||This article may be too technical for most readers to understand. (December 2010)|
Proto-Balto-Slavic is reconstructed proto-language descending from Proto-Indo-European (PIE) and out of which all later Balto-Slavic languages (represented by Baltic and Slavic branches) and dialects descended, such as modern Latvian, Lithuanian, Polish, Russian and Serbian.
The Proto-Balto-Slavic language is not directly attested by any surviving texts but has been reconstructed using the comparative method. There are number of isoglosses that Baltic and Slavic languages share in phonology, morphology and accentology, which represent common innovation from Proto-Indo-European times, and which can be chronologically arranged.
Proto-Indo-European phonological system has exhibited several significant changes in Balto-Slavic period:
- the three series of PIE stops were reduced to two series (voiced and unvoiced)
- PIE syllabic sonorants were substituted with sequences of a short vowel (*i or *u) and a non-syllabic sonorant
- the three PIE laryngeals merged into one (*H), which may have disappeared even during the Balto-Slavic period
- the complex system of PIE dorsals was simplified due to the delabialization of labiovelars and the change of PIE palatovelars into fricatives
PIE voiced and unvoiced stops were preserved in Balto-Slavic and Proto-Slavic, and aspirated series was deaspirated.
Winter's law was still operable when there was phonemic distinction between the series of plain and aspirated voiced stops. As a result of Winter's law, the distinction between those two series has been indirectly preserved in Proto-Balto-Slavic, because Balto-Slavic vowel would lengthen before a plain voiced stop, but not before an aspirated stop, this occurring probably only if the stop was in syllable coda (i.e. in closed syllable).
On the basis of relative chronology of sound changes it has been ascertained that Winter's law acted rather late, after some other less prominent Balto-Slavic changes occurred, such as after the disappearance of laryngeals in prevocalic position. Compare:
Therefore, the merger of PIE aspirated and plain velar stop series was one of the last common Balto-Slavic sound changes.
Three series of PIE dorsal stops (velars, palatovelars and labiovelars) merged to two series in Balto-Slavic: velars and palatovelars. PIE labiovelars lost their labialization in Balto-Slavic, just like they did in the Indo-Iranian, Armenian and Greek branches. Unlike some other Indo-European languages, Balto-Slavic labiovelars were delabialized unconditionally and at once, leaving no noticeable direct or indirect traces.
There are a number of words in Balto-Slavic which show Centum reflex of PIE patalalized dorsals. A number of these can be explained by regular sound laws, although some of these laws have been obscured by numerous analogical developments. Others are argued to be borrowings from Centum languages, e.g. Proto-Balto-Slavic *kárwā 'cow' (Lith. kárvė, OCS krava, Russ. koróva) was likely borrowed from Proto-Celtic *karawā, which in turn is a regular reflex of PIE *ḱerh₂weh₂.
PIE palatovelars could also depalatalize in Balto-Slavic. Several depalatalization rules for Balto-Slavic have been proposed. According to Matasović, the depalatalization of palatovelars occurred before sonorant followed by a back vowel: K' > K/_RVback. That would explain Centum reflexes such as:
- Lithuanian akmuõ, Latvian akmens and OCS kamy would have regular /k/ as opposed to Sanskrit áśmā < PIE *h₂eḱmōn 'stone'
- OCS svekry < PIE *sweḱruh₂ 'mother-in-law'
- Old Prussian balgnan < PIE *bʰolǵʰno- 'saddle'
PIE palatovelars */ḱ/, */ǵ/, */ǵʰ/ turned to Balto-Slavic fricatives: */ś/, */ź/ and */źʰ/, the last two merging when contrastive aspiration was lost. They probably had an intermediate stage of affricates */ć/, */đ/, */đʰ/, and it's conceivable that they may have still been affricates at the Balto-Slavic stage. However, by applying usual methods of reconstructions on Baltic and Slavic languages, fricatives */ś/ and */ź/ represent the most likely phonological interpretation of the reflexes of PIE palatovelars.
Another view is that satemization occurred in Proto-Baltic and Proto-Slavic independently, after the split between Proto-Baltic and Proto-Slavic.
Reflexes of PIE laryngeals */h₁/, */h₂/, */h₃/, which represented 3 different phonemes in PIE, became merged in Balto-Slavic to a single */H/. Laryngeals disappeared in the Balto-Slavic period over a very long period. No Balto-Slavic language has preserved them, but relative chronology of sound changes shows that they were not lost at once in all positions in a word.
The Balto-Slavic laryngeal was especially durable in a position before a vowel; PIE *tn̥h₂u- 'thin' (Latin tenuis, Sanskrit tanú) was in Balto-Slavic reflected as *tunHu-, and only then as Proto-Slavic *tunu-ku/*tinu-ku (OCS tьnъkъ, Russ. tónkij, Pol. cienki), which shows that the loss of laryngeals in Balto-Slavic occurred after the development of vocalic prothesis in Balto-Slavic syllabic sonorants.
In a syllabic position (between consonants), laryngeal disappeared if it was in the second syllable, but in the first syllable it was preserved as */a/. Compare:
- PIE *(h₁)rh̥₃deh₂ 'heron, stork' > (Ancient Greek erōdiós, Latin ardea) Proto-Slavic *radā > Common Slavic *roda (Serbo-Croatian róda)
- PIE *sh̥₂l- (oblique case stem of *seh₂ls 'salt') > OPr. sal, Proto-Slavic *sali (OCS solь, Pol. sól, Russ. sol´)
Loss of laryngeals in syllabic position occurred probably in early Balto-Slavic period. Compare:
- PIE *dʰugh₂tḗr > PBSl. *duktēr > Lith. duktė, OPr. duckti, Proto-Slavic *duktī or *duktej (OCS dъšti, Russ. doč’, Bulg. dъšterja)
The same phenomenon happened in Germanic and Celtic, which indicates that it might have been a dialectal isogloss in Late Proto-Indo-European.
PIE */s/ has been preserved in Balto-Slavic and Proto-Slavic in most of the positions; it changed to Balto-Slavic */š/ according to the RUKI law, and in Proto-Slavic it was probably lost word-finally. No attested Slavic language has word-final *-s preserved.
Before voiced stops */s/ manifested as */z/ in Balto-Slavic. That */z/ came to be phonologically distinctive in Slavic after the transition of Balto-Slavic */ź/ (a reflex of PIE */ǵ/ and */ǵʰ/) > Proto-Slavic */z/.
As a result of RUKI law, Proto-Slavic has */š/ before front vowels (*/e/, */i/), */x/ before back vowels and */s/ before consonants. That distribution is most probably a result of series of changes:
- PIE */s/ > */š/ after */r/, */u/, */k/, */i/
- */š/ > */s/ before consonants, */š/ > */x/ before vowels
- */x/ > */š/ before front vowels (Slavic first palatalization of velars)
RUKI rule also operated if there was a laryngeal after */u/ or */i/, i.e. */s/ changes to */š/ after *uH and *iH, but it remains open to debate whether the laryngeal was already lost in that environment, i.e. are we dealing with the change of */s/ to */š/ after Balto-Slavic */ū/ and */ī/.
In Baltic languages the evidence of RUKI rule is recognizable only in Lithuanian, because in Latvian and Old Prussian a merger occurs of Balto-Slavic */š/ (< PIE */s/ by RUKI rule), */ś/ (< PIE */ḱ/) and */s/ (< PIE */s/). In Lithuanian, Balto-Slavic */š/ and */ś/ are merged to /š/, which remains distinct from /s/ so the effect of RUKI rule is still evident in Lithuanian.
Most handbooks, on the basis of Lithuanian material, state that in Baltic RUKI law has been applied only partially. The most common claim is that Balto-Slavic */s/ turned to */š/ in Baltic unconditionally only after */r/, while after */u/, */k/ and */i/ we have both */s/ and */š/. Compare:
- Lith. aušrà; Latv. ausma < PIE *h₂ewsro- (cf. Latin aurōra, Sanskrit uṣás) with RUKI applied vs.
- Lith. ausìs; Latv. auss < PIE *h₂ews- (cf. Latin auris, OCS uxo) with */s/ unchanged.
Similarly, Lith. maĩšas/Latv. maiss "sack" completely matches etymologically with OCS měxъ and Sanskrit meṣá, but in the word teisùs "correct" */s/ has been preserved while in Slavic there is */x/ < */š/ in accordance with RUKI rule (OCS tixъ, Russ. tíxij 'quiet, peaceful').
There is no simple solution to such double reflexes of PIE */s/ after */r/, */u/, */k/, */i/ in Baltic, and thus no simple answer to the question of whether RUKI law is a common Balto-Slavic isogloss or not. The most probable seems the assumption that PIE */s/ was changed to */š/ after */r/, */u/, */k/, */i/ completely regularly in Baltic, just like in Slavic, but the traces of the effect of RUKI law were erased by subsequent changes, such as the change of word-final *-š to *-s.
Generally it can be ascertained that Baltic shows the effect of RUKI law only in old words inherited from Balto-Slavic period, meaning that Lithuanian /š/ will come after /r/, /u/, /k/, /i/ in words that have complete formational and morphological correspondence in Slavic (ruling out the possibility of accidental, parallel formations).
Unlike Indo-Iranian, where the change */s/ > */š/ also occurred after the palatovelar */ḱ/, it is possible that palatovelars yielded fricatives in Balto-Slavic even before the effect of RUKI law. Compare:
- PSl. *desnu 'right' (OCS desnъ, Russ. désnyj, SCr. dèsnī, Lith. dešinys) < PIE *deḱs-no- (Lat. dexter, Skt. dákṣiṇas)
- PSl. *asi 'axle, axis' (OCS osь, Russ. os’, SCr. ȏs, Lith. ašis) < PIE *h₂eḱsi- (Lat. axis, Skt. ákṣas)
By satemization of PIE dorsals and the merger of PIE laryngeals, Balto-Slavic has significantly modified the system of PIE fricatives. After the merger of PIE voiced and aspirated stop series, Balto-Slavic system of fricatives had the following shape:
Phonological relevance of the consonants */š/ and */z/ is disputed; it cannot be known whether they were phonologically predictable allophonic variants of */s/ and */z/ in all environments.
Phonological interpretation of the laryngeal */H/ is also disputed; on the basis of typological considerations it can be ascertained that the Balto-Slavic laryngeal was probably a voiceless glottal fricative [h] or a glottal stop [ʔ].
Proto-Indo-European sonorants */w/, */j/, */l/, */r/, */m/, */n/ were preserved in Balto-Slavic and Proto-Slavic, as they were in most other Indo-European branches. From context-conditioned sound laws, notable is the disappearance of word-initial PIE */w/ before */r/ and */l/ (so-called Lidén's law).
PIE */w/ was retained in Balto-Slavic and Proto-Slavic as a bilabial semivowel (glide), but in Lithuanian and most Slavic languages it has eventually changed to labiodental fricative /v/.
PIE */m/ changes to */n/ word-finally in Balto-Slavic period; in Old Prussian there is a clear attestation of that change e.g. in nominoaccusative of neuters (cf. OPr. assaran 'lake' < PIE *eǵʰerom). In Slavic however that change of *-m > *-n is indirect because in Common Slavic period all word-final vowels were dropped. It becomes more clear in sentence sandhi conditions due to which earlier *kom emōj yielded Proto-Slavic *kan jemъ (OCS kъ n'emu), and not **ka memō.
Syllabic sonorants 
PIE */i/ and */u/, syllabic allophones of PIE glides */y/ and */w/, have been preserved as vowels in Balto-Slavic. Before laryngeals they yielded long vowels *iH > */ī/, *uH > */ū/.
PIE */u/ has lengthened into Balto-Slavic */ū/ when followed by */n/ which was followed by a stop. In Slavic *-n- later drops regularly. Compare:
- PIE *Hunk 'to get used to' > PBSl. *ūnk > Lith. jùnkti, Latv. jûkt, OCS vyknǫti, Upp. Sorb. wuknyć
PIE */i/ did not exhibit lengthening in such conditions, as older literature often states.
In a syllabic position, PIE sonorants */l/, */r/, */m/, */n/ have twofold reflexes in Balto-Slavic, differing in a prothetic vowel: *iR and *uR (where symbol R denotes any of aforementioned sonorants). Analysis of their distribution has shown that the reflexes of type *iR are much more common. It has remained an unsolved mystery to this day which exact phonological conditions trigger which reflex.
Several theories have been proposed, most notable being one by André Vaillant. According to him, *uR reflexes arose after PIE labiovelars. If this was true, it would be the only trace of PIE labiovelars in Balto-Slavic.
Similarly, Jerzy Kuryłowicz thought that *uR reflexes arose after PIE velars, and also notable is also older opinion of Jānis Endzelīns and Reinhold Trautmann according to whom *uR reflexes are the result of zero-grade of morphemes that had PIE */o/ (> PBSl. */a/) in normal-grade. Matasović (2008) proposes following rules:
- PIE syllabic R > PBSl. *əR
- *ə > */i/ in a final syllable
- *ə > */u/ after velars and before nasals
- *ə > */i/ otherwise
Balto-Slavic preserved Late PIE vowel system, after the effect of "laryngeal colouring". The only exception is the change of PIE */o/ > PBSl. */a/, which is an isogloss shared with the Germanic and Anatolian branches.
Proto-Slavic preserved Balto-Slavic system of short vowels unchanged, but merged PBSl. */ō/ and */ā/ yielding PSl. */ā/, while the difference between these long vowels was preserved in Baltic.
PIE */e/ changes to PBSl. */a/ before */w/ in heterosyllabic position, i.e. */e/ > */a/ / _wV. Compare:
It appears that in some cases in Balto-Slavic period initial *(H)e- and *(H)a- were mixed. That change, called Rozwadowski's rule by some, is based on the cases where Balto-Slavic has initial *e- in etymons which in PIE had initial *(H)a-, *(H)o-, *h₂e-, *h₃e-. Slavic has preserved some relics of initial *e-, *a- alternations. Compare:
- PSl. *elawa, *alawa (Common Slavic *olovo) 'lead' > Middle Bulg. élav, Pol. ołów, Russ. ólovo as opposed to OPr. elwas 'tin'
Similar to vowels, PIE diphthongs were preserved in Balto-Slavic, with the exception of *ow (and also *h₂ew and *h₃ew), which yielded PBSl. *aw. Later in Proto-Slavic PBSl. *aw (< PIE *ow, *aw, *h₂ew, *h₃ew, *How) > PSl. */ō/, which was reflected as /u/ in all Slavic languages.
Relative chronology of sound changes 
Austrian Balto-Slavist Georg Holzer has reconstructed a relative chronology of 50 Balto-Slavic sound changes (just phonology, no accentuation), from Proto-Balto-Slavic down to the modern daughter languages. However, only the first 12 are Common Balto-Slavic, and thus relevant for this article:
- change of PIE */o/ > PBSl. */a/
- RUKI law
- PIE *CHC > PBSl. *CC
- Winter's law
- PIE. */Cʰ/ > PBSl. */C/
- PIE */Cʷ/ > PBSl. */C/
- PIE *ewC > PBSl. *jawC (e.g. PIE *h₁lewdʰ- > OCS ljudьje, Lith. liáudis)
- PIE *ewV > *awV
- PIE syllabic. */R̥/ > PBSl. *iR (*uR)
- PIE *#wr-, #wl- > *#r-, #l- (Lindemann's law, e.g. *wronkeh₂ > OCS rǫka)
- PIE *sr > PBSl. *str (Slavic, Latvian, but not in Lithuanian)
Balto-Slavic accentual system 
The Proto-Indo-European accent was completely reworked in Balto-Slavic, with far-reaching consequences for accentual systems of the modern daughter languages. For the reconstruction of the Balto-Slavic accent the most important are the Balto-Slavic languages that have retained tonal oppositions, these being Lithuanian, Latvian, (probably) Old Prussian and the West South Slavic languages of Slovene and Serbo-Croatian. However, one should keep in mind that the prosodical systems of dialects in the aforementioned languages are sometimes very different from those of standard languages. For example, some Croatian dialects like Čakavian and Posavian dialects of Slavonian Štokavian are especially important for Balto-Slavic accentology as they retain more archaic and complex tonal accentual system than the Neoštokavian dialect on which modern standard varieties of Serbo-Croatian (Bosnian, Croatian and Serbian) are based. On the other hand, many dialects have completely lost tonal oppositions (e.g. some Kajkavian varieties, the Zagreb spoken nonstandard idiom).
To this day, there is no consensus among Balto-Slavists on the precise details of the development of Balto-Slavic accentual system. All modern research is based on the seminal study of Stang (1957), which basically instituted the field of comparative Balto-Slavic accentology. However, a number of laws and correspondences have been discovered that are nowadays held to be true by the majority of researchers, even though the exact details sometimes remain in dispute.
What follows is a short overview of the commonly used diacritical marks for Balto-Slavic (BSl.) accents and/or prosodic features, all based on the example letter a. In each case, there is a crude characterization of the pronunciation in terms of High, Mid, and Low-tone sequences.
- Lithuanian: "falling"/HL (acute) á, "rising"/H(L)H (circumflex) ã, "short"/H à
- Latvian (on all syllables): "falling"/HL à, "rising"/LH (or "lengthened") ã, "broken"/L'H â
- Slovenian: "falling"/HL â, "rising"/LH á, "short"/H ȁ (sometimes also à)
- Serbo-Croatian: "short falling"/HL ȁ, "long falling"/HML ȃ, "short rising"/LH à, "long rising"/LMH á, "posttonic length" ā
- Common Slavic: "short falling"/HL (short circumflex) ȁ, "long falling"/HML (long circumflex) ȃ, "acute"/LH (old acute, old rising") a̋, "neoacute"/L(M)H (old acute, old rising") á or ã
In Croatian dialects, especially Čakavian and Posavian, the "new acute" (neoacute, the "new rising") is usually markied with tilde, as ã. Short neoacute ("short new rising") is marked as à. Neoacutes represent post–Proto-Slavic development.
Here is a reverse key to help decode the various diacritical marks:
- acute accent (á): Usually long rising and/or BSl. acute. Neoacute in some Slavic reconstructions. The default accent when a language has only one phonemic prosodic feature (e.g. stress in Russian, length in Czech). Marks long falling in Lithuanian because this derives from BSl. acute.
- grave accent (à): Usually short rising, or simply short.
- circumflex accent (â): BSl. circumflex in reconstructions. Broken tone in modern Baltic (Latvian and Žemaitian Lithuanian), i.e. a vowel with a glottal stop in the middle (derives from BSl. acute!). Long falling in modern Slavic languages.
- tilde (ã): Alternative notation for BSl. circumflex in reconstructions. Long rising in various modern languages (Lithuanian, Latvian, archaic Serbo-Croatian dialects such as Chakavian), deriving from diverse sources: Lithuanian < BSl. circumflex, Latvian < BSl. acute, Serbo-Croatian dialects < long Common Slavic neoacute (from accentual retraction).
- double grave accent (ȁ): Usually short falling (mostly in Slavic). Mnemonic: Derived from circumflex (= long falling) by converting the "acute" portion of the accent to a grave, much as a simple acute (= long rising) is shortened by conversion to a grave.
- double acute accent (a̋): Old acute in some Slavic reconstructions. (As opposed to single acute for Slavic neoacute in reconstructions. Based on the fact that the old acute was shortened in Common Slavic.)
- macron (ā): Vowel length, particularly in syllables without tone (e.g. unstressed syllables in Slavic).
- breve (ă): Vowel shortness.
Note that there are multiple competing systems used for different languages and different periods. The most important are:
- Three-way system of Proto-Slavic, Proto-Balto-Slavic, modern Lithuanian: Acute tone (á) vs. circumflex tone (â or ã) vs. short accent (à).
- Four-way Serbo-Croatian system, also used in Slovenian and often in Slavic reconstructions: long rising (á), short rising (à), long falling (â), short falling (ȁ).
- Two-way length: long (ā) vs. short (ă).
- Length only, as in Czech and Slovak: long (á) vs. short (a).
- Stress only, as in Russian, Ukrainian and Bulgarian: stressed (á) vs. unstressed (a).
Numerous non-prosodic marks are also found in various languages in combinations with certain letters. The various combinations of letter and diacritic should normally be viewed as single symbols, just like simple symbols (a, b, c).
Examples on vowels:
- ogonek (ą): With a rightward-curving hook, cf. the leftward-curving cedilla (ç): Vowel nasalization. In standard Lithuanian, the nasalization is historical and the vowels are nowadays simply reflected as long vowels; but some dialects still preserve nasalized vowels. Occasionally used to indicate low-mid quality in e, o.
- overdot (ė ȯ), underdot (ẹ ọ): High-mid vowel quality [e o], distinguished from plain e o indicating low-mid vowels [ɛ ɔ]. Overdot is normally found in Lithuanian, underdot in Slovenian.
- inverted breve below (e̯ i̯ o̯ u̯), indicating non-syllabic vowels (often, the second part of a diphthong).
- haček (ě): With pointed v shape, rather than the rounded u shape of the breve: ě in Slavic reconstructions is a vowel known as yat, distinct in length and later quality from simple e (originally longer and lower; later, longer and higher in many dialects); however ě in Czech indicates a simple e with palatalization of the preceding consonant.
- ô, ó, ů originally indicated a high-mid [o] or diphthongized [uo] in various Slavic languages (respectively: Slovak/dialectal Russian; Polish/Upper Sorbian/Lower Sorbian; Czech). It now indicates [u] in Polish and long [uː] in Czech.
Examples on consonants:
- Most diacritics on consonants indicate various sorts of palatal sounds, e.g. by use of an acute accent (ć ǵ ḱ ĺ ń ŕ ś ź), a comma (ģ ķ ļ ņ), a haček (č ď ľ ň ř š ť ž) or an overbar (đ). These indicate either:
- palatoalveolars (č š ž): These have a "hushing" pronunciation [tʃ ʃ ʒ], as in English kitchen, mission, vision, less palatal than the sounds indicated by ć ś ź;
- alveolopalatals (ć đ ś ź, e.g. in Polish and Serbo-Croatian);
- palatal stops (voiceless ḱ/ķ/ť and voiced ǵ/ģ/ď" in Macedonian, Latvian and Czech, respectively);
- a palatal nasal (ń ņ ň);
- a palatal lateral (ĺ ļ ľ); or
- a palatalized trill (ŕ, also ř in Czech specifically for a fricative trill).
- In Slovak, ń and ŕ indicate doubled rather than palatal(ized) consonants.
- In western West Slavic (Polish, Cassubian, Upper Sorbian and Lower Sorbian), ż indicates a voiced retroflex sibilant [ʐ]. (Other such sibilants are indicated by digraphs, e.g. cz, sz.)
- In western West Slavic, ł indicates a sound that was once a dark (velarized) l, but now is mostly pronounced [w].
Balto-Slavic acute and circumflex 
The development of the Proto-Indo-European accent in Proto-Balto-Slavic was conditioned by several delicate factors, such as the syllable length, presence of a laryngeal closing the syllable, and the position of PIE ictus.
In short syllables PIE tone became short accent in Balto-Slavic, and was preserved as such in both Baltic and Slavic branch, although its lengthening could be triggered by certain conditions. For example, in Lithuanian vowels /a/ and /e/ were lengthened when they initially bore short accent in open syllable, and rising tone emerged that is marked with tilde sign ã. Compare:
- PIE *kʷékʷlo- 'circle, wheel' > PBSl. *kákla- > Lith. kãklas 'neck', SCr. kȍlo
- PIE *déḱm̥t 'ten' > PBSl. déśimt > Lith. dẽšimt, SCr. dȅset
In long syllables, however, opposition between tones emerged, which are called acute and circumflex, i.e. acute (á) and circumflex (ã) accent. In this context, Balto-Slavic long syllables encompasses the following cases:
- syllables with long (lengthened-grade) PIE vowels */ē/ and */ō/, and long vowels which emerged in Balto-Slavic (e.g. by means of Winter's law)
- syllables with short vowels closed by a laryngeal (which merged to one Balto-Slavic laryngeal */H/)
- syllables with PIE diphthongs (i.e. all clusters of short vowels followed by a sonorant, including */y/ and */w/)
Balto-Slavic acute emerges in the following cases:
- in all syllables which were closed by a laryngeal in PIE, probably also when PIE laryngeal closed syllable with lengthened-grade vowel
- in all syllables which were closed by a voiced stop in PIE, and were lengthened in PBSl. according to the Winter's law
- in all cases of Balto-Slavic vrddhi, i.e. apophonical lengths (including new alternations *u/ū and *i/ī) which emerged only in Balto-Slavic period and have no PIE correspondences
- on long PBSl. */ū/ which was lengthened before *nC (this can be considered a case of new Balto-Slavic length, and grouped under the preceding case)
Balto-Slavic circumflex emerges in all the other syllables, and these are:
- PIE ablaut lengths
- PIE diphthongs (which were not followed by a laryngeal), i.e. all sequences of PIE short vowels and the sonorants (*/m/, */n/, */l/, */r/, */y/, */w/)
As one can see, rules governing the development of Balto-Slavic acute and circumflex accents seem to be very complicated, when formulated within the framework of "classical" Proto-Indo-European laryngeal theory. Dutch Indo-Europeanist Frederik Kortlandt has proposed an alternative, more elegant and economic rule for the derivation of Balto-Slavic acute: acute is a reflex of a glottal stop, which has two sources - merger of PIE laryngeals and the dissolution of PIE pre-glottalized stop ("voiced stops" in traditional reconstruction) to glottal stop and voiced stop, according to the Winter's law. Kortlandt's formulation appears very elegant initially, and seems to be confirmed independently by a glottal stop in Latvian as a reflex of Balto-Slavic acute in words in which accent was retracted, and is in accordance with the typological universal according to which in most languages high tone is developed in syllables closed with a glottal stop. Rising tone can then be explained as a result of the development of high tone on the second mora of a long syllable.
Though elegant, Kortlandt's theory also has some problems. Glottalic theory of Proto-Indo-European reconstruction which was proposed in the 1970s is not generally accepted among linguists, and today only a small minority of linguists would consider it a reliable and self-supportive framework onto which base modern Indo-European research. Also, there is a number of Balto-Slavic lexemes which point to acute accent but that are provably not of PIE laryngeal origin, and some of which were are result of apophonical lengthenings occurring only in Balto-Slavic period.
Matasović (2008) lists the following scenario as the most probable origin of Balto-Slavic acute:
- Rising tone, which we call Balto-Slavic acute, initially arose in the syllables closed by a laryngeal, partly due to the retraction of word-final accent onto such syllables which were phonologically long (Hirt's law). Other long syllables, if they bore the accent, were circumflexed (with falling tone).
- In later period new Balto-Slavic lengths were acuted.
- That younger acute has been largely eliminated in Slavic due to the effect of Meillet's law.
Reflexes in Balto-Slavic languages 
In (at least) the East Baltic languages, the acute register developed into the so-called "broken tone", a long vowel with a glottal stop in the middle of it, typically denoted by a circumflex accent, e.g. â [aʔa]. (Note that technically this is not a tone but a register distinction, much like the ngã tone in Northern Vietnamese.) This broken tone is preserved in syllables in certain dialects of Latvian and Lithuanian. Elsewhere, the difference between acute and circumflex is often continued as a tonal distinction.
The most direct continuation of the acute is in Latvian, particularly in the three-tone central dialects. In thee dialects, the Baltic broken tone of the acute register is directly continued as a broken tone (lauztā) in originally unstressed syllables, marked with a circumflex accent, e.g. luôgs "window". In originally stressed syllables, the acute register is continued as a rising or lengthened intonation (stieptā), marked with a tilde, e.g.luõks "spring onion". The circumflex register is generally continued as a falling intonation (krītošā), marked with a grave accent, e.g. lùogs "arch, bow". Note that can occur on all syllables, e.g. locative plural gal̂vâs "on the heads" (cf. Lithuanian galvosè with stress on a short final vowel that was deleted in Latvian), including monosyllables, e.g. dêt "to lay eggs" < *dêtì.
In Lithuanian, the distinction between acute and circumflex is only preserved on stressed syllables. In standard Lithuanian, based on the Aukštaitian dialect, the acute becomes a falling tone (so-called "Lithuanian metatony"), and is marked with an acute accent, while the circumflex becomes a rising tone, marked with a tilde. In diphthongs, the acute accent is placed on the first letter of the diphthong while the tilde marking rising tone (i.e. original circumflex) is placed on the second letter. In diphthongs which have a sonorant as a second part, the same convention is used, but the acute accent is replaced with a grave accent (e.g. Lith. pìlnas 'full' < PIE *plh₁nos). Word-finally the acute was regularly shortened: gerà 'good' (indefinite adjective) : geróji 'the good' (definite adjective). That rule is called Leskien's law, after the German neogrammarian August Leskien. Shortening operated according to Leskien's law after the Lithuanian metatony. In monosyllabic words the acute became circumflexed. Metatonical retraction of the accent from the final syllable to the penultimate syllable also created a circumflex automatically.
In Old Prussian the acute was reflected probably as a rising tone and circumflex as a falling tone. The marks on long vowels and diphthongs in Abel Will's translation of Martin Luther's Enchiridion point to that conclusion, which is the only accented Old Prussian text preserved. Diphthongs that correspond to a reconstructable Balto-Slavic acute are generally long in the second part of the diphthong, while those corresponding to a Balto–Slavic circumflex are generally long in the first part.
In the Slavic languages, the acute produced a rising tone and the circumflex a falling tone, as in Latvian and Old Prussian. In all Slavic languages the acute has been shortened, and a new rising accent (the neoacute), generally long, developed from retraction of the stress from an ultrashort yer vowel (later usually lost). The short rising accent from the old acute (and in some circumstances, the neoacute) was later lengthened again in a number of Slavic languages (e.g. Russian, Czech, Slovenian). The circumflex was shortened in some dialects as well (e.g. Polish, Russian, Czech, Slovak). Direct continuation of the acute vs. circumflex difference as a tonal distinction occurs only in archaic Serbo–Croatian dialects (e.g. Chakavian) and to some extent in Slovenian (although the relationship between Slovenian and Proto-Slavic tones and accent position is complex). In addition, the Proto-Slavic tonal distinction on liquid diphthongs is reflected fairly directly in Russian as a multi-syllable accent shape (pleophony), e.g. *ôr (falling) > óro, *ór (rising) > oró. In some other languages (most notably, Czech and standard neoshtokavian Serbo-Croatian), the acute vs. circumflex distinction is continued as a length distinction (although in all these languages, both long and short vowels have other sources as well). The same length-from-tone distinction once existed in Russian but has since been lost.
Here is a table of basic accentual correspondences of the first syllable of a word:
|Balto-Slavic and Proto-Slavic||Lithuanian||Old Prussian||Latvian||Serbo-Croatian||Slovenian||Czech||Russian|
Balto-Slavic fixed and mobile paradigms 
Proto-Balto-Slavic had, just like Proto-Indo-European, a class of nouns with so called "mobile" accentuation in which accent alternated between the word stem and the ending. These classes of nouns are usually reconstructed on the basis of Vedic Sanskrit and Ancient Greek, which have retained the position of the original PIE accent almost unchanged. However, by comparing the Balto-Slavic evidence, it was discovered that the PIE rules on accent alternations, devised on the basis of Vedic and Greek, do not match. Moreover, nouns that belong to mobile paradigms in Balto-Slavic belong to declension classes that had strictly fixed accent in PIE paradigms, i.e. ā-stems and o-stems. So for a long time the exact relationships between the accentuation of nouns in Balto-Slavic and PIE was one of the most mysterious questions of Indo-European studies, and some parts of the puzzle are missing to this day.
Research conducted by Christian Stang, Ferdinand de Saussure, Vladislav Illich-Svitych and Vladimir Dybo has led to a conclusion that Balto-Slavic nouns, with regard to accentuation, could be reduced to two paradigms: fixed and mobile. Nouns of the fixed paradigm had accent on the root, and in the nouns of the mobile paradigm the accent alternated between the root and the ending.
As shown by the Illič-Svityč, Balto-Slavic nouns of the fixed paradigm correspond to the PIE nouns with accent on the root (PIE barytones), the only exception being nouns with the accent on the ending (PIE oxytones) which was shifted onto the root in Balto Slavic in accordance with Hirt's law: such nouns also have fixed accent in Balto-Slavic.
The origin of the Balto-Slavic nouns of the mobile paradigm has not been completely determined, but the most probable and prevalent possibility is that proposed by Illič-Svityč, according to whom they originate as an analogical development from fixed-accent PIE oxytones. Nevertheless, it remains unclear why PIE nouns with fixed accent on the ending would become mobile, as analogies usually lead to uniformity and regularity.
The Balto-Slavic accentual system was further reworked during the Proto-Slavic and Common Slavic period (Dybo's law, Meillet's law, Ivšić's law etc.), resulting in 3 Common Slavic accentual paradigms (conveniently marked with letters as A, B, C) to correspond to 4 Lithuanian accentual paradigms (marked with numbers 1, 2, 3, 4) in a simple scheme:
|Balto-Slavic acute on the root|
|fixed accent||yes||a.p. 1/a.p. A||a.p. 2/a.p. B|
|no||a.p. 3/a.p. C||a.p. 4/a.p. C|
Common Slavic accentual paradigm a 
The simplest accentuation is that of nouns which were acuted on the root in Balto-Slavic. They remain accented on the root throughout the paradigm in Baltic (Lithuanian first accentual paradigm) and Slavic (accent paradigm a). In the same time, Both Baltic and Slavic have expected reflexes of Balto-Slavic acute:
In Russian the Balto-Slavic acute yielded expected reflex with "polnoglasie". In Serbo-Croatian the short falling accent in genitive plural has been substituted with long falling due to the loss of the yer.
Common Slavic accentual paradigm b 
In the nouns with non-mobile initial accent, which did not have an acuted root syllable, in both Lithuanian and Slavic an independent accent shift occurred from the root to the ending. In Lithuanian these are the nouns of the second accent paradigm, and in the Slavic of accent paradigm b.
Lithuanian noun rankà 'hand' etymologically corresponds to Russian ruká and Serbo-Croatian rúka, but both of these became mobile in a later Common Slavic development. So the reflexes of the Proto-Slavic noun *jōxā́ 'soup' are listed instead.
In Lithuanian the initial accent was preserved in all cases in which ending did not contain syllable with Balto-Slavic acute. In these cases (NVI sg, A pl) accent shifted onto the acuted ending, in accordance with the rule discovered by F. de Saussure. Later that acuted syllable was shortened due to the Leskien's law.
In Slavic the accent shifted from the root onto the ending in accordance with the Dybo's law, regardless of the syllable nature (i.e. whether it contained Balto-Slavic acute or not), so the nouns of the a.p. b are consistently accented on the ending (oxytonic, except in the I pl). In Neoštokavian dialects, which is used as a basis for standard Bosnian, Croatian and Serbian, accent was retracted from the ending onto the root syllable and became rising (so called "Neoštokavian retraction"). Old Štokavian and Čakavian dialects preserved the original ending-stressed paradigm.
Common Slavic accentual paradigm c 
Nouns with mobile accent had in some cases accented root, on the others the ending.
Lithuanian distinguishes two accent paradigms of these nouns: if they had acuted root, they belong to the third accent paradigm, and if the root was not acuted, by the operation of de Saussure's law the accent shifted onto the all acuted endings in the paradigm, so these nouns belong to the fourth accent paradigm.
In Proto-Slavic the acute has been eliminated in the nouns with mobile accentuation by the operation of Meillet's law, so therefore all the nouns with mobile accentuation belong to one accent paradigm, the so called accent paradigm c.
|Lithuanian||Russian||Neoštokavian Serbo-Croatian||Čakavian Serbo-Croatian||Common Slavic|
|D||gálv-ai||golov-é (OESl. gólov-ě)||glȃv-i||glāv-ȉ||*gȏlv-ě → *golv-ě̀|
Lithuanian has preserved the best Balto-Slavic mobile paradigm. In Neoštokavian the final accent has been retracted and gained rising intonation, and the Proto-Slavic initial accent is preserved as a circumflex.
Balto-Slavic apophony 
Indo-European ablaut has been significantly reworked in Balto-Slavic. Prominence of lengthened-grade has been significantly increased, as opposed to PIE in which it was used only for rare vrddhi-formations, nominative singulars of some consonant-stem nouns and sigmatic aorist.
Proto-Slavic abundantly used lengthened-grade in morphology. For example:
- PSl. *slāwā 'fame, glory' (OCS slava) vs. PSl. *slawa 'word' (OCS slovo)
- PSl. *twāri 'substance' (OCS tvarь) vs. PSl. *twarītej 'to form, create' (OCS tvoriti)
Similarly, in Lithuanian:
- Lith. prõtas 'intellect, mind' (< *prāt) vs. Lith. pràsti 'to understand'
- Lith. gė̃ris 'goodness' (< *gēr-) vs. Lith. gẽras 'good'
On the basis of already-present apophonic oppositions beween Balto-Slavic the long */ā/, */ē/, */ō/ and the short */a/ and */e/, new oppositions in arose between the long */ī/, */ū/ and the short */i/, */u/. This latter type of apophony was not productive in PIE. Compare:
- Lith. mū̃šis 'battle' vs. mùšti 'to kill, hit'
- Lith. lỹkis 'remainder' vs. lìkti 'to stay, keep'
This new type of apophonic length was especially used in Proto-Slavic in the formation of durative, iterative and imperfective verbs. Compare:
- PSl. *dirātej > OCS dьrati vs. PSl. *arz-dīrātej 'to tear' > OCS razdirati
- PSl. *birātej 'to pick' > OCS bьrati vs. PSl. *bīrātej 'to choose' > OCS birati
See also 
- For an alternative formulation, see Kortlandt (1978:12–24)
- Matasović (2008:86). For a more precise formulation of the rule, see Matasović (2005)
- Kortlandt (2002:8), rule 5.8.
- Matasović (2008:87)
- Matasović (2008:96)
- These were phonetically really vowels, but phonologically sonorants, as the syllabicity of the PIE */j/ and */w/ was predictable by a rule.
- Matasović (2008:109)
- Vaillant, André, "Grammaire comparée des langues slaves", (I-IV), IAC, Lyon 1950-77
- Holzer 2001, 2007
- This change was common for East Baltic and Slavic, but not for West Baltic.
- According to the data taken from Anatole V. Lyovin, An Introduction to the Languages of the World, Oxford University Press, New York - Oxford, 1997.
- All examples given for Serbo-Croatian are based on the standard language, i.e. stylised Neoštokavian dialect, and are accented accorcding to the Rječnik hrvatskoga jezika, F. Broz and I. Iveković, Zagreb 1901 and Akademijin Rječnik hrvatskoga ili srpskoga jezika, XXIII volumes, 1880-1976
- Some obsolete literature claims that Balto-Slavic acute is on all PIE long vowels, but this has been proven false.
- E.g. in Vietnamese and Old Chinese. There are, however, cases when syllables closed with glottal stop yielded low tone, e.g. in Tibetan and Halkomelem.
- Matasović (2008:136)
- Derksen, Rick (2008). Etymological Dictionary of the Slavic Inherited Lexicon. Leiden: Brill. P. 12.
- root is here understood in Proto-Slavic, not PIE sense