The Accentuation of Balto-Slavic Vrddhi Formations

Ranko Matasović University of Zagreb E-mail: [email protected] THE ACCENTUATION OF BALTO-SLAVIC VRDDHI FORMATIONS1 1. Theories of the origin of Balto-Slavic acute There are three current theories about the origin of the Balto-Slavic acute intonation: a) the traditional theory: acute develops on all long vowels, whether apophonic or secondarily lenghtened after the loss of laryngeals (e.g. Carlton 1991). b) The Leiden school approach (Kortlandt 2011, Derksen 1996, Pronk 2012): the acute develops from glottalization, which in turn occurs on vowels preceding laryngeals and voiced (< glottalized) stops (by Winter's law). All apophonic long vowels are circumflexed. c) Villanueva-Svensson's theory: apophonic long vowels are circumflexed in non-initial syllables and monosyllables, but acuted in initial syllables. Vowels lengthened by the loss of laryngeals are acuted (Villanueva Svensson 2011). d) The present writer's opinion (Matasović 2008): the Leiden school is correct with respect to PIE lengthened grade vowels, which are circumflexed. However, new, morphologically derived lengthened grades in Balto-Slavic receive the acute. This is the ‘Balto-Slavic Vrddhi’, to which this paper is dedicated. 2. Vrddhi in PIE? Vrddhi is a formation of denominal adjectives by lengthening the root syllable. The process is best attested in Indo-Iranian, especially in Sanskrit, cf. Skt. sákhi- ‘follower, friend’ vs. sākhyá- ‘society’, víś- ‘village, settlement’ vs. váiśya- ‘member of the vaiśya caste’, mānu a‘connected with men’ vs. mánu a- ‘man’. 1 This paper was read ad the conference of the Indogermanische Gesellschaft in Leiden, in July 2013. Thanks to Tijmen Pronk, Frederik Kortlandt and Sasha Lubotsky for their comments. 1 A possible PIE instance of Vrddhi is the word for ‘egg’: PIE *h2ewi- ‘bird’ > Lat. avis, Skt. ví- : PIE *h2ōwyom (‘that of the bird, bird’s’ >‘egg’) > Lat. ōvum, OHG ei, ON egg (< PGerm. *ajjaz- n), Croat. jâje (Novi), Pol. obs. jajo, jaje, ULus. jejo < PSl. *âje, (AP c, Derksen 27, ESSJa I: 61-2). Alternative reconstruction of the word for ‘egg’ may be PIE *h2oh2w(y)o-, with the first reduplicated syllable, as in PIE *h1oh1k’u- ‘quick’ vs. *h1ek’wo- ‘horse’, *kwekwlo- ‘wheel’ (< *‘turning’) from *kwel(H)‘turn’. If this is accepted, there is no need to posit a Vrddhi formation. 3. Balto-Slavic Vrddhi? Collection of (mostly implausible) examples of the derivational lengthened grade can be found in Gołąb (1967). More probable cases have been collected below, and they share three defining features: firstly,they contain either an acuted long vowel, or an acuted diphthong, and are derived from roots that do not end in laryngeal. Secondly, they usually have a derived meaning with respect to the meaning of the base noun. Their meaning is usually possessive, relational, or collective: ‘belonging to X’, ‘pertaining to X’, ‘descending from X’, or ‘a collection of X’es’, where X is the base noun. Finally, formations with Balto-Slavic Vrddhi usually do not show evidence for lengthened grade except in Balto-Slavic. Here is the relevant material: 1. PSl. *bérmę (a) ‘load, burden’ (OCS brěmę, Russ. dial. berémja, Pol. brzemię, Croat. brȅme) < PIE *bher- ‘carry, bear’ (Lat. fero, Gr. phérō, Skt. bhárati, etc.), cf. Skr. (L sg.) bhārman ‘by bringing’, RV 8,2,8 < *bhērmen- (NIIL 16). Alternatively: PSl. *bérmę, Skr. bhárīman- ‘burden, maintenance’ < *bher-H-men- (but the suffix –īman- in Skr. cannot be used as evidence for the laryngeal, cf. dhárman- besides dhárīman- ‘support’ from *dher‘support, fix’, IEW 252-3; Wackernagel-Debrunner 1954: 756). 2. Lith. káimas, káima ‘village’ vs. ki mas ‘farmstead’, from the root of Goth. haims ‘village’. Ultimately, they may be from the same root as the preceding etymon. Derksen (1996) explains the acute of káima(s) by metatony, pointing out that the reflex –ai- for original *-oymeans that the root was originally unstressed, so that the acute may be attributed to the retraction of the stress from the last syllable. If this is correct, the circumflex of ki mas would 2 be original, and Lith. ki mas and káima(s) would represent PIE *kóymo- and *koymórespectively. 3. PSl. *kórsta (a) ‘crust’ (Russ. korósta, Pol. krosta ‘pustule’, Croat. krȁsta) may be a Vrddhi derivative from the root of Lith. kar ti, kar iu ‘card, comb (wool)’, Latv. kārst ‘id.’ < PIE *sker- ‘cut, scratch’ (OHG skerran ‘scratch’, Lat. carro ‘card (wool)’, De Vaan 95). 4. PSl. *lápa (a) ‘paw’ (Russ. lápa, Pol. łapa, Croat. dial. lȁpa, Slov. lápa 'snout', ESSJa XVI: 26-28), Lith. dial. lópa (1) ‘paw’, Latv. l pa ‘paw’. Like Goth. lofa ‘flat of the hand’, this appears to go back to PIE *leh2p-, or *lōp-, which may be a Vrddhi formation to *lop- seen in OCS lopata ‘shovel’, Russ. lopáta, Croat. lòpata (ESSJa XVI: 39-43), Lith. l pas ‘leaf’. However, Lith. lópeta ‘shovel’ and Latv. lâpsta show the word for ‘shovel’ with the long vowel and the acute (Smoczyński 363). Derksen (268-269) proposes that there were two different roots, *leh2p- (PSl. *lápa, Lith. dial. lópa) and *lop- (> OCS lopata, Lith. l pas), but this seems like an ad hoc solution. Latv. lępa ‘paw’ proves that we are indeed dealing with a PIE lengthened grade (*ē) rather than a root in laryngeal. 5. PSl. *lípa (a) ‘lime-tree’ (Russ. lípa, Cz. lípa, Pol. lipa, Croat. lȉpa, Bulg. lipá, Slov. lípa, Derksen 279, ESSJa XV 114-116), identical to Lith. líepa, Latv. li pa. PIE *leyp- ‘smear, glue’ (Skt. limpáti, Lith. lìpti, limpù): *loypos ‘glue’ (PSl. *lěpъ (c), cf. CSl. lěpъ, Cz. lep, Croat. lijêp, Derksen 273), and a derivative *lēypos ‘sticky’, substantivized as *lēypā ‘sticky one’ > ‘sticky tree’. 6. Lith. lúobas ‘bark’, Latv. luobas ‘id.’: Vrddhi derivatives of the root *lewbh- / *lubh- > Goth. laufs ‘leaf’, Lith. lubà ‘plank’, Latv. luba ‘linden bark’, Lat. liber ‘bark’, perhaps also in ORuss. lъbъ ‘front of the head, skull’. A Slavic parallel *lúbъ (with the acute) is found in Croat. dial. lȕb ‘bark’ (Vodice, Istra), lȕba ‘lump’ (Istra, Rijeka), Russ. lúb (G sg. lúba). Croat. lȗb (G sg. lȗba) points to a circumflex. PIE *lowbho- ‘bark’: BSl. Vrddhi *lōwbho- > *lāuba- ‘(made of) bark’ > Lith. lúobas and PSl. *lúbъ. 7. PSl. *pálica ‘stick, staff'’ (OCS palica, Russ. pálica ‘club’, Cz. palice ‘baton’, Croat. pȁlica, Derksen 390) vs. PSl. *políca ‘shelf’ (CSl. polica, Russ. políca,Cz. police, Croat. pòlica, Derksen 410), cf. also Russ. pol ‘floor’ (< *‘plank’) < PIE *(s)pol- ‘plank, staff’ (OIc. fjǫl ‘plank’, Latv. spals ‘handle’, perhaps also Skt. phálakam ‘plank’. Another Vrddhi3 derivative could be PSl. *pálьcь ‘finger’ (CSl. palьcь, Russ. pálec, Cz. palec ‘thum’, Croat. pȁlac ‘thumb’, Derksen 390), which has been related to Lat. pollex ‘thumb’; *polo- (> PSl. *polъ) ‘staff, plank’ : *pōlo- (> *palica, perhaps *palьcь). 8. PSl. *pítja (a) ‘nourishment, food’ (OCS pi ta, Russ. pí ča, Croat. dial. pȉća, Cz. píce ‘fodder’, Derksen 401) < PIE root *peyt- (Lith. pi t s (N pl.) 'dinner', OIr. ithid ‘eats’, Skr. pitú- ‘nourishment’). Derksen's assumption that the acute is due to the contamination with the root *peyH- (Skt. pīvan- ‘fat’) is ad hoc. 9. Latv. si va ‘wife’ vs. OHG hīwo ‘husband’, Lat. cīvis ‘citizen’ and Skt. śivá- ‘dear’; further connections to the root *k'ey- ‘lie’, or the deictic particle *k'i- ‘this, here’ are possible, but problematic. ? *k'ey-wo- ‘local, member of the local community’ (De Vaan 116) : *k'ēywo‘belonging to the local community, own’ > ‘(own) wife’. 10. PSl. *sláva (a) ‘glory’ (OCS slava, Russ. sláva, Croat. slȁva, Cz. sláva, Pol. sława, Derksen 453), Lith. dial. (Žemaitian) lóvė. As Pronk (2012: 18-19) points out, many (but not all) derivatives from the root *k'lew- in Balto-Slavic are acuted, so one must count with the possibility that a laryngeal was added to that root as a dialectal innovation. In Lith. kláusti 'ask' the laryngeal may be a part of the desiderative suffix *-Hs-, and this may also be the source of the acute in PSl. *slúšati ‘listen’ (OCS slu ati, Russ. slú at', Croat. slȕ ati, Pol. słuszać, Derksen 455), *slýšati ‘listen’. Note, however, Latv. klàust ‘ask’ without the acute, pointing to the conclusion that the acute in Lith. kláusti is secondary, as assumed by LIV. It may have been introduced to avoid homophony with klausýti ‘listen’, where the root is not acuted, cf. 3 sg. pres. kla sė ‘he listened’ vs. kláusė ‘he asked’. BSl. *k'lowo- ‘fame’ (East Lith. l vė, lavė, Latv. slava, slave ‘fame’, unless these were influenced by –a- in Slavic cognates) vs. *k'lōwo- ‘having fame’ >> *k'lōwā > PSl. *sláva ‘famous deeds, glory’. 11. Lith. sróvė (1) ‘stream’ (besides srovė (4), both forms in LKŽ): a Vrddhi derived from the same root as Lith. sra ja ‘stream’ (OCS struja, Croat. strúja, Russ. strujá, etc., Vasmer III: 32-33), from PIE *srew- ‘flow’ (Skr. srávati, Gr. rhéō, etc.)? *srōwo- ‘flowing (water)’: *srowo- ‘flow’ (Gr. rhóos, Skt. srāva-, OCS –strovъ in ostrovъ ‘island’). Derksen (1996: 59) considers this to be an instance of métatonie douce and derives srovė (the only form he cites) from the root of sr ti ‘flow’. However, there is no evidence for a laryngeal in the root *srew-. 4 12. PSl. *tą a (a) ‘(snow-)storm’ (Russ. túča ‘dark cloud’, OCS tąča ‘snow-storm’, Croat. tȕča ‘hail’, Vasmer III: 158-159), derived from the same root as Lith. tánkus ‘thick’, Skr. ta c- ‘be solid’, MHG dīhte ‘thick’. The acute in Balto-Slavic seems to point to Vrddhi (*tānk-jā > PSl. *tą a). LIV reconstructs the root as *temk- because of Hitt. tamekzi ‘fixes’. 13. PSl. *ú(s)tro (a) ‘morning’ (OCS utro, Russ. útro, Pol. jutro, Croat. jȕtro) may be a Vrddhi formation built to the same root as Lith. au rà ‘dawn’ (PIE *h2ewsōs > Gr. Hom. s, Lat. aurōra, etc.). 14. Lith. vìlkė ‘she-wolf’. Pronk (2012) justly points out that this cannot be an old formation because of its fixed initial acute in light of Skt. vrkī. Precisely: if this is an instance of Vrddhi with respect to Lith. vilkas 'wolf', it has to belong to a younger stratum, cf. also Lith. žùikė ‘she-hare’ vs. žu kis ‘hare’). It probably belongs to the same, Balto-Slavic stratum as the following item. 15. PSl. *vórna ‘crow, corvus corone’ (Russ. voróna, Bulg. vrána, Croat. vrȁna, Cz. vrána, Vasmer I: 229), Lith. várna appear to be Vrddhi formation with respect to PSl. *vornъ ‘raven, corvus corax’ (Russ. vóron, OCS vranъ, Croat. vrȃn, Cz. vran, Vasmer I: 228) and Lith. varnas. The similarity with Gr. kóraks ‘raven’ and kor n ‘crow’ is probably accidental, and does not testify to the difference of suffixes (masc. –no- vs. fem. *-Hno-). In Baltic (though not in Slavic) this pattern of opposing masculines to feminines derived from the same root must have been productive, cf. Lith. ernas ‘wild boar’ vs. érnė ‘wild sow’, ántis ‘duck’ vs. a tinas ‘drake’, cf. Petit 2004: 174-176). 16. PSl. * ętva ‘harvest’ (OCS žętva, Russ. žátva, Cz. žatva, Croat. žȅtva, Cz. žatva, Vasmer I: 411) vs. * ęti ‘reap, mow’ (OCS žęti, Croat. žȅti), Lith. genėti ‘prune, hem’, Derksen 561. The PIE root is *gwhen- ‘strike’ (Hitt. kuenzi, Skt. hánti, Gr. theínō). ? *gentwo- ‘striking, mowing’ : *gēntwo- ‘(time) of the mowing’ > ‘harvest’. Any objective discussion of this material would have to admit 1) that the instances of acuted lengths in possible Vrddhi derivatives are not numerous, 2) there are very few exact lexical cognates in Baltic and Slavic (*vórna and várna, *sláva and lóvė, *lápa and *lópė, *lípa and líepa), and 3), as noted by Petit (2004: 179ff.), most of the examples of alleged BSl. Vrddhi do not involve long vowels, but rather diphthongs opposing acute intonation to the 5 circumflex. The explanation of this opposition offered by Petit for Baltic might work for Balto-Slavic as well: En d’autres termes, un degré long morphologique ē peut avoir en baltique, dans les voyelles, une existence distincte la fois du degré bref e et du degré long d’origine glottale , tandis que, dans les diphtongues, un degré long morphologique ne peut se distinguer du degré bref, s’il aboutit une intonation douce er , ou du degré long d’origine glottale, s’il aboutit une intonation rude ér . Il me semble que, dans ces conditions, l’économie du système favorise l’assimilation du degré long morpholoqigue plut t au degré long d’origine glottale (d’o degré bref (d’o ér dans les deux cas) qu’au er dansl es deux cas) : cette dernière hypothèse emp cherait toute possibilité d’un degré long morphologique distinct du degré bref, dans les radicaux diphtongue du baltique. (Petit 2004: 180-181). Here is how we can represent this development: short (base) CeC CeRC long (derived) acute (base or derived) CēC C C CéRC Roots with vowels (CEC) Roots with diphthongs (CERC) If this is correct, then the acute, e.g. in Lith. líepa and PSl. *lípa does not imply a proto-form *l ypā; rather, the derived form *léypa- ‘sticky’ was opposed to the base form of the root *le p- *la p- ‘glue’. After this pattern, the association of the acute with the derived morpheme was transferred to proper vowels, so that the pattern *a (base) : *ā (derived) was established, e.g. in PSl. *lopa vs. *lápa, or *polica vs. *pálica. 5. Root nouns with lengthened grade in Balto-Slavic? Most root-nouns in BSl. with cognates in other branches of IE are i-stems, cf. OCS myšь ‘mouse’ vs. Lat. m s, m ris, OCS no tь ‘night’, Lith. naktìs vs. Lat. nox, noctis, etc. (Larsson 2001). 6 In some original i-stems we have the lengthened grade in BSl., and the root vowel is regularly non-acuted; however, in words belonging to AP (c) the acute may have been eliminated by Meillet’s law. 1. PSl. *dalь ‘distance’ (Russ. dal’, Pol. dal, Croat. dâlj, ESSJa IV: 186-7), from the root of *dьliti ‘last’ (Russ. dlít’sja, Cz. dlíti, Derksen 133). The connection with PIE *dlh1gho- ‘long’ (Skr. dīrghá-, Gr. dolikhós) is possible, but uncertain. 2. PSl. *granь ‘edge, boundary’ (Russ. gran’, Pol. grań, Vasmer I: 304) and *grana ‘branch (Croat. grána ‘branch’, ULus. hrana ‘edge’, ESSJa VII: 106-107); the lengthened grade points to the vocalism of the Nom. sg. (PIE *grōn-); the o-grade is preserved in OHG grana ‘beard’, and the e-grade in OIr. grend ‘beard’. 3. PSl. *rě ь (c) ‘speech’ (OCS rěčь, Russ. reč’, Pol. rzecz ‘thing’ Croat. rijêč ‘word’, Derksen 434). From the root of *rekti ‘say’ (OCS re ti, Croat. rèći, etc.). 4. PSl. *mělь (beside *mělъ) ‘sand bank’ (Russ. mél’, SerbCSl. mělь ‘chalk’, Pol. miel ‘shallow water’ Vasmer II: 115), Lith. smėlis, smėl s ‘sand’, cf. also Germ. dial. māle ‘der feine Staub der Landstrassen’ < *mēl- and ON melr ‘sand bank’ < *mel-. 5. PSl. *tvarь (c) ‘creation, creature’ (OCS tvarь, Russ. tvar’, Pol. twarz ‘face’, Croat. tvâr), parallel to Lith. tvorà ‘fence’. The same root is attested in *tvorъ ‘creation’ with the full grade. 6. PSl. * alь ‘grief, pity’ (OCS žalь ‘tomb’, Russ. žal’, Pol. żal, Croat. žȁo, Derksen 554), Lith. gėlà ‘acute pain’, from the root of OHG quāla ‘violent death’ (< *gwēlH-) and OIr. atbaill ‘dies’ < *gwelH- (IEW 471). However, we also find some ā-stems that are good candidates for root-nouns,2 cf. OCS gora ‘mountain’ vs. Lith. girià (2) ‘wood’ < PIE *gworH- (Skt. girí- ‘mountain’). These do not appear to have direct cognates in Baltic. Some such ā-stems show the long vowel in the root syllable. 2 See Matasović 2014: 21-24. 7 7. PSl. *trav (b) ‘grass’ (OCS trava, Russ. travá, Croat. tráva, Cz. tráva, Derksen 496) < PIE *trewH- ‘rub, spend’ (Gr. trýō ‘wear down, exhaust’, OCS tryti ‘rub’, Lith. tr nėti ‘spoil, putrefy, decay’, Beekes 1514, LIV *trewH-). Perhaps *trava is a deverbal formation based on *travìti (b) 'digest, feed with grass' (Russ. travít' 'exterminate by poisoning', Pol. trawić 'digest', Croat. tráviti 'feed with grass'), rather than vice versa as assumed by Derksen 496. In its turn, *traviti is an intensive/iterative of *truti 'feed' (OCS natruti, ORuss. truti 'consume', Pol. truć 'poison'), and Slavic intensives/iteratives regularly have the circumflex root (cf. PSl. *davìti 'suffocate', Russ. davít', Croat. dáviti < PIE *dhōh2u- , ESSJa IV: 198-199, Derksen 97, PSl. *dirati 'touch' (Croat. dírati, originally an intensive formation from the same root a s *derą, *dьrati). 8. PSl. *děra ‘crack, hole’ (ORuss. děra ‘opening’, Cz. díra ‘hole’, ESSJa V: 12), from the root *der- ‘flay’ (OCS dьrati,Lith. dìrti, Gr. dérō). 9. PSl. *dira ‘crack, hole’ (OCS dira, Russ. dirá, Croat. dial. dìra, Derksen 107, ESSJa V: 3031); from the root *der- ‘flay’. The AP cannot be determined. 10. PSl. *kara ‘punishment’ (Russ. kára, Pol. kara, Croat. kára, ESSJa IX 151); derived from the root of *karati ‘punish’ (Russ. karát’, Pl. karać, Croat. kárati), which is from *koriti (Croat. kòriti ‘reproach’). In Baltic, we find several ā-stems with long non-acuted vowel, e.g. Lith. lomà (2/4) ‘hollow’, bylà ‘case, speech’, gėlà ‘pain’, etc. Pronk (2012: 9) thinks these are best derived from old collectives. However, this type is very rare in other IE languages. Pronk cites only Lat. cella (derived by ‘littera-rule’ from *cēla, from PIE *k’el- ‘hide’, cf. OIr. celid), which he thinks is the regular development of the collective *k’el-h2 > *k’ēl-h2, with a regular lengthening in monosyllables. I find this too speculative, not only because there are too few lengthened grade ā-stems in other IE languages (even cella could represent *kelsā or *kelnā, among other things), but also because the lengthening in monosyllables is not a sound law established beyond doubt. Larsson (2004a-b): the long circumflex vowel in Lithuanian ē-stems is often the result of a retraction from the following syllable (*VC-íyā > 8 :C-iyā), and this pattern was extended analogically to many ā-stems, which are often parallel formations to ē-stems, with little difference in meaning, cf. Latv. tvāre ‘fence’ vs. Lith. tvorà ‘id.’ (from tvérti ‘close’), Lith. bėgė ‘run’ vs. Lith. begà ‘id.’ (from bėgti ‘run’), Lith. piovė ‘cutting’ vs. piovà ‘id.’ (from piáuti ‘cut’), Lith. k vė ‘fight’ vs. kovà ‘id.’ (from káuti ‘strike’), cf. Larsson 2004b: 166. However, Larsson’s lengthening (and the analogical spread to ā-stems) will allow us to explain the long vowel in Lith. lomà, tvorà, etc., but not in Slavic *lamъ ‘hollow, bend’ (Russ. dial. lam ‘wasteland’, Pol. łam ‘quarry, bend’, Croat. dial. lâm ‘knee-joint, underground passage’, Slov. dial. lam ‘quarry’, Derksen 268), *tvarь ‘creature’, etc., since there was no parallel retraction of the ictus in Slavic that would trigger the analogy. The Slavic nouns thus probably represent old root-nouns. It is also possible that Balto-Slavic preserved a number of root-nouns with lengthened grade in the Nom. sg. and that some such nouns became ā-stems either in Balto-Slavic, or separately in Baltic and Slavic. In Slavic, these nouns mostly belong to AP c), so it is impossible to establish whether they were originally acuted or circumflexed. In Baltic, long vowels in the root of circumflexed ā-stems can always be the result of Larsson’s lengthening. Therefore, it is impossible to establish the original accentuation of root nouns in BSl. In our opinion, then, Balto-Slavic Vrddhi is not inherited from PIE; rather, it is a parallel innovation in word-formation, similar to, but independent of Indo-Aryan and, possibly, PIE Vrddhi. REFERENCES Burrow, Thomas 1968. The Sanskrit Language, London: Faber and Faber. Carlton, Terence 1991. Introduction to the phonological history of the Slavic languages. Columbus: Slavica. Darms, Georg 1978. Schwäher und Schwager, Hahn und Huhn. Die rddhi-Ableitungen im Germanischen. München: Münchener Studien zur Sprachwissenschaft. Derksen, Rick 1996. Metatony in Baltic. Smsterdam: Rodopi. 9 Gołąb, Zbigniew 1967. The traces of vrddhi in Slavic , in: To honor Roman Jakobson: Essays on the occasion of his seventieth birthday, 11 october 1966, The Hague & Paris: Mouton, 770-784. Kortlandt, Frederik 2011. Selected Writings on Slavic and General Linguistics, Amsterdam & Atlanta: Rodopi. Larsson, Jenny Helena 2001. Proto-Indo-European root nouns in the Baltic languages , in: M. E. Huld et alii (eds.) Proceedings of the 12th annual UCLA Indo-European conference, Los Angeles May 26-28, 2000. Washington: Institute for the study of man, 50-64. Larsson, Jenny Helena 2004a. Metatony and length in Baltic , in: A. Hyllested et al. (eds.) Per aspera ad asteriscos. Studia Indogermanica in honorem Jens E. Rasmussen, Innsbruck: IBS, 305-322. Larsson, Jenny Helena 2004b Length and Métatonie douce in Baltic Deverbative Nouns , in: J. Clackson and Birgit Anette Olsen (eds.) Indo-European Word Formation, Copenhagen: Museum Tusculanum Press, 159-170. Matasović, Ranko 2008. Poredbenopovijesna gramatika hrvatskoga jezika, Zagreb: Matica hrvatska. Matasović, Ranko 2012. Dybo’s Law in Celtic , Zeitschrift für celtische Philologie 59/2012: 129-141. Matasović, Ranko 2014. Slavic Nominal Word-Formation, Heidelberg: Winter. Petit, Daniel 2004. Apophonie et catégories grammaticales dans les langues baltiques. Leuven: Peeters. Pronk, Tijmen 2012. Indo-European long vowels and Balto-Slavic accentuation , Baltistica 47/2, 205-247. Rasmussen, Jens E. 2007. The Accent on Balto-Slavic Monosyllables , in: M. Kapović & R. Matasović (eds.) Tones and Theories. Proceedings of the International Workshop on Balto-Slavic Accentology, Zagreb: IHJJ, 29-38. Schwyzer, E. 1939. Griechische Grammatik, Bd. I, München: Beck. Villanueva-Svensson, Miguel 2011. Indo-European long vowels in Balto-Slavic , Baltistica 46/1, 5-38. Wackernagel, Jakob & Debrunner, Albert. 1954. Altindische Grammatik, II. 2. Die Nominalsuffixe. Göttingen: Wandenhoeck & Ruprecht. 10 11