Proto-Slavic is the unattested, reconstructed proto-language of all the Slavic languages. It represents Slavic speech approximately from the 5th to 9th centuries AD. As with most other proto-languages, no attested writings have been found; scholars have reconstructed the language by applying the comparative method to all the attested Slavic languages and by taking into account other Indo-European languages.
Rapid development of Slavic speech occurred during the Proto-Slavic period, coinciding with the massive expansion of the Slavic-speaking area. Dialectal differentiation occurred early on during this period, but overall linguistic unity and mutual intelligibility continued for several centuries, into the 10th century or later. During this period, many sound changes diffused across the entire area, often uniformly. This makes it inconvenient to maintain the traditional definition of a proto-language as the latest reconstructable common ancestor of a language group, with no dialectal differentiation. (This would necessitate treating all pan-Slavic changes after the 6th century or so as part of the separate histories of the various daughter languages.) Instead, Slavicists typically handle the entire period of dialectally-differentiated linguistic unity as Common Slavic.
One can divide the Proto-Slavic/Common-Slavic time of linguistic unity roughly into three periods:
Authorities differ as to which periods should be included in Proto-Slavic and in Common Slavic. The language described in this article generally reflects the middle period, usually termed Late Proto-Slavic (sometimes Middle Common Slavic) and often dated to around the 7th to 8th centuries. This language remains largely unattested, but a late-period variant, representing the late 9th-century dialect spoken around Thessaloniki in Greek Macedonia, is attested in Old Church Slavonic manuscripts.
The ancestor of Proto-Slavic is Proto-Balto-Slavic, which is also the ancestor of the Baltic languages, e.g. Lithuanian and Latvian. This language in turn is descended from Proto-Indo-European, the parent language of the vast majority of European languages (including English, German, Spanish, French, etc.). Proto-Slavic gradually evolved into the various Slavic languages during the latter half of the first millennium AD, concurrent with the explosive growth of the Slavic-speaking area.
There is no scholarly consensus concerning either the number of stages involved in the development of the language (its periodization) or the terms used to describe them. For consistency and convenience, this article adopts the following scheme (as does the article History of the Slavic languages, which see for further discussion of the historical and linguistic development of Proto-Slavic from Proto-Balto-Slavic, and the further development of Proto-Slavic into the modern Slavic languages).
Proto-Slavic is divided into periods. One division is made up of three periods:
Another division is made up of four periods:
This article considers primarily Middle Common Slavic, noting when there is slight dialectal variation. It also covers Late Common Slavic when there are significant developments that are shared (more or less) identically among all Slavic languages.
Two different and conflicting systems for denoting vowels are commonly in use in Indo-European and Balto-Slavic linguistics on one hand, and Slavic linguistics on the other. In the first, vowel length is consistently distinguished with a macron above the letter, while in the latter it is not clearly indicated. The following table explains these differences:
|Short close front vowel (front yer)||i||ĭ or ь|
|Short close back vowel (back yer)||u||ŭ or ъ|
|Short open back vowel||a||o|
|Long close front vowel||ī||i|
|Long close back vowel||ū||y|
|Long open front vowel (yat)||ē||ě|
|Long open back vowel||ā||a|
For consistency, all discussions of words in Early Slavic and before (the boundary corresponding roughly to the monophthongization of diphthongs, and the Slavic second palatalization) use the common Balto-Slavic notation of vowels. Discussions of Middle and Late Common Slavic, as well as later dialects, use the Slavic notation.
For Middle and Late Common Slavic, the following marks are used to indicate tone and length distinctions on vowels, based on the standard notation in Serbo-Croatian:
There are unfortunately multiple competing systems used to indicate prosody in different Balto-Slavic languages (see Proto-Balto-Slavic language#Notation for more details). The most important for this article are:
The following is an overview of the phonemes that are reconstructible for Middle Common Slavic.
Middle Common Slavic had the following vowel system:
The columns marked "central" and "back" may alternatively be interpreted as "back unrounded" and "back rounded" respectively, but rounding of back vowels was distinctive only between the vowels *y and *u. The other back vowels had optional non-distinctive rounding. Thus:
The vowels described as "short" and "long" were simultaneously distinguished by length and quality in Middle Common Slavic. Vowel length evolved as follows:
Some authors avoid the terms "short" and "long", using "lax" and "tense" instead.
Middle Common Slavic had the following consonants:
The phonetic value (IPA symbol) of most consonants is the same as their traditional spelling. Some notes and exceptions:
In most dialects, non-distinctive palatalization was probably present on all consonants that occurred before front vowels. When the high front yer *ь/ĭ was lost in many words, it left this palatalization as a "residue", which then became distinctive, producing a phonemic distinction between palatalized and non-palatalized alveolars and labials. In the process, the palatal sonorants *ľ *ň *ř merged with alveolar *l *n *r before front vowels, with both becoming *lʲ *nʲ *rʲ. Subsequently, some palatalized consonants lost their palatalization in some environments, merging with their non-palatal counterparts. This happened the least in Russian and the most in Czech. Palatalized consonants never developed in Southwest Slavic (modern Croatian, Serbian, and Slovenian), and the merger of *ľ *ň ř with *l *n r did not happen before front vowels (although Serbian and Croatian later merged ř with r).
As in its ancestors, Proto-Balto-Slavic and Proto-Indo-European, one syllable of each Common Slavic word was accented (carried more prominence). The placement of the accent was free and thus phonemic; it could occur on any syllable and its placement was inherently part of the word. The accent could also be either mobile or fixed, meaning that inflected forms of a word could have the accent on different syllables depending on the ending, or always on the same syllable.
Common Slavic vowels also had a pitch accent. In Middle Common Slavic, all accented long vowels, nasal vowels and liquid diphthongs had a distinction between two pitch accents, traditionally called "acute" and "circumflex" accent. The acute accent was pronounced with rising intonation, while the circumflex accent had a falling intonation. Short vowels (*e *o *ь/ĭ ъ/ŭ) had no pitch distinction, and were always pronounced with falling intonation. Unaccented (unstressed) vowels never had tonal distinctions, but could still have length distinctions. These rules are similar to the restrictions that apply to the pitch accent in Slovene.
In the Late Common Slavic period, several sound changes occurred. Long vowels bearing the acute (long rising) accent were usually shortened, resulting in a short rising intonation. Some short vowels were lengthened, creating new long falling vowels. A third type of pitch accent developed, known as the "neoacute", as a result of sound laws that retracted the accent (moved it to the preceding syllable). This occurred at a time when the Slavic-speaking area was already dialectally differentiated, and usually syllables with the acute and/or circumflex accent were shortened around the same time. Hence it is unclear whether there was ever a period in any dialect when there were three phonemically distinct pitch accents on long vowels. Nevertheless, taken together, these changes significantly altered the distribution of the pitch accents and vowel length, to the point that by the end of the Late Common Slavic period almost any vowel could be short or long, and almost any accented vowel could have falling or rising pitch.
Most syllables in Middle Common Slavic were open. The only closed syllables were those that ended in a liquid (*l or *r), forming liquid diphthongs, and in such syllables, the preceding vowel had to be short. Consonant clusters were permitted, but only at the beginning of a syllable. Such a cluster was syllabified with the cluster entirely in the following syllable, contrary to the syllabification rules that are known to apply to most languages. For example, *bogatĭstvo "wealth" was divided into syllables as *bo-ga-tĭ-stvo, with the whole cluster -stv- at the beginning of the syllable.
By the beginning of the Late Common Slavic period, all or nearly all syllables had become open as a result of developments in the liquid diphthongs. Syllables with liquid diphthongs beginning with an o or e had been converted into open syllables, e.g. *tort became *trot, *trat or *torot. The main exception are the Northern Lekhitic languages (Kashubian, extinct Slovincian and Polabian) only with lengthening of the syllable and no metathesis (*tart, e.g. PSl. *gord > Csb. gard; > Plb. *gard > gord). In West Slavic and South Slavic, liquid diphthongs beginning with ĭ or ŭ had likewise been converted into open syllables by converting the following liquid into a syllabic sonorant (palatal or non-palatal according to whether an ĭ or ŭ preceded). This left no closed syllables at all in these languages. The South Slavic languages, as well as Czech and Slovak, tended to preserve the syllabic sonorants, but in the Lekhitic languages (e.g. Polish), they fell apart again into vowel-consonant or consonant-vowel combinations. In East Slavic, the liquid diphthongs in ĭ or ŭ may have likewise become syllabic sonorants, but if so, the change was soon reversed, suggesting that it may never have happened in the first place.
Proto-Slavic retained several of the grammatical categories inherited from Proto-Indo-European, especially in nominals (nouns and adjectives). Seven of the eight Indo-European cases had been retained (nominative, accusative, locative, genitive, dative, instrumental, vocative). The ablative had merged with the genitive. It also retained full use of the singular, dual and plural numbers, and still maintained a distinction between masculine, feminine and neuter gender. However, verbs had become much more simplified, but displayed their own unique innovations.
As a result of the three palatalizations and the fronting of vowels before palatal consonants, both consonant and vowel alternations were frequent in paradigms, as well as in word derivation.
The following table lists various consonant alternations that occurred in Proto-Slavic, as a result of various suffixes or endings being attached to stems:
|+t (in infinitive)||t||t||t1||t2||st||st||st||st||t2||lt3||rt3||ť||ť||?||t1|
Vowels were fronted when following a palatal or "soft" consonant (*j, any iotated consonant, or a consonant that had been affected by the progressive palatalization). Because of this, most vowels occurred in pairs, depending on the preceding consonant.
|After hard consonants||o||e||ь||ъ||a||ě₁||i||y||ǫ||ę||ę, ь||ǫ, ъ||y||u||ě₂||i|
|After soft consonants||e||ь||a||i||ǫ||ę||ę, ь||ę̇, ь||ę̇||u||i|
Most word stems therefore became classed as either "soft" or "hard", depending on whether their endings used soft (fronted) vowels or the original hard vowels. Hard stems displayed consonant alternations before endings with front vowels as a result of the two regressive palatalizations and iotation.
As part of its Indo-European heritage, Proto-Slavic also retained ablaut alternations, although these had been reduced to unproductive relics. The following table lists the combinations (vowel softening may alter the outcomes).
|zero grade||?||ь||ъ||ьl, ъl||ьr, ъr||ę, ǫ|
Although qualitative alternations (e-grade versus o-grade versus zero grade) were no longer productive, the Balto-Slavic languages had innovated a new kind of ablaut, in which length was the primary distinction. This created two new alternation patterns, which did not exist in PIE: short *e, *o, *ь, *ъ versus long *ě, *a, *i, *y. This type of alternation may have still been productive in Proto-Slavic, as a way to form imperfective verbs from perfective ones.
Most of the Proto-Indo-European declensional classes were retained. Some, such as u-stems and masculine i-stems, were gradually falling out of use and being replaced by other, more productive classes.
Adjective inflection had become more simplified compared to Proto-Indo-European. Only a single paradigm (in both hard and soft form) existed, descending from the PIE o- and a-stem inflection. I-stem and u-stem adjectives no longer existed. The present participle (from PIE *-nt-) still retained consonant stem endings.
Proto-Slavic had developed a distinction between "indefinite" and "definite" adjective inflection, much like Germanic strong and weak inflection. The definite inflection was used to refer to specific or known entities, similar to the use of the definite article "the" in English, while the indefinite inflection was unspecific or referred to unknown or arbitrary entities, like the English indefinite article "a". The indefinite inflection was identical to the inflection of o- and a-stem nouns, while the definite inflection was formed by suffixing the relative/anaphoric pronoun *jь to the end of the normal inflectional endings. Both the adjective and the suffixed pronoun were presumably declined as separate words originally, but already within Proto-Slavic they had become contracted and fused to some extent.
The Proto-Slavic system of verbal inflection was somewhat simplified from the verbal system of Proto-Indo-European (PIE), although it was still rich in tenses, conjugations and verb-forming suffixes.
The PIE mediopassive voice disappeared entirely except for the isolated form vědě "I know" in Old Church Slavonic (< Late PIE *woid-ai, a perfect mediopassive formation). However, a new analytic mediopassive was formed using the reflexive particle *sę, much as in the Romance languages. The imperative and subjunctive moods disappeared, while the old optative came to be used as the imperative instead.
In terms of PIE tense/aspect forms, the PIE imperfect was lost or merged with the PIE thematic aorist, and the PIE perfect was lost other than in the stem of the irregular verb *věděti "to know" (from PIE *woyd-). The aorist was retained, preserving the PIE thematic and sigmatic aorist types (the former is generally termed the root aorist in Slavic studies), and a new productive aorist arose from the sigmatic aorist by various analogical changes, e.g. replacing some of the original endings with thematic endings. (A similar development is observed in Greek and Sanskrit. In all three cases, the likely trigger was the phonological reduction of clusters like *-ss, *-st that arose when the original athematic endings were attached to the sigmatic *-s- affix.) A new synthetic imperfect was created by attaching a combination of the root and productive aorist endings to a stem suffix *-ěa- or *-aa-, of disputed origin. Various compound tenses were created, e.g. to express the future, conditional, perfect and pluperfect.
The three numbers (singular, dual and plural) were all maintained, as were the different athematic and thematic endings. (Only five athematic verbs exist: *věděti "to know", *byti "to be", *dati "to give", *ěsti "to eat" and *iměti "to have". dati has a finite stem *dad-, suggesting derivation by some sort of reduplication.) A new set of "semi-thematic" endings were formed by analogy (corresponding to modern conjugation class II), combining the thematic first singular ending with otherwise athematic endings. Proto-Slavic also maintained a large number of non-finite formations, including the infinitive, the supine, a verbal noun, and five participles (present active, present passive, past active, past passive and resultative). In large measure these directly continue PIE formations.
Proto-Indo-European had an extensive system of aspectual distinctions ("present" vs. "aorist" vs. "perfect" in traditional terminology), found throughout the system. Proto-Slavic maintained part of this, distinguishing between aorist and imperfect in the past tense. In addition, Proto-Slavic evolved a means of forming lexical aspect (verbs inherently marked with a particular aspect) using various prefixes and suffixes, which was eventually extended into a systematic means of specifying grammatical aspect using pairs of related lexical verbs, each with the same meaning as the other but inherently marked as either imperfective (denoting an ongoing action) or perfective (denoting a completed action). The two sets of verbs interrelate in three primary ways:
In Proto-Slavic and Old Church Slavonic, the old and new aspect systems coexisted, but the new aspect has gradually displaced the old one, and as a result most modern Slavic languages have lost the old imperfect, aorist, and most participles. A major exception, however, is Bulgarian (and also Macedonian to a fair extent), which has maintained both old and new systems and combined them to express fine shades of aspectual meaning. For example, in addition to imperfective imperfect forms and perfective aorist forms, Bulgarian can form a perfective imperfect (usually expressing a repeated series of completed actions considered subordinate to the "major" past actions) and an imperfective aorist (for "major" past events whose completion is not relevant to the narration).
Proto-Slavic also had paired motion verbs (e.g. "run", "walk", "swim", "fly", but also "ride", "carry", "lead", "chase", etc.). One of the pair expresses determinate action (motion to a specified place, e.g. "I walked to my friend's house") and the other expressing indeterminate action (motion to and then back, and motion without a specified goal). These pairs are generally related using either the suffixing or suppletive strategies of forming aspectual verbs. Each of the pair is also in fact a pair of perfective vs. imperfective verbs, where the perfective variant often uses a prefix *po-.
Many different PIE verb classes were retained in Proto-Slavic, including (among others) simple thematic presents, presents in *-n- and *-y-, stative verbs in *-ē- (cf. similar verbs in the Latin -ēre conjugation), factitive verbs in *-ā- (cf. the Latin -āre conjugation), and o-grade causatives in *-éye-.
The forms of each verb were based on two basic stems, one for the present and one for the infinitive/past. The present stem was used before endings beginning in a vowel, the infinitive/past stem before endings beginning in a consonant. In Old Church Slavonic grammars, verbs are traditionally divided into four (or five) conjugation classes, depending on the present stem, known as Leskien's verb classes. However, this division ignores the formation of the infinitive stem. The following table shows the main classes of verbs in Proto-Slavic, along with their traditional OCS conjugation classes. The "present" column shows the ending of the third person singular present.
|*nestì, *nesȅtь "carry"
*mę̀ti, *mьnetь "crumple"
*peťì, *pečetь "bake"
*žìti, *živetь "live"
*bьrati, *beretь "take"
*zъvati, *zovetь "call"
|PIE primary verbs, root ending in a consonant. Several irregular verbs, some showing ablaut. Not productive. Contains almost all infinitives in -Cti (e.g. *-sti or *-ťi), and a limited number of verbs in -ati. In verbs with an infinitive in -ti, various changes may occur with the last consonant.|
|(ę)-e-tь||-ti||*leťi, *lęžeti "lie down"
*stati, *stanetь "stand (up)"
|PIE nasal-infix presents. The infinitive stem may end in either a vowel or a consonant. Not productive, only a few examples exist.|
|2nd||-ne-tь||-nǫ-ti||*rìnǫti, *rìnetь "push, shove"||From various PIE n-suffix verbs, the nasal vowel was a Slavic innovation. Two subclasses existed: those with -nǫ- also in the aorist and participle, and those without.|
|*bìti, *bь̏jetь "beat"
*myti, *myjetь "wash"
*duti, *dujetь "blow"
*dajati, *dajetь "give"
|PIE primary verbs and presents in -ye-, root ending in a vowel. -j- is inserted into the hiatus between root and ending. Verbs with the plain -ti infinitive may have changes in the preceding vowel. Several irregular verbs, some showing ablaut. Not productive.|
|-je-tь||-a-ti||*sъlàti, *sъljȅtь "send"||PIE presents in -ye-, root ending in a consonant. The j caused iotation of the present stem.|
|-aje-tь||-a-ti||*dělati, *dělajetь "do"||PIE denominatives in -eh₂-ye-. Remained very productive in Slavic.|
|-ěje-tь||-ě-ti||*uměti, *umějetь "know, be able"||PIE stative verbs in -eh₁-ye-. Somewhat productive.|
|-uje-tь||-ova-ti||*cělovàti, *cělùjetь "kiss"||An innovated Slavic denominative type. Very productive and usually remains so in all Slavic languages.|
|-ęje-tь||-ę-ti||In some cases.|
|-C-tь||-C-ti||In some cases, where not class I.|
|4th||-i-tь||-i-ti||*prosìti, *prõsitь "ask, make a request"||PIE causative-iteratives in -éye-, denominatives in -eyé-. Remained very productive.|
|*mьněti, *mьnitь "think"
*slỳšati, *slỳšitь "hear"
|A relatively small class of stative verbs. The infinitive in -ati was a result of iotation, which triggered the change *jě > *ja. In the present tense, the first-person singular shows consonant alternation (caused by *j): *xoditi "to walk" : *xoďǫ, *letěti "to fly" : *leťǫ, *sъpati "to sleep" : *sъpľǫ (with epenthetic *l). The stem of the infinitives in *-ati (except for *sъpati) ends in *j or the so-called "hushing sound".|
|5th||-(s)-tь||-ti||*bỳti, *ȅstь "be"
*dàti, *dãstь "give"
*ě̀sti, *ě̃stь "eat"
*jьměti, *jьmatь "have"
*věděti, *věstь "know"
|PIE athematic verbs. Only five verbs, all irregular in one way or another, including their prefixed derivations.|
Originally in Balto-Slavic, there were only two accent classes, barytonic (with fixed stem accent) and mobile (with mobile accent), corresponding to Slavic classes A and C. There was no class with fixed accent on the ending. Both classes originally had both acute and circumflex stems in them. After the operation of Dybo's law, three basic accent classes emerged for nominals (nouns, adjectives, pronouns, participles):
For this purpose, the "stem" includes any morphological suffixes (e.g. a diminutive suffix), but not generally on the inflectional suffix that indicates the word class (e.g. the -ā- of feminine ā-stem nouns), which is considered part of the ending. Verbs also had three accent classes (A, B and C) with similar characteristics to the corresponding noun classes. However, the situation is somewhat more complicated due to the large number of verb stem classes and the numerous forms in verbal paradigms.
Due to the way in which the accent classes arose, there are certain restrictions:
Some nouns (especially jā-stem nouns) fit into the class A pattern but have neoacute accent on the stem, which can have either a short or a long syllable. A standard example is *võľa "will", with neoacute accent on a short syllable. These nouns earlier belonged to class B; as a result, grammars may treat them as belonging either to classes A or B.
During the Late Common Slavic period, the class B paradigm became mobile as a result of a complex series of changes that moved the accent leftward in certain circumstances, producing a neoacute accent on the newly stressed syllable. The paradigms below reflect these changes. All languages subsequently simplified the class B paradigms to varying degrees; the older situation can often only be seen in certain nouns in certain languages, or indirectly by way of features such as the Slovene neo-circumflex tone that carry echoes of the time when this tone developed. See History of Proto-Slavic#Accentual developments for more details.
The following tables are examples of Proto-Slavic noun-class paradigms, based on Verweij (1994). There were many changes in accentuation during the Common Slavic period, and there are significant differences in the views of different scholars on how these changes proceeded. As a result, these paradigms do not necessarily reflect a consensus. The view expressed below is that of the Leiden school, following Frederik Kortlandt, whose views are somewhat controversial and not accepted by all scholars.
Class A nouns
|Masc. long -o||Neut. long -o||Masc. long -jo||Fem. long -ā||Fem. long -jā||Masc. long -i||Fem. long -i||Masc. long -u||Fem. long -ū||Fem. long -r||Masc. long -n||Neut. long -n||Neut. long -s||Neut. long -nt|
Note that all class A stems are long. This is because all such stems had Balto-Slavic acute register in the root, which can only occur on long syllables. (Short syllables, and long syllables with Balto-Slavic circumflex register, became class B nouns in Common Slavic.)
The distribution of short and long vowels in the stems without /j/ reflects the original vowel lengths, prior to the operation of Van Wijk's law, Dybo's law and Stang's law, which led to class B nouns and the differing lengths in /j/ stems.
Class B nouns
|Masc. long -o||Neut. long -o||Masc. short -jo||Neut. short -jo||Fem. short -ā||Masc. long -i||Fem. short -i||Masc. short -u||Fem. short -ū||Masc. short -n||Neut. short -n||Neut. long -nt|
Class B jā stem nouns are not listed here. The combination of Van Wijk's law and Stang's law would have originally produced a complex mobile paradigm in these nouns, different from the mobile paradigm of ā-stem and other nouns, but this was apparently simplified in Common Slavic times with a consistent neoacute accent on the stem, as if they were class A nouns. The class B jo stem nouns were also simplified, but less dramatically, with consistent ending stress in the singular but consistent root stress in the plural, as shown. Class B s stem noun are not listed here, because there may not have been any.
Class C nouns
|Masc. short -o||Neut. long -o||Masc. long -jo||Neut. short -jo||Fem. short -ā||Fem. long -jā||Masc. long -i||Fem. short -i||Masc. long -u||Fem. nonsyllabic -ū||Fem. short -r||Masc. short -n||Neut. short -n||Neut. short -s||Neut. long -nt|
The accent pattern for the strong singular cases (nom., acc.) and all plural cases is straightforward:
For the weak singular cases, it can be observed:
Note also that the long-rising vs. short-rising accent on ending-accented forms with Middle Common Slavic long vowels reflects original cirumflex vs. acute register, respectively.
The same three classes occurred in verbs as well. Middle Common Slavic class B verbs in *-ī́tī had a neoacute retraction in Late Common Slavic in the present tense; that is, these verbs had original acute accent on the *-i- inflectional suffix in the infinitive, but neoacute accent on the stem in the present tense. This is due to the same process that caused neoacute retraction in class B jā-stem nouns (see above).
August Schleicher wrote a fable in the PIE language he had just reconstructed, which though it has been updated a few times by others still bears his name. Below is a rendering of this fable into Proto-Slavic.
Its immediate successors were Proto-East Slavic, Proto-South Slavic, and Proto-West Slavic. The Proto-Slavic era itself is often divided arbitrarily into three periods: (1) early Proto-Slavic, until about 1000 B.C.; (2) middle Proto-Slavic, during the next millenium; (3) late Proto-Slavic, from the 1st to the 6th century A.D., although it was not until the 12th century that Slavic linguistic unity actually ceased to function.