaviti 232 otevříti 59n47 otýniti OCz. 25n47 perna, pera Slk. 165n19 piroh Slk. 160n14 rešeto Slk. 57n40 řiť 123 rouno 267n73 rysý 180n18 sporý 281 střemcha, střenka 247 střenka 247n27 stříbro 225, 261 tenatá Slk. 57n40 tis 265 tlumač 236, 236n6 tlumočiti 236n6 topoľ Slk. 274 treska 255 třtina 201n63 tur 250 týň Slk. 25 úhoř 93 včela 227 víko 102 vlha Slk. 179 vrabec Slk. 168 vřes, vřesk 215 výheň 181n23 žabokřeky 261 Sorbian kobušk USrb. 168 kołp USrb. 189 krušej USrb., kšuška LSrb. 197 kula, kulka USrb. 257n53 lemić USrb. 214n96 lom LSrb. 235 łomić USrb., łomiś LSrb. 214n96 rysy LSrb. 180n18 wósa LSrb. 219 Slovene bȃsta 27 blẹ̀n 163 bòb 228 brȃtər 194n50 bŕna 165n18 brzdit, bŕzda 222n110 čəbę̑la 227 čẹ́maž 246 čésən 247 čmeríka 245 čremha 247 črę̑moš 246 črę̑msa 247 drǫ̑zd, drǫ̑zg 191n41, 204 drǫ̑zga 191n41 dúpa 210 gnída golbica grȃbər grúda grȗška 183 190n38 194 69n78 197 hr̀t 26 hrȗška 197, 197n58 jáblana 269 jábołkọ 268 jasíka, jásen 272n83 jéčmen 295 ję́łša 277 jerę̑b 174, 241, 304 jesíka, jésen 219, 272n83 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 402 word index kóbəc 168 krmẹ́žəlj, krmę́lj 258n54 kŕplja 259 labǫ́ d 176, 234, 257n51 lȃn 295 lebed 176 lebę́da 177 lę̑mez 214 lémež 214 lìm 234 lȋnj 93n126 lobod 177n11 lóboda 177 lǫ́ ka 198 lom 235n5 lúža 198n61 màk 253, 296 mílọ 26 molẹ́ti 122n11 nȃt 203 rę̑b, rebíca 175 rȋdž, rȋdžək 295, 296n17 rȋs 180, 271 ȓž 191, 295 sáje 136 skóbəc 168 srebrọ̑ 225, 261 srẹ̑n 59n50 sŕna 221 stółb 211 svínəc 301 ščę̑t 127 šlẹ̀m 26 tȋs, tȋsa 180, 265 tréska 255 trǫ̑ ska 191n41 trǫ̑ t 200 ugọ́ r 93 ȗlj 279 ógənj 4 ǫ̑ ł 49 órẹh 238 ostrǫ́ ga 160n14 otȃva, otáviti 232 óvəs 212, 239, 295 vółga 178, 179, 195, 257 vrábəc 164, 168 vŕša 213 píra 294 prẹ́sən 271 želẹ́zọ 274 zȋd 300 zǫ̑ bər 242 Serbo-Croatian bȍb 295 brázda 224 brȅđa 225 bȑnjica, brnjȕse 165n18 bȓz 222 br̀zār, brzdar 222 brzdica 222 cremoš 246 crìjemuša 246 čȅmēr 245 čèpac 122 dòspijeti 101n144 drȍzd 203 drȍzg 204 drsk 205n77 drȗg, druk 187 dȕti, dmēm 182 gnȏj 184 grȁb, gràbar 194 grȁh 295 gȗb ‘goby’ 190n38 gȗb ‘swan’, guf 189, 190 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 403 word index hlȁp 26 hȑt 8, 26 hruška 197n58 jȁsēn 271, 272n83 jàsika 272n83 jelha 277 jèmlješ 214 jȅsēn Kajk. 272n83 jèsetra 218, 236 jóha 277 júha 169n25 kȉla 26 klȅn 260 kònoplja 206 kòsa 248 kȑbulja 17n24 kȑmež, kr̀meželj, kr̀melj 258n54 krȕška 197 kȗf 189 kȕk 257 kȗn 260 kúna 257n50 kȗp 189 labòda 177 lȁbūd 171, 176 lȅbūt 176 léća 201, 296 lèmeš 213 lèmež, lemȅž 214 lijèmati 214n96 lijèska 266 lȉko 125, 171, 181 lȉpa 89, 171 lobòda 177 lòpīž 238 lȗb 161 mȃh mȁša mèđa mȏst mȑkva 273 265n71 54 281 229 ȍlovo 301 pȉr 294, 295 prȍso 295 rȅpa 237, 296 rȋđ 180n18 rȉpa See rȅpa róda 237, 238 ròditi 238 rúno 267n73 sȉt 26 sȋv 55n30 sljȅme 128 sȍm 93n126 srȉjemuž 246 stáblo 60 stȕba 211 šljȅm 26 tíniti, tȉn 25n47 topòla 274 trȅska 255 trȍska, trȕska 191n41 vȇz 182 vrȉjes 215 vȕga 179 želȅzo 275 Church Slavic Unmarked = Old Church Slavic агнѧ 56 брада бразда брьзо бьчела 224 213 222 227 васъ 2n5 вепрь 243 верѣꙗ 24n43 весна 245 вечеръ 228, 245 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 404 влъга CS See ивлъга воскъ 217 врьхъ 213 вътрь CS 301 въı 2n5 въıкнѫти 181 вьчера 228 вѧзати 183 глава 196 глаголати 213n92 гласъ 213n92 голъ 196 голѫбь 178, 187, 188, 304 грабити 269n78 градъ 128 дѫбъ 229, 231 дебель → See *лебедь 176 дльгъ 169 дрождьѩ 190 дрѫгъ* 187 елень 188 емешь CS 214 желѣзо 274 жрънъı* 181n21 знакъ CS, знати 171n30 зракъ 171n30 зрьно 84, 294 зълъ, зълоба 188 зьрѣти 171n30 ивлъга CS 179, 179n15 исто* 181 кедрi 265n70 крѫгъ 168n22, 249 кѫсъ* CS 213n92 ладии 80n105 лакъть 272 *лебедь CS 176 лемешь* CS 213 листъ 265 лоза 266n72 лопата 101 word index льмъ 235n5 лѣсъ 266 лѧгѫ (лещи) 198 лѧща CS 202n67 млатъ 239n12 млънии 202 мостъ 281 мъногъ 195 мъноѭ 196n55 мъı 2n5 мьнѣ 196n55 мѣсѧць 10n10 наврапити 61n58 насъ 2, 2n5 носъ CS 281n101 обрѣсти 238 олово 301 омлатъ CS 239n12 оскръдъ 239n12 папежь 201n64 папрьтъ 30 плъкъ 8 подобати 60 пробрѣзгъ 191 проврѣти* 59n47 пьрѣти сѧ 30n51 пьрꙗ 30n51 пьшеница 295 пѫта 135 рить SCS 123 сапогъ 160 село 220n105 слабъ 270n79 смиꙗти сѧ 213n92 смѣхъ 213n92 снѣгъ 2 соль 281n101 срѣнъ CS 59n50 стлъпъ 211 столъ 281n101 стражь 220 стѣна 70, 220 съзьдати 300n22 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 405 word index съто 196n53 сьребро 221, 225, 245, 261 сѣно 70, 78 сѣти* 96n132 тебѣ 2 тіса 265n70 тлъкъ 8 тръгъ 8 тръсть, трьсть 201, 224 тоуръ 250 тъıсѫщи, тъıсѧщи 106, 222 тѧ 2 оу 58 оульмъ CS 235n5 оучити 70n84, 181 хлъмъ 235n4 чловѣкъ 14 чрѣмошъ 246 ю See оу ѫзъкъ 83, 182n24 Bulgarian (Macedonian) авлѝ га 178, 179, 195 брна Mac. 165n18 бъ̀рна, бъ̀рла 165, 165n18 вра 59n47 ду̀ пе 210 елша̀ , елха̀ 277 емѐш 214, 214n95 ѐрбица, ѐребица 174n3 ѕид Mac. 300n23 и́ волга 179n15 кру̀ ша 197 къ̀жел 236 къ̀лка 257 лабед Mac. 176 лѐбед 176, 234 лемѐж 214 лѐща 201 ло̀ бод 177n11 ло̀ бода 177 лъка̀ 131, 198 лю̀ пя 162 мо́лзѫ 179n15 мо́ рковъ 179n15 о̀ сен, осѝ ка 271 полѐ, пойѐ 214n95 рж, ἄρжυ Mac. 192n43 стена Mac. 300n23 стъ̀лба 211 чемер Mac. 245 чѐмер, чемерѝ ка 245 я̀ сен, ясѝ ка 272 Germanic Languages Gothic aleina asilus auhns* aukan 272n85 10 231, 254 170 -ba 188 bairhts 169n27 bars Crimean 224 baurd See fotubaurd Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 406 word index brukjan 169 brunjo 10, 42 laufs* 162 liuhaþ 250n34 digan* 300n22 diups 210, 211 maihstus* 191 manags 195 midjis 80 miliþ 41, 184n28 faihu 37, 41, 41n9, 231n127 fon 268 fotubaurd 224n113 gadaban 60 gatwo 37, 38 guta* 42 haban 186 haihs 170 haims 40, 50 hairdeis 41, 80n105 harjis 80 hauns 40n7 hilms 43, 43n13 hlaifs 37, 38 hlain(s)* 235n4 ƕaiteis* 40 jabai 33n55 ju 58 katils* 10, 42 niujis 80n104 rauþs* 250n35 sarwa 13, 37 silubr 225, 261 sparwa 259n60 stains 220, 300n23 stikls 10, 42 stilan 220 stiur 251 stols 220, 281n101 tulgus 169 þaurp 270 þiuda 249 þusundi 106, 222 wraiqs* 169n26 North Germanic Unmarked = Old Norse al OSw. 234 almr 234 alr 39n4 ankare OSw. 18, 25n46 apaldr 269 ár 90 arta 237 askr 234, 272 Ạ́ smundr 18 asp OSw. 234 aurr 87 barkanrot, barkenrot Sw. 32 baun 228 bíldur Ic. 96n132 Biærkö OSw. 13 bjǫrk 169 borð 224 brekanne-rod Danish 32 brisk Nw. 271n82 broddr 224 brún 58 byrr 90 dóttir 126n22 dráttr 126n22 dregg 190, 191 drått Nw. 126n22 dúfa 210 duppe Nw. 210 dýfa 210 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 407 word index enni 71n86 fé 41n9 fit 72n86 fley 90 gata 37 geirr 72 gler 39 gløggr 251n38 gómr 267 greve Sw. 68 hagre OSw. 195n52 hallr 106n163 hani 106n163 haukr 168 heimr 50 Helga 235n4 hinna 88n120 hiupon OSw. 168 hjalmr 43 hleifr 37, 38, 65 hlynr, Ic. hlynur 260 hnot 96 hosor Sw. 69 hringr 249 hrogn 261 hveiti 37 hý Ic. 49n12 hässja Sw. 75 Inguar OSw. 18 jarpr 174n2 jerpe Nw. 174 júgr 170 järpe Sw. 174 kringr 249 kverk, kvern Nw. 126 nafli 52n20 nate, nata Gutnish 203 naust 268n75 nór, Ic. nótrog 268n75 nätla, nätsla OSw. 203, 273 nǫf 52n20 omn Sw. 231 pipar Sw., OSw. 14n17, 25n46 poik Sw. 9n4 pund 18 pörte, pyrte Sw. 141 reyfa 162n16 rjúpa 241 ropa Ic. 241 ros Sw. 69 rugr 191 rype Nw. 241 sáld 96n132 saumur Ic. 110 selr 112 síkr, Nw. Sw. sik 94, 95 síld, OSw. sīldh 96 silfr 225, 261 silki 14 sill, sill-lake Sw. 96, 97, 97n134 slo Nw. 171 sléttr 126n22 slå Sw. 171 smæra Ic. 242 sonur Ic. 4 stafr 280 stelpe Nw. 211n91 stjǫlr 220 stolpi 211 strǫnd 50n13 styrja Ic. 236 lax 98, 258 luo Elfdalian 181 telgja 191 tjara 49 torg 10n7 mjǫðr 37 mjǫllnir 202 mäld OSw. 72n86 mǫl 122n11 þéttr 126n22 þjórr 250, 251, 251n37 þófi 39 þorskr 255 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 408 word index þrǫstr, Ic. þröstur 49, 203 þús-hund 106n162 þúsund 106, 222 þömb Ic. 58n42 vǽringjar 18n27 vömb Ic. 58n42 urt Ic. 237 ǽs 57n37 ässja Sw. 72 vagnsne Nw. 56n36 valmogha OSw. 253 vangsni, Nw. vangsne 56n36 vax 106n163, 217 vegg Nw. 50 Vermundr 18 årta Sw. 237 ǫl 37, 49 ǫlpt 176, 234, 304 ǫlr 276 ǫnd 169n25 ǫsp, Ic. ösp 219, 278 English Unmarked = Old English al ME 39n4 ār 90 æl 39n4 æppel 268 æpse* 219n102 bār 264 beam MoE 59n49 beard 224 beolone 163 bord 224 brord 224 brūcan 169 calu 196 cīþ 59n48 clīfe, clīfan 81 culufre, culfre 178, 178n13, 187, 304 dærste 191 doder 100 dran 200 dræst 191 drone MoE 200, 201 drōs 191 ealgian 170 ealu 37, 49 ēam 50 ēar 87n116 egeþe 50, 273 elm 234 eofor 243 eorp 174n2 fæt 270n81 fearh 78 fersc 271 glas ME 39 glær, glæs 39 gōma 267 hār 59n50 hænep 206 hīwan 51n16 hlidgeat 192 hnitu 183, 184 horefrost ME 59n50 hramsa 246 hran, hron 267 ielfetu 176 klutz MoE 166n20 lāurice, lāwerce 92, 237n8 lǣs 107 lox 180, 271 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 409 word index magu 192n45 mæst 281 medu 37 mēos 273 meox 191 mīgan 191 moru, ME more 229, 230 mos 273 sōt 136 start MoE 100 stær 204n74 stela, ME stele 220 sterten ME 100 strosle 204n72 stulpe ME 211 syrfe* 204n74 netele 203, 273 tare ME 31 teoru 49 thresh MoE 204n74 ōsle 233 papig, ME papig 192 rex MoE 204n74 rūh 170 ryge 191, 192 scearu, sceran 120 schlep, schmuck, schnozz MoE 166n20 searo 13 secg 240 share MoE 120 sherve, sharve MoE 204n74 siolfor, siolufr- 178, 178n13 slāwyrm 171 þel 57n41 þēod 249 þorp 270 þræsce See þrysce þrop 270 þrostle 204 þrysce* 204, 204n74 warp 61n58 wealhmore 230 wesend, weosend 220, 242 wice, MoE wiech 182 wodewale ME 178, 180, 195, 257 Dutch / Frisian Unmarked = Middle Dutch arut ODu. 233 bij Du. 227 crinc 249 cringhen 249 dobbe 210 dorne 200 dorpel 164 dorsch* 255 dreppel, drempel 164 droes, droesem Du. 191 dronen 201 drumpel 164 heen Du. 70n82 hoon 40n7 lepel 43 lins 201 lomre 215 male 180 mies Du. 273 oom 50 opruspen 241 rinc 249 rob Du. 160n13 rogge, OFri. rogga 192 rove 237 ruppen 241 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 410 word index stelpen 211n91 stolpe 211 stuur 251n41 tarwe 31 vlotscip, Du. vlot 14 zeelt Du. 96n133 zegge Du. 240 Low German Unmarked = Middle Low German bart 225n113 bene 227 bilina OS 163 brackannyen 32 brōk 40n5 dran, drano, dreno* OS, Drahn MoLG 200 dorsch 255 dumpeln 210 magonhouut OS 41 maldia OS 177, 231 male 180 pavos OS 201n64 penning OS 10n10 gers, gersele 277 glar, glar(r)en 39 gume, MoLG gūmen 268 ramese 246 rap 241 raphōn 241, 259 reddik, MoLG Röddick 14 rokko* OS 192 Rubbe MoLG 160n13 rūm 21n37 Heers MoLG 277 holm* OS 235n4 slagge 165 sluse 21n37 korf OS 17n24 tūn 21n37 lēs OS 107 lönenholt, MoLG Löhn 260 Wieke MoLG 182 High German Unmarked = Old High German Ahle, Alle* MoHG 39n4 ala 39 albiz 176 Alp MoHG 245n22 amsla 233, 304 apful 268 arawīz 153, 173, 229 aruz 233 aspa 219, 219n102 bābest 201n64 bart 224 Beie MoHG 227 bīa 227 Biene MoHG 227 bīhal 96n132 bilisa 163 bini 227 Bock MoHG 11 Bolwan BaltG 9 bōna 228 Borkane See Burkane brunna, MHG brünne 10n10 bruoch MHG 40n5 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 411 word index Bulwan BaltG 9 Burgund MoHG 230n124 Burkane, Borkane BaltG 32, 229, 230n124 dahs 218 diot, diota 249 dorf 270 drosca 204 ebur* 243 egen, eckan 273 egida 50, 273 elbiz 176, 234 elira 276 elm 234, 235n4 enker MoHG 18n28 erpf * 174n2 essa 72 farah 78 Floß MoHG 14n18 folc 10 frisc 105n159, 271 furh, furuh 208 Gatt PrG 37 Gaumen, gaumo MoHG 267, 268 gaumo 268 gerrich MoHG 277 Gewalt MoHG 7 gierisch, giersig, Giersch MoHG 277 gires 248, 277 glas 39 goumen 268n76 goumo 267, 268, 268n76 Greck, Grick PrG 15 Grecken, greckisch MoHG 15 gūme MoHG, gūmə PrG 268 guome MHG 268 guomo See goumo habuh 168 hammer MoHG 245 hanaf 206 harmo 143 hasal 280 hemera, MHG hemere 245, 245n23 hemern, hemerwurz MoHG 245 hiufo 168 hōn MHG 40n7 hulis 178 hūt 169n25 ilme MHG 234 kalb 78 kalo* 196 kīnan* 59n48 klība, klīban 81 klioban 248 klobalouh* 248 korb 17n24 korb, korw PrG 9 kranc MHG 249 Kricken BaltG 15 krinc MHG 249 kübel MHG 11n12 lahs 98, 258 lamb 78 leffil 43 Lehne MoHG 260 linsī 201 līso 107 louft 162 luhs 180, 271 mahen, mān MHG 254 maho, mago 253 manag 195 mast 281 melta 177, 180, 231, 304 misken 218 mist 191 Mohn MoHG 253 Möhre MoHG 230 moraha 229, 230 mos 273 nezzila 203, 273 niz 183, 184 nuz 96 ōheim 50 ovan 231, 254 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 412 word index Pleskau MoHG 14 pund BaltG 19 Purkahne See Burkane treno 200 trestir 191 trostel MHG 204 rebahuon 241, 259, 304 rocko 192 rogo 261 ropfezzen 241 ruoba 237 ūter 170 sahar 240 Schlack MoHG 165n17 schuole MHG 21n37 selah 112 Siek BaltG 95 silabar* 225, 261 šlǖse BaltG 21n37 šoil BaltG 21n37 sparo 259n60 stahal, MoHG Stachel 254n47 Stahl (stol) MoHG 254n47 stior 251 stiuri 251n41 sturio 236 tanna 229 toter, totter MHG vorwërc MHG 229n120 wägese MoHG 56n36 wahs 106n163, 217 wahsan 170 waraf 61n58 wefsa 219n102 weggi 50, 79n102 Weiche MoHG 182 Wenter PrG 33 wīda 169n25 Wiedewalch MoHG 180 Wieken PrG 182 wisa 242n19 wisc 239 wisunt 220, 242, 257, 304 Wisent MoHG 242n19 witewal MHG 178, 180, 195, 257, 304 100 Other Indo-European Languages Goidelic Unmarked = Old Irish á 272n84 aball 269 argat 302 Craumthann MIr. 246 crem MIr. 246, 247 crim, Crimthann MIr. 246 bech 227 brú 122n11 dluigid 191 domain 211 do-tluchethar 10n8 druid MoIr. 205 cáech 170 cainnenn 247 cairem 169n26, 260n61 camm 282 cóim 51 coll 280 corca MIr. 195n52 fear MoIr. 4 ferb 251 ferg 169n26 fróich 216 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 413 word index glúair MIr. 251n38 gúaire 55, 251n38 gulban 190n39 inis 123n14 lem MIr. 234, 235 lóichet 250n34 lug 180n19 sab 280 seamair 242 seisc 240 sine 248 slat, MoIr. slat 266n72 sníid 57 suide*, MoIr. súiche 136n38 nead MoIr. 266n72 nenaid 203, 240n17 Tadhg MIr. 218 talam 58n41 tarb 251 taul, tel MIr. 246, 247n26 tluchethar See do-tluchethar treb 248, 270 truit 203, 204n74, 205 túath 249 raith MIr. 240n17 rón ‘horse-hair’ MoIr. 267 rón ‘seal’ 112, 266 ubull 268, 269 uilen 272n85 uinnius 271 maide 281 mala 122n11 meinic 195 British Celtic Unmarked = Middle Welsh adar 248 alarch 234, 287n8 aval 268 avall 269 baed 264 begegyr 227 berth 169n27 cam 282 carw 55, 55n30, 251 celyn 178, 260n62 cen 88n120 cennin OW 247 clyd 54 craf 246 cryd 169n26, 260n61 danat 203 drask, draskl Bret. 205 drydw 205 duw 192n45 dynat 203 erfin 207, 237 eskit 169n25 gar 167 gruc 216 haearn 205n75 hesc 240 hoern OCo. 205n75 hoet OCo. 205n75 hwyat 205n75 irvin Bret. 207, 237 iwrch 287n8 kelin OCo. 260n62 kenn Bret. 88n120 Letau OBret. 155n4 llidiat 192 llwyf 234, 235 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 414 word index maw MCo. 192n45 mellt 202 meudwy 192n45 moelh OCo. 205n75 moelrawn 266 moualc’h Bret. 233 mwyalch 205, 205n75, 233 mynych 195 oget 273 onn 271 pabi 192 pryf 63n62 rawn 267 reunig Bret. 112, 266 ryc 191, 192 rych 208 tarw 251 trascl OBret. 205 tref 270 tresglen MoW 205 tret MBret. 205 troet OCo. 205n75 trot OBret. 205 trydw, trudw 205 tut 249 Latin agnus 56 albus 177 alnus 276 alvus, alveus 251n39, 279 anguīlla 54 anguis 93n125 ānsa 57 aper 243 aperō 59n47 āra 301 ardea 237 argentum 302 arō 293 ascia 218 augeō 170 aurum 301 avēna 212, 213, 239 avus 254n48 barba 224 Birca ML 13 bōs 251 caecus 170 calvus, calva 196 calx 257 carpinus 194, 195, 206 carpō 195 caulis 251n39 caurus 251n39 cēra 248, 249 cervus 221 chenopus 177n12 cloāre, cluere 60n51 columba 178, 187, 188, 209, 210, 304 corium 260n61 cornū 188n33 corylus 280 crābrō 55 crēscō 195 crux 13n16 deus 162 domus 300 equa 55 ervum 153, 173 excetra 218, 219 faba 228 faber 60n56 falx 167n22, 191 fax 167, 280n99 fingō 300n22 fodiō 57 forda 225 forum 280n99 fracēs 167n22, 190 fruor, frūctum 169 glaesum 39 glamae 258n57 gōbius 190n38 grāmae 258 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 415 word index grandō 54 grānum 56, 293 habeō 186 hiems 186n30 hirundō 185, 189, 304 īnstaurō 251n39 īnsula 123n14 jūs 169n25 lacus 227n116 lapideum 238 lēns ‘lentil’ 201 lēns ‘nit’ 183, 184 liber 162 līlium 189n37 loquor 10n8 mālum 269n77, 298 mālus ‘mast’ 281 mare 284n2 marītus 51 merula 233, 304 missa ML 265n71 mūscus 273 nervus 251n39, 251n40 nōs 3n5 nux 96 occāre, occa 273 olor 234, 287n8 ornus 271 ōs 12, 272n84 ōstium 47n6 palleō 209 palumbēs 186–188, 209, 210, 304 parvus 251n39 piper 14n17 pirum 298 pōpulus 274 porca 208, 224 porcus 73, 78, 284n3 portione 195 prōcērus 195 pulvis 56 quiēs 40 rāpum 207, 237 raudus 233 rota 57 rumpō 162n16 runcō 181n22 sēmen 56, 293 solium 220n105 taurus 251 taxeus 265n71 taxus 180, 265 tellūs 58n41 tenuis 260 trabs 270 trutius, truitius ML 204n74 turdus 203, 204, 204n74 ūber 170 ulmus 234 ulna 272n85 ungarus ML 19n30 ūrīna 87n117 vapor 126n21 vespa 227n116 vieō 169n25 virga 239 viscum 218 vōs 3n5 Fragmentary Italo-Celtic alauda Gaul. 237n8 krapuvi Umbr. 195n51 feíhúss Osc. 300 *Letavia Gaul. 155 Leucetius Gaul. 250n34 haba Faliscan 228n119 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 416 word index Neuiodunon, Nouiodunum Gaul. 250n34 śilaᴘuṙ Celtiberian 226 ταυρομ Osc. 251 Teut- Gaul. 250n34 teuta Venetic 249 touto Osc. 249 trebeit Umbr. 270n80 trííbúm Osc. 248, 270 uisumarus Gaul. 242 Romance (It. = Italian, Occ. = Occitan, Sp. = Spanish) aloe OFr. 237n8 basta It. 27 bèc, bèca Occ., bièco Creuse 227, 227n116 canapa It. 207 cânepă Romanian 207 colombla Old Occ. 178 crespa Occ., crêpe Fr. 227n116 embigo Galician 117n178 foule, fouler Fr. 196 gyepo Creuse 227n116 lac Occ. 227n116 lostrița, lostița Romanian 258n58 mùschio It., musgo Sp. 273n86 ombre Fr. 215 pescha Occ. 227n116 sesca Occ. 240n16 sisca Sp., Cat. 240n16 sutge Cat. 136n38 tordo It. 204n71 xisca Sp., Cat. 240n16 Albanian ah 272 ãnkth 182 arrë 238 bardhë 169n27 bletë 184n28 mjaltë 184n28 mollë 298 rjep, rrjep 270n78 taroç, tarok 251n37 ter 251, 251n37 treg, tregë 10n7 dallëndyshe 185, 304 dimër 186n30 drā 191 thëri 184 gjatë 169 vidh, vĩdh 182 labë 162, 162n15 lajthi, lakthi 266 lapë 162n15 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 417 word index Greek ἀέξω 170 ἀθρήνη 200n62 αἰεί 58 ἄκαρα 167 hακόσια 254 ἀλέξω 170 αλεπού MoGr., ἀλωπού MGr. 246 ἀλκή 170 ἅλς 59, 137 ἀλφούς, ἀλωφούς 234n2 ἀμνός 56 ἄνηθον, ἄνητον 206 ἀνθηδών, ἄνθος 200 ἀνθρήνη 200 ἀξίνη 218 ἄπιον 298 ἄρκευθος 244 ἀρόω 293 ἄρυα 238 αὐλός 251n39, 279 -άφιον 188 ἔχιδνα 219 ἕψω 254 ζορκάς 287n8 θέλω 238n9 θεός 162 θρήνη, θρηνῶδες θρῆνος 201 θρύον 201 θρώναξ 200 Ἴγγωρ MGr. 18 ἱμάσθλη 237n9 ἰξός 218 hιπνε[ύεσθαι] 254 ἰπνός 231, 254 ῑτ̓ έα 169n25 ἰχθῡ́ς 93 βάραγγοι MGr. 18n27 βλίτον 177 βλίττω 184n28 βρίζα MoGr. 192n43 Γαλίνδοι 29 γλάμος, γλαμυρός 258n57 γλεῖνος 260n64 γράμματα 18n29 δέμω 300 δένδρεον 200 δολιχός 170 δορκάς 287n8 ἔαρ 245 ἐθέλω 238n9 ἔλαφος 188 ἐλεφαίρομαι, Myc. erepa(i)ro ἐνδελεχής 169 ἐρέβινθος 153, 173, 229 ἐρείκη 216 ἐρῳδιός 237 εὑρύς 137 200n62 169n26 καίω 108n168 κάμαρος, κάμμαρος 245 κάμπτω 282 κάνναβις 206, 207 καπνός 126n21 καρβάτιναι 259 κάρυον 238 κέδρος 265n70 κείρω 57 κελαινός 188 κέλευθος 244 κηκίς 100 κήλεος 108n168 κήρινθος 249 κηρός 133, 155, 248 κλινότροχον Macedonian 260, 261 κόγχη 173 κόλυμβος 188n31 κονίς 184 κόραφος 188n34 κόχλος 173 κρέμυον, MoGr. κρεμμύδι 246 κρηπίς 260 κρόμμυον, κρόμυον 246 κώμη 40 λάθυρος 201 λάπη 156n10 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 418 word index λείριον 189n37 Λενζανηνοί MGr. 19n31 λευκός 250n35 λιαρός 107 λοιπός 58, 68n76 λύγξ 180, 181 ῥαιβός 169n26 ῥαπάνια, ῥάπυς 208n81 ῥάφανος 207, 237 ῥάφυς 208n81 ῥόβιλλος 168n24 ῥῳδιός 238n9 μάσθλης 238n9 μείγνυμι 170 μεῖραξ 51 μέλι 184n28 μήκων 253 μῆλον 298 Μῑνώταυρος 253 σκῦτος, σκῡτεύς 169n25, 260n61 σπαράσιον, σπέργουλος 259n60 στατός 281 στελεά 220 στέργω 220 στόμα, στόμαχος 160 νεῦρον 251n39 οἶσος 169n25 ὀλέκρανον 272 ὀργή 169n26 ὀρθός 169n26 ὄρνις 244 ὄρυζα 192n43 ὀρύσσω 181n22 ὅτε 33n55 οὖθαρ 170 οὐρά, οὐραχός 160 ὀφνίς 56, 293 παλκός 257n50 παπάς MGr. 202n64 παῦρος 251n39 πέλεια, πελιός 209 πέργουλον 259n60 πηλός 257n50 ποιμήν 56 πόλις 300 πῡροί 294 Ταυλάντιοι 186n30 ταῦρος 58, 251 τείνω 57 τεῖχος 300 τενθρήνη 200, 201 τενθρηνιῶδες 200n62 τέρμινθος 173 τίφη 192n43 τοῖχος 300 τρέμιθος 173 φάρυγξ, φάρυξ 173 χαλκός 275 χανδάνω 282 χάσκω 267n74 Χελιδόνιοι 186n30 χελῑδών 185, 186, 189, 304 χέλυμνα 161 ψάρ 259n60 ὠλέκρανον 272n85 ὠλένη 272 Armenian aławni 209 anic 184n28 anjuk 183n26 arcatʿ 302 awłi 41 čeł 197 darbin 60n56 ełn 188 epʿem 254 eri 123 hacʿi 272 henum 135 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 419 word index jukn 93 ǰowr 87 opʿi 219, 278, 279n98 kanepʿ, kanapʿ 207n80 kełem 55 salamb 188 sołim 220 stełn 220 losdi 258n58 lusann* 180 tordik 204n71 tun 300 Iranian ʿsṯwrʾn Parth. 251 amāč, amāǰ NP 215 aršan- Av. 188 ąiθiiā- YAv. 169n25 æxsīn, æxsīnæg Oss. 209n85 æxsirf Oss. 193 ərəzata YAv. 302 baś- Khot. 183n25 bązaiti YAv. 183n25 bælon, bælæw Oss. 209n83 bærz, bærzæ Oss. 169 bæzz- Oss. 183n25 bz- Parth. 183n25 idawæg Oss. 193n48 insæj Oss. 183n25 kʾnb Pahlavi 207n80 kaṃha- Khot. 207 kanab, kanaf NP 207 karəna- YAv. 53 kauruua- YAv. 196 kært, kærtæ Oss. 134 kinif Kurdish 207n79 korēshī Kurdish 197n59 kunæg Oss. 207n78 kynpʾ Sogdian 207n80 kʷynæg Oss. 207n78 læsæg Oss. 258n58 cā́x̌ay Pashto 146n58 maṇá Pashto 269n77 δʾn Sogdian darəɣa- YAv. dombaj Oss. drājiiō OAv. nʾy MP 240 nāfa- YAv. 52 nāv Kurdish 207n79 293 169 221n107 169 fštāna- YAv. 248 pāeman- Av. 68 pasa- Khot. 146 gaδα- YAv. 146n60 gazīdan NP 99n139 gæn, gænæ Oss. 207 grd-, grdn MP 122n10 gzəm Bakhtiari 182 rbʾy MP 162 rrv- Khot. 162 rsn Parth. 121n6 ruvas Oss. 146n60 rwb’s Parth. 146n60 ɣal Pashto 146n60 ɣuzbe Zaboli 182 sæl- Oss. 54 staora- YAv. 251 stura- Khot. 251n41 haētu- YAv. 68n75 hnzwg MP 183n26 šīša NP 143 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 420 word index tāb- NP 58 taδarv NP 146 tašta- YAv. 146n60 tng Parth. 53 tštˈ Pahlavi 146n60 ttatara-, ttara- Khot. 146 ttura- Khot. 252 uruša- OAv. 170 vaēiti- YAv. 169n25 vārənjana-, vārəɣna- YAv. 179n16 vərəzuuaṇt- YAv. 169n26 vīsaiti YAv. 183n25 vizvā Khunsari, vızm Talysh 182 wʾrɣnʾk Sogdian, wʾrɣnyk Khwarezmian 179n16 wiz Gorani 182 wpc Pahlavi 55 wríže Pashto 192n43 wyst MP 183n25 xor Oss. 262n66 ᵆxsyrf Oss. 193 xwʾr MP 262n66 xwar Oss. 262n66 zairita- YAv. 52 ʒæbæx Oss. 253n45 ʒæbīdyr, ʒæbodur Oss. 253 Indo-Aryan Unmarked = Sanskrit agní- 3 ajá- 56, 284n3 aṃhú- 52, 83, 183n26 aratní- 272 ávi- 56, 284n3 āg Hindi 4 ā́rā- 39n4 ā́s- 272n84 ā́tā- 169n25 bīr Hindi 4 bhrā́jate 169, 191 bhrū́- 58, 58n45 bhūrjá- 169 bhr̥ ṣṭí- 224 cakrá- 122n10 dáhati 54 dárvi- 146n60 dehī-́ 300 dīrghá- 169 drā́ghīyas- 169 duhitár- 52 éti 51 gardabhá- 188n33 gáv- 284n3 haṃsá- 56 hári- 52 jámbha- 52 kekara- 170 kravyá- 80 kr̥ mí- 55 kr̥ ntáti 57n38 kṣobhate 220n104 kulva- 196 lumpati 162 luñcati Pāli 181n22 lupyáte 162 mádhya- 54, 80 mánthā- 57 myákṣati 170 nagná- 270n81 nápāt- 52 nás- 281n101 palā́va- 56, 293 párśāna- 208 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 421 word index paśú- 41, 41n9 pátra- 146n58 plúṣi- 184n29 pŕ̥ t- 30n51 psāti 262 púr- 300 tárati 78 tavás- 106 tīrthá- 78 tsárati 220 tsáru- 220, 220n103 tucchyá- 53 rákṣati 170 raśmā́ 121n6 rátha- 57 rāsabhá- 188n33 rūkṣá 170 ukhā́- 254 úpa-valhati 169n26 r̥ ṣabhá- 188 sásti 254n48 sphirá- 281 stambha- 60n55 sthūrá- 251n41 sūnā Hindi 4 sū́nu- 3 svásar- 52, 269 śarabhá- 188n33 śā́khā- 53 śéva- 40, 51n16 śvā́ 3 śvítna- 40 ū́dhar/n- 170 ūrdhvá- 169n26 ū́rj- 169n26 ū́rṇā- 56 vavákṣa 170 vāhaka- 99n139 vā́r- 87n117 viṃśatí 183n25 viṣā́ṇa- 242n19 vīrá- 3 vrīhí-, Kati wriċ 192n43 vr̥ ṣ́ an-, vr̥ ṣabhá- 188 yábhati 33n55 yáva- 293 yúdhyate 121 yuváti 169n25 yū́ṣ- 169n25 Tocharian laks B 259 tsaikaṃ B 300n22 mit B 141n48 wäs A 301n24 śak B 141n48 śeriye B 248n31 yasa B 301n24 yutk- A 121 tapre, tewpe B 211n90 tpär, top A 211n90 Anatolian Unmarked = Hittite ais 272n84 āssa CLuw. 272n84 hāran- 244 hāssā- 301 huhha- 254n48 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 422 word index karpina- 195n51 kis-ᶻⁱ 248 samlu- 298 ses-ᶻⁱ 254n48 maliddu- 184n28 milit 184n28 tama- HLuw., tãma Lycian 300 tarah-ᶻⁱ 78 tuhhae-ᶻⁱ 58 nāta- 240, 272 Uralic Languages Finnish (Karelian) ahdas 52 ahingas 105n160 ahjo 72 ahta- 52n21 airo 90 aita 66n70 aivot 66n70 angerva 104 ankerias 54, 103 ankkuri 25n46 ansa 57 apila, apelias 100 ativo, atima 51, 51n14, 81 buitto K 20 čirkku K 74 čivissä K 74 čongie K 74 ehkonen 55 haapa 288, 289n10 haara 59, 144 hako F, K 53, 53n25 haljakka, Ingr. haljas 52 halla 54 halli 54n26, 59n50 hammas 52 hanhi 56 harakka 95n128 harava 144 harmaa 59, 64 hehvo 55 heimakunda K 81 heimo 51 heinä 65, 126 heiti- 59 helle, K helleh 54 helma 128 helmi 92 herhiläinen 55 herne 56, 72 hieho 55 hihna 87, 88 hiiva 46 hiki 92 hinta 130 hirvas, K hirvaš 124 hirvi 55, 124 hiähky K 85 hoaśśa K 75 holkka 9n3 hosia 90 housut 69 hulkka 9n3 huone 46 hurtta 8 huttu F, K 76n95 huuhto- 88 hylje 112 hähnä 240n15 Häme 129 härkä 59, 113 härmä 59 inhiminen 130 Inkeroin Ingr. 18, 299n21 jo 58 jousi 66n69 jälki 249n33 järvi 86 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 423 word index kaari 99 kaarne 72 kaiho 48 kaima 50 kaiva- 48, 66n70 kala 94 kallas 106n163 kana 106n163 kansa 124n16 karva 55 kataja 84 kaula 122 kausta 46 keihäs 72 kelles, K kelleš 64n65 kelo 142 kelta 52n22 keltainen 52 keltalieko 54 kerta 119n2, 140n44 ketara 46 kevät 82 kieli 119 kieral K 53n23 kiiliäinen 55 kinnas 49n9 kirves 57 kreivi 68 kukka 76n95 kulo 108 kumpu- 48 kute- 94 kutsu- 120 kuurna 53n23 kuurne, K kuurnis 53n23 kuuru F 53n23 kuusi 84n109 kypärä 122 käki 120 kääppä 76, 77 käärme 55, 72 mahla 92 malo F, K 122, 135 maltsa 71n86 mehiläinen 107n163 mela 105n159 mene- 48 metsä 54, 67, 220n104 mettä 220n104 miero K 24 morsian, K moršien 51, 75 muila K 20 muna 4 muta 54n27 muula 20n34 muurahainen 55n31 mäntä 57 mäyrä 134n34 lahti lahto laiha laiva lampi lansi lapa napa 52, 119 nepaa 52 netäli K 24n43 nevat 52, 82 -niekka K 24 niided 57 niini 135n35 121 107 107 90 47, 156n10 48 101n149 lapio, lapia 101 lappa 76n95 laukama 77 laukas, K laukka 76 laukki 76, 77 laukko 76 lava 101 lavea, lavia 53n24 lehti 265 leinä K 65 leipä 65 leivo 67n73, 92 leppä 89, 102n153, 288 leuka 70n83 liha 92 liika 58, 68, 119 lohi 98 lohke-, lohjeta 60n54 loma 120n2 louhi- 60n54 lovi 130n26 lunka 125 luuta 59, 108 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 424 niisi 57 noki 126 norppa 112 nyhtää 139 nysi 48 ńorppi K 159n13 odelma, Ingr. oelma 104n158 oinas 56 olut 49 orehka 25n46 orsi 144 ota 104n158 otsa 71n86 pahla 57 paimen 56 papu 228 pato 94 peippo 67n73 pelu, pelut 56 pertti K 30, 141 pesä 4, 263 petäjä 85 pihlaja 85 piiraa 8 pippuri 25n46 pirtti F, K 76, 77, 141, 142 polkka F, K 9n3 populi 20n34 porkkana 32 puitto K 20 pulkka 8n3 puna, punainen 53n22 purje 90 puuro 60 pyörä 86n112 pähkinä 288 pälvi 120 račoi K 104 rahko 131 rako 60 ranne 121 ranta 50n13 rastas 49, 65, 103 ratas, rattaat 57 reikä 66 reisi 66, 69, 122 word index reki 104 reuna 58 rieska 105n159 roakie K 53n24 rouhi- 60, 60n54 routa 54 ruusu 69 rästäs 65, 103 saari 91, 92 sakara 46 salakka 97 salo 123 sammal 110 sampi 46, 219 sapa 74n91 saparo 74 seimi 68 seinä 66 seipi 74, 97 seiso- 66 seitsen 120n3 seiväs 60, 66, 74 seura 119n2 siemen 56, 119 siika 94 siipi 74 siivatta 8 sika 73 silakka 95 silli 97 silta 57, 73, 116, 119 sinä 73 sisar 52n19 sitkeä 73 siula 57 soappoa K 8 soimi 68n74 sora 262 suitset 120n3 suka 127 sulkku 17 suola 59 Suomi 129 suomus 94 suu 120n3 syvä 69, 73 särki 94 säyne 94 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 425 word index šikla K 57 šilta K 119 taaja 53, 119 takiainen, takkiainen 54, 81 takkisheinä K 81 talja 120 tammi 231 tapa 60 taula 54 tavea 53, 53n24 tavi 92 tavia 53n24 teeri 146 tempa- 58 terva 49 tetri 146 tiine 68 toakie K 53 toe 97 torvi 58 tuhat 105 tuhkuri 25n46 tuohi 48 turku 8 turska 255 tuulas, tuulaalla 58, 108 tuura 60, 108 tyhjä 53 tytär 52 tyttö, K tyttö, tytöi 52n18 uksi 47 vaaja 50, 119n1 vaapsainen 55, 55n31 vaha 106n163, 218n98 vahtera 288 vakka 102 vanha 72, 146 vannas 56 varas 100 varhainen 100 varpunen 103n154 varsi 143 vehnä F, K 239 veitsi 120n3 veräjä 24n43 viehkuri K 24, 25n46 vielä 61 viha 72 vihuri K 25n46 villa 56 voakie K 53n24 vohla, vohli 62 vuapsahane K 55n31 vuitti K 20 vuohi 56 vuona 56 vuonna 75 vuoro 59 vuota 49n9 ynnä 56n34 äes 50 ängeriäs K 103 ätälä 104 Veps adiv 51 aivod 66n70 bapshaine 55, 55n31 čiraita 74 hago 53n25 hähk 85 inehmoi 130 kaput 20n34 karni̮š 72 kauh, kouvaz 78n101 kurdeh 53 küud 54 lahk 107 lapak 76n95 pelu 56 perť 76, 141 pihľ 85n110 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 426 word index rouhi- 60 röun 58 silöi 69n80 sizar 52n19 tineh 68 tuľhuuda 58, 108 värpitada, värbitada 61 šuuk 17 Votic ahaz 52 apila 100n140 borkkana 32 harmaa 59 iiva 46 jarvi, järvi 87n115 kurrõ 53 kõlta 52n22, 54 leipä 67 paalis, pahlis 57n39 paimõõ 56 pihľpuu 85n110 põippõ 67n73 päälin, pählin 57n39 rassa 103 rõuta 54 räsäz 103 seiväz 74 siipi 74 siira 20n34 soola, śohla 57n39 suura, syyru 20n34 sõiso- 67n73 sõõra 20n34 takkiaz, takkiain 81 toho 56n33 tšääppä 76 tõgõ 97 vaapsaz 55n31 vaapsia 55 vadnaz 56 voho 56 värpo 103n154 võdna 56 õimo 67n73 õlud 49 ängeriä 103 ätälä 104 Estonian (Võro) aas 57 aed 66n70, 91 aer 90 ahing 105n160 aid Võ. 66n70 aim Leivu 65, 67n73 aivõq Võ. 66n70 angerjas 54, 103 arak 95n128 atma Võ. 52n21 eerus, Võ. eherüs 73 ere 128 haab 288 haah́ Võ. 56, 72 haar 59 hagu, haga 53, 82 hain Võ. 65 hakkama 100 haljas 52 hall 54, 54n26, 59n50 hammas 52 hani 56 harakas, Võ. harak 95n128 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 427 word index haŕm Võ. 59, 64 haro Võ. 59 hein 65, 126 heis 59 here 128 herilane 55 hernes 56 hiitä- Võ. 68n74 hind 130 hirv 55 hoonõh Võ. 46 hurt 8 hõim Võ. 51, 67n73, 68n73 hõimanõ Võ. 81 hõlm 128 hädal 104 häitse- Võ. 59 härg 59 härm 59 hüljes 112 ihn 87 iive 46 inemine Võ. 130 jo, ju 58, 71 judõr, jutr Võ. 100 jälg 249n33 järv 86 kadakas 84 kael 122 kaeva- 66n70 kaim 50 kala 94 kard 90n121 karv 55 kaust 46 kiil, Võ. kiińläne 55 kinnas 49n9 kirves 57 kodar 46 kold 54 kollane 52 kollõq Võ. 9 kulu 108 kurre, kurt 53, 77 kurt 99n138 kutsu- 120 kõllanõ Võ. 52 kõno Võ. 299 kõrts, Võ. kõrtś 32n54 kärg, käri, kärv 249, 249n33 kärm, kärv 55, 77, 99n138 kägi, kägu 120 kääbas 76 kübar 122 labidas 101 laev 90 lahja 107 laht 121 laib Leivu 65 lainalinõ Seto 65 lava 101 leht 265 leib 65, 67 leika- 67n73 lein 65 lepp 89, 102n153 liig 58, 68 luud 59, 108 lõhe, lõhi 98 lõiv Võ. 67n73 lõoke 92 lõug 70n83 lõõnõq Võ. 58n44 lõõts 105n159 lääts 202 maasikas, Võ. maaśk 72 malts 71n86 mets 54, 67 mugõl, mukl Võ. 20n34 mõla 105n159 mõrsja, mõrsija 51 mõts Võ. 54, 67 mäger 134n34 mähk 99n138 mänd 57 määr 134n34 naba 52 niied 57 nõbu 52 nõgi 126 nännüt Võ. 56n34 nätäľ Võ. 24n43 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 428 oinas, Võ. oinas, oonas 56 osi 90 ots 71n86 pahl 57 pahr, parh Võ. 56, 73, 78 pihl E, Võ. 85n110 pihlakas 85 pi̬it̬ ä- Võ. 68n74 pipar 14n17 pobul 20n34 poeg, Võ. poig 9 polk 9n3 porgand 229n122 pragu 60 puder 60 puri 90, 249n33 põrḱnas Võ. 32 pähkel 288 pähn Võ. 240n15 raastas 103 raig Võ. 66 raiź Võ. 66, 69, 122 rand 50n13 ranne 121 ratas, rattad 57 regi 104 reig 66 reis 66, 122 rõhu- 60 rõõnõq Võ. 58 rõõsk 105n159 rästas, Võ. rästäs 49, 103 saba 74n91 sagar 46 sain Võ. 66 saisa- Võ. 66 saivas Võ. 60, 66, 74, 74n92 salak 97 salu 123 samb 46 sammal 110 seeme 56 seiḿ Võ. 68 sein 66 seisa- 66 siga 73 word index siib Võ. 74 siig 94, 95 siil 69n80 siis 52n19 sikkõ Võ. 73 silakas 95 sild E, Võ. 57, 73 silk 95, 95n129 sina E, Võ. 73 sitke 73 soime 68n74 sool 59 suga 127 suidsõq Võ. 120n3 suu 120n3 suurmaq Võ. 261 sys Võ. 52n19 sysalik Võ. 52n19 sysaŕ Võ. 52, 52n19 sõim 68n74 sõisa- 67n73 sõra 262 sõsar 52 süvä Võ. 69, 73 taba 60 tael 54 taivikas 74n90 takjas, takijas 54, 80 tamm 231 tava 60 teder 146 teib 74, 97 teivas 60, 66, 74 teivikas 74n90 tiib 74 toht 48 tsibisemä Võ. 74 tsiga Võ. 73 tsirisemä Võ. 74 tsirk Võ. 74 tsunǵma Võ. 74 tuhat 105 tuhkur 15n20, 25n46 turg 8 tursk 255 tuur 60, 108 tõmba- 58 tõri 58 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 429 word index tõrv 49 täivikas 74n90 tötkes 94 tühi 53 tütar 52 uks 47 vaha 106n163 vaher 288 vahn Leivu 72 vai 50 vakk 102 vana 146 vannas 56 vapsik 55, 55n31, 77 varane, Võ. varahinõ 100 varblane 103n154 vars 143 vasak 146 veel 61 veer 137n40 vi̬il̬ Võ. 61 vill 56 vohl 62n61 voor 59 vu̬ u̬n Võ. 56 vähk 99 väi Leivu 99n138 värav 24n43 värb, värblane 103n154 õhv 55 õis, õitse- 59 õrs 144 ädal 104 äke 50 änger( jas) 103 ätäl Võ. 104 Livonian a’ddõl 104 a’g 53 aim 51, 65 āina 65, 126 akkõ 100 aļ 54n26, 59n50 āmbaz 52 aņgõrz 54, 103 a’ņgõz 105n160 a’r 59 arāgõz 95n128 borkõn 32 da’ggõl 54 īlgaz 112 imi Salaca 130 īnda 130 īņõ 56n34 īra 55 ja’mdõ 87n115 jērnaz 56 jǭra 87n115 ju 58 ju’ddõr 100 jõ 58 jämde Salaca 87n115 järu Salaca 87n115 ka’ggõl 122 kāima 50 kalā 94 kārda 90n121 kārnaz 72 ke’g 120 kibār 122 kīermõz 55 kil, kiļ 50 kīndaz 49n9 kīraz 57 kǭla 50 kōsta 46 kōva- 66n70 kriev Salaca 261 ku’l 108 kutsõ 120 kõ’ddõrz 46 kȭidaz 68n77 käpā 141n48 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 430 laib* Salaca 65 laik 76 lajā 107 lēʾḑ 265 liepā 89, 102n153 lǭja 90 lovā 101 lūdõ 59, 108 lȭga 70n83 lȭiga 68n77 lä’bḑi 101 madāl 104n157 mõltsõz, mõltsi 71n86 mõtsā 54, 67 mä’ggõrz 134n34 nabā 52 nīdõd 57 nīʾņ 87 nädīļ 24n43 ǭʾdõz 52 ǭla 54 ǭļaz 52 ǭz 57 paḑā 141n48 paint 56 palāndõks 50 piedāg 85 pǭʾlõz 57n39 pǭtõg 165 pubā 228 pūŗaz 90 rānda 50n13 rastā 49, 103 rat Salaca 57 rattõd 57 re’ggõz 104 rȭskõ 105n159 sāina 66 sārmā 129n25 sīemt 56 sīgõz 94 sīlda 57 siļk 95 word index sǭmal 110 sū 120n3 sugā 127 suiksud Salaca 120n3 sūol 59 sūrmed Salaca 262 sūrmõd 261 sõzār 52 tabār 74 tāibaz 60, 66, 74 te’ddõr 146 teib 74, 97, 165 tībõz 74 tidār 52 tijā 53 tīn Salaca 68 tǭgõz 97 tsīļ 69n80 tūʾoigõz 48 tūʾontõ 105 tūrska 255 tȭmbõ- 58 tȭra 49 täm 231 ukš 47 ūoņõz 56 ūrtapi’ņ 8 vaigā 50 vakā 102 vanā 146 vaps 55n31 varāz, va’rri 100 vaŗž 143 vēʾjõz 99 vēl, ve’l 61n59 vērbikšõ 61 vie’ddõl 104n157 vīla 56 vǭ’ 106n163 vȯʾl 49 vȱntsa 71n86 vȯŗž 144 vȯžā 90 vägāli 99 väji* Salaca 99n138 vǟrbõkš 61 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 431 word index õ’v 55 ǖnis Salaca 56n34 ä’ggõz 50 ǟrga 59 ǟrma 59 Western Sámi Unmarked = North Sámi áldu 86n114 barta 141 biebmat 120n2 bievla 120 boaris 101n146 buošši 125 čohkut 127 čuonjá 138n43 čuorpmas 127 čuožžut 66 daajvaj S 119 dađgat 73 daktere S, dakter Pite 126n22 darvi 49n12 dávjá 119 dealljá N 112n172 divttis 126n22 dullja Nw. Sámi 112n172 duollji 120 duovli 124 faggi 121 fihčču 72n86 gahpir 122 galda 64n65 gamhtse S 49n9 gaskkas, S gasnges 84n109 gavja 49n12 gáranas 72 geardi 119n2 gearhka S 126 giehka 120 giehpa 125 giella 119 gierdnas Torne 126 goahcci 85 gohččut guoibmi guosˈsi guossa guovllas 120 50 124n16 69n79 122 heargi 113 jávri 86 johtit 121 jorbmi 142n51 lasta 265 leaibi, S liejpie 89, 126, 288 liigi, S lijkie 119 livttis 126n22 loapmi 119n2 loggut 125 luokta 121 luossa 98 luovvi 101 mádjit 96 meahccánit, S meahtsanidh 124n17 mealli 105n159 miehtjiedidh, miehtsie S 124n17 miehttjen L, miehttjiedit Ume 124n17 muolos L 122, 135 náhpi, S naepie 119 neahpát, S neapede 82 nåervie 112 raaktse S 126n22 riessan, riessat, S rïesedh 121 ruoida 122 saajmie S 110 saertie S 121 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 432 word index sallit, L sallet 96 sarvva 123 sarvvis 124 sassne L 87 sáhppasat 88n118 sápmi 129 sárdi, L sárdde 122, 122n7 sealma S 128 seamul 110 sesnie S 87 siebrre L 119n2 siekkis 88n118 siepman 119 sjerrat L 128 sjåavonje S 128 slikte S 126n22 suhpi 288 suoidni 67n72, 126 suoldni 123 suolu 123 suorgi 144 suorri 123, 144 sämol L 110 šaldi 119 šearrat 128 šielbmá 128 šūvon 128 tjalle S 142 tjetskie S, tjaskie Ume 143 urries S 124 vagŋa L 121n5 vašši 72 veardi 144 vietka 125 vievgŋa 121n5 vuessie S 124 vuevjie S, vuojvve L 50, 119n1 vuoras 100 vuosˈsi 124 Eastern Sámi Unmarked = Skolt Sámi čõrmm, K čirrm 142n50 čuõrmâs 127 čååkkad 127 jåå´tted 121 jäu´rr 86 keä´pˈper 122 kõbjj 49n12 kõldd 64n65 kuei´mm 50 kū´ss K 124n16 kuvlâs 122 kuä´cˈcev 85 kåččad 120 lue´vv 101 luhtt 121 luõss 98 lyepi i 101n149 lå´ŋŋge K 125 meččin, mecci i 124n17 meä´cˈc 124n17 muâlas 122 nap̜p̜e Ter 119 nue´rjj 112 nää´pp 119 ǩiõkk 120 ǩiõpp 125 peâmˈmad 120n2 piõull 120 põrtt 141 puä´res 101n146 leä´pˈp, K lie´hp 89 li´jjg K 119 lõstt 265 riõzzâm 121 ruõddâs, i ruodâs 121 ruõidd 122 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 433 word index saa´rrd K 122 saa´vvŋ K 110 sââ´rves 124 sââu´ŋel 110 si̮elaj Ter 123 siävŋul i 110 sõrvv 123 sõvŋal K 110 suei´nn 126 sue´lnn 123 sue´rr 123, 144 sä´rdd 122 säu´nnj, i sävŋi 110 šaahpreš K 88n118 šâ´ldd 119 še´šnn, 87n118 ši´ŋŋg K 88n118 šišne i 87n118 taj̄va Ter 119 tâ´rvv 49n12 tue´lˈlj 120 tu´vll 124 täujja 119 vâ´ǧǧ 121 verdi i 144 viõtkk 125 võŋŋ 121n5 võhč K 72n86 vuõrâs, K vūras 100, 101n144 åå´res 124 Mordvin al EM 130, 130n27 al- EM 130n27 avkə̑s M 102n152 avto- E 139 äj M 86 äŕʿkä M 86 äźńä M 86 čana, čando E 130n28 čonda E 130 ej E 86 eŕke E 86 eźńe E 86 gžniva E 88 inže E, inži M 130 jäj, jäľ, jäľďä, jäŕʿkä M 86n114 kaldaz M 134 kardaz, kardo E, karda M 134 karks EM 130 karman M 230 kavto E 139 känǯä E 263 käŕaz M 133 kenže E 263 kerš E, kerži M 137 keŕaz E 133 kijov E 69n80 kirda E 140n44 kodorks E, kodərks M 46n4 kšna E 87, 135 kšta-, kšta E 88 kućkan EM 102n152, 230 kərda M 140n44 lango E, langa M 131 luŋgəďə- M 135n35 luv M 130n26 luvoďe- E 135n35 ľejks E 135n35 ľenge E, ľengä M 135 ľepe E, ľepä M 89, 135, 288 ľevš E 135n35 ľija E, ľijä M 136 ľijado- E, ľijadə- M 136 ľišme E, ľišmä M 113 malaso E, malasa M 135 miľä M 105n159 miŕkou̯ E 230n123 moda M 54n27 morkov, moŕʿkov E 230n123 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 434 word index ńäŕgaz M 134 ńefťə- M 139 ńeŕgaz E 134 ńevťa- E 139 paro E 132 panct, panst E, pandəz M 135 päkšä, päšä M 240n15 päšťä M 288 pejeľ E, pejəľ M 135 penge E, pengä M 130 piŕgińe E 136 potmo, potso E, potma, potsa M 132 praban M 230 pŕä M 136n36 puŕgińe E, puŕgəńä M 136 puŕʿkä M 32, 229 pusmo E, pusma M 35, 130 raśke E 132 raško E, raška M 131 roź EM 192 rudaz E 131 ŕeďa- E, ŕäďa- M 132 ŕiśme E, ŕiśmä M 121n6 sad E 145n56, 262n68 sal EM 137, 137n41 simeń E, śiməń M 132 sod EM 136 sokś E 109 stiŕ E 139 suro E, sura M 136, 262 śeď EM 136n37 śeńä M 86 śiməń M 132 śiŕe E 132 śiśem E, śiśəm M 136 śive E, śivä M 132 śokś E 109 śťiŕ EM 139 šäkši M 240n15 šenže, šenš E 138 šenžej, šenžij E 138 šna M 87, 135 šta-, šta M 87 talaj EM 133 truba EM 137 tumo E, tuma M 231 turba EM 137 turtov, turtoŋ E 132 tuvo E, tuva M 73 ťejťeŕ E 139 ťerďe- E, ťeŕďə- M 131 ťožəm E 109 ťožań E, ťožäń M 108, 135 uďem E, uďəm M 109n169 ukštor E 288 vakan E 102n152 veŕe E, väŕä M 137n40 viŕ EM 137 viš EM 239 ərdaz M 131 Mari ertak E 140n44 nerɣe E, nerɣə W 134 jer E, jär W 86 karaš E, käräš W 133 kož 84n109 küδər E 146n59 ńəkta- 139 lume-ɣož, W lə̑me-kož 84n109 ləštaš E, ləštäš W 265 pište E 240n15 pondo E 144 pört EW 140, 142 pŭrɣo Volga 144n55 püläɣerδə W 140n44 pükš EW 288 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 435 word index rüδaŋa- E 131n30 rə̑ža W 193n46 sə̑rmə̑ W 142 sə̑ts W 136 šaške E, šäškə W 85, 143 šište E 240n15 šopke E 288 šor-wondo E 144 šuk E 140n44 šukerte, W šukerδə 140n44 šurmaŋše E, šŭrmŏ 142 šü W 136n37 šüć E 136 šüδür E 145n56 šüj E 136n37 šülə W 69n80 šüštö E, šəštə W ‘strap’ 88 šəštə W ‘wax’ 88, 88n119 teŋə̑z E, taŋə̑ž W 133 tić 86, 137n37 tul EW 69n79 tumo E, tum W 231 tüž, tü.üž E 70 tüžem E, təžem W 109 urža E 193n46 užar E 142 waštar EW 288 wiste, wiśte E, wištə NW 239 wondo E 144 wurɣo E, wurɣə̑ W 144n55 wurδo E, wurδə̑ W 143, 144n55 wurɣem E, wə̑rɣem W 144 ərδäŋgeš W 131n30 ə̑rža W 193n46 ə̑žar W 142 Permic (K = Komi, U = Udmurt) ami̮ś, ameʒ́ K, ameź U 215 be̮ž K, bi̮ž U 262n68 bo̮ rd K, burd U 146n58 će̮rs K, ćers U 146n58 ćuš K 143 če̮ž KU 138 dar K, duri̮ U 146n60 ež K 88n119 e̮n K 69n79 e̮s- K, e̮ski̮- U 262n68 gu- K 146n60 jir, jər K 142n51 kač-pomeľ K 85 karas U 133 karta K 134 ke̮z K 69n79 ki̮rni̮š K 72 komiʒ́, ku·mić K, kumi̮ź U 248n29 lo̮ - K 145n57 ma K, mu U 145 ńin K 135n35 o̯ K 145n56 omeź U 215 pašpu U 288 puz U 263 rode̮g K 131n30 ruć K 146n60, 192 ruʒ́eg̮ K 192 sa K 136 si̮lal, slal U 137, 192, 193 so̯ l K 137 sot- K 136n37 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 436 word index sov K 137 su U 136 suti̮- U 136n37 śer K śiś K śor U śuś U 142 88n119 142 88n119 tar K taśti K te̮v K tur U tuśti̮ U 146 146n60 69n79 146 146n60 va K 145 wapum U 145n56 važ K 145 vež K 142 ve̮ź- K vo̮ - K vo K vož U vu U vur- K vuž U 120n3 145n57 145n56 142 145 144 145, 146n62 za K 145n56, 262n68 ze̮r KU 262 ziľ K 262n68 zu U 145n56 ʒ́aźeg U 138n43 ʒ́eg U 192 ʒ́ići̮ U 146n60, 192 ʒ́iźeg U 192 ʒ́oʒ́eg̮ , ʒ́u̇·ʒ́ok K 138n43 Other Uralic (Hu. = Hungarian, Kh. Khanty, Ms. = Mansi) åml Ms. 69n79 basa Ngan. 301n24 burgonya Hu. 230n124 ća’n Kamas 86 čač Kh. 138n42 hagyma Hu. 248n29 joŕa Nenets 87 jur, jǫr Kh. 142n51 kȧ̆ɣʷi Kh. 120 kajbu Ngan. 48 käp, kämp Ms. 77n98 kǟšäp Ms. 85 kε̮p, kε̮mp Ms. 77n98 kē̮si̮ Selkup 301n24 koče Ngan. 48 koŋhu Ngan. 48 kori̮ Selkup 87 kŏs-, χŏs- Kh. 66n69 kɔśśm, χōsman Ms. 248n29 χɔwt Ms. 69n79 köäsp Ms. 85 köɣi Kh. 120 kuns Kh. 263 kutər Kh. 146 kwə̑dəgej Selkup 146 kwə̑ndəj, kwə̑ndəga Selkup 146 lengyel, lengyen Hu. 19n31 li̮ntə Ngan. 48 li̮ńť-, ᴧɔńś- Kh. 66n69 lūp Ms. 101n149 ľüŋhə Ngan. 48 mači̮ Selkup 54n27 mi̮nsi̮ Ngan. 48 mumus Hu. 29 ńir Ngan. 48 poč, pŏč Kh. 138n42 rég, régi Hu. 100n144 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 437 word index same Enets, sarḿikᵊ Nenets 142n50 sās Ms. 138n42 szül, sün Hu. 69n80 śurəm Kh. 142 šē̮š Ms. 138n42 šɔš Kh. 138n42 tit Mator 86 tȫt Selkup 86 tȫti̮- Selkup 86 tuńś- Ms. 66 ťurəm Kh. 142 ügyes Hu. 262n68 varr Hu. 144 vés Hu. 120n3 wöɣ, wȧ̆ɣʷ Kh. 120n4 Other Languages Turkic (Mongolian, Tungusic) amač Turk., Uighur 215 anja Manchu, anjis Mongolian 215 apsaq, ausaq Siberian Tat. 278 arïš Tat., Bashkir 193 avăt- Chuv. 278 ăvăs Chuv. 278 barq OTur. 140, 140n46 bāš OTur. 141 bïrt Yakut 140 čäške Tat. 143 čiša Tat. 143 čišta Tat. 143 dalaj Kyrgyz 134 diŋgez Tat., deŋiz Kyrgyz 133 dommaj Karachay 221n107 elmä Tat. 235 elme Kumyk, elmen Noghai 235 emen Noghai 235 gömbä Tat. 236n7, 256 gün Turk. 236n7 izep Khakas 242 ïhïr̄ Yakut 242 ïraš Chuv. 193 ̄ Yakut 141n47 ïrt- jĕlme Chuv. 235 karas Chuv. 133 karta Chuv. 134 kămpa Chuv. 236n7, 256 kăn Chuv. 236n7 kănčala Chuv. 236 käräz Tat., käräδ Bashkir 133 kärtä Tat., Bashkir 134 kĕnčele Chuv. 236 kĕrpe Chuv. 133, 256 kirtä Tat. 134 kïrt- Yakut 140n47 ̄ Yakut 141n47 kïrt kŏmpa Chuv. 256 könǯälä Tat. 236 kör̆ pe Chuv., körpä Tat. 133, 256 Kureyş Turk. 197n59 kürtük Shor, kürtkü Khakas 146n59 os Khakas 278 öt- OTur. 278 papaz Turk. 202n64 parmak Turk. 141 părăx Chuv. 144n55 porńa Chuv. 141 pört Chuv. 140 pürńe Chuv. 141 pürt Chuv. 140 püśek Chuv. 141 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 438 word index salat Chuv. 193n47 sarā- Yakut 141n47 sïrdā- Yakut 141n47 sof Turk. 242 śurt, śort Chuv. 141 šaškĕ Chuv. 143 šäške Bashkir 143 šešä Tat. 143 šišta Tat. 143 talaj Tat., Kazakh 134 tilmač, OTur. 236 tïlbās Yakut, tïlmač Tat. 236 usaq Tat. 278 zerbaf Turk. 242n21 zümrüt Turk. 242n21 ǯuğutur Karachay 253n45 Afroasiatic *a-baw Berber 228n118 bjt Egyptian 227 qnpʾ Syriac 207 qunnabu Akkadian 207 churú Ma’a 252 croompə, crampə Coptic 189 *sin Berber 252n43 sor Geʿez 252 ebiô Coptic 227 šūru Akkadian 252 gr-(n)-p.t Egyptian 189 twr Aramaic 252 hrêrə Coptic 189n37 ḥrr.t Egyptian 189n37 ṯawr Arabic 252 imlīq Arabic 117n178 *a-zgăr Berber 252n43 ‘Caucasian’ domba Georgian 221n107 a-domp’éj Abkhaz 221n107 kart Ingush 134 marχ̄ Lak, marqʷa Dargwa 231n126 msxali Georgian 298n20 šəqʷłtər Kabardian 253n45 ǯixvi Georgian 253n45 Other aruda Sumerian 233 uruda Sumerian 233 kuʾs Ket, Yugh 113n174, 114n174 zilhar Basque 226 θevrumineś Etruscan 253 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0 Anthony Jakob - 978-90-04-68647-2 Downloaded from Brill.com 12/30/2023 02:55:17PM via Open Access. https://creativecommons.org/licenses/by-nc-nd/4.0