Slavic Accentology and Accentual Laws (Kapovic, ESLL 2020)

Encyclopedia of Slavic Languages and Linguistics Online Accentology (6,270 words) Slavic accentology concerns the synchronic and diachronic aspects of Slavic accentual (wordprosody) systems – from accentual systems in modern Slavic languages and dialects to the reconstruction of the Proto-Slavic accentual system. Among Indo-European languages, Slavic (and Balto-Slavic) historical accentology is known for its complexity. Article Table of Contents Modern Slavic accentual systems Modern Slavic accentual systems The Proto-Slavic accentual system The accentual (word-prosody) systems of modern Slavic languages are rather diverse – western South Slavic languages (Slovene, BCMS) have pitch accent (limited-tone accent), most West Slavic languages (e.g., Polish and Czech) and South Slavic (Macedonian) have xed (predictable) dynamic stress, while others (e.g., East Slavic and Bulgarian) have free and mobile dynamic stress (accent place is not predictable and can shift in a paradigm). Dialects sometimes signi cantly di fer from standard varieties, e.g., some BCMS dialects are not tonal, while some Macedonian dialects have free stress. With regard to vowel quantity, a feature closely related to prosody, Slavic languages also di fer: some have no distinctive length (e.g., Russian and Polish), while others do Accentual paradigms Valence theory From Proto-Slavic to modern Slavic accentuation Re exes of Proto-Slavic prosodemes in Slavic languages Vowel length in Slavic / (e.g., BCMS and Czech). However, some languages without distinctive length and tone (and some that have it) preserve traces of old quantitative and tonal distinctions in their vowel systems, e.g., Po <ó> [u] from the old *ō or Sk <ô> [uo] from PSl *ò. Re exes of Proto-Slavic paradigmatic accent in Slavic languages The Proto-Slavic accentual system Balto-Slavic and IndoEuropean accentuation The Proto-Slavic accentual system is reconstructed by using material (including dialects and old accentuated manuscripts) from all Slavic languages that preserve traces of older stages of the accentual system –in tones, stress position, quantity (which is interconnected with tones), or tone/quantity relicts in vocalism. The only Slavic language yielding no information for the reconstruction of Proto-Slavic accentuation is standard Macedonian. The following prosodemes are usually reconstructed for late Proto-Slavic (or Common Slavic): Slavic and Baltic The problem of Balto-Slavic accentual mobility Further reading Bibliography Proto-Slavic prosodemes and their main re exes symbol traditional name probable phonetic value example most important re exes * ̋ old acute rising and/or glottal *kőrva ‘cow’ BCMS krȁva, Sn/Cz kráva, Ru koróva * ̏ short circum ex short falling *slȍvo ‘word’ BCMS slȍvo ‘letter’, Sn slovọ̑ * ̑ long circum ex long falling *zȏlto ‘gold’ BCMS zlȃto, Sn zlatọ̑ , Cz zlato, Ru zóloto *ˋ short neo-acute short rising *bòbъ ʿbeanʾ BCMS/Sn bȍb, Sk/Ru dial. bôb * ̃ long neo-acute long rising *kõrļь ‘king’ BCMS dial. krãlj, Sn králj, Cz král, Ru koróľ / The two circum exes and neo-acutes are usually distinguished by length, but could theoretically be marked with one symbol respectively (e.g., * ̑ and * )̃ with an additional mark for length. Some scholars use * ´ instead of * ̋ for the old acute, and some use * ´ instead of * ̃ for the long neo-acute (di fering symbols are often used even for modern languages/dialects). The accent in nal open syllables is often marked with neutral * ̍ (marking just the stress position) because the accent there cannot always be reconstructed directly; sometimes it is not clear what should be reconstructed, and various hypotheses on accentual developments in that position exist. Short prosodemes (* ̏ and * ˋ) were possible only on originally short vowels (*e, *o, *ъ, *ь), while long prosodemes (* ,̋ * ,̑ * )̃ occurred only on originally long vowels (all others). These were the possible positions of the prosodemes in a word in Proto-Slavic: Distribution of Proto-Slavic prosodemes in a word initial syllable medial syllable nal (open) syllable old acute + + + circum ex + – – neo-acute + + + As already mentioned, the reconstruction of the prosodic values in nal (open) syllables is always indirect and partly controversial. Many researchers reconstruct medial falling syllables prior to Stang’s law, but these were, if they existed, not identical to initial circum exes. Words with initial circum ex are widely regarded as phonologically unaccented, i.e., obtaining the automatic falling tone on the rst syllable (typologically comparable to Tokyo Japanese, in which unaccented forms have an automatic rst low and all other high syllables). That would leave Proto-Slavic with the tonal opposition of just the “old acute” and “neo-acute” (both terms are somewhat of traditional misnomers); if both * ̋ and * ̃ were phonetically rising (as traditionally assumed), these would have been di ferent types of rising tones (cf. typologically two types of rising tones, ̃ and ´, in many Štokavian and Čakavian dialects). Accentual paradigms / Since Stang (1957), Proto-Slavic has been analyzed as a language with paradigmatic accent. This means that all in ected words (nouns, adjectives, pronouns, numerals, and verbs) belonged to a certain accentual paradigm (a.p.) – a pattern of accentual behavior in a speci c paradigm (i.e., in declensions and conjugations). These paradigms governed how the accent in a word will change (or not) in paradigms and derivation (when a xes were added to a root). Thus, accentuation in (Balto-)Slavic applies not only to prosody (phonology) but also in morphology (and derivation). Three basic accentual paradigms are usually reconstructed for the last stage of Proto-Slavic: a, b, and c (its re exes in later Slavic languages/dialects are written as A, B, and C here). The prototypical accentual paradigm a has a xed (immobile) * ̋ on a root syllable (*bra̋trъ ‘brother’, *boga̋tъ ‘rich’, *kőrva). Words with immobile neo-acute (*zakònъ ‘law’, *sũša ‘drought’, *gotòvь ‘ nished’) are usually also considered a subtype of a.p. a due to being immobile, though their historical origin is often completely di ferent than the one in words with the xed * .̋ Accentual paradigm b has either the neo-acute accent on the nal syllable of the root (*bòbъ, *kõrļь, *dòbrъ ‘good’, *žènъ ‘women’. , *mòžete ‘you can’- .2 ) or the neo-acute or old acute (or the nonspeci ed accent in case of some nal open syllables) on the rst syllable of the ending (*ženòj- . , *dobra̋- , *žena̋. , *mogǫ̋ ‘I can’-1 , *boba̍. , *korļa̍. ; the only exception is probably the vocative, see below). Only the rst syllable of the ending (i.e., the rst post-stem syllable) can bear the stress in a.p. b. Accentual paradigm c has the circum ex on the initial syllable of the root (*slȍvo, *zȏlto, *bȍgъ ‘god’, *pȅkǫ-1 ‘I bake’) or the accent (old acute or neo-acute) on one of the syllables, rst or second, on the ending (*slova̋ . , *zolta̋ . , *bogỹ- . , *noga̋ mъ. ‘legs’, *pečetè-2 ‘you bake’). The initial circum ex shifted to the absolute rst syllable, i.e., to (what would otherwise be) a proclitic: *nȃ zlato ‘on gold’ (: *zȏlto), *pȇrko noťi ‘during the night’ (: *nȍťi), *nȋ otь boga ‘not even of god’ (*bȍga), *nȅ pekǫ ‘I don’t bake’ (*pȅkǫ). The accent can also shift to (what would otherwise be) an enclitic: *pekǫ sę̍ ‘I bake myself’ (this phenomenon is called VasilevDolobko’s rule). Distribution of prosodemes in Proto-Slavic accentual paradigms initial/only root syllable medial/last root syllable ending (non nal syllable) ending ( nal opne syllable) prototypical a.p. a * ̋ * ̋ – – a.p. a sub-type * ˋ, * ̃ * ˋ, * ̃ – – / a.p. b * ˋ, * ̃ * ˋ, * ̃ * ˋ, * ,̃ * ̋ * ,̋ * ̍ (?) a.p. c * ,̏ * ̑ – * ̋ * ˋ, * ,̃ * ̋ The medial * ̃ in the a.p. a subtype (like *mokrĩna ‘wetness’) would probably not be accepted by all scholars. The nature of the stress in open nal non-acute syllables in a.p. b is not clear and is hard to mark properly due to inconsistencies in Proto-Slavic formal reconstruction and problems with various chronological layers: it was probably identical to * ˋ (e.g., *bobà. , and the ending most likely obtained the stress only after it was shortened). Some scholars would reconstruct only the place of stress in nal open syllables (i.e., using an ictus marker * )̍ in a.p. c, since the accentual developments in this position are complex and controversial. I shall illustrate the accentual paradigms with nominal ā-stems and verbal i-stem present tense as usually reconstructed (the accent in the nal syllables is marked as ictus only, * )̍ : Proto-Slavic nominal ā-stem accentual paradigms a.p. a ‘cow’ . a.p. b ‘woman’ a.p. c ‘head’ *kőrva . *kőrvy . *žena̍ . *ženy̍ . *golva̍ . *gȏlvy . *kőrvy . *kőrvъ . *ženy̍ . *žènъ . *golvy̍ . *golvъ̍ . *kőrvě . *kőrvamъ . *ženě ̍ . . *žena̋ mъ *gȏlvě . *golva̋ mъ . *kőrvǫ . *kőrvy . *ženǫ̍ . *ženy̍ . *gȏlvǫ . *gȏlvy . *kőrvo . *kőrvy . *žȅno . *žȅny . *gȏlvo . *gȏlvy . *kőrvě . *kőrvaxъ . *ženě ̍ . . *žena̋ xъ *golvě ̍ . *golva̋ xъ / . *kőrvojǫ . *kőrvami . *ženòjǫ . . *žena̋ mi *golvojǫ̍ . *golva̋ mi In a.p. b, we can reconstruct the accent on the ending in most cases (as a result of a progressive stress shift from the originally accented root) – except in vocatives (which seem to have been unaccented, as in a.p. c, cf. Štokavian žȅno! ‘woman’, pȍpe! ‘priest’, Uk žóno!, pópe!, Bg žéno!) and in front of a weak yer in the genitive plural. The Moscow accentological school (MA;S, see Schools of Balto-Slavic accentology) believes that some Slavic dialects had an unshifted accent on the root in some forms, e.g., *žènǫ- . and not *ženǫ̍ (cf. Dybo’s law). In the a.p. c genitive plural, we reconstruct the original *golvъ̃ (cf. Štokavian dial. -ã, Sn -á), though some scholars would reconstruct both *žènъ (b) and *gõlvъ (c) or *ženъ̍ (b) and *golvь̍ (c), depending on their stances on the quantity of *-ъ and stress shifts/retractions. The a.p. c locative plural probably had a variant *golvãxъ < *golvaxъ̍ as well (cf. Cz -ách, Čakavian both -ȁh and -ãh and Li galvosè- . ). Proto-Slavic i-stem present tense accentual paradigms a.p. a *čı̋stiti ‘to clean’ a.p. b *nosı̋ti ‘to carry’ a.p. c *lovı̋ti ‘to hunt’ 1 *čı̋sťǫ 1 1 2 *čı̋stišь 2 *nòsišь 2 *lovĩšь 3 *čı̋stitь 3 *nòsitь 3 *lovĩtь 1 *čı̋stimo/e/ъ 1 1 2 *čı̋stite 2 *nòsite *nošǫ̍ *nòsimo/e/ъ *lȍvļǫ *lovimo̍ /e̍/ĩmъ 2 *lovite̍ / 3 *čı̋stętь 3 *nòsętь 3 *lovę̃tь The neo-acute 2nd- and 3rd-person singular, (1st-person plural), and 3rd-person plural in a.p. c is due to retraction of the nal stress from a weakened yer (original *lovišь̍ , etc.). The a.p. b has stem stress in all forms except in 1st-person singular, in the open nal acute ending (*nošǫ̋ ). MAS also supposes the existence of a separate a.p. b2 (mostly in causatives and denominatives) with the xed neoacute accent on the thematic vowel *-i- after a progressive stress shift in some Slavic dialects (as opposed to a.p. b1 with stem stress everywhere): *lòžišь/ložĩšь-2 ‘you lay’, *lòžite/ložĩte-2 ‘you lay’ (Dybo et al. 1990: 112–121). Forms like *ložĩšь later usually mix with re exes of a.p. c. The accent of derived forms depends on the accentual paradigm of the root word (i.e., on the prosodic properties of the root) and the ̃ ‘sin’ (a.p. b), *gȏlsъ ‘voice’ (a.p. c), and the accentual characteristics of a xes/endings: cf. *ga̋ dъ ‘snake, scoundrel’ (a.p. a), *grěxъ ̃ derived adjectives *ga̋ dьnъ – *ga̋ dьna ( ) ‘loathsome’, *grěšь̀nъ – *grěšьna ( ) ‘sinful’, *gȏlsьnъ – *golsьna̍ ( ) ‘loud’. The three Slavic accentual paradigms did not arise all at the same time. The rst two, a.p. a and b, originate from a single, immobile, accentual paradigm and were opposed to a mobile accentual paradigm (i.e., a.p. c). The immobile (a/b) paradigm originally had a xed stem stress; it was clear that this was the case in a.p. b as well from cognates in Baltic and Indo-European languages: cf. end stress in PSl *blъxa̍ ‘ ea’ –*blъxy̍ . , or *dvora̍ . ‘court(yard)’ with stem stress in Li blùsos. (also Gr psýllai. ) and OIn dvā ́ra ‘door’. The original stem stress was preserved when the root was “acute” (i.e., in a.p. a), while in a.p. b, where the root was not “acute” (i.e., not glottalized), the accent tended to shift to the next syllable: *blъ̀xy > *blъxy̍ , *dvòra > *dvora̍ . This progressive shift is usually called Dybo’s law. The shift is seen in some forms (e.g., *blъxy̍ , *dvora̍ , *nošǫ̍ ) but not in others (e.g., *dvòrъ, *nòsišь. ). That is how a.p. b came into existence (see further under Dybo’s law and Stang’s law). Besides the three major accentual paradigms (a, b, and c), another possible accentual paradigm is often referred to, a.p. d (which is, in fact, more of a lexical class than an actual a.p.), usually with regard to masculine o-stems. Originally proposed by Illich-Svitych (1979: 103–104), and later heavily criticized (most famously by Vermeer [2001]), the theory based on data from the most widely recognized a.p. d-preserving Čakavian dialect of the island of Susak (Croatia) appears to have been vindicated (Shrager 2011; Kapović 2015: 171–175; 2020 for a.p. d in East Slavic; cf. Nikolaev 2012). A.p. d is a supposedly mixed accentual paradigm in o-stems (e.g., *nȍsъ/ . / ‘nose’ as in a.p. c, but oblique case *nosa̍ . , *nosu̍ - . as in a.p. b): where one would expect a.p. B (according to external data), one nds a.p. C in most Slavic dialects (except in relicts, in which one nds either a.p. B or a kind of synchronic “mixed” a.p. D). All Slavic a.p. b masculine stems (e.g., *dvòrъ ‘courtyard’) are originally old neuter stems (cf. OIn dvā ́ram- ). Valence theory The Moscow accentological school derives the late Proto-Slavic system with the old acute (* ̋ ) and the neo-acute (* ,̀ * ;̃ the circum ex being the surface realization of accent in phonologically unaccented words) and three accentual paradigms from an older system of high and low tones (e.g., Dybo 2000: 11–14). Before the accent shift in a.p. b (see Dybo’s law), Slavic had a system of immobile accentual paradigm (with acute roots like *bőlto ‘mud’, later to be a.p. a, and non-acute ones like *sèlo ‘village’, *vĩno ‘wine’, later to be a.p. b) and mobile accentual paradigm (e.g., *zvȍno ‘bell’, *bȏlgo ‘good’, phonologically unaccented). Roots in the immobile accentual paradigm had an inherent accent, which remained xed throughout the paradigm and in derivation (*bőlta. , *bőlta. , adjective *bőltьnъ ‘muddy’, verb *bőltiti ‘to muddy’, etc.). Roots in the mobile accentual paradigm did not have an inherent accent – so the forms were either unaccented (with automatic initial circum ex, e.g., *zvȍno, if the endings did not “attract” the stress) or had end stress (in case the ending “attracted” the stress, e.g., *zvona̋ . ). Unaccented words (without an inherent accent on the root and without the ending that “attracted” the stress) are called enclinomena; words with an inherent accent (on the root or ending) are called orthotona. Immobile stems had only orthotona (with root stress), while mobile stems had alternation of enclinomena (unaccented forms) and orthotona (end stress). According to the valence theory, as stated earlier, all morphemes (syllables) in Slavic were either dominant/high (+) or recessive/low (–). All words could thus have all (+) tones, all (–) tones, or mixed (+) and (–) tones. The rule was that the ictus (stress) always fell on the rst high tone in a word, e.g., *bő̟ lto̠ . , *bő̟ lta. , *zvo̠ na̟̋ . . The same occurs in derivation: *bő̟ lti̟ti̟ ‘to muddy’, *zvo̠ nı̟t̋ i̟ ‘to ring a bell’. If all syllables were recessive, the form was an enclinomenon (unaccented, i.e., with an initial circum ex): *zvȍ̠ no̠ , *sy̠̑nъ̠ ‘son’. Dominant morphemes were roots of immobile a.p. (a/b; e.g., *kő̟ rva̟ ‘cow’, *sès̟ tra̟ ‘sister’), stress-“attracting” endings in mobile a.p. c (e.g., *-a̟ in *no̠ ga̟̋ ‘leg’), and stress-“attracting” a xes in combination with recessive roots (e.g., *no̠žı̟c̋ a̟ ‘little leg’). Recessive morphemes were roots of mobile a.p. c (e.g., *no̠ g- ‘leg’), unaccented endings in a.p. c (e.g., *-ǫ̠ in *nȍ̠ gǫ̠- . ), and unaccented a xes in combination with recessive roots (e.g., *sy̠ ̑ no̠vъ̠ ‘son’s’). In forms with dominant roots, the ̠ of other morphemes determined the accent. valences of other morphemes were irrelevant. In forms with recessive roots, the valences The paradigmatic (im)mobility of a word thus depended on the inherent prosodic characteristics of all its morphemes. According to MAS, this system of high and low tones was still a real phonetic phenomenon in Proto-Slavic (together with the separate prosodemes: acute, neo-acute, and circum ex), and they in uenced various phonological and morphological changes in Slavic – e.g., the development of post-tonic and nal-syllable length (see below). / From Proto-Slavic to modern Slavic accentuation The following table shows which features of Proto-Slavic accentuation were preserved and which were lost in major Slavic languages: Preservation of Proto-Slavic prosodic features in modern Slavic languages Proto-Slavic tone free stress phonetic stress shifts distinctive length traces of old length in vocalism BCMS – (dial. +1) + + (dial. –) + + Slovene – + + – – Bulgarian – + + – – Macedonian – ‒ (dial. +2) – – Czech/Slovak – – (Sk dial. +3) + + Polish – – – – Upper/Lower Sorbian – – – + Kashubian – +/– 4 + – + East Slavic – + – (dial. +) – – (Uk/dial. +) 1 Dialects that preserve the neo-acute. 2 Eastern dialects near Bulgaria. 3 In the Sotak dialect. 4 Free stress is preserved in North Kashubian (and in the extinct Slovincian). Re exes of Proto-Slavic prosodemes in Slavic languages We have already seen the reconstructed Proto-Slavic prosodemes and some re exes. In what follows I present brie y their development to modern Slavic languages (for the rst syllable): / Development of Proto-Slavic prosodemes in Slavic languages Proto-Slavic *̋ * ̏ * ̑ *ˋ * ̃ BCMS (NeoŠtokavian) monosyllabic | ̏| | ̑| | ̑| | ̏| | ̑| disyllabic | ̏| | | ̏| | | ̑| | | ̏| | | ̑| | trisyllabic | ̏| | | | ̏| | | | ̏| | | | ̏| | | | ̑| | | Slovene monosyllabic | ̏| | ̑| | ̏| |´| disyllabic |´| | | | ̑| |´| | |´| | trisyllabic |´| | | | | ̑| | |´| | | |´| | | monosyllabic | ̄| | ̆| | /̆ ̄ |3 | ̄| disyllabic | ̄| | | ̆| | | ̄ | |4 | ̄| | trisyllabic | ̆| | | | ̆| | | | ̄ | | |4 | ̄| | | monosyllabic | ̆| | ̆| | /̆ ̄ |5 | ̄| disyllabic | ̆| | | ̆| | | /̆ ̄ | |5 | ̄| | trisyllabic | ̆| | | | ̆| | | | /̆ ̄ | | |5 | ̄| | | *e/oR oRo̍ dial. ô oRo̍ Czech1 Slovak1 Russian2 o̍ Ro 1 Length orthographically written as <´>. 2 Only in PSl *CerC, *CorC, *CelC, *ColC (polnoglasie) syllables. 3 Only *ò > ů (sporadically) in monosyllables. 4 Long re exes are never found before a long syllable (cf. Cz nosíte ‘you carry-2pl’ < PSl *nòsite). 5 Only *ò > ô [uo] (old length) sporadically. The following are some examples: / * ̋ – Neo-Štokavian grȁh ‘bean’, vrȁna ‘crow’, vrȁnamavrana, vranami- . ; Ru goróx, voróna. ; . ; Sn grȁh, vrána, vránami- * ̏ – Neo-Štokavian nȏs ‘nose’, pȍlje ‘ eld’, ȍblāka ‘cloud- . ’; Sn nọ̑ s, poljẹ,̑ oblȃka- . ; Cz hrách, vrána, vranami. ; Cz/Sk nos, pole, oblaku- * ̑ – Neo-Štokavian grȃd ‘city’, glȃvu ‘head’- . , prȁseta ‘piglet’. ; Sn grȃd, glavọ̑ hlavu- . , Cz prasete. , Sk prasaťa. ; Ru górod, gólovu- . ; . ; Sk hrach, . ; . , sadȋlo ‘(it) planted’; Cz/Sk hrad, * ˋ – Neo-Štokavian snȍp ‘bundle’, kȍža ‘skin’, mȍžete-2 ‘you can’; Sn snȍp, kọ́ža, mọ́ rete-2 ; Cz snop/kůň ‘horse’, kůže, můžete-2 ; Sk snop/kôň, koža/vôľa ‘will’, môžete-2 (cf. Kashubian mȯžeš-2 , US móžeš-2 ); Ru dial. snôp, kôža [uo] (also korôľ and korôva in those dialects); ̆ * ̃ – Neo-Štokavian krȃlj ‘king’, strȃža ‘guard’, mlȃtīte-2 ‘you thresh’ (archaic Old Štokavian/Čakavian/Kajkavian krãlj, strãža, mlãtı̄te2 ); Sn králj, stráža, mlátite-2 ; Cz král, stráže. , mlátíte-2 ; Sk kráľ, mlátite-2 (cf. Po król, stróża, młócicie-2 ; US stróža, młóćiće-2 ); Ru koróľ, storóža. Most re exes are well established and indisputable. However, re exes in Czech/Slovak (and West Slavic in general) are more problematic and controversial. Traditionally, it is assumed that Czech preserves the old acute length in monosyllables (e.g., mák ‘poppy’) but with a lot of secondary short vowels (e.g., rak ‘crab’), probably due to analogy to polysyllabic plural forms. Kortlandt (e.g., 2011: 174, 262) takes the short variant to be the original one. The most problematic is the short neo-acute. In monosyllables, Slovak has ô [uo] in most words (usually considered regular – but not by Kortlandt 2011: 345–346), while the situation in Czech is muddied due to analogical short re exes and general secondary lengthening before voiced consonants (e.g., Cz dům < *dȍmъ ‘home’). In polysyllables, there are many levelings, cf. Cz méně ‘less’, hůře ‘worse’ (cf. US bóle ‘better’, hórje ‘worse’) but Sk menej, horšie; OCz péřeš ‘you wash’-2 but Cz pereš. Some researchers choose to take another approach and interpret all long re exes of the short neo-acute as sporadic and secondary (Carlton 1991: 202–205). Both * ̏ and * ̑ occur only in initial syllables. In BCMS and Slovene, the re exes of the other prosodemes (old acute and neo-acute) are originally the same in medial syllables, but the accent is retracted (always so in Neo-Štokavian, in Slovene only ̏ from the last syllable): cf. BCMS pòtok, Sn pótok < potȍk < PSl *potòkъ ‘creek’; BCMS lòpata (Sn lopáta – no retraction) < PSl *lopa̋ ta ‘shovel’. In Czech, the acute is always short in medial syllables, cf. Cz lopata. Neo-acutes behave the same as in initial syllables, cf. Cz devátý < PSl *devę̃tъjь ‘ninth’ (the same in Sk deviaty). / The old initial circum ex accent shift is best preserved in BCMS (Štokavian/Čakavian) and East Slavic, cf. BCMS nȁ glāvu ‘on the head’, Ru ná golovu < PSl *nȃ golvǫ. Vowel length in Slavic In Proto-Slavic, all vowels were originally either short (*e, *o, *ъ, *ь) or long (all others – *a, *ę, *ě, *i, *ǫ, *u, *y + diphthongs *ъl, *ъr, *ьl, *ьr, *el, *er, *or, *ol; see Proto-Slavic phonology). Originally long vowels were sometimes shortened in later (South and West) Slavic languages and sometimes remained long. We have already seen what happens with long syllables under stress in Slavic. The development of old length in pre-tonic and post-tonic, and the length in general (stressed and post-tonic), in nal open syllables is complex and controversial. Dybo (2000: 84–92) maintains the traditional tenets on the development of pre-tonic length (cf. also Stang 1957: 40–42; Carlton 1991: 208–209): length is preserved before a medial (and nal) short accent (PSl *kǭkòľь > BCMS kúkolj, Cz koukol) but shortened before a medial old acute (PSl *języ̋ kъ ‘tongue’ > BCMS jèzik, Cz jazyk). However, the West Slavic pattern of preserving length in a.p. b lparticiples (Slovincian stȯ ų́ ̯ pjėl ‘stepped in’) but not in a.p. b in nitives (Slovincian stą̃ pjı̆c ‘to step in’) brings Dybo to assume that the shortening occurs only before a dominant internal old acute (e.g., in in nitives, cf. *či̠nı̟t̋ i̟ ‘to do’ in a.p. c), while the length before a recessive internal old acute is preserved (e.g., in l-participles, cf. *č n̠̑ i̠lъ̠ ‘I did’ in a.p. c). Kapović (2015: 416–501) analyzes all conditions of pre-tonic length shortening/preservation in great detail. Kortlandt (e.g., 2011: 263–264, 273) believes that all pre-tonic long vowels shortened in a.p. c (like in Cz hlava ‘head’), while the long vowels in a.p. b (like in Cz tráva ‘grass’) where preserved because they were still stressed at the time of the shortening of all pre-tonic vowels. Kapović (2017: 380–386; 2019: 79‒84, 88‒91) criticizes Kortlandt’s approach for its failure to explain the material – e.g., Slovene a.p. c length in forms like trẹ́sti ‘to shake’ – trẹ́seš-2 ‘you shake’; West Slavic length in forms like Cz třásti ‘to shake’ (a.p. c); Cz devíti. ‘of nine’ (a.p. c); Štokavian pattern rúka. ‘arm’, rùkū – rùkama- . (a.p. C); archaic Čakavian pattern trēsȅš-2 – trĕsemȍ-2pl ‘we shake’ (a.p. C); OPo przystępić ‘to begin’ (no length) – przystąpisz-2 ‘you begin’ (re ex of length; a.p. b). In South Slavic, post-tonic length is generally preserved in mono-, di-, and trisyllabic words, cf. BCMS mjȅsēc ‘moon’ – mjȅsēca. , but mjȅsĕčina ‘moonlight’. In West Slavic, there is no post-tonic length in a.p. c, cf. Cz holub ‘pigeon’ (but BCMS gȍlūb), while old a.p. a words preserve it in some cases (cf. Cz měsíc ‘moon’, but havran ‘raven’) – perhaps only in dominant syllables (see Kapović 2015: 502– 525 for a detailed treatment of post-tonic length, now also Kapović 2019: 84‒88). / Traditionally, length shortens in nal open syllables, and all long vowels in that position are taken to be later and secondary (cf. Stang 1957: 38–40; Carlton 1991: 212). However, the problem is that there are too many examples of long nal open syllables for such an explanation. As Dybo (2000: 37) has observed, long endings occur almost always in cases where a.p. c has end stress, i.e., in dominant syllables (thus, one could claim that the nal length is preserved under high tone). However, even in dominant endings, one nds both the short (e.g., Cz města ‘towns. ’, BCMS mjȅsta ‘places. ’) and the long (e.g., Sk mestá ‘towns. ’, BCMS dial. mjȅstā. ) re exes. Kapović (2015: 526–550) expands on Dybo’s hypothesis by claiming that the length was originally preserved in triand polysyllables and shortened in disyllables (cf. some Čakavian and Štokavian dialects with drvȁ ‘wood’, but nebesã ‘heaven’, or Sn bíla ‘was’ < *by̋ lă, but nosȋla ‘carried’ < *nosı̋lā), with the merger of the original nal * ̋ (as in *polja̋ . ‘ elds’) and * ̃ (as in *golvỹ. ‘head’) in the shortened * ̋ in disyllables (e.g., drvȁ) and in * ̃ in trisyllables (e.g., nebesã), and later with subsequent generalizations. Re exes of Proto-Slavic paradigmatic accent in Slavic languages Languages that have preserved the free stress can also preserve at least some of the basic original accentual paradigms (a, b, and c) in their modern form (A, B, and C – sometimes with various subparadigms). Of course, accentual paradigms change, and some words transfer from one a.p. to another. Thus, in BCMS, the o-stems a.p. C locative singular (but only in inanimate nouns) has a stress-“attracting” -u ending (appropriated from the original u-stems) instead of the old recessive *-ě ending (cf. BCMS grádu- . with PSl *gȏrdě > North Čakavian grȃde), while the verb govòriti ‘to talk’ transferred from the old a.p. c to modern a.p. B ( gòvorīte-2 ‘you talk’). I shall illustrate paradigmatic developments with a few forms of nominal ā-stems and verbal i-stem present (forms in [] brackets do not have a direct re ex of the Proto-Slavic accent): Re exes of nominal ā-stem accentual paradigms in Slavic languages BCMS Slovene Russian krȁva kráva koróva . krȁvē kráve koróvy . krȁvu krávo koróvu krȁve kráve koróvy a.p. A . / . / žèna žéna 3 žená . žènē ženẹ́ žený . žènu [ženọ̑ ] ženú žène 1 [ženẹ]̑ [žëny] 4 gláva gláva golová . glávē 2 glavẹ́ golový . glȃvu glavọ̑ gólovu glȃve glavẹ̑ gólovy a.p. B . / a.p. C . . / . 1 In dialects without phonetic retraction, ženȁ – ženẽ- retraction, glāvȁ – glavẽ- . – ženȕ- . – ženȅ/ . . 2 In dialects without phonetic . . 3 In more archaic dialects, ženȁ. 4 Older and dialectal Russian preserves the original žený. Standard Slovene shows no di ference of the old a.p. b and c in ā-stems (it has only a.p. C).Standard Bg has kráva [a], but žená [b], glavá [c] (most old a.p. c stems have -á, like the old a.p. b stems, but srjáda ‘Wednesday’ and zíma ‘winter’ have the accent like kráva. However, conservative Bulgarian/Macedonian dialects preserve a three-way opposition in the combination of noun and noun with a de nite article: kráva – krávata (a), žená – ženáta (b), gláva – glaváta (c) (with gláva originating in the old *gȏlvǫ- . ). Re exes of i-stem present tense accentual paradigms in Slavic languages a.p. A 2 a.p. B 2 BCMS Slovenian Bulgarian Russian čȉstiti čístiti – čístit´ čȉstīš čȋstiš čístiš čístiš nòsiti 1 nosīti 3 – nosít´ nȍsīš nọ́ siš nósiš nósiš / a.p. C 2 lòviti 1 lovīti 3 – gostít´ 4 lòvīš 2 lovíš lovíš gostíš 1 In dialects without phonetic retraction, nosȉti, lovȉti. 2 In more archaic dialects, lovĩš. 3 ī = í or ȋ (both variants exist). 4 ‘to live somewhere as a guest’. In Russian, many old a.p. C verbs shift secondarily to a.p. B (e.g., lóviš ⇐ lovíš), which is still an ongoing process. In BCMS (most Štokavian and Čakavian, but not Kajkavian, dialects), most long-vowel c verbs shift secondarily to long-vowel a.p. B (e.g., sȃdīš or dial. sãd š̄ ̆ ‘you plant’ instead of the older dial. sadĩš with the original accent). (For a synchronic description of all free-stress accentual systems of modern Slavic languages, see Stankiewicz 1993.) Slavic and Baltic Slavic and Baltic are closely related (see Balto-Slavic). This is especially apparent when it comes to accentuation, e.g., Lithuanian āstem accentual paradigms: Lithuanian nominal ā-stem accentual paradigms a.p. 1 a.p. 2 a.p. 3 a.p. 4 várna ‘crow’ blusà ‘ ea’ galvà ‘head’ žiemà ‘winter’ . várnos blùsos galvõs žiemõs . várną blùsą gálvą žiẽmą várnos blùsos gálvos žiẽmos várnas blusàs gálvas žiemàs . . . / A.p. 1 and 2 are originally immobile and correspond to Slavic a.p. a and b respectively. A.p. 2 had a nonacute root, and the accent shifted to the acute ending in *blùsā. > blusà and *blùsās- . > blusàs (the endings were acute) by de Saussure’s law (cf. the immobile PSl *vőrna [a] and pre-Dybo’s law [or pre-de Saussure’s law] PSl *blъ̀xa – after the rightward shift *blъxa̋ [b]; see Dybo’s law and de Saussure’s law). A.p. 3 and 4 are mobile stems that correspond to Slavic a.p. c – cf. PSl *golva̋ (c) and *zima̋ (c). A.p. 3 has the acute root (mobile stems in Slavic always have the circum ex, e.g., *gȏlvǫ- . , not **gőlvǫ), while a.p. 4 has the nonacute root, which causes de Saussure’s law in *žiẽmās- . > žiemàs. The characteristics of the Lithuanian a.p. are presented in the following table: Characteristics of Lithuanian accentual paradigms accentual paradigm original mobility acute in the root de Saussure’s law 1 – + – 2 – – + 3 + + – 4 + – + The following table shows the relation of Lithuanian accentual paradigms with the Slavic ones: The relation of Lithuanian and Slavic accentual paradigms Lithuanian Slavic (originally) immobile 1 a 2 b1 mobile 3 c 4 / 1 Except in masculine o-stems, where one usually a.p. d; see above). nds a.p. c instead of the expected a.p. b (with the intermediate phase in the mixed Lithuanian has three prosodemes: ̀ (short accent), ́ (falling accent), and ˜ (rising accent). The rst one appears on short vowels and the last two on long vowels and diphthongs ( ̃ appears on secondarily lengthened à > ã, è > ẽ as well, e.g., Li rãsą below). Latvian has three prosodemes (all on long vowels and diphthongs only): ` (falling accent), ˜ (rising accent), and ̂ (glottalized accent, “broken tone”). Unlike Lithuanian, in which the prosodemes appear only in stressed syllables (the stress is free), in Latvian they appear on all long syllables (the stress is always on the rst syllable – in standard Latvian, ` appears only in the stressed syllable; cf. Lt rùokā ̂s- . ‘(in the) hands’ (~ Li rañkose- . ). Lithuanian and Latvian prosodemes have regular correspondences to each other and to Slavic: Li mótė (1), Lt mãte ~ PSl *ma̋ ti ‘mother’ (a); Li ùgnis (2; the variant ugnìs [4] is younger) ~ PSl *ògńь ‘ re’ (b); Li skiẽtas ‘reed (in the loom)’ (2) ~ PSl *ščĩtъ ‘shield’ (b); Li gálvąLi rãsąLi žiẽmą- . (3), Lt galvu ~ PSl *gȏlvǫ (c); . < *ràsą (4) ~ PSl *rȍsǫ ‘dew’ (c) . (4), Lt zìemu ~ PSl *zȋmǫ ‘winter’ (c) The relation of Proto-Slavic, Lithuanian, and Latvian prosodemes Slavic immobile paradigm mobile paradigm Lithuanian Latvian (+)acute ̋ ´ ˜ (+)circum ex ̀/ ˜ ̀/ ˜ ` (–)acute ͂ ´ ̂ / (–)circum ex ̏/ ̑ ̀/ ˜ ` Balto-Slavic and Indo-European accentuation The connection of Balto-Slavic accentuation with the accentuation of other Indo-European languages that preserve Proto-IndoEuropean free stress (or traces of it) was rst explained by Illich-Svitych (1979). His major insight was that Balto-Slavic immobile a.p. corresponds to Vedic-Greek barytona (root stress) and that Balto-Slavic mobile a.p. corresponds to Vedic-Greek oxytona (end stress). Balto-Slavic immobile accentual paradigm ~ IE barytona: PSl *blъxa̋ (b) ‘ ea’, Li blusà (2) ~ Gr psýlla, Pashto wrə́žạ ; PSl *sestra̋ (b) ‘sister’, older Li sèsuo (1; sesuõ [4] is younger) ~ OIn svásā; PSl *zǫ͂ bъ ‘tooth’ (d > c), Li žam̃ bas (2 > younger 4) ‘sharp edge’ ~ OIn jámbhas ‘tooth’, Gr gómphos ‘bolt’; PSl *tь̃ rnъ ‘thorn’ (b) ~ OIn tŕ̥ṇam ‘(blade of) grass’. Balto-Slavic mobile accentual paradigm ~ IE oxytona: PSl *cěna̋ ‘price’ (c), Li kainà (4) ~ Gr poinḗ ‘price paid, ne’; PSl *snъxa̋ ‘daughter-in-law’ (c) ~ OIn snušā ́, Gr nyós, OEn snoru < PGe*snuzṓ(n); PSl *dъťı̋ ‘daughter’ (c), Li duktė ̃ (4) ~ OIn duhitā ́; PSl *žȋvъ ‘alive’ (c), Li gývas (3), Lt dzîvs ~ OIn jīvás; In some cases, Balto-Slavic has root stress compared to end stress in other Indo-European languages when the root was closed by a Proto-Indo-European laryngeal (yielding Balto-Slavic acute; cf. PSl *grı̋va ‘mane’ [a], Lt grĩva ‘river mouth’ ~ OIn grīvā ́ ‘nape of neck’ [PIE *gwriHweh2]). This is called Hirt’s law (its regularity is somewhat obscured by subsequent generalizations and analogies). / The problem of Balto-Slavic accentual mobility Traditionally, the reconstruction of Proto-Indo-European accent is very simple – one just compares the accent of Vedic, Greek, and Germanic (where traces of Proto-Indo-European accent can be seen through the development of voiceless medial stops via Grimm’s/Verner’s law). Thus, one reconstructs PIE *bhréh2tēr ‘brother’ on the basis of OIn bhrā ́tā, Gr phrā ́tēr ‘member of a fraternity’, Go broþar < *βrṓþēr but PIE *ph2tḗr ‘father’ on the basis of OIn pitā ́, Gr patḗr, Go fadar < *faðḗr. This reconstructed accent was usually analyzed in correlation to Proto-Indo-European ablaut changes, i.e., *bhreh2- had a vowel and thus the accent, while *ph2t- did not and therefore the accent was at the end. The Balto-Slavic accent was usually somehow derived from this “Greco-Aryan” reconstruction of Proto-Indo-European accentuation. One of the basic characteristics of the Balto-Slavic accentual system is that all in ecting words belong to either immobile or mobile stems, with the latter having the accent alternating between the root and ending. This was especially interesting considering that Vedic and Greek, thought to be the most archaic in terms of accentuation, only had very limited accentual mobility – mostly in nominal consonantal stems with monosyllabic roots (e.g., OIn pā ́t – padás. , Gr poũs – podós. ‘foot’). Most accentologists have tried to derive the Balto-Slavic system from the supposed (“Greco-Aryan”) Proto-Indo-European system –via various analogies either to the consonantal stems or to various proposed sound laws like “Pedersen’s law” (the supposed *dhugh2térm̥ - . > BSl *duktèrin > BSl *dùkterin > PSl *dъ̏ťerь, Li dùkterį ‘daughter’), etc. (see the overview of various hypotheses in Olander 2009: 14–46). The MAS, however, claims that the Balto-Slavic system is too complex to be derived from a far simpler “Greco-Aryan” type of accentuation and that it should be the other way round: the Balto-Slavic high-/low-tone system is actually the most archaic and should be derived from a similar Proto-Indo-European high/low system. For example, instead of PIE *ǵómbhos ‘tooth’, one should reconstruct *ǵo̟ mbho̠ s, and instead of PIE *yugóm ‘yoke’ (OIn yugám, Gr zygón), one should reconstruct *yu̠ go̠ m (Dybo et al. 1993: 69–73). The major evidence for this is the accentuation in Balto-Slavic derivation. If one accepts the “Greco-Aryan” Proto-Indo-European accentuation, that would mean that the accent of derivatives in Balto-Slavic abstractly depends on the accentual pattern of the basic form, i.e., that the accent pattern of *žena̍ – *ženǫ̍ - . ‘woman’ (b) abstractly causes the derived accent *žènьskъ –*žènьska ( ) –*žènьskъjь ( ) ‘female’, while the pattern of *mǫ̑žь – *mǫ̑ža. ‘man, husband’ abstractly causes the derived accent *mǫ̑žьskъ – *mǫžьska̍ ( ) – *mǫžьskъjь̍ ( ) ‘male’ and thus in all derivative processes with remarkable regularity. If one supposes a phonetically real high-/low-tone system (which is, however, not derivable from a “Greco-Aryan” type of Proto-Indo-European accentuation), derivation is easy to understand since it depends on the prosodic characteristics of all morphemes, which simply remain unchanged, and the stress is predictable based on them. / Further reading Stang’s work (1957) was revolutionary in its time and is still useful, but also dated in many respects. The MAS (represented primarily by Dybo and Nikolaev) is best approached through its major monographs (Dybo 1981; Dybo et al. 1990; 1993; Dybo 2000). An overview of MAS is provided by Lehfeldt (2001), while Hendriks (2003) provides a good survey of major changes in the MAS approach in the 1990s. Kapović (2015) is an extensive monograph on Slavic accentuation, often critically following the MAS approach. Most of the accentological works of Kortlandt, the main proponent of the Leiden accentological school, are collected by Kortlandt (2011). Jasano f (2017) is another example of the classical approach to Balto-Slavic accentuation (i.e., deriving it from a “Greco-Aryan” Proto-IndoEuropean system; see Oslon 2017, for an insightful critique from the MAS perspective. Much recent scholarship on Slavic accentology is available in numerous proceedings from the International Workshop on Balto-Slavic Accentology (IWoBA) conferences. Mate Kapović Bibliography Carlton, Terence R. 1991. Introduction to the phonological history of the Slavic languages. Columbus. Dybo, Vladimir A. 1981. Slavjanskaja akcentologija: Opyt rekonstrukcii sistemy akcentnyx paradigm v praslavjanskom. Moscow. Dybo, Vladimir A. 2000. Morfonologizovannye paradigmatičeskie akcentnye sistemy: Tipologija i genezis. Moscow. Dybo, Vladimir A. et al. 1990. Osnovy slavjanskoj akcentologii. Moscow. Dybo, Vladimir A. et al. 1993. Osnovy slavjanskoj akcentologii: Slovar´. Moscow. Hendriks, Pepijn 2003. A Note on Stang’s Law in Moscow Accentology. In: Schaeken, Jos et al. (eds.), Dutch Contributions to the Thirteenth International Congress of Slavists, Ljubljana: Linguistics (SSGL 30), 107–123. Illich-Svitych, Vladislav M. 1979. Nominal accentuation in Baltic and Slavic. Cambridge MA. Jasano f, Jay H. 2017. Prehistory of Balto-Slavic Accent. Leiden. / Kapović, Mate. 2015. Povijest hrvatske akcentuacij: Fonetika. Zagreb. Kapović, Mate 2017. On shortening, lengthening, and accent shifts in Slavic. Rasprave Instituta za hrvatski jezik i jezikoslovlje 43/2, 379– 400. Kapović, Mate 2019. Shortening, Lengthening, and Reconstruction: Notes on Historical Slavic Accentology. Rasprave: Časopis Instituta za hrvatski jezik i jezikoslovlje 45/1, 75–133. Kapović, Mate 2020. Accentual Paradigm D on Susak: New Data. Proceedings of IWoBA XII (forthcoming). Kortlandt, Frederik. 2011. Selected writings on Slavic and general linguistics. Amsterdam. Lehfeldt, Werner. 22001. Einführung in die morphologische Konzeption der slavischen Akzentologie. Munich. Nikolaev, Sergei L. 2012. Vostočnoslavjanskie re eksy a.p. d i indoevropejskie sootvestvija slavjanskim akcentnym tipam suščestvite´nyx m. r. s o- i u-osnovami. Karpato-balkanskij dialektnyj landšaft: Jazyk i ku´tura 2, 32‒189. Olander, Thomas. 2009. Balto-Slavic Accentual Mobility. Berlin. Oslon, Mikhail 2017. Review of Jasano f (2017). Journal of Language Relationship 15/4, 299‒311. Shrager, Miriam-Maria. 2011. Accentuation of masculine monosyllabic nouns of Susak speakers in New Jersey. In: Rinkevičius, Vytautas (ed.), Proceedings of the 6th IWoBA (Baltistica VII Priedas). Vilnius, 207–225. Stang, Christian S. 1957. Slavonic accentuation. Oslo. Stankiewicz, Edward. 1993. The accentual patterns of the Slavic languages. Stanford. Vermeer, Willem. 22001. Appendix: Critical observations on the modus operandi of the Moscow Accentological School. In: Lehfeldt, Werner, Einführung in die morphologische Konzeption der slavischen Akzentologie. Munich, 131–161. Cite this page / Kapović, Mate, “Accentology”, in: Encyclopedia of Slavic Languages and Linguistics Online, Editor-in-Chief University of Kansas General Editor University of Chicago Associate Editors University of Kansas Brown University University of Amsterdam University of Warsaw University of Zagreb University of Mainz Stockholm University Marc L. Greenberg, Lenore A. Grenoble, Stephen M. Dickey, Masako Ueda Fidler, René Genis, Marek Łaziński, Anita Peti-Stantić, Björn Wiemer, Nadežda V. Zorixina-Nilsson. Consulted online on 28 April 2020 10.1163/2589-6229_ESLO_COM_032115> First published online: 2020 / Encyclopedia of Slavic Languages and Linguistics Online De Saussure’s Law (459 words) De Saussure’s law normally denotes a progressive accentual shift from a nonacute to an acute syllable in various interpretations of historical Schools of Balto-Slavic Accentology – cf. Li *blùsā (with the rst syllable short/circum ex and the second acute) > *blusā ́ (with the nal syllable accented and acute), which is later shortened to the attested blusà ‘ ea’ by Leskien’s law. Article Table of Contents Further reading Bibliography In pre-Stangian accentology (and still today in some nonmainstream approaches to Balto-Slavic accentology that reject Stang), de Saussure’s law was often taken to operate not only in Lithuanian but also in Slavic. Christian S. Stang (1957: 15–20) refuted the law for Slavic, pointing to forms like Štokavian zȋme ‘winters - ’ but Li žiemàs (one expects an acute ending from PIE *‑eh2[n]s), which was subsequently accepted by the majority of accentologists (e.g., by the Moscow and Leiden accentological schools, cf. “Schools of Balto-Slavic Accentology”). However, the situation changes with the “new approach” of the Moscow accentological school (MAS) in the 1990s, which again reintroduces de Saussure’s law into Slavic. MAS now operates with it in two sets of forms. First, it is considered to be an early shift to the right (part of the so-called multiphasal rightward accentual shift [pravostoronnij drejf]; cf. “Dybo’s law”) in accentual paradigm b forms like Slavic *blъxa > *blъxa̋ ‘ ea’, where the shift to dominant acute syllables (e.g., in - *blъxa̋ ) would be earlier than the shift to the nonacute ones (e.g., in - *blъxǫ, with the nonacute ending, which remained unchanged at rst), as indicated by such Czech forms as sova/sůva ‘owl’, the rst variant supposedly stemming from forms like - *sova̋ with the shift and the other from / forms like - *sòvǫ without it (Dybo et al. 1993: 13–15). Second, de Saussure’s law is discovered in remnants of a shift to recessive acute endings in a.p. c forms like Štokavian vrijéme ‘time’ but nȁ ̮vrijēme ‘on time’, which were subsequently often eliminated by analogy (Dybo et al. 1993: 30–31). Further reading Neville Edgar Collinge’s (1985: 149–152) overview of de Saussure’s law is useful, but unfortunately somewhat confusing. Thomas Olander (2009: 45–46) o fers a brief overview of various stances on de Saussure’s law in di ferent approaches to Balto-Slavic accentology. Vladimir A. Dybo (2006) o fers a MAS view on Stang’s refutation of de Saussure’s law in Slavic. (For de Saussure’s law in Lithuanian, cf. e.g., Zinkevičius 1998: 103–105; Dini 2014: 127–129.) Mate Kapović Bibliography Collinge, Neville Edgar 1985. The laws of Indo-European. Amsterdam. Dini, Pietro U. 2014. Foundations of Baltic languages. Vilnius. Dybo, Vladimir A. et al. 1993. Osnovy slavjanskoj akcentologii: Slovar´. Moscow. Dybo, Vladimir A. 2006. Sravnitel´no-istoričeskaja akcentologija, novyj vzgljad (po povodu knigi V. Lefe´dta “Vvedenie v morfologičeskuju koncepciju slavjanskoj akcentologii”). Voprosy jazykoznanija 2, 3–27. Olander, Thomas 2009. Balto-Slavic accentual mobility. Berlin. Stang, Christian S. 1957. Slavonic accentuation. Oslo. Zinkevičius, Zigmas 1998. The history of the Lithuanian language. Vilnius. Cite this page / Kapović, Mate, “De Saussure’s Law”, in: Encyclopedia of Slavic Languages and Linguistics Online, Editor-in-Chief University of Kansas General Editor University of Chicago Associate Editors University of Kansas Brown University University of Amsterdam University of Warsaw University of Zagreb University of Mainz Stockholm University Marc L. Greenberg, Lenore A. Grenoble, Stephen M. Dickey, Masako Ueda Fidler, René Genis, Marek Łaziński, Anita Peti-Stantić, Björn Wiemer, Nadežda V. Zorixina-Nilsson. Consulted online on 28 April 2020 10.1163/2589-6229_ESLO_COM_032117> First published online: 2020 / Encyclopedia of Slavic Languages and Linguistics Online Dybo’s Law (759 words) Dybo’s law is a name given to a rightward accentual shift that occurs in Proto-Slavic accentual Article Table of Contents paradigm b or in nonacute accentual paradigm a (a.p. a; see Slavic accentology), where the original *ˋ and *˜ (nonacute prosodemes in immobile accentual paradigm; “dominant Further reading circum exes” in the terminology of the Moscow accentological school) shift to the next syllable Bibliography (the recessive circum exes, * ̏ and * ,̑ which are phonologically unstressed, do not shift, nor does the old acute *˝), e.g., *bòba ‘bean. ’ > *boba̍ . The law is named after the two great Russian accentologists who discovered it (cf. Dybo 1962; Dybo and Illič-Svityč 1963). It is usually called Dybo’s law, more rarely Dybo-Illič-Svityč’s law or Illič-Svityč’s law. This law is, in one interpretation or another, nowadays generally accepted by mainstream Slavic accentology. Accentual paradigm b after the operation of Dybo’s law had the stress either on the last syllable of the stem (immediately before the ending, e.g., *žènъ ‘women. ’, *nòsite ‘you carry-2. ’) or on the rst syllable of the ending (e.g., present *nošǫ̍ ‘I carry-1. ’, *ženǫ̍ ‘woman- . ’, *travǫ̍ ‘grass- . ’). The comparison of Slavic with Baltic and other Indo-European data has shown that in a.p. b Slavic originally had stem stress, which was then shifted via Dybo’s law to the following syllable (be it part of the stem, su x, or ending) in certain conditions: *nòšǫ > *nošǫ̍ , *žènǫ > *ženǫ̍ , *trãvǫ > *travǫ̍ . The existence of such a progressive stress shift in Slavic is apparent from comparisons such as Slavic *blъxy̍ ‘ eas. ’ ~ Li . blùsos (Gr . psýllai), Slavic *dvora̍ ‘court(yard). ’ ~ OIn dvā ́ra- ‘door’, etc. The neo-acutes shift both in a.p. b and in b-forms that are part of other paradigms, e.g., in *bòdla > / *bodla̍ ‘stabbed’ ( , active), *bodèna > *bodena̍ ‘stabbed’ ( , passive; *bostı̍ ‘to stab’ is a.p. c). The end result after the shift is not always a.p. b, cf. the type *dòbrota > *dobròta ‘goodness’ (nonacute a.p. a), where the accent shifts to the su x (where it remains in all forms) and not to the ending. The accent type of the newly shifted accent depended on the nature of the newly stressed syllable (cf. *dobròta but *žènica > *ženı̋ca ‘wife’ [ ], while the nature of the nal accent in *bodla̍ is unclear), and some of the results are disputed (cf. Stang’s law for the *čь̀ rnina > *čьrnĩna ‘blackness’ type). Dybo’s law is usually presented as a uni ed all-encompassing shift. However, some forms, where it apparently did not occur, are problematic: pan-Slavic de nite a.p. b adjectives like *nòvъjь ‘the new’ (the contraction of the ending did not occur in East Slavic, so it cannot be used as an explanation via the supposed retraction), participles like *nòšenъ ‘carried’, and comparative adverbs like *mь̀ ńe ‘less’ (cf. Kapović 2017: 390n21). (For a.p. b e-presents like *mòžete ‘you can’, see Stang’s law, and for the Czech sova/sůva ‘owl’ type see “Fortunatov–de Saussure’s law.”) After 1993, the Moscow accentological school no longer believes in Dybo’s law as one simple uni ed shift that occurred in the same way in all positions in the whole of Slavic, but interprets it as a rightward accentual shift (pravostoronnij drejf udarenija), which occurred in phases, did not occur in all positions and conditions, and exhibited di fering results in di ferent Slavic dialects (cf. Dybo et al. 1993: 16, 18–21). Further reading Neville Edgar Collinge’s (1985: 31–33) overview of Dybo’s law, though awed, is still useful (cf. also Feldstein 1990: 43–45). Thomas Olander (2009: 45–46) provides a brief overview of various stances on Dybo’s law in di ferent approaches to Slavic accentology. Pepijn Hendriks (2003) provides a useful overview of the changes in the approach of the Moscow accentological school in the 1990s. A detailed overview and discussion of the rightward accentual shifts in Slavic are given by Mate Kapović (2015: 95–97, 103–134). Mate Kapović Bibliography Collinge, Neville Edgar. 1985. The Laws of Indo-European. Amsterdam. / Dybo, Vladimir A.. 1962. О rekonstrukcii udarenija v praslavjanskom glagole. Voprosy slavjanskogo jazykoznanija 6, 3‒27. Dybo, Vladimir A., & Vladislav M. Illič-Svityč. 1963. K istorii slavjanskoj sistemy akcentuacionnyx paradigm. In: Vinogradov, V.V. (ed.). Slavjanskoe jazykoznanie: V Meždunarodnyj s´´ëzd slavistov. So ja, sentjabr´ 1963. Doklady sovetskoj delegacii. So a, 70‒87. Dybo, Vladimir A. et al. 1993. Osnovy slavjanskoj akcentologii: Slovar´. Moscow. Feldstein, Ronald F.. 1990. On the structural motivation for Dybo’s law. Indiana Slavic studies 5, 43–60. Hendriks, Pepijn 2003. A Note on Stang's Law in Moscow Accentology. In: Schaeken, Jos et al. (eds.), Dutch Contributions to the Thirteenth International Congress of Slavists. Ljubljana, 107–123. Kapović, Mate 2015. Povijest hrvatske akcentuacije: Fonetika. Zagreb. Kapović, Mate. 2017. On shortening, lengthening, and accent shifts in Slavic. Rasprave Instituta za hrvatski jezik i jezikoslovlje 43/2, 379– 400. Olander, Thomas 2009. Balto-Slavic Accentual Mobility. Berlin. Cite this page Kapović, Mate, “Dybo’s Law”, in: Encyclopedia of Slavic Languages and Linguistics Online, Editor-in-Chief University of Kansas General Editor University of Chicago Associate Editors University of Kansas Brown University University of Amsterdam University of Warsaw University of Zagreb University of Mainz Stockholm University Marc L. Greenberg, Lenore A. Grenoble, Stephen M. Dickey, Masako Ueda Fidler, René Genis, Marek Łaziński, Anita Peti-Stantić, Björn Wiemer, Nadežda V. Zorixina-Nilsson. Consulted online on 28 April 2020 10.1163/2589-6229_ESLO_COM_032120> First published online: 2020 / Encyclopedia of Slavic Languages and Linguistics Online Stang’s Law (524 words) Stang’s law (more rarely called Ivšić’s law) is the supposed retraction of the long internal circum ex in Common Slavic that, according to many accentologists of the post-Stang era, accounts for most neo-acute stem-stressed forms in accentual paradigm b (a.p. b). Thus, attested a.p. b forms like present *mõltite ‘you thresh-2. ’ and *pòpěxъ ‘priests- . ’ (see “Accentology”) would originate in preforms ̂ like *moltîte, *popěxъ. Stjepan Ivšić (1911: 169–170, 177–182) was the rst one to propose a retractional nature of the Slavic neo-acute in forms like *mõltite. However, it was Christian S. Stang (1957: 13, 21–22, 70, 108–109, etc.) who formulated it in a modern manner and as part of his new paradigmatic approach to Slavic accentuation (cf. Dybo 2000: 73–74 on Ivšić and Stang). Stang’s law was at rst accepted by both major accentological schools, Moscow and Leiden (see “Schools of Balto-Slavic accentology”), but as an accent shift that occurs after the rightward shift in a.p. b (see “Dybo’s law”) – i.e., forms like *moltîte were to be derived from an even older *mõltite. Though Stang’s law is widely accepted in modern mainstream Slavic accentology, there are some serious problems with it. First of all, the supposition that the rightward shift (see “Dybo’s law”) onto a nonacute long vowel yields an internal falling intonation (which supposedly yields forms in which Stang’s law should occur) is very much in question, since there seem to be many instances proving that a neo-acute is actually the result of such shifts – cf. e.g., the Čakavian type črnĩna ‘blackness’ (cf. Dybo 1981: 145–146; 2000: 203– 204), ravnĩca ‘plane’, dvorĩšće ‘courtyard’, popĩć ‘little priest’, Old Štokavian (Posavina) sestrĩn ‘sister’s’, etc. (Kapović 2015: 184–195; 2017: / 395, 2019: 113‒116). Secondly, not all stem-stressed forms in accentual paradigm b are explained easily by it – the greatest problem being the e-presents like *mòžete ‘you can-2. ’, where the thematic *-e- is originally and in many dialects short, which hardly merits a panSlavic retraction via Stang’s law (cf. Kapović 2017: 391n22). (For other problematic forms, see “Dybo’s law”.) Since 1988, the Moscow accentological school changed its approach to Stang’s law (for a good overview, cf. Hendriks 2003), in the end concluding that the supposition of a rightward and then a leftward accentual shift in cases like *mõltite (the original form) > *moltîte (form produced by Dybo’s law) > *mõltite (form produced by Stang’s law) is super uous and that forms like *mõltite just preserved their original accent (Dybo et al. 1993: 15). Mate Kapović Bibliography Dybo, Vladimir A. 1981. Slavjanskaja akcentologija: Opyt rekonstrukcii sistemy akcentnyx paradigm v praslavjanskom. Moscow. Dybo, Vladimir A. 2000. Morfonologizovannye paradigmatičeskie akcentnye sistemy: Tipologija i genezis. Moscow. Dybo, Vladimir A. et al. 1993. Osnovy slavjanskoj akcentologii. Slovar´. Moscow. Hendriks, Pepijn 2003. A note on Stang’s law in Moscow accentology. In: Schaeken, Jos et al. (eds.), Dutch contributions to the thirteenth international congress of Slavists, Ljubljana: Linguistics, 107–123. Ivšić, Stjepan. 1911. Prilog za slavenski akcenat. Rad JAZU 187, 133–208. Kapović, Mate 2015. Povijest hrvatske akcentuacije: Fonetika. Zagreb. Kapović, Mate. 2017. On shortening, lengthening, and accent shifts in Slavic. Rasprave Instituta za hrvatski jezik i jezikoslovlje 43/2, 379– 400. Kapović, Mate. 2019. Shortening, lengthening, and reconstruction: Notes on historical Slavic accentology. Rasprave: Časopis Instituta za hrvatski jezik i jezikoslovlje 45/1, 75–133. / Stang, Christian S. 1957. Slavonic accentuation. Oslo. Cite this page Kapović, Mate, “Stang’s Law”, in: Encyclopedia of Slavic Languages and Linguistics Online, Editor-in-Chief University of Kansas General Editor University of Chicago Associate Editors University of Kansas Brown University University of Amsterdam University of Warsaw University of Zagreb University of Mainz Stockholm University Marc L. Greenberg, Lenore A. Grenoble, Stephen M. Dickey, Masako Ueda Fidler, René Genis, Marek Łaziński, Anita Peti-Stantić, Björn Wiemer, Nadežda V. Zorixina-Nilsson. Consulted online on 28 April 2020 10.1163/2589-6229_ESLO_COM_036054> First published online: 2020 / Encyclopedia of Slavic Languages and Linguistics Online Ivšić’s Retraction (426 words) In Slavic, there is a general tendency for newly arisen noninitial falling accents (called neoArticle Table of Contents circum ex) to retract to a preceding syllable with the neo-acute as the result on newly stressed syllables. For the most known, but also the most questionable, of such retractions see Stang’s law. Further reading The dialectal Common Slavic post-contractional retraction of the 2 present *stǫpa̋ ješь ‘you Bibliography tread’ > *stǫpa̋ ešь > *stǫpâšь > *stǫ̃ pašь (Cr dial. stũpāš, Sk stúpaš, Po stąpasz) type (after a long pretonic syllable) is well known. In Štokavian and Čakavian, neo-circum ex retracts to a preceding short syllable as well, in multiple phases, with regional di ferences and with later generalizations – cf. the retracted accent in Štokavian lȍpātā ‘shovels. ’, nȅ ͜znām ‘I don’t know’, colloquial ȉskopō ‘dug out’ ( ; Kapović 2015: 343–354; 2018; 2020). One of the established retractions of neo-circum ex, called Ivšić’s retraction ( rst mentioned by Ivšić 1937: 188), occurs in Kajkavian and northern dialects of Slovene (Carinthian and Pannonian). There, the neo-circum ex regularly retracts to a preceding long vowel, e.g., *zābȃva ‘party’ (with a typical Kajkavian-Slovene neo-circum ex in the ictus-ceding position) > zãbava. In some types, the retraction appears analogically with short preceding syllables as well, e.g., nȍsila ‘carried’ ( ) by analogy to brãnila ‘defended’ ( ; from nosȋla, *brānȋla). In the southern Kajkavian dialect of Turopolje, the (internal) neo-circum ex is generally retracted (to both long and / short preceding syllables), e.g., nȅ ͜pušim ‘I don’t smoke’ (elsewhere ne ͜pȗšim). In Kajkavian (and more rarely in Slovene), there are also some cases of retraction of neo-circum exes to preceding short syllables that are not completely clear, e.g., Kaj mȍtika ‘hoe’ but standard Sn motȋka (cf. Kapović 2015: 354–363). Further reading Tijmen Pronk (2007) discusses some Slovene forms (with a critique in Kapović 2015: 357n1320). It is di cult to connect Ukrainian forms like bésida ‘conversation’ and Štokavian/Čakavian forms like zȁstava ‘ ag’ (cf. Greenberg 2000: 111–112; see Kapović 2015: 454–458 and Štokavian variant forms like zástava). Mate Kapović Bibliography Greenberg, Marc L. 2000. A historical phonology of the Slovene language. Heidelberg. Ivšić, Stjepan 1937. Osnovna hrvatska kajkavska akcentuacija u Pergošića (1574). In: Zbornik lingvističkih i loloških rasprava A. Beliću o četrdesetogodišnici njegova naučnog rada posvećuju njegovi prijatelji i učenici. Belgrade, 183–195. Kapović, Mate 2015. Povijest hrvatske akcentuacije: Fonetika. Zagreb. Kapović, Mate. 2018. Povijest glagolske akcentuacije u štokavskom (i šire). Rasprave Instituta za hrvatski jezik i jezikoslovlje 44/1, 159–285. Kapović, Mate. 2020. On the retraction of neo-circum ex in Čakavian. Festschrift for Josip Lisac. Pronk, Tijmen 2007. The retraction of the neocircum ex in the Carinthian dialects of Slovene (on Ivšić’s retraction). In: Kapović, Mate, and Ranko Matasović (eds.), Tones and theories: Proceedings of IWoBA 2005. Zagreb, 171–183. Cite this page Kapović, Mate, “Ivšić’s Retraction”, in: Encyclopedia of Slavic Languages and Linguistics Online, Editor-in-Chief University of Kansas General Editor University of Chicago Associate Editors University of Kansas Brown University University of Amsterdam University of Warsaw University of Zagreb University of Mainz Stockholm University Marc L. Greenberg, Lenore A. Grenoble, Stephen M. Dickey, Masako Ueda Fidler, René / Genis, Marek Łaziński, Anita Peti-Stantić, Björn Wiemer, Nadežda V. Zorixina-Nilsson. Consulted online on 28 April 2020 10.1163/2589-6229_ESLO_COM_032118> First published online: 2020 /