3. Slavic historical phonology

3.1. Sound changes from PIE to OCS: overviews

The development of the individual consonants from PIE to CS:

Table 5: Overview of the development of consonants from PIE to Common Slavic.
PIE PBS PS CS
*‍p *‍p *‍p *‍p
*‍b *‍b (+ Winter’s Law 3.2.8) *‍b *‍b
*‍ *‍b (3.2.9)
*‍t *‍t *‍t *‍t
*‍d *‍d (+ Winter’s Law 3.2.8) *‍d *‍d
*‍ *‍d (3.2.9)
*‍ *‍ś (3.2.1) *‍s *‍s
*‍ *‍ź (3.2.1; + Winter’s Law 3.2.8) *‍z *‍z
*‍g̑ʰ *‍ź (3.2.1; 3.2.9)
*‍k *‍k *‍k *‍k
*‍ *‍k (3.2.2) *‍č (1st pal.; 3.4.1)
*‍c (2nd pal.; 3.4.4)
*‍g *‍g (+ Winter’s Law 3.2.8) *‍g *‍g
*‍ *‍g (3.2.9) *‍ ž (1st pal.; 3.4.1)
*‍ *‍g (3.2.2; + Winter’s Law 3.2.8)
*‍gʷʰ *‍g (3.2.2; 3.2.9)
*‍ (d)z (2nd pal.; 3.4.4)
*‍s *‍s *‍s *‍s
*‍š (ruki; 3.2.3) *‍x *‍x
*‍š (1st pal.; 3.4.1)
*‍ś (2nd pal.; 3.4.4)
*‍h₁ *‍∅ (3.2.4) *‍∅ *‍∅
*‍h₂
*‍h₃
*‍r *‍r *‍r *‍r
*‍l *‍l *‍l *‍l
*‍m *‍m *‍m *‍m
*‍n *‍n *‍n *‍n
*‍ *‍ *‍ *‍j (3.4.7)
*‍ *‍ *‍ *‍v

The development of the vowels from PIE to CS:

Table 6: Overview of the development of vowels from PIE to Common Slavic.
PIE PBS PS CS
*‍i *‍i *‍i *‍ĭ
*‍e *‍e *‍e *‍e
*‍a (before heterosyll. *‍; 3.2.14) *‍a *‍o
*‍a *‍a (3.2.10)
*‍o *‍e (after *‍j; 3.4.7)
*‍u *‍u *‍u *‍ŭ
*‍ĭ (after *‍j; 3.4.7)
*‍ih *‍ī (3.2.4) *‍ī *‍i
*‍ē *‍ē *‍ē *‍ě
*‍eh₁ *‍ḛ̄ (3.2.4)
*‍a (after *‍j; 3.4.7)
*‍ah₂ *‍ā̰ (3.2.4) *‍ā *‍a
*‍ō *‍ō *‍ā (3.3.3)
*‍oh₁/h₂/h₃ *‍ō̰ (3.2.4)

The development of diphthongs from PIE to CS (cf. Andersen 1998a: 426–427):

Table 7: Overview of the development of diphthongs from PIE to Common Slavic.
PIE PBS PS CS
*‍ei̯ *‍ei̯ *‍ei̯ *‍i
*‍ai̯ *‍ai̯ *‍ai̯ *‍ě
*‍oi̯
*‍eu̯ *‍i̯au̯ (3.2.14) *‍i̯au̯ *‍u
*‍au̯ *‍au̯ *‍au̯ *‍u
*‍ou̯

The development of syllabic nasals from PIE to CS:

Table 8: Overview of the development of syllabic nasals from PIE to Common Slavic.
PIE PBS PS CS
*‍ *‍in (3.2.6) *‍in *‍ĭn (before vowel)
*‍ę (before consonant and word-finally)
*‍un (3.2.6) *‍un *‍ŭn (before vowel)
*‍ǫ (before consonant and word-finally)
*‍ *‍im (3.2.6) *‍im *‍ĭm (before vowel)
*‍ę (before consonant)
*‍um (3.2.6) *‍um *‍ŭm (before vowel)
*‍ǫ (before consonant)

The development of syllabic liquids from PIE to CS:

Table 9: : Overview of the development of syllabic liquids from PIE to Common Slavic.
PIE PBS PS CS
*‍ *‍ir (3.2.6) *‍ir *‍ĭr
*‍ur (3.2.6) *‍ur *‍ŭr
*‍ *‍il (3.2.6) *‍il *‍ĭl
ul (3.2.6) *‍ul *‍ŭl

For the further changes of the syllabic liquids in OCS see 2.2.

3.2. Sound changes from PIE to Proto-Balto-Slavic

Traditional presentations of the historical phonology of any language are usually organised around the development of specific sounds, e.g. the outcome of PIE *‍ in OCS in different positions. It may be more useful to see the sound changes in their relative chronology, however, so this is what we are going to try here (cf. Holzer 2007; Wandl 2020; Olander 2015: 47–67; for thoughts on the visualisation of sound changes in a relative chronology see Wandl & Theliz 2024; cf. also Wandl, Olander & List forthcoming). The chronology below more or less follows that of Olander (2015: 47–67).

As the relative order of several (groups of) sound changes cannot be determined with certainty, the order given below should be taken with a grain of salt.

3.2.1. Satəm assibilation

In the satəm branches of Indo-European (Armenian, Albanian, Indo-Iranian, Baltic and Slavic) the palatal stops series (whatever its age, whether PIE or post-PIE) *‍k̑ g̑ g̑ʰ undergoes assibilation to PBS *‍ś ź ź (see Clackson 2007: 49–53 for a discussion of the satəm change). In Slavic, these sibilants develop further to PS *‍s z z, while in Lithuanian they become š ž ž.

3.2.2. Delabialisation of labiovelars

The PIE labiovelar stops *‍kʷ gʷ gʷʰ lose their labial articulation and merge with the plain dorsals *‍k g gʰ. This innovation also takes place in Indo-Iranian and possibly also in Armenian and Albanian.

3.2.3. Ruki change

At a pre-stage of Proto-Balto-Slavic and Proto-Indo-Iranian, PIE *‍s develops an allophone *‍[ʂ] (> PBS *‍š; PIIr. *‍ʂ > Ved. , Av. š) after *‍ī̆ i̯ ū̆ u̯ r K (where *‍K represents voiceless, voiced unaspirated and voiced aspirated palatal, plain and velar dorsals). PBS *‍š yields PS (CS) *‍x, Lith. š, and Latv. s, OPr. s (merging again with the reflex of PIE *‍s).

Before an obstruent PBS *‍š reverts to *‍s in Slavic:

In Slavic, ruki *‍x is subsequently generalised almost everywhere there is an alternation between *‍s and *‍x; but in Lithuanian it is the non-ruki s that is generalised

According to Andersen (1968), the different trajectories taken by Slavic and Lithuanian can be explained by the outcome of PIE *‍:

3.2.4. Loss of laryngeals

The PIE laryngeals *‍h₁ h₂ h₃ were lost to different degrees in all Indo-European subgroups. Balto-Slavic has no traces of laryngeals (a) word-initially before a consonant, as opposed to Anatolian, Greek, Armenian and Phrygian; (b) word-initially before a vowel, as opposed to Anatolian; (c) between vowels.

Syllable-final laryngeals are also lost in pre-Proto-Balto-Slavic, but when this happens, a preceding vowel is lengthened and becomes acute (see 9.3.1), also across an intervening resonant (Olander 2009: 151–152; 2015: 47–48; cf. Jasanoff 2017: 74–103; Villanueva Svensson 2023).

Another possible source of acute tone are PIE plain long vowels, but this is disputed (see 9.3.1).

Between consonants it seems likely that the three laryngeals yielded PBS *‍a, although examples are difficult to find.

3.2.5. Hirt’s Law

In pre-Proto-Balto-Slavic the accent is retracted from the accented syllable to an immediately preceding syllable which ends in a postvocalic laryngeal (Olander 2009, xxx).

3.2.6. Diphthongisation of syllabic sonorants

The Baltic and Slavic languages have two different outcomes of PIE *‍ (*‍l̥ r̥ m̥ n̥): *‍iR and *‍uR (see the overview in Hock 2004: 6–7). The normal outcome is PBS *‍iR, which is the most frequent outcome in roots and the only outcome in all inflectional endings. We also find PBS *‍iR when the ablaut alternations are preserved, e.g. Lith. kir̃pti ‘cut’, prs. ker̃pa, past kir̃po.

In some cases we find *‍uR.

The double outcome has not found a good explanation (cf. e.g. Stang 1966: 77–82; Hock 2004: 6–7; Matasović 2004). It has been suggested that the u-outcome is due to dialect mixture with an extinct Indo-European dialect that had *‍uR from PIE *‍; this dialect would have been a centum dialect since the Balto-Slavic words with centum reflexes as a rule also have uR. Others assume that i is the regular zero-grade outcome in roots with e in the full grade, whereas u appears in the zero grade of roots with o in the full grade. Still others ascribe u to expressive formations or loanwords.

3.2.7. Accentual mobility

In pre-Proto-Balto-Slavic certain word-forms accented on the ending in PIE receive initial accent (or, perhaps more precisely, become unaccented). The formulation is disputed (see Olander 2009; Jasanoff 2017). Some authors consider it to be in principle inherited from PIE, others purely analogical, still others as the result of a sound law plus systematic analogical processes, still others as the result of a sound law.

My own suggestion (the “Mobility Law”) is that at some point in pre-Proto-Balto-Slavic a high tone became low (i.e. accented words became unaccented) if it was on a short final vowel or on a final hiatal structure: μ[+H] > [-H] / C₀#.

Two synchronic prosodic rules of Proto-Slavic find a natural diachronic explanation if one accepts that the Proto-Balto-Slavic accentual mobility is the result of accent loss (Andersen 2009a). The example word without clitics is PBS acc.sg *‍ˌgā̰lu̯ām > PS *‍ˌgālu̯ān (CS *‍gȏlvǫ) > Ru. gólovu, BCMS glȃvu ‘head’.

Šaxmatov’s Law in Slavic: “initially accented” word-forms in the mobile accent paradigms (i.e. unaccented) give up their accent to a preceding preposition.

Vasil´ev–Dolobko’s Law in Slavic: “initially accented” word-forms in the mobile accent paradigms (i.e. unaccented) give up their accent to a following enclitic – the accent is not lost by the Mobility Law since it is on a non-final syllable.

3.2.8. Winter’s Law

Before the deaspiration of voiced aspirated stops in pre-Proto-Balto-Slavic (3.2.9) a vowel is lengthened and becomes acute, also across a resonant, in front of voiced unaspirated stops (PIE *‍b d g gʷ g̑), but not in front of voiced aspirated stops (PIE *‍bʰ dʰ gʰ gʷʰ g̑ʰ).

3.2.9. Deaspiration of voiced aspirated stops

Like in Albanian and Anatolian, at a pre-stage of Balto-Slavic the voiced aspirated stops (*‍bʰ dʰ g̑ʰ gʰ gʷʰ) lose their aspiration and merge with the voiced unaspirated stops (*‍b d g̑ g gʷ). This change is later than Winter’s Law (3.2.8), which requires a distinction between voiced unaspirated and voiced unaspirated stops.

3.2.10. Delabialisation of *‍o to *‍a

PIE *‍o and *‍a merge in PBS *‍a, like in Germanic and Albanian. Since Winter’s Law (3.2.8) produces different outcomes of lengthened PIE *‍o and *‍a in Baltic, it must precede their merger.

3.2.11. Dissimilation of dental stops to *‍s plus stop

At a pre-stage of Proto-Balto-Slavic, the first dental stop of dental stop clusters becomes a sibilant: PIE *‍TT > PBS > PS *‍sT. The same outcome is seen in Tocharian, Greek and Iranian.

For Slavic see Andersen (1998a), 425–426.

3.2.12. Loss of word-final stops

Word-final stops are lost in pre-Proto-Balto-Slavic (but cf. 3.3.5).

The last example, with a short vowel before original *‍‑d, shows that that before Winter’s Law (lengthening before PIE *‍b d g̑ g gʷ; see 3.2.8) final voiced stops either became devoiced, or had already been lost. If not, we would have had PS *‍*‍ˌtā (CS *‍*‍ta).

3.2.13. Regressive voicing assimilation of obstruents

Balto-Slavic has regressive voicing assimilation of obstruents.

In Ukrainian, a voiced obstruent appears before a voiceless obstruent in cases like vezti [ʋezˈtɪ]. This is not an archaism (Andersen 1969), but the result of an analogical reintroduction of the voiced obstruent from prevocalic allomorphs, e.g. prs.3sg vezé. The reintroduction was made possible when new clusters of voiced plus voiceless obstruent arose following the loss of reduced vowels (3.4.12), e.g. Ukr. dúžka [ˈduʒkɐ] ‘handle’ (from CS *‍dǫ́žĭka, from CS *‍dǫgà a.p. b ‘arc’).

3.2.14. PIE *‍ē̆u̯ in Balto-Slavic

The development of PIE *‍ĕ̄u̯ in Balto-Slavic is complicated. When the sequence *‍ĕ̄u̯ occurred before a consonant or in final position, it apparently yielded *‍iau̯ in all cases.

Otherwise *‍eu̯ became *‍au̯ before a vowel.

In endings *‍ was lost again in analogy with endings with preserved *‍ĕ̄u̯ (before a vowel).

3.3. Sound changes from Proto-Balto-Slavic to Proto-Slavic

3.3.1. Dybo’s Law

By Dybo’s Law the accent is advanced from an accented non-acute syllable to a following syllable.

Dybo’s Law does not affect the non-desinentially accented word-forms of accent paradigm c lexemes since such word-forms with “initial” accent were in fact phonologically unaccented.

Some authors assume that Dybo’s Law did affect accent paradigm c lexemes – however, it did not target the unaccented word-forms, but such desinentially accented word-forms that had a disyllabic ending where the first syllable was non-acute (Rasmussen 1999: 478; Olander 2004; 2009: 142–143).

Cf. Saussure’s Law in Lithuanian (9.6.1), according to which the accent is advanced from a non-acute syllable to a following acute syllable.

Contrary to Dybo’s Law, Saussure’s law does affect the initial syllable of originally unaccented word-forms, indicating that when Saussure’s Law operated, these words had become accented. For the similarities and differences between Saussure’s Law and Dybo’s Law see Olander (2009), 116.

3.3.2. Meillet’s Law

In Slavic the distinction between acute and non-acute syllables is lost in unaccented position, including the first syllable of unaccented word-forms.

This leads to a merger of Proto-Balto-Slavic mobile words with acute and non-acute roots (i.e. words with a.p. 3 and 4 in Lithuanian) as a.p. c in Slavic (9.3.2).

3.3.3. Delabialisation of *‍ō to *‍ā

In Slavic, but not in Baltic, PBS *‍ō is delabialised to *‍ā, thereby merging with PBS *‍ā.

3.3.4. Loss of word-final fricatives

PIE final *‍s is preserved in Baltic, but not in Slavic. According to many authors, final *‍s disappeared tracelessly in Slavic; others maintain that a final *‍s affected the quality of a preceding vowel in pre-Proto-Slavic. It is difficult to speak of a communis opinio.

A limited version states that only *‍oi̯ was affected by a word-final *‍s, *‍‑oi̯s becoming CS *‍‑i (not *‍‑ě) (Kortlandt 2011a: 177–178):

According to this view, forms like OCS o-stem nom.sg vlĭkŭ ‘wolf’ must be analogical for *‍vlĭko; the ending ‑ŭ may have been taken over from u-stem nom.sg (e.g. OCS synŭ ‘son’_ < PIE *‍suHnús) or from o-stem acc.sg (e.g. OCS vlĭkŭ < PIE *‍u̯ĺ̥kʷom), where ‑ŭ < *‍‑om is regular (see 3.3.5).

A relatively widespread view maintains that *‍o was also affected, *‍‑os regularly becoming CS *‍‑ŭ:

Within this framework it is problematic that neuter s-stem nom–acc.sg shows -o, not as expected.

These forms may have analogical -o, ultimately from pronouns (*‍tód > OCS to etc.).

An extended version claims that pre-PS *‍ā (from PIE *‍ah₂) was also affected, pre-PS *‍-ās becoming CS *‍-y (Meillet 1897: 96, 104, 125).

An even more extended version claims that any pre-PS *‍ā̆ in a final syllable ending in *‍s is affected, and that the Old Novgorod dialect has a different outcome (Olander 2012; 2015: 56–57; Viredaz 2009; partly Zaliznjak 2004: 148–149): pre-PS *‍a ā > PS *‍ə ə̄ > Classical CS *‍ŭ y (merging with PS *‍u ū) || Old Novgorod dialect e ě (merging with PS *‍e ē).

3.3.5. Loss of word-final nasals

Final *‍m is lost after a short vowel in pre-Proto-Slavic; a preceding PIE *‍o becomes PS *‍u > CS *‍ŭ.

According to the communis opinio final *‍n also disappears after a short vowel in Slavic, except if preserved by a final dental.

This would mean that the loss of word-final stops took place independently in Baltic and Slavic. An alternative hypothesis would be to ascribe the loss of word-final stops to the pre-Proto-Balto-Slavic period, and to assume that the different outcomes of PIE *‍‑V̆m and *‍‑V̆nd in Slavic is conditioned by the different nasals (*‍m vs. *‍n) (Olander 2010; but cf. Hill 2013: 172 n. 7). In the alternative scenario, *‍m disappears after a short vowel in Slavic, whereas final *‍n is preserved and eventually forms a nasal vowel with the preceding vowel.

3.3.6. Loss of syllable-final stops

Before a back vowel PS *‍kt and *‍gd yield CS *‍t and *‍d. However, *‍kt gd before front vowels merge with *‍ti̯ di̯ in CS *‍tʲ dʲ

3.4. Common Slavic phonological developments

3.4.1. First Palatalisation

Before front vowels and *‍, PS *‍k g x are palatalised to CS *‍č ž š.

Note that in Klotz (2023) the First Palatalisation is considered to be pre-Proto-Slavic.

3.4.2. Fronting of vowels after palatal consonants

One of the defining changes of the Slavic vowel system was the fronting of non-front vowels after palatal consonants: PS *‍a ā̄ u ū ə ə̄ appear as CS *‍o a ŭ y ŭ||e y||ě after non-palatal consonants, but *‍e ě ĭ i ĭ ě(?) after palatal consonants.

3.4.3. Monophthongisation of oral diphthongs

The Proto-Slavic oral diphthongs *‍ā̆i̯ ē̆i̯ ā̆u̯ ē̆u̯ are monophthongised to CS *‍ě i u (j)u.

3.4.4. Second Palatalisation

After the monophthongisation of PS *‍ai̯ əi̯ to *‍ē ī, the Slavic velars (PS *‍k g x) underwent a second round of palatalisations, this time to CS *‍c ʒ ś, in two positions:

  1. before the new front vowels CS ě and i
  2. between PS *‍ī̆(n) and a non-high vowel (*‍ā̆ or *‍ə̄̆)

Formula: PS *‍k g x > CS *‍c ʒ ś / _(u̯)V[+front] (“Second Palatalisation”) and ī(n)_V[–high] (“Third Palatalisation”)

Note that CS *‍ś is a shorthand for the outcome *‍s in South and East Slavic and *‍š in West Slavic.

In most Slavic dialects CS *‍ʒ is simplified to z, but in Polish it is retained as dz.

The palatalisation across *‍ (PS *‍ku̯, *‍gu̯, *‍xu̯ > CS *‍cv, *‍ʒv, *‍śv) takes place in South Slavic and some East Slavic dialects, but not in West Slavic and other East Slavic dialects.

The Old Novgorod dialect is only partly affected: there is no regressive palatalisation and probably no progressive palatalisation of *‍x and perhaps *‍g (see 1.1.3).

Some authors consider the Second and Third Palatalisations to be different processes (mostly with the Third Palatalisation preceding the Second Palatalisation) (e.g. Pedersen 1905; Wandl 2020).

Others consider them to be the same process as they yield exactly the same results in the Slavic dialects (e.g. Andersen 1998b: 588–589; Vermeer 2003; Olander 2015: 63–64 with further references).

3.4.5. Liquid Metathesis

The Liquid Metathesis refers to the process whereby inherited sequences of PS *‍ā̆ or *‍ē̆ followed by *‍r or *‍l undergo various changes in the Slavic languages. It is one of the diagnostics in the traditional subgrouping of Slavic (see 1.1.2). The general outcomes in the Slavic languages are shown in Table 10.

Table 10: The outcomes of the liquid diphthongs in the Slavic languages.
PS CS SSl. ESl. Po., LSorb., USorb. Cz. Slk.
*‍-ā̆r- -ā̆l- *‍-or- -ol- ra la -oro- ‑olo- -ro- -lo- -ra- -la- -ra- -la-
*‍ar- al- *‍ȏr- ȏl- ro- lo- ro- lo- ro- lo- ra- la-
*‍ār- āl- *‍őr- ől- ra- la- ra- la- ra- la- ra- la-
*‍ē̆r ē̆l *‍er el rě lě ere olo re le rě lě re le

Examples:

Fun fact: Russian toponyms in ‑grad – e.g. Saint Petersburg’s former names Petrográd (1914–1924) and Leningrád (1924–1991); Volgográd and its earlier name Stalingrа́d (1925–1961) – have the South Slavic version. The East Slavic outcome is seen in toponyms in ‑gorod (e.g. Nóvgorod).

3.4.6. Common Slavic vowel contractions

The Slavic languages, including OCS, have varying degrees of vowel contractions (Andersen 2014).

3.4.7. Loss of postconsonantal *‍

PIE (and Proto-Balto-Slavic) consonants followed by *‍ underwent significant changes in Slavic. The first step was loss of the *‍ with palatalisation of the preceding consonant: PIE *‍Ci̯ > PS *‍Ci̯ > CS *‍ (see e.g. Andersen 1998a: 431; Olander 2015: 65).

The outcome of dental stops plus *‍ – PS *‍ti̯ di̯ (CS *‍tʲ dʲ; also written *‍t’ d’) – is one of the main diagnostics for the internal subgrouping of Slavic (see 1.1.2), although the South Slavic languages have diverging results.

Table 11: Loss of postconsonantal *‍
PS *‍ti̯ *‍di̯
CS *‍ *‍
OCS št žd
Bulg. št žd
Mac. ǵ
BCMS ć đ
Sln. č j
East Sl. č ž
West Sl. c dz / z

The reflexes of PS *‍su̯ai̯ˈti̯ā (CS *‍světʲà) f ‘light’ and *‍meˈdi̯ā (CS *‍medʲà) f ‘border’ are shown in Table 12.

Fun fact: The preposition CS *‍medʲù ‘between’ is originally a locative dual form of *‍medjà ‘border’; the -žd- of Ru. méždu shows it is a borrowing from Church Slavonic.

Table 12: Loss of postconsonantal *‍: examples
PS *‍ti̯ *‍su̯ai̯ˈti̯ā ‘light’ *‍di̯ *‍meˈdi̯ā f ‘border’
CS *‍ *‍světʲà *‍ *‍medʲà
OCS št svěšta žd mežda
Bulg. št svéšt žd meždá
Mac. sveḱa ǵ meǵa
BCMS ć svéća, svijèća đ mèđa
Sln. č svẹ́ča j méja
ESl. č Ru. svečá ž Ru. mežá
WSl.: Po. c świeca dz miedza
WSl.: Cz. svíce z meze
WSl.: USrb. swěca mjeza

PS *‍kt followed by a front vowel merges with PS *‍ti̯ as CS *‍.

Before a non-front vowel PS *‍kt yields CS *‍t.

If the consonant was a labial, *‍l was inserted after the consonant in Common Slavic.

A following front vowel may have prevented the insertion of *‍l.

The alternation of allomorphs with and without *‍ was not tolerated, an *‍ was normally generalised, e.g. Ru. zemljá, loc zemljí.

3.4.8. Backing of *‍ē to *‍ā after palatalised consonants

After a palatalised consonant *‍ē becomes *‍ā.

The change is later than the first palatalisation (3.4.1), the vowel fronting after palatal consonants (3.4.2), and the monophthongisation of oral diphthongs (3.4.3); cf. CS *‍zemʲì above, not *‍*‍zemlʲě̀. It takes place before the monophthongisation of nasal diphthongs (3.4.9); cf. CS *‍pišǫ̀ above, not *‍*‍pišę̀.

3.4.9. Monophthongisation of nasal diphthongs

In the Common Slavic period nasal diphthongs were monophthongised to nasal vowels: *‍ā̆n ē̆n ī̆n ū̆n > CS *‍ǫ ę ę ǫ. For the hypthesised reflexes of *‍ə̄̆n see Olander (2015: 66).

3.4.10. Reinterpretation of vowel quantity as quality

In all Slavic languages, at some point vowel quantity was reinterpreted as vowel quality: *‍ī ē ā ū ȳ i e a u > CS *‍i ě a u ĭ e o ŭ. This systemic change affects all Slavic words.

The change is referred to as the “Second CS Vowel Shift (VS₂)” by Andersen (1998c).

3.4.11. Stang’s Law (Slavic)

The accent was retracted to a preceding syllable from, from reduced vowels in weak position, from short (i.e. non-acute) medial diphthongs and from contracting syllables. The syllable receiving the accent acquired neoacute tone.

3.4.12. The loss of the reduced vowels (jers)

The Common Slavic vowel system included the two so-called “reduced vowels”, or jers, written *‍ĭ ŭ or *‍ь ъ and normally reflecting PS *‍i u. In “Urkirchenslavisch”, the original language of Cyril and Methodius, they may have been pronounced [ɪ ʊ].

The reduced vowels are to some extent preserved as ĭ ŭ in the most conservative OCS and Old East Slavic manuscripts. In most manuscripts, however, and in the remaining Slavic languages, the reduced vowels have two different types of reflexes, depending on their position in the word.

A reduced vowel is in weak position if it is the final vowel of the word, or if it is followed by a non-reduced vowel; and in strong position if it followed by a reduced vowel in weak position. This rule is referred to as “Havlík’s Law”.

In weak position the reduced vowel disappears, although in many cases ĭ causes palatalisation of the preceding consonant. In strong position, the reduced vowel becomes a full vowel, with different qualities in the various Slavic languages (see Table 13).

At the time of the writing of the actual extant OCS manuscripts, the speakers had probably already lost the weak reduced vowels and replaced the strong ones with the full vowels e o.

In all Slavic languages reduced vowels in strong position merge with other vowels.

Table 13: Reflexes of the reduced vowels in strong position in the Slavic languages.
CS *‍ĭ *‍ŭ
OCS ĭ (e) ŭ (o)
Bulg. e ă [ə]
Mac. e o
BCMS a a
Sln. a e [ə]
East Sl. e o
West Sl. e e

Examples of reduced vowels in strong and weak position (an underlined ŭ or ĭ in the Common Slavic reconstruction indicates strong position).

Reduced vowels in weak position are lost in all Slavic languages, with one exception: Polabian treats a weak reduced vowel as strong if it occurs in the word’s first syllable in accented or immediately pretonic position