B A L T Ų I R S L A V Ų K A L B Ų R Y Š I A I L I E T U V O S T S R M O K S L Ų A K A D E M I J A

SIMAS KARALIŪNAS

SOME QUESTIONS CONCERNING THE EARLIEST RELATIONS BETWEEN THE BALTIC AND SLAVIC LANGUAGES

The problem of the earliest relations between the Baltic and Slavic languages arose in linguistics long ago. For many years, researchers studied and debated it, but reached no consensus; indeed, the solution to the problem became even more complicated. Discussions and disputes on this subject continue today with the same fervor as many years ago. The continuing relevance of the problem of the earliest relations between the Baltic and Slavic languages is shown not only by the fact that, as one of the most important problems in Slavic linguistics, it was widely discussed at the Fourth International Congress of Slavists in Moscow (four papers were delivered on the subject), but also by the fact that, by resolution of the International Committee of Slavists, an International Commission for the Study of Baltic-Slavic Relations was established in 1962, bringing together a dozen or so prominent specialists in Slavic and Baltic studies from eleven European countries¹. The problem of the earliest relations between the Baltic and Slavic languages is precisely one of the areas of work of that commission. The problem is also relevant because it forms part of the wider range of questions in Indo-European studies. The resolution of many questions in the history of the Baltic and Slavic languages depends, in part, on how the problem of their earliest relations is resolved. It is clear that the problem of the earliest relations between the Baltic and Slavic languages is also connected with particular questions in the history of Lithuanian.

The fact that “sharp discussions on the problem of the earliest relations between the Baltic and Slavic languages are taking place in scholarship today, in some cases even acquiring an urgent ideological and political significance”^² also places upon Lithuanian linguists the task of stating their views on this question without delay.

The Baltic and Slavic languages, as members of the Indo-European language family, descended from a linguistic grouping that existed several millennia ago, variously called the Indo-European proto-

¹ At the Fifth International Congress of Slavists in Sofia, the following composition of the commission was approved: K. Korsakas (chairman of the commission, USSR), J. Safarevičius (vice-chairman, Poland), S. Bernšteinas (USSR), I. Duridanovas (Bulgaria), V. Falkenhanas (GDR), K. O. Falkas (Sweden), V. Georgijevas (Bulgaria), V. Holtumas (England), Č. Kudz inovskis (Poland), V. Kiparskis (Finland), B. Larinas (USSR), Chr. Stangas (Norway), R. Šmitelinas (France), V. Toporovas (USSR), P. Trostas (Czechoslovakia), A. Vračius (Romania).

² K. Korsakas, Lyginamoji kalbotyra ir tarybinės lietuvių filologijos uždaviniai, LKK, 6, 1963, p. 13.

7

language, language, linguistic unity, linguistic community, or common epoch³. Because the Baltic and Slavic languages descended from an Indo-European linguistic community, they share many features inherited from ancient times, which are also found in other Indo-European languages. Alongside these, however, the Baltic and Slavic languages also share features that cannot be explained by their descent from an Indo-European linguistic community. The problem of the earliest relations between the Baltic and Slavic languages consists precisely in explaining how the innovations of the Baltic and Slavic languages arose—that is, phonetic, morphological, and lexical features shared by these languages that are not found beyond their linguistic area.

Several different theories have been developed in response to the questions of how the innovations of the Baltic and Slavic languages arose and what kind of relations existed between them in the first centuries after they separated from the Indo-European linguistic community. The first chapter of this work is devoted to the history of research on the problem of the earliest Baltic-Slavic relations and to an account of the theories proposed.

The second chapter, owing to limitations of space, is confined to an analysis of phonetic phenomena; it gives an account of the history of their study and sets out conceptions of their origin and development.

The third chapter attempts to address certain questions concerning the chronology of the earliest relations between the Baltic and Slavic (as well as other Indo-European) languages.

In order to clarify as far as possible the origins of the features shared by the Baltic and Slavic languages and the time at which they arose, the author draws on material from Baltic, Slavic, and other Indo-European languages. Among the Baltic languages, the discussion relies chiefly on Lithuanian data, which are often archaic and have not always been given sufficient consideration in addressing this problem.

In presenting the material, especially in the first part of this work, a chronological principle has been followed: researchers’ works are discussed one after another in the order in which they appeared. A researcher’s work is assigned to a single period—usually the period in which that researcher’s most important work on the problem of the earliest relations between the Baltic and Slavic languages appeared—even though some of their articles on the subject may have been published in other periods. For example, Chr. Stang’s views are reviewed in the chapter on works of the 1920s and 1930s, although one of his articles appeared in 1963; J. Kurilowicz’s views, meanwhile, are discussed in the chapter on postwar works, despite the fact that his first article on the subject appeared as early as 1934.

The author of this work does not in any way seek to resolve the problem of the earliest relations between the Baltic and Slavic languages as a whole—it remains far too complex for now. The principal aim is simply to shed light on the current state of research on the problem, identify the strengths and weaknesses of the theories proposed, and present arguments for the theory that the author considers most reliable.

³ The terms “protolanguage” (праязык, Ursprache), “linguistic unity” (языковое единство, Spracheinheit), and “linguistic community” (языковая общность, Sprachgemeinschaft) are used interchangeably in this work, but each is assigned a different meaning. The term “protolanguage” is appropriate for a linguistic unit with little dialectal differentiation that existed in prehistoric times. The term “linguistic community” is used for a prehistoric linguistic unit with greater dialectal differentiation. When discussing the conception of researchers who recognize a Baltic-Slavic language that existed after the dissolution of the Indo-European linguistic community, their own term “Baltic-Slavic linguistic unity” is used.

8

HISTORY OF THE STUDY OF THE OLDEST RELATIONS BETWEEN THE BALTIC AND SLAVIC LANGUAGES

The history of the study of the oldest relations between the Baltic and Slavic languages has already been covered quite extensively. A broad article by N. Bogoliubova and T. Jakubaitė¹ is devoted to the development of general theoretical propositions concerning the oldest relations between these languages. V. Toporov describes the history of research into these relations in detail in three substantial articles². Here, attention is focused chiefly on the more important aspects of this research, especially its theoretical moments. Nevertheless, repetition has proved unavoidable in this and the work mentioned above.

The Question of the Relations between the Baltic and Slavic Languages in the Works of the Founders of the Comparative-Historical Method

The scholarly investigation of the kinship relations between individual languages began in the early nineteenth century, when the works of F. Bopp, R. Rask, and J. Grimm¹ laid the foundations of Indo-European comparative-historical linguistics.

The first comparativists took a new approach to the kinship of languages, basing their views on comparative studies of languages.

Rasmus Rask (1787—1832) devoted considerable attention to the analysis of the Baltic and Slavic languages in his works. He was the first to highlight the importance of the Baltic languages for scholarship, and from that time onward these languages have occupied an important place in comparative linguistics. R. Rask pointed out that the Baltic languages constitute an independent branch of the Indo-European languages and are related to the Slavic and Germanic languages.

Franz Bopp (1791—1867) held a similar view. In his opinion², Prussian, Lithuanian, and Latvian formed a group of closely related languages allied with the Slavic languages. F. Bopp noted that the Baltic languages should not be considered an integral part of the Slavic languages. He attempted to establish a relative chronology of the origin of languages, based on the principle that closer or more distant kinship between languages depends on whether they separated later or earlier. On the grounds that the Baltic languages are more closely related to Slavic than to other languages, F. Bopp believed that these languages had apparently diverged considerably later. The Balto-Slavic language separated from its Asiatic sister, which could supposedly be called Sanskrit or left unnamed, later than the classical, Germanic, and

¹ Н. Д. Боголюбова, Т. А. Якубайтис, История разработки вопроса о балто-славянских языковых отношениях, „Rakstu krājums (Veltījums akadēmikim profesoram Dr. Jānim Endzelīnam viņa 85 dzīves un 65 darba gadu atcerei)“, Rīga, 1959, p. 331—375.

² V. N. Toporov, Recent Works on the Study of Balto-Slavic Linguistic Relations, ВСЯ, 3, 1958, p. 134—161; An Outline of the History of the Study of the Oldest Balto-Slavic Linguistic Relations, УЗИСл, 17, 1959, p. 248—274; From the History of the Study of the Oldest Balto-Slavic Linguistic Relations, УЗИСл, 23, 1962, p. 3—43; see also O. Szemerényi, The Problem of Balto-Slav Unity—A Critical Survey, „Kratylos“, 2, 1957, p. 97—123; A. Vraciu, Slava si baltica. Problema vechilor raporturi dintre ele în lumina noilor cercetari, „Romanoslavica“, 12, 1965, p. 283—297, which surveys the most recent research in this field.

¹ F. Bopp, On the Conjugation System of the Sanskrit Language in Comparison with That of the Greek, Latin, Persian and German Languages, Frankfurt am Main, 1816; R. Rask, Investigation of the Origin of the Old Norse or Icelandic Language, Copenhagen, 1818; J. Grimm, German Grammar, Göttingen, 1819 (2nd ed. 1822).

² F. Bopp, Über die Sprache der alten Preussen, Berlin, 1853, p. 80.

9

the Celtic languages, but before the division of the Asiatic branch into the Iranian and Indic language branches. Thus, F. Bopp was the first to put forward the idea of a Balto-Slavic proto-language and of close relations between the Baltic and Slavic languages and the languages of the Indo-Iranian group.

Jacob Grimm (1785–1863) also recognized the Baltic languages as having equal standing among the other Indo-European languages. In his view, frequent lexical correspondences indicate an ancient relationship between the Slavic and Germanic languages. The Baltic and Slavic languages are much closer to one another than either is individually to the Germanic languages. To some extent, the Baltic languages serve as intermediaries between the Slavic and Germanic languages. However, this does not mean that the Baltic languages are a mixture of the latter languages. Likewise, in J. Grimm’s view, one cannot suppose that the Slavic and Germanic languages borrowed the oldest elements of their systems from the Baltic languages. According to J. Grimm, the commonalities and similarities among the Germanic, Slavic, and Baltic languages arose because a community (Gemeinschaft)⁴ had long existed among them.

August Schleicher (1821–1868) summarized and further developed the views of the early comparativists, devoting much attention in his research to questions concerning the kinship of the Baltic languages. A. Schleicher’s understanding of linguistic relationships is connected with his general conception of the origins of the Indo-European languages.

A. Schleicher formulated a coherent hypothesis of genealogical language classification, called the “family-tree theory” of languages (Stammbaumtheorie). According to this theory, the unified Indo-European proto-language (indogermanische Ursprache) first split into two branches: one comprising the Baltic, Slavic, and Germanic languages (Slawodeutsch), and the other comprising the Indo-Iranian, Greek, Italic, and Celtic languages. The large Baltic-Slavic-Germanic branch later split into two smaller ones—the Germanic (Deutsch) and Balto-Slavic (Slawolitauisch) branches. At the same time, the second branch of the Indo-European proto-language also split, first into Indo-Iranian and Greco-Italo-Celtic. Later, after the Greeks separated, the Italo-Celtic proto-language remained. The branch of the Baltic and Slavic languages, which A. Schleicher called the Balto-Slavic proto-language (Slawolitauische Ursprache), later split into the Slavic and Baltic parent languages (Grundsprache), which in turn would split again into individual languages⁵.

Thus, A. Schleicher developed F. Bopp’s idea of a Balto-Slavic proto-language and incorporated it into a coherent theory of the division of the Indo-European proto-language. He clearly identified the place of the Balto-Slavic proto-language among the Indo-European languages and recognized its genetic relationship with the Germanic proto-language. A. Schleicher’s authority was so great that most linguists adhered to the theory of a Balto-Slavic proto-language in the second half of the 19th century and even at the beginning of the 20th.

During the formative period of comparative-historical Indo-European studies, the debate was not about whether a Balto-Slavic proto-language existed—all without exception recognized it—but about its place among the other Indo-European languages. R. Rask considered the Baltic and Slavic languages more closely related to those of the Indo-Iranian (Aryan) group, whereas F. Bopp, J. Grimm, and A. Schleicher argued that the Baltic and Slavic languages were genetically related to the Germanic languages. The question of the place of the Baltic and Slavic languages remained a subject of discussion in the second half of the 20th century as well.

³ J. Grimm, Geschichte der deutschen Sprache, Aufl. III, Leipzig, 1868, I, p. 9–10, 119; II, p. 711–712, 715.

⁴ J. Grimm, op. cit., I, p. 119.

⁵ A. Schleicher, Compendium der vergleichenden Grammatik der indogermanischen Sprachen, Aufl. III, Weimar, 1871, p. 6–7.

10

Baltų ir slavų kalbų seniausiųjų santykių tyrinėjimas XIX a. antrojoje pusėje

A. Schleicher died convinced that he had succeeded in constructing a perfect and complete edifice of Indo-European linguistics. However, in the second half of the 19th century, that edifice began to be demolished. And it was A. Schleicher’s own pupils who set about demolishing it. For instance, Johan Schmidt (1843–1901) critically assessed the theory of the “genealogical tree of languages.” In his view, the facts show that the Baltic-Slavic (Lituslawische) languages are closely related both to the Germanic languages and to the Indo-Iranian languages. Therefore, J. Schmidt points out, the proto-language of the Baltic-Slavic-Germanic group may be recognized on an equal footing with the proto-language of the Baltic-Slavic-Aryan group. Thus, the position of the Baltic and Slavic languages remains unclear. It will become clear, in J. Schmidt’s view, when it is recognized not only that the Baltic and Slavic languages separated from neither the Germanic nor the Indo-Iranian languages, but also that they organically unite these latter two language groups. Just as there are no geographical borders between Europe and Asia, so strictly drawn lines of demarcation should not exist between the languages of Europe and Asia either. J. Schmidt adds that everywhere one can observe an uninterrupted transition (continuierlicher Übergang) from one language to another, and that there had never been any strict boundaries in the Indo-European linguistic area: two dialects distant from one another, for example A and X, were continuously linked by the dialects B, C, D, etc. lying between them.

With all this in mind, J. Schmidt proposes the “wave” theory (Wellentheorie) in place of the “genealogical tree of languages” theory: “In its place I would like to present the image of a wave which, moving away from its center, spreads in concentric, continually weakening circles”². Thus, according to him, linguistic changes spread from one language to another like waves.

J. Schmidt’s new ideas represented a considerable step forward. Some of them (Indo-European dialects, the idea of geographical continuity, etc.) were developed further in the 20th century.

It is important to note that J. Schmidt had no doubts about the theory of the Baltic-Slavic proto-language. His research concerned the position of the Baltic and Slavic languages within the Indo-European language family.

Kitas A. Šleicherio mokinys Augustas Leskynas (1840–1916) bandė suderinti „genealoginio kalbų medžio“ ir „bangų“ teorijas. Jis manė³, kad baltų, slavų ir germanų kalbos galėjusios pergyventi bendrą istorinį etapą nepriklausomai nuo kitų ide. kalbų, taigi ir skilimą, kaip jis suprantamas A. Šleicherio teorijoje. Be to, visai tikėtina, kad baltų ir slavų kalbos buvusios ta grandimi, kuri jungusi germanų ir indų-iranėnų kalbas. A. Leskyno nuomone, nesą būtino reikalo pripažinti baltų-slavų-germanų prokalbę, nes tų kalbų linksniavimo sistemos, išskyrus kai kuriuos linksnius su bendru formantu -m-, nesančios jau tokios panašios⁴. Taigi „Vadinamoji „genealoginio kalbų medžio“ teorija visai neprieštarauja „perėjimo“ teorijai, jei taip trumpai pavadintume J. Šmito hipotezę“⁵.

¹ L Schmidt, Die Verwandtschaftsverhältnisse der indogermanischen Sprachen, Weimar, 1872, p. 16–27.

² J. Schmidt, op. cit., p. 27.

³ A. Leskien, Die Declination im Slawisch-litauischen und Germanischen, Leipzig, 1876, p. XII.

⁴ A. Leskien, op. cit., p. 157–158.

⁵ A. Leskien, op. cit., p. XII.

11

Neither J. Schmidt nor A. Leskien was interested in the mutual relations between the Baltic and Slavic languages. J. Hanuš drew attention to them. Jan Hanuš’s (1858—1887) article⁶, published in 1886, was the first to be devoted specifically to the problem of the earliest relations between the Baltic and Slavic languages.

J. Hanuš raised two questions: 1) What allows us to recognize a common Balto-Slavic linguistic period? 2) How can we determine more precisely when the Balts and Slavs spoke one language? In answering the first question, J. Hanuš presented a dozen or so common features of Baltic and Slavic phonetics and morphology: the fate of IE sonants, the declension of pronominal adjectives, and so on. His second argument for a common Balto-Slavic linguistic period was the loans of Baltic origin in the Finnic languages of the Baltic region, which he considered Balto-Slavic (at the time, scholars had not yet established their exclusively Baltic origin). In answering the second question, J. Hanuš conjectured that the Balto-Slavic tribe must have split into Balts and Slavs approximately between the 5th century BCE and the 2nd century CE, since, in his view, Herodotus’s Neuri were Balto-Slavic. Later researchers did not confirm this dating.

J. Hanuš’s article was significant because it posed explicitly the question of the mutual relations between the Baltic and Slavic languages, and because, for the first time, all the common features of Baltic and Slavic languages known at the time were presented in one place.

Soon, questions concerning not only the Baltic and Slavic languages but also the relationship among all IE languages receded into the background. Entirely new questions arose (sound laws, analogy, loans), which for many years were at the center of researchers’ attention⁷. An article⁸ by Karl Brugmann (1849—1919), one of the most prominent representatives of the Neogrammarian (Junggrammatiker) school, appeared; it took a skeptical view of the theory of proto-languages: shared innovations in individual languages cannot serve as evidence of close linguistic relationship, since such innovations could have arisen independently in each language. K. Brugmann considered the totality of innovations in phonetics, inflection, syntax, and lexicon⁹ to be the only argument for linguistic relationship with probative force.

Although K. Brugmann viewed theories of linguistic relationship skeptically, he nevertheless believed that the Baltic and Slavic languages had earlier experienced an era of community (Zeit der baltisch-slawischen Urgemeinschaft)¹⁰. In his “Outline of the Comparative Grammar of the Indo-European Languages,” K. Brugmann presented seven common features of the Baltic and Slavic languages which, in his view, indicated close relations between them¹¹:

1) the development of IE sonants n, r̥ into in, ir (cf. Lithuanian at-mintis and Old Church Slavonic mьnǫ, pa-mętь);

2) simplification of intervocalic geminates;

3) the formation of pronominal adjective forms with the pronoun *jo- (cf. Lithuanian gerasis and Old Church Slavonic dobrъjь);

⁶ J. Hanusz, O dobie litewsko-słowiańskiej w stosunku do pra języka indoeuropejskiego, „Rozprawy i sprawozdania z posiedzeń wydziału filologicznego Akademii Umiejętności“, 11, Kraków, 1886, p. 249—272.

⁷ Plačiau žr. F. Specht, Die „indogermanische“ Sprachwissenschaft von den Junggrammatikern bis zum ersten Weltkriege, „Lexis“, 1, 1948, p. 229—263.

⁸ K. Brugmann, Zur Frage nach den Verwandtschaftsverhältnissen der indogermanischen Sprachen, „Internationale Zeitschrift für allgemeine Sprachwissenschaft“, 1, 1884, p. 226—256.

⁹ K. Brugmann, op. cit., p. 253.

¹⁰ K. Brugmann Grundriss der vergleichenden Grammatik der indogermanischen Sprachen, I, Strassburg, 1897, p. 146.

¹¹ K. Brugmann, op. cit., p. 20—21.

12

4) the transition of nt-stem participles and consonant-stem nouns to the i-stem in certain cases (cf. Lithuanian vėžančio, Old Church Slavonic vezǫšta from *vezontio, and Lithuanian vežàs, Old Church Slavonic vezi from *vezonts; Lithuanian akmenìmis, Old Church Slavonic kamenьmi, and Lithuanian akmuõ, Old Church Slavonic kamy from *akmōn¹²;

5) the replacement of the Indo-European pronouns *so, *sā by Lithuanian tas, tā, Old Church Slavonic to, ta;

6) the formation of the singular dative forms of the pronouns Lithuanian mānie, Old Prussian mennei, Old Church Slavonic mьnē on the model of the singular genitive;

7) the replacement of the singular genitive of o-stem nouns by the ablative (cf. Lithuanian vilko, Old Church Slavonic vlьka, and Sanskrit vṛkāt).

These common features of the Baltic and Slavic languages presented by K. Brugmann sufficed to convince most linguists of the late nineteenth and early twentieth centuries that a Balto-Slavic proto-language had existed.

In the second half of the nineteenth century, as earlier, one of the most important questions concerning the Baltic and Slavic languages was their place in the Indo-European language family. It was, however, precisely at this time that attention turned to the relations between the Baltic and Slavic languages: the first special article devoted to this problem, by J. Hanuš, appeared, as we have seen, and K. Brugmann presented common features characterizing the relations between the Baltic and Slavic languages.

Although most scholars of this period did not doubt the existence of a Balto-Slavic proto-language, it was precisely at this time, at the end of the nineteenth century, that skeptical voices began to be heard. These voices grew especially strong in the early twentieth century.

The 1908–1912 debate on the earliest relations between the Baltic and Slavic languages

At the beginning of the twentieth century, the theory of a Balto-Slavic proto-language began to be seriously questioned. All the common features of the Baltic and Slavic languages, which had generally been regarded as arguments for the theory of a Balto-Slavic proto-language, were critically examined.

From the Neogrammarian period onward, scholars began to view all proto-languages skeptically. B. Delbrück considered all groupings of Indo-European languages into dialects doubtful, with the exception of the Indo-Iranian proto-language. P. von Bradke also spoke skeptically about a Balto-Slavic proto-language². In 1903 Jan Baudouin de Courtenay (1845–1929) wrote that there were no more, and perhaps even fewer, arguments in favor of the “Lithuanian-Slavic epoch” than against it, and that we could therefore regard it as a scientific fiction³. Kazimieras Jaunius (1848–1908) maintained that “it is not true that Slavic and Aestic are twin sisters: the Aestic language is the sister of Slavic, just as it is the sister of the other Indo-European languages”⁴.

The scholars mentioned above, without presenting arguments, confined themselves to general observations directed against the theory of a Balto-Slavic proto-language.

¹²In K. Brugmann’s book „Kurze vergleichende Grammatik der indogermanischen Sprachen“ (Berlin—Leipzig, 1922, p. 18), participles and consonant stems are treated separately, and therefore eight features are listed there.

¹ B. Delbrück, Einleitung in das Sprachstudium, 3. Aufl., Leipzig, 1893 (1st edition 1880).

² P. von Bradke, Contributions to the Knowledge of the Prehistoric Development of Our Language Family, Giessen, 1880.

³ I. Baudouin de Courtenay, Linguistic Notes and Aphorisms, ZhMNP, 346, 1903, p. 330.

⁴ K. Jaunius, Grammar of the Lithuanian Language, Petrograd, 1908—1916, p. 27.

13

The first to argue against the theory of a Balto-Slavic proto-language was Antoine Meillet (1866—1936).

As early as 1905, A. Meillet wrote that the Baltic and Slavic languages did not have as many innovations as were attributed to them, and that there was no basis for speaking of an epoch of Balto-Slavic unity (d’une époque d’unité balto-slave)⁵. In the book “Indo-European Dialects”⁶, published in 1908, A. Meillet devoted a large chapter to characterizing the oldest relationships between the Baltic and Slavic languages.

Before beginning to examine the common features of the Baltic and Slavic languages identified by K. Brugmann, A. Meillet noted that the similarity between the Baltic and Slavic languages is evident. In his view, the similar character of the Baltic and Slavic language systems was determined by the following factors: 1) the Baltic and Slavic languages descended from identical dialects of the Indo-European linguistic community (l’unité indo-européenne) (de parlers indo-européens sensiblement identiques), which were not separated by any major isogloss; 2) in their development, the Baltic and Slavic languages did not undergo radical breaks in their systems (aucune fracture brusque du système), and are therefore conservative and archaic, thus preserving many ancient features; 3) the Balts and Slavs lived at all times in neighboring geographic regions under similar conditions of civilization. In addition, A. Meillet pointed out that borrowings strengthened the similarity between the Baltic and Slavic languages—especially borrowings into the Baltic languages from Slavic⁷.

A. Meillet then examined in detail the common features of the Baltic and Slavic languages listed by K. Brugmann.

1. Apart from the Baltic and Slavic languages, the similar development of the Indo-European sonants l, m, n, r also occurred in Germanic and Celtic. In certain positions, u arose next to a sonant in Greek, Latin, and Armenian. Therefore, the appearance of anaptyctic vowels i, u next to sonants is a dialectal phenomenon of the Indo-European linguistic community, not an innovation of the Baltic and Slavic languages.

2. The simplification of geminates is the result of the general tendency toward simplification in the Indo-European languages, and it began as early as the Indo-European linguistic community. In the Baltic and Slavic languages, as in Armenian, geminates were simplified, whereas this did not occur in the other languages. The presence of geminates in the Indo-European linguistic community itself is doubtful, since they occurred there mostly in hypocoristic and child-language vocabulary.

3. The types of pronominal adjectives in the two language groups differ in detail, and their meanings do not coincide. The pronoun *jo- is used very similarly alongside nouns and adjectives in Avestan (A. Meillet particularly emphasizes that in both Avestan and the Baltic and Slavic languages, the pronoun *jo- agrees with the words alongside which it is used).

4. The masculine present active participles shifted to the jo stem on the model of feminine participles (the jā stem), which existed already in the Indo-European linguistic community. A. Meillet pointed out that adjectives tend to shift to a vowel stem much faster than nouns (Lithuanian has only adjectives with o and u stems; the Slavic languages have only those with the o stem). This shift is an expression of a general innovation in the Indo-European languages, since an exact equivalent to the Baltic and Slavic pattern is found in the West Germanic languages (Old English, Old Saxon, and Old High German).

5. The insertion of -i- in forms such as Lithuanian akmen-i-mis, Old Church Slavonic kamen-b-mi, is characteristic not only of the Baltic and Slavic languages, since Latin exhibits exactly the same phenomenon (cf. ped-i-

⁵ A. Meillet, Etudes sur l’étymologie et le vocabulaire du vieux slave, II, Paris, 1905, p. 201.

⁶ A. Meillet, Les dialectes indo-européens, Paris, 1908.

⁷ A. Meillet, Les dialectes indo-européens, Nouveau tirage, Paris, 1922, p. 40—41.

14

bus), Armenian (cf. ot-i-wkh), and other languages. Moreover, the convergence of these two types was facilitated by the phonetic identity of both the singular accusatives of i-stems and consonant stems and their plural accusatives.

6. The old nominative of the demonstrative pronoun *so-, *sā- (cf. Sanskrit sá, sā, Greek ὁ, ἡ, Gothic sa, so) was replaced by the stem *to-, *tā- not only in Baltic and Slavic, but also in the West Germanic languages (cf. Old Saxon thē, thia, Old High German dēr, diu). A. Meillet emphasized that this is a very simple innovation, since all forms of the paradigm, except the singular nominative, began with the consonant t-, which could very easily have been transferred by analogy to the sole case form beginning with the consonant s-.

7. The dative of the pronoun Lithuanian manie, man, Prussian mennei differs from Old Church Slavonic mъně in the vowels of both its stem and its ending. A. Meillet added that these Baltic and Slavic forms are an excellent example of parallel but independent innovations characterizing both language groups.

8. The singular genitive-ablative Lithuanian vilko, Old Church Slavonic vlъka corresponds to the Sanskrit ablative vṛkāt. The genitive and ablative merged because, in all stems except the o-type, the singular genitive and ablative had one and the same form. Greek used the genitive form to express the ablative, whereas in Baltic and Slavic, conversely, the ablative acquired the function of the genitive. A. Meillet stresses that this phenomenon does not appear to have arisen during the period of Balto-Slavic unity (à une période de communauté), since Prussian has the genitive deiw as ‘of the god’. Thus, he added, the generalization of the ablative is readily explained by independent linguistic development.

Having examined these common features of Baltic and Slavic, A. Meillet concluded that “the development of the Baltic and Slavic languages proceeded in parallel; this parallelism naturally gave rise to some identical forms, but these innovations do not warrant recognizing a common period in the development of these languages”⁸. As a striking example of parallel innovation, A. Meillet cited de Saussure–Fortunatov’s law, correctly asserting that this law operated independently in Lithuanian and Slavic.

By presenting analogous evidence from other Indo-European languages and adducing further similar phenomena, A. Meillet showed that the aforementioned common features of Baltic and Slavic are either archaisms inherited from Proto-Indo-European times or innovations. If a phenomenon or fact is archaic, then Baltic and Slavic, like the other Indo-European languages, have retained it. Nor can innovations constitute conclusive evidence for a Balto-Slavic proto-language or common period, since they could have arisen either independently in Baltic, Slavic, and other Indo-European languages, or during a period of contact among these languages. Moreover, in A. Meillet’s view, the Balto-Slavic innovations are similar but not identical (cf., for example, the dative of the personal pronoun).

A. Meillet’s bold new ideas shook the foundations of the theory of a Balto-Slavic proto-language. His critique showed that the existing arguments were insufficient to prove the existence of a Balto-Slavic proto-language. The proponents of the proto-language theory were therefore compelled to seek new evidence in support of their claims.

The first to respond was Viktor Porzeziński (1870—1929). The conclusion of his work devoted to the question of Baltic and Slavic was: “the Baltic and Slavic languages underwent a common period, during which a considerable number of innovations arose”⁹.

⁸ A. Meillet, op. cit., p. 45.

⁹ W. Porzeziński, Die baltisch-slavische Sprachgemeinschaft, RSI, 4, 1911, p. 2—3.

15

Further in his work, V. Porzeziński strives to identify as many innovations as possible that, in his view, arose in the Proto-Balto-Slavic language, which he calls the “Balto-Slavic period.” In order to defend the common features of the Baltic and Slavic languages against A. Meillet’s criticism, V. Porzeziński analyzed in detail the first and last of the common features listed by K. Brugmann. He considers these common features very important for resolving the question of the earliest relations between the Baltic and Slavic languages. V. Porzeziński also presented new facts, treating them as Balto-Slavic innovations. He considered the past tense (preterite) of verbs to be such an innovation, cf. Lithuanian nešė, vedė and Old Church Slavonic nešė-acho, vedė-acho. In his view, the difference between the Baltic preterite and the Slavic imperfect was that the preterite in the Baltic languages arose on the basis of the asigmatic aorist¹⁰.

V. Porzeziński defended a common period of the Baltic and Slavic languages (eine baltisch-slavische Periode), while Jan Rozwadowski (1867—1935) devised a fairly complex scheme to elucidate the earliest relations between the Baltic and Slavic languages¹¹. He distinguished three periods in the development of relations between the Baltic and Slavic languages:

1) a period of commonality (doba prawspólności), probably lasting through the third millennium BCE;

2) a period when relations between the Baltic and Slavic languages had ceased and their development proceeded independently (the second and first millennia BCE);

3) a period of renewed rapprochement and relations as neighbors, which began around the start of the Common Era and continues to the present day¹².

Most scholars regarded J. Rozwadowski’s theory with indifference. Only the Czech linguist O. Hujer considered it “entirely reliable”¹³. Moreover, B. Gornung also holds the view that contacts between the Baltic and Slavic languages had been interrupted for a long time and resumed anew around the start of the Common Era¹⁴.

J. Rozwadowski’s theory raises serious doubts. It is doubtful that for two or two and a half thousand years the Baltic and Slavic languages maintained no contacts at all, only for these to resume around the start of the Common Era. Toponymy, material culture, and fragments of historical sources alike suggest that the Balts and Slavs were probably neighbors from time immemorial and therefore, naturally, could have maintained certain relations that were reflected in their languages as well.

Moreover, it is unlikely that relations between the Balts and Slavs were very close in the third millennium BCE, since there is reason to think that at that time Baltic dialects may have been closer to Germanic dialects.

J. Rozwadowski did in fact defend the theory of a Proto-Balto-Slavic language, but was forced to modify it and place it in the very distant past, because the old, unmodified version of this theory had already become difficult to defend. J. Rozwadowski’s example shows how influential A. Meillet’s ideas were at the time.

Following the articles by V. Porzeziński and J. Rozwadowski, resolving the question of the earliest relations between the Baltic and Slavic languages began to become more complicated.

¹⁰ W. Porzeziński, op. cit., p. 24–25.

¹¹ J. Rozwadowski, O pierwotnym stosunku wzajemnym języków bałtyckich i słowiańskich, RSI, 5, 1912, p. 1–36.

¹² J. Rozwadowski, op. cit., p. 24.

¹³ O. Гуйер, Введение в историю чешского языка, Москва, 1953, р. 42.

¹⁴ Б. В. Горнунг, Из предыстории образования общеславянского языкового единства, Москва, 1963, р. 95.

16

Janis Endzelīns (1873—1961) undertook a comprehensive study of the oldest relations between the Baltic and Slavic languages, devoting an entire monograph to this question¹⁵.

First, J. Endzelīns raised a fundamental question: what linguistic facts and phenomena demonstrate the close kinship between the Baltic and Slavic languages? It is understood that such linguistic facts and phenomena must be innovations in those languages. However, for innovations to have probative force, in J. Endzelīns’s view, they must be numerous¹⁶.

Guided by these principles, J. Endzelīns examined the common features of Baltic and Slavic phonetics, grammar, syntax, and lexicon, as well as their differences. He analyzed phonetic phenomena in especially great detail. J. Endzelīns clearly showed that such shared phonetic phenomena in the Baltic and Slavic languages as the simplification of geminates, the merger of voiced consonants with voiced aspirated consonants, the change of final -m to -n, and the loss of consonants -d, -t at the end of a word are either dialectal phenomena of the Indo-European linguistic community or arose independently in the separate Indo-European languages, and therefore have no probative force.

J. Endzelīns identified many new facts common to the Baltic and Slavic languages, as well as many long-standing differences between them, in morphology and syntax. He also presented a substantial list of common Baltic and Slavic vocabulary (about 200 words).

Having examined the commonalities and differences between the Baltic and Slavic languages with extraordinary care, J. Endzelīns described their oldest relations as follows: “Even in the common Indo-European proto-language, the ‘Slavic’ dialect already differed in some respects from the ‘Baltic’ dialect, occupying approximately the middle position between the ‘Aryan’ and ‘Baltic’ dialects. When the Indo-Iranians became territorially separated from the Slavs as the Indo-European proto-language split up, the latter, being the closest neighbors of the Baltic tribes, experienced a common period with them, during which their languages were still so close that the Slavs understood the Baltic language, and the Balts the Slavic language. At that time their languages acquired many new words in common, and some (not many!) common phonetic, morphological, and syntactic changes arose in them. Thus, one may speak of a ‘Slavic-Baltic period’¹⁶; it is better not to use the term ‘Slavic-Baltic proto-language,’ since it readily suggests complete linguistic unity. The fact that Slavic differed from Baltic already in the Indo-European proto-language is also indicated by the circumstance that the degree of closeness of the individual Slavic languages to the Baltic languages appears to be uniform (although some Baltic dialects disappeared without leaving any written records). It is hardly possible to doubt that all Slavic languages are equally distant from the Baltic languages. Prussian differs in some respects from Lithuanian and Latvian and is closer to the Slavic languages. Apart from Prussian twais, swais (Old Church Slavonic tvoi, svoi), where the correspondence is probably accidental, in particular cases of similarity in the grammatical sphere we encounter not shared innovations but the preservation of archaisms (compare, for example, Prussian noūson, tebbei, wans: Old Church Slavonic nasъ, tebě, vy, and the neuter gender), and therefore these correspondences do not indicate a closer kinship of Prussian with the Slavic languages (compared with Lithuanian and Latvian)”¹⁷.

On the basis of what he considered very ancient differences between the languages (for example, the change of Indo-European s to š after r, k in Baltic and ch after k, r, i, u in Slavic; the change of Indo-European kh to k in Baltic and ch in Slavic, etc.), J. Endzelīns, unlike proponents of the theory of Baltic-Slavic linguistic unity, held that the Baltic and Slavic languages descended from different dialects of the Indo-European linguistic community. In contrast to A. Meillet’s theory of parallel development, J. Endzelīns argued,

¹⁵ И. Эндзелин, Славяно-балтийские этюды, Харьков, 1911.

¹⁶ И. Эндзелин, op. cit., p. 3.

¹⁷ И. Эндзелин, op. cit., p. 200—201.

2. Baltų ir slavų kalbų ryšiai

17

that, after the disintegration of the Indo-European linguistic community, the Baltic and Slavic languages drew closer and passed through a common period. J. Endzelin was led to this latter conclusion by the following innovations in the Baltic and Slavic languages: the Indo-European r̥, l̥, m̥, n̥ becoming ir, il, im, in and ur, ul, um, un; the dative plural ending *-mus; the development of Indo-European eu and ēu into jau in the Baltic and Slavic languages, from which Slavic ju later arose; the formation of pronominal adjectives; the replacement of the old genitive singular of o-stem nouns by the ablative; the emergence of the pronoun *mun-; the predicative use of the instrumental; the genitive of negation; and, finally, the similarity of a large part of the vocabulary¹⁸.

Thus, to explain the common innovations of the Baltic and Slavic languages, J. Endzelin developed a theory of convergence. J. Endzelin held that linguistic similarities do not necessarily imply a common proto-language that existed after the Indo-European linguistic community had broken up: some of these similarities may be due to a common origin in the Indo-European linguistic community; others may arise when languages draw geographically closer and establish contacts.

J. Endzelin’s idea that geographical proximity, neighborly relations, and contacts of various kinds may give rise to common new phenomena in languages, or create conditions for the transfer of a neologism from one language into another, was relatively new in linguistics. In fact, the geographical aspect of J. Endzelin’s conception recalled J. Schmidt’s idea of geographical continuity. This idea of J. Endzelin’s was consistent with the facts of the natural life of languages: geographical proximity leads to shared changes in different languages and brings them closer together (cf. the languages of the Balkan Peninsula).

J. Endzelin’s new theory was assessed controversially. Proponents of the theory of Baltic-Slavic linguistic unity (a proto-language) considered J. Endzelin to be on their side. J. Rozwadowski wrote that, in his view, the Baltic-Slavic proto-language theory, hitherto accepted by everyone, had been proved, and that J. Endzelin had strengthened it with new arguments¹⁹. Reviewing J. Endzelin’s book, A. Meillet stated that J. Endzelin’s conclusions, apart from subtle nuances, did not differ from his own²⁰.

This conflicting assessment was prompted by the reviewers’ insufficient engagement with the new theory and, moreover, by a certain lack of precision in J. Endzelin’s statements. In drawing general conclusions, J. Endzelin was very cautious. As if wishing to clarify and explain his position, which others had not quite understood, J. Endzelin returned to the question of the earliest relations between the Baltic and Slavic languages in 1931.

In an article devoted to the verb in the Baltic and Slavic languages²¹, J. Endzelin particularly emphasized the great similarity of the Indo-European dialects from which the Baltic and Slavic languages developed. Even after the Indo-European proto-language had split up, the ancestors of the Baltic and Slavic peoples probably remained neighbors for some time and maintained linguistic contacts. Some linguistic innovations arose during this period of linguistic community (in dieser Periode der Sprachgemeinschaft). However, they were very few, since at that time the Baltic and Slavic languages were very

¹⁸ И. Эндзелин, op. cit., p. 200. Besides these common features of the Baltic and Slavic languages, in 1938 J. Endzelīns noted three more: frequent cases of stress shift; the genitive of the 1st person plural pronoun *nōs(s)ēm; the infinitive with *-et(e)i alongside present-tense forms with the 1st stem, see J. Endzelīns, Latviešu valodas skaņas un formas, Rīgā, 1938, p. 9–10; cf. J. Endzelynas, Baltų kalbų garsai ir formos, Vilnius, 1957, p. 7.

¹⁹ J. Rozwadowski, op. cit., p. 6–7. Plg. dar A. Šachmatovo ir N. Joklio recenzijas, žr. A. Шахматов, ЮРЯС, 17, 1912, p. 281–290; N. Jokl, AfslPh, 35, 1913, p. 307–317.

²⁰ A. Meillet, RSl, 5, 1912, p. 153.

²¹ J. Endzelin, Zur slavisch-baltischen Konjugation, APh, 2, 1931, p. 38–46.

18

conservative. J. Endzelin considers that this term can be used to refer to the Balto-Slavic linguistic community (slavisch-baltische Sprachgemeinschaft), and, if desired, also to a proto-language (Ursprache)²².

This article clearly reveals J. Endzelin’s views on the most ancient relations between the Baltic and Slavic languages: after the breakup of the Indo-European linguistic community, the Baltic and Slavic languages formed separate linguistic units whose relations were characterized by contact.

In this article, drawing a conclusion about the most ancient relations between the Baltic and Slavic languages, J. Endzelin relied on the common features of their verbs, since the verb was scarcely touched upon in the monograph “Studies in the Baltic and Slavic Languages.” Contrary to A. Meillet’s view that the verbal structures of the Baltic and Slavic languages were very different, J. Endzelin sought to show that there were many common features in this area as well. For example, in J. Endzelin’s view, the Indo-European optative came to be used in Slavic languages with imperative meaning. The same fate befell it in the Baltic languages: in Prussian it means the imperative, while the Lithuanian permissive (for example, tenesė) derives from an earlier optative.

The Baltic languages have a sigmatic future tense. In J. Endzelin’s view, the participles byšęšt- or byšąst- (cf. Lithuanian būsiančio), etc., show that the Slavic languages had one too.

J. Endzelin held this view of the most ancient relations between the Baltic and Slavic languages until the end of his life. This is shown by his article “The Ancient Balto-Slavic Linguistic Connections”²³, in which he once again repeated his earlier claims.

The 1908–1912 debate²⁴ played an important role in the history of research into the problem of the most ancient relations between the Baltic and Slavic languages. It showed that the problem of the most ancient relations between these languages was not as simple and easy to resolve as had been thought in the nineteenth century. The theory of a Balto-Slavic proto-language, universally accepted by researchers of the previous century, proved to be only one of several hypotheses, which, like the others, had to be substantiated and defended.

The 1908–1912 debate also revealed the weak points of linguistics at the time and the shortcomings of its methods. On the basis of analyses of almost the same linguistic facts, four theories were developed, each interpreting differently the origin and character of the common features (innovations) of the Baltic and Slavic languages. After this debate, the solution to the problem of the most ancient relations between the Baltic and Slavic languages became even more complicated, and this apparently caused interest in theoretical research to wane. This makes understandable the skeptical attitude of some researchers toward the 1908–1912 debate. Aleksander Brückner (1856–1939) called it “a futile affair”²⁵, since A. Meillet did not deny the similarity between the structures of the Baltic and Slavic languages, while the proponent of the theory of Balto-Slavic unity (a proto-language) had honestly indicated the features that distinguished the Baltic and Slavic languages. Brückner himself argued for the unity of the Balto-Slavic languages on the basis of vocabulary. The Baltic and Slavic languages

²² J. Endzelin, op. cit., p. 45.

²³ Ya. M. Endzelin, The Ancient Slavic-Baltic Linguistic Connections, “Proceedings of the Academy of Sciences of the Latvian SSR,” 3 (56), 1952, p. 33—46; “Proceedings of the Institute of Language and Literature of the Academy of Sciences of the Latvian SSR,” 2, 1953, p. 67—82.

²⁴ It has been described in detail by B. Gornung; see Б. В. Горнунг, Из истории изучения балтийско-славянских языковых отношений (научная полемика 1911—1912 г.г.), „Rakstu krājums“, p. 109—132.

²⁵ A. Brückner, Die lituslavische Spracheinheit, KZ, 46, 1914, p. 217—239.

2*

19

stem-formation correspondences, in his view, being more important than all the points indicated by K. Brugman²⁶.

This situation, in which four different hypotheses were formulated on the basis of one and the same linguistic facts, most likely arose because at the time the Slavic languages, and especially the Baltic languages, had not been sufficiently studied, the chronology of the comparative phenomena was unknown, and there was a lack of factual material for more or less convincing theoretical generalizations. K. Būga was therefore entirely right when he wrote that “this question (i.e. the question of the earliest relations between the Baltic and Slavic languages — S.K.) can be resolved only after a comprehensive study of the Baltic and Slavic languages”²⁷.

Despite the fact that research on the Baltic languages has intensified considerably today, many questions in the history of the Baltic languages still need to be resolved by the historical-comparative method or revisited in light of the latest advances in linguistics (for now, we are not speaking of structural research on the Baltic languages, which must still be left to the future). The circumstances mentioned here, together with the late appearance of written monuments in the Baltic languages, the fragmentary texts of Old Prussian, and the absence of written monuments in other Baltic languages—Curonian, Selonian, and others—also hinder successful resolution of the problem of the earliest relations between the Baltic and Slavic languages.

Research into the Earliest Relations between the Baltic and Slavic Languages in the Second Decade of the 20th Century

Kazimieras Būga (1879—1924) did not take part in the discussion of 1908—1912, but he expressed his opinion on the earliest relations between the Baltic and Slavic languages on several occasions.

Already in “Aistiški studijai,” influenced by K. Jaunius’s scholarship, K. Būga regarded the proto-language of the Aists as one of the daughters of the Indo-European proto-language, “which, like a true child, retained very many of its mother’s qualities and characteristics”². K. Būga did not even consider the possibility of a common proto-language or period for the Baltic and Slavic languages.

In an article on the Lithuanian language, written in 1910 for a Polish encyclopedia³, K. Būga called the theory of Baltic-Slavic unity a dogma that had arisen for two reasons: 1) the close proximity of the Baltic and Slavic peoples; 2) the fairly numerous Slavic borrowings and other foreign elements in the Lithuanian language, since linguists of the previous century had been unable to distinguish what was Lithuanian from what was not. Here K. Būga makes a very interesting supposition: if the Aists lived not by the Baltic Sea but somewhere farther away from the Slavs, for example in Iran or on the Iberian Peninsula, it would never occur to anyone to look for a close kinship between the Balts and the Slavs⁴.

K. Būga went on to write that contemporary proponents of the theory of Baltic-Slavic unity had not put forward a single convincing argument in its support. The seven points of K. Brugman,

²⁶ A. Brückner, op. cit., p. 236.

²⁷ K. Būga, RR, I, p. 433.

¹ The terminology used by K. Būga.

² K. Būga, Aistiški studijai, I, Peterburgas, 1908, p. 10.

³ Kaž. Būga, Litewsk i język, „Wielka encyklopedja powszechna ilustrowana“, serja I, 43—44, Warszawa, 1910, p. 624.

⁴ Loc. cit.

20

In K. Būga’s view, these in no way indicate a closer kinship (powinnowactwa) between the Baltic and Slavic languages. Therefore, K. Būga emphasizes, the Baltic language group must be distinguished from the Slavic language group and regarded as a separate member of the Indo-European family ⁵.

Straipsnyje „Kalbų mokslas bei mūsų senovė“ (1913) seniausiesiems baltų ir slavų kalbų santykiams charakterizuoti K. Būga pacituoja J. Endzelyno išvadą⁶, padarytą monografijoje „Baltų ir slavų kalbų etudai“. Iš to matyti, kad K. Būga buvo J. Endzelyno konvergencijos teorijos šalininkas.

Straipsnyje „Šis-tas iš lietuvių ir indoeuropiečių senovės“ (1924) K. Būga rašė: „Kiti kalbininkai tiki, kad kalbos bendrybės, kurių turi gana daug aisčių ir slavėnų kalba, rodančios aisčius ir slavėnus turėjus dar ir atskirą prokalbę po indoeuropiškosios. Tuo aš netikiu, nes nerandu nė vieno kalbos fakto, kuris būtinai verstų mus pripažinti tariamąją aisčių-slavėnų prokalbę.

All the common features of the Baltic and Slavic languages show only that the Balts lived as neighbors with the Slavs (emphasis added in the quotation – S.K.) from time immemorial, that is, since the Indo-European period.

Neighboring languages, even when they are not related to one another, often share phonetic features: having spread from a single center, they can, like waves, inundate the whole area or part of the area of a neighboring language⁷.

K. Būga’s statements on the common Baltic and Slavic lexicon are especially important.

On the sole basis that no correspondences with other Indo-European languages are found in the common Baltic and Slavic lexicon, one cannot, in K. Būga’s view, claim that the Baltic and Slavic languages also had a separate proto-language. Since the Balts and Slavs had been neighbors since ancient times, it is therefore not at all surprising that their languages contain many shared words. This is one reason why the Baltic and Slavic languages have a common lexicon. K. Būga regards the origin of the Baltic and Slavic languages in the Indo-European linguistic community as the second reason: the words shared by Baltic and Slavic may have been “words of the Slavic and Baltic dialect groups of the Indo-European proto-language.” In K. Būga’s view, all these words shared by the Baltic and Slavic languages show only “that in ancient times the two dialects of the northeastern Indo-Europeans (i.e., the Baltic and Slavic dialects – S.K.) had other names for those things, names that did not coincide with those of the western and southern Indo-European proto-language dialects”⁸.

For example, the Balts and Slavs use one word for a crow or raven, whereas the Italics and Greeks had an entirely different word for this, cf. Lithuanian várna, Latvian vārna, Old Prussian warne, Slavic *varnā (Russian воро́на, Polish wrona, Czech vrána, Serbo-Croatian vrȁna); Lithuanian varnas, Old Prussian warnis, Slavic *varnъ (Russian во́рон, Polish wron, Czech vran, Serbo-Croatian vrȁn), and Latin cornix “crow,” Greek κορώνη “crow”; Latin corvus “raven,” Greek κόραξ “raven”⁹. K. Būga especially emphasizes the idea that “Shared words do not yet show that the Slavs and Balts originated not directly from neighboring dialects of the Indo-European proto-language, but through an intermediate stage, the so-called common ‘Slavic-Baltic proto-language’”¹⁰.

⁵ K. Būga, loc. cit.

⁶ K. Būga, RR, I, p. 407–408.

⁷ K. Būga, RR, III, p. 595.

⁸ K. Būga, RR, II, p. 169.

⁹The names for crow and raven are not a feature shared specifically by Baltic and Slavic, since Tocharian also has the same word for this, cf. Tocharian B wrauna (see E. Schwentner, IF, 63, 1958, p. 167).

¹⁰ K. Būga, loc. cit.

21

Thus, in K. Būga’s view, the common vocabulary of the Baltic and Slavic languages arose because:

a) the Baltic and Slavic languages descended from the Indo-European linguistic community, in which the Baltic and Slavic dialects already had words peculiar to them alone;

b) in the course of history, the Baltic and Slavic languages bordered one another directly and were neighbors.

This interpretation of the origin of the common vocabulary of the Baltic and Slavic languages has not lost its validity even today. Prolonged contact between related languages can indeed produce not only analogous phonetic and morphological phenomena, but also lead to the emergence of a common vocabulary.

Įga found not a single fact indicating the unity of the Baltic and Slavic languages (in his terminology, a proto-language), and therefore was firmly opposed to it, whereas the Russian linguists A. Shakhmatov and A. Sobolevsky, working at the same time, had no doubts about the unity of the Baltic-Slavic languages.

Aleksejus Šachmatovas (1864–1920), pažymėjęs, kad baltų ir slavų kalbų ypatingu artumu vargu ar galima abejoti, rašo: „Šis artumas negali būti atsitiktinis ir priklausyti tiktai nuo bendros jų kilmės iš indoeuropiečių prokalbės; ir baltų, ir slavų kalbos turi daug bendrų nukrypimų nuo senųjų santykių, ir šitie bendrieji nukrypimai liudija jų bendrą gyvenimą tuo metu, kada jos jau atsiskyrė nuo kitų indoeuropiečių kalbų. Vadinasi, reikia pripažinti baltus ir slavus kilus iš bendrų protėvių (от одного общего родоначальника), kurie jau buvo atsiskyrę nuo giminės, kalbėjusios indoeuropiečių prokalbe. Baltų ir slavų kalbų protėviu, kurį galima vadinti baltų-slavų prokalbe, buvo kalba tos indoeuropiečių giminės, kuri, išsiskyrusi iš indoeuropiečių šeimos, tik paskui suskilo į dvi — baltų ir slavų šakas“¹¹.

The quotation reveals A. Shakhmatov’s view of the most ancient relations between the Baltic and Slavic languages: he asserts without hesitation that there was a Baltic-Slavic unity, both linguistic and ethnic, which existed after the dissolution of the Indo-European linguistic community.

Aleksey Sobolevsky (1856–1929) developed an original and intriguing hypothesis about the genesis of the Slavic peoples and languages. In addressing the problem of the formation of the Proto-Slavic language, he paid particular attention to the reflexes of Indo-European s in the Slavic languages. After i, u, r, and k in Slavic, Indo-European s became ch; in all other positions it remained unchanged. A. Sobolevsky absolutized this phonetic law and concluded that Proto-Slavic was a combination of two languages: an s-language and a ch-language¹².

The ch-language belonged to the Iranian branch of languages and was most likely one of the other languages’ pronunciations. Words derived from that language have ch instead of s. The s-language, in A. Sobolevsky’s opinion, was “close to the Baltic languages, having earlier formed with them a unified Slavic-Baltic, or Slavic-Lithuanian, language”¹³. Words and forms with s instead of the old s belonged to it; this change to ch never occurred in Baltic-Slavic.

Thus A. Sobolevsky was convinced that the Baltic and Slavic peoples had once spoken a single language. In his view, this unified Baltic-Slavic language had existed until the beginning of our era, that is, until the time when the Scythians encountered the Baltic-Slavs on the shores of the Baltic Sea.

A. Sobolevsky could say nothing more about the hypothetical Baltic-Slavs and their language: “From where did the ancestors of the peoples with the s-language come to the Baltic coast, the Baltic-Slavs’

¹¹ А. Шахматов, Очерк древнейшего периода истории русского языка, „Энциклопедия славянской филологии“, вып. II, Петроград, 1915, p. IX.

¹² А. Н. Соболевский, Русско-скифские этюды, ИОРЯС, 27, 1924, p. 334.

¹³ А. Н. Соболевский, op. cit., p. 334.

22

people, we do not know; scarcely from the west, southwest, or east. We have no reason to seek it among the peoples mentioned by Herodotus either.” In his view, the people speaking the ch-language are an ancient inhabitant of the Baltic coast, who also gave its name to the Nemunas River¹⁴.

A. Sobolevsky’s hypothesis was sharply criticized and rejected¹⁵. Both A. Shakhmatov’s and A. Sobolevsky’s hypotheses lacked argumentation based on analysis of factual material from the Baltic and Slavic languages. The example of A. Shakhmatov and A. Sobolevsky seemed to confirm K. Būga’s words that, when addressing questions connected with the genesis of languages, it is first of all necessary to study the languages themselves thoroughly.

The problem of the oldest relations between the Baltic and Slavic languages in the 1920s and 1930s

During the third and fourth decades of the twentieth century, scholars studied primarily the Baltic and Slavic languages themselves. In examining particular questions of their phonetics, accentology, morphology, and lexicology, researchers also drew more general conclusions concerning the most ancient relations between these languages. This inductive approach to resolving the question of the most ancient relations between the Baltic and Slavic languages, based on the analysis of abundant factual material, was methodologically of exceptional significance.

N. van Wijk’s study on the origins of the accentual and intonational systems of these languages¹, and R. Trautmann’s dictionary of the Baltic and Slavic languages², were of great importance for research into the most ancient relations between the Baltic and Slavic languages.

In the introduction to his book, Nikolai van Wijk (1880—1941) stated that there was no fundamental difference between A. Schleicher’s and J. Rozwadowski’s theories: both scholars recognized a “Balto-Slavic period.” N. van Wijk went on to add explicitly that A. Meillet’s theory, too, differed little from A. Schleicher’s, although at first glance the difference between them might seem very great. For, according to A. Meillet, the Baltic-Slavic dialects (which might also be said to constitute a “proto-language”) existed within the Indo-European linguistic community until it split, whereas A. Schleicher, as his “genealogical tree of languages” diagram shows, recognized their existence even after the Indo-European linguistic community (in A. Schleicher’s terminology, the proto-language) had split. N. van Wijk himself held that the dialects of the Indo-European linguistic community from which the Baltic and Slavic languages later developed formed a fairly close-knit (zusammenhängendes), though dialectally differentiated, whole (Ganzes) even when other groups of Indo-European languages, such as Indo-Iranian, had already separated. This continuum apparently did not last long, since it left no notable traces in the languages known to us. Therefore, says N. van Wijk, we can in practice disregard the theoretically possible Baltic-Slavic linguistic unity³.

Having reconstructed two intonations for the Proto-Baltic and Proto-Slavic dialects—the circumflex and the acute—N. van Wijk noted that “this in no way indicates

¹⁴ A. H. Соболевский, op. cit., p. 332.

¹⁵ Kritišką A. Sobolevskio hipotezės įvertinimą yra davęs A. Pogodinas, žr. А. Л. Погодин, Теория акад. А. Н. Соболевского о двояком происхождении славянского племени, „Slavia“, 9, 1930—1931.

¹ N. van Wijk, The Baltic and Slavic Accentual and Intonational Systems (A Contribution to the Study of Baltic-Slavic Relations), Amsterdam, 1923.

² R. Trautmann, Baltisch-slavisches Wörterbuch, Göttingen, 1923.

³ N. van Wijk, The Baltic and Slavic Accentual and Intonational Systems, 2nd ed., ’s—Gravenhage, 1958, p. 13.

23

a closer kinship between the two language groups, because: 1) the quality of Proto-Baltic and Proto-Slavic intonations can be determined only very imprecisely; 2) Greek, too, has similar intonation relationships”⁴.

Reinhold Trautmann’s (1883—1951) “Dictionary of the Baltic and Slavic Languages” is the largest and most comprehensive collection of Baltic and Slavic vocabulary. It consists mainly of words shared by the Baltic and Slavic languages, although it also includes words that exist only in one of the language groups, or even in just one language. This apparently happened because R. Trautmann was convinced that a unity of the Baltic and Slavic languages (baltisch-slavische Sprachgemeinschaft) had existed⁵: in his view, words of the Baltic languages that now have no counterparts in the Slavic languages might once have existed in the Slavic languages as well.

A large part of the vocabulary presented by R. Trautmann cannot be considered Balto-Slavic: some of the words shared by the two language groups have counterparts in other Indo-European languages and therefore were most likely inherited from the Indo-European linguistic community; others are found in only one language group, or even in just one language.

R. Trautmann’s dictionary was important for the study of the vocabulary of the Baltic and Slavic languages, as well as of the earliest relations between those languages. However, it also had a certain negative effect on the question of the earliest relations between the Baltic and Slavic languages. Researchers were influenced by the very fact that such a dictionary had been compiled: if such a large dictionary could be written from the vocabulary of the Baltic and Slavic languages, then there could be no doubt that these languages had once been united! An obvious example is the following statement by F. Specht: “That the Baltic and Slavic languages once underwent a common development, which can be explained only by the theory of a Baltic and Slavic proto-language, should no longer be in doubt after the appearance of R. Trautmann’s ‘Dictionary of the Baltic and Slavic Languages’”⁶.

The question arises: what is the actual significance of vocabulary in the study of linguistic relationships of kinship? In this connection, A. Meillet’s statements about the development of the vocabulary of the Baltic and Slavic languages are worth recalling. Raising questions about the earliest relations between the Baltic and Slavic languages and the significance of their vocabulary, and referring more than once to R. Trautmann’s dictionary, in his article “The Origins of the Slavic Vocabulary: I. — The Problem of Balto-Slavic Unity”⁷ he pointed out:

1. Lost elements characterize the vocabulary just as much as retained ones. The Baltic and Slavic languages lost many old Indo-European words, replacing them with new ones. For example, the Baltic and Slavic languages lost the old word for “father” (Sanskrit pitá, Greek πατήρ, Latin pater). In its place, the Balts began using a neologism: Lithuanian tėvas, Latvian tēvs and Prussian tāws, while the Slavs used the expressive formation Old Church Slavonic ot-bьcь (for the suffix, compare the adjective Old Church Slavonic ot-bьjь “paternal”). The fact that the Baltic and Slavic languages lost old words and created different substitutes in their place shows that the Balts and Slavs had similar living conditions and that their languages changed in parallel.

2. The lexical correspondences between the Baltic and Slavic languages do not show that the Baltic and Slavic languages had a period of common life after the Indo-European period, since this shared Baltic and Slavic vocabulary was inherited from the dialects of the Indo-European linguistic community; compare, for example, Lithuanian miglà, Old Church Slavonic mbgla and Greek ο-μίχλη⁸.

⁴ N. van Wijk, op. cit., p. 46.

⁵ R. Trautmann, op. cit., p. V.

⁶ F. Specht, Zur baltisch-slavischen Spracheinheit, KZ, 62, 1934, p. 248.

⁷ A. Meillet, Les origines du vocabulaire slave: I. — Le problème de l’unité balto-slave, RÉSI, 5, 1925, p. 5—13.

⁸ A. Meillet, loc. cit.

24

3. The parallel development of the Baltic and Slavic languages is reflected in lexical derivational formatives. For instance, in instrument nouns Baltic has the suffix *-tlo- (Lith. -kla-), whereas Slavic has the suffix *-dhlo; cf. Lith. árklas and OCS ralo, Czech rádlo. These suffixes are similar but not identical, and both are ancient: the first corresponds to Lat. -culum (cf. Lat. pōculum ‘cup’), while the second corresponds to Lat. -bulum (cf. Lat. pābulum ‘fodder’) and Gr. -θλον (cf. Gr. γένεθλον).

4. A marked difference between the Baltic and Slavic languages is that the Baltic languages have preserved the Indo-European lexicon better than the Slavic languages. For example, the Baltic languages have well preserved the words for ‘human being’ and ‘man’ (cf. O. Lith. žmuo, Pr. smoy and Lat. hemō or homō, Goth. guma; Lith. vyras, Latv. vīrs, Pr. wijrs and Skt. vīráḥ, Lat. vir, O. Ir. fer, Goth. wair), whereas the Slavic languages have lost them.

Concluding his article, A. Meillet writes: “The Indo-European part of the vocabulary of the Baltic and Slavic languages derives from one and the same group of Indo-European dialects. However, one need only leaf through Trautmann’s *Baltisch-slavisches Wörterbuch* to become convinced that between the Indo-European linguistic community and the historical period of the Baltic and Slavic languages there was no interval during which a body of vocabulary common to both language groups took shape. There is parallelism, not unity”⁹.

Interestingly, A. Meillet’s view of the lexicon of the Baltic and Slavic languages is very close to K. Būga’s: both scholars consider the common lexicon of these languages to be (1) derived from the dialects of the Indo-European linguistic community and (2) to have arisen independently, in parallel, while the two language groups had long lived as neighbors.

In fact, there is absolutely no need to acknowledge Baltic-Slavic linguistic unity in order to account for the common lexicon of the Baltic and Slavic languages: the various contacts between the Balts and Slavs throughout their history suffice to explain its emergence. Moreover, when considering the question of the earliest relations between the Baltic and Slavic languages, the significance of the lexicon should not be overstated, since vocabulary, as is well known, is the most changeable part of a language system.

To substantiate the theory of Baltic-Slavic unity, Franz Specht (1888—1949), in addition to the lexical argument, drew on several grammatical commonalities he had observed¹⁰, such as the replacement of the locative singular of demonstrative pronouns with the accusative plus a postposition (Lith. tamē=OCS tomb from *tamen), the strict distinction between men and n stems, and the loss of masculine agent nouns with the formative *-tēr- (cf. Indo-European *patēr and Lith. tėvas, OCS отbь), among others.

N. van Wijk, R. Trautmann, and F. Specht addressed the question of the earliest relations between the Baltic and Slavic languages on the basis of the conclusions of their own specific research. The Italian Neolinguists took a quite different approach. In keeping with their theoretical principles of general linguistics, they mainly sought to determine the character of the Baltic and Slavic languages.

At first, Vittore Pisani acknowledged¹¹ that, following the breakup of the Indo-European linguistic community, the Baltic and Slavic languages underwent a common period of development. In V. Pisani’s view, the isoglosses of the Baltic and Slavic languages, which cross the area occupied by those languages, suggest that the territories of the Baltic and Slavic languages remained united for a fairly long time after the breakup of the Indo-European linguistic community. V. Pisani called the common period of development of these languages a period of community

⁹ A. Meillet, op. cit., p. 13.

¹⁰ F. Specht, op. cit., p. 248—259.

¹¹ V. Pisani, Balto e slavo, StB, 2, 1932, p. 1—22 (= Linguistica generale e indoeuropea, Milan, 1947, p. 65—82).

25

an epoch (but not a proto-language), which he dated to approximately the fifth century BCE and the fourth century CE. However, V. Pisani later changed his views: he thought that, on the basis of the system of isoglosses, it was scarcely possible to speak of a Balto-Slavic community, and likewise scarcely possible to speak of a Balto-Slavic proto-language, since it apparently had never existed¹².

Another Italian neolinguist, Matteo Bartoli (1874—1946), having analyzed several correspondences between the Baltic and Slavic languages, concluded that the chronological position of the Baltic languages is determined by the fact that they supposedly contain more prehistoric innovations than the Germanic languages, and considerably fewer than Old Indian¹³. Disagreeing with M. Bartoli’s conclusion, Giuliano Bonfante expressed the view¹⁴ that the Baltic languages, especially Lithuanian, are both archaic and conservative, whereas the Slavic languages are conservative but not archaic. G. Bonfante pointed out several archaisms, adding that the Baltic languages, too, had not escaped the waves of innovation that came from the center (Greek, Armenian, and other languages)¹⁵.

The Norwegian linguist Christian Stang not only analyzed the oldest relations between the Baltic and Slavic languages, but also wrote major works illuminating the history of the verb and accentology in these languages.

In one of his articles¹⁶ Chr. Stang posed the question: what, after all, does the concept of “linguistic unity” mean? In his view, linguistic unity can be understood in two ways:

1) a unity characterized by individual (neither dialectal nor social) divergences;

2) a unity in which speakers feel that their language is one and the same, even though it is divided into dialects (according to this definition, for example, German is one language, while Swedish and Norwegian are two, although the difference between the latter is smaller than that between northern and southern German dialects; in Chr. Stang’s view, this kind of linguistic unity is based on political and cultural relations).

There could have been no unity of the first type between the Baltic and Slavic languages after the dissolution of the Indo-European linguistic community, since, for example, the fate of IE s after k, r, i, u brings the Slavic languages closer to the Indo-Iranian languages. The existence of unity of the second type between these languages after the aforementioned dissolution cannot be demonstrated on the basis of linguistic evidence.

Chr. Stang portrays the oldest relations between the Baltic and Slavic languages as follows: the Baltic and Slavic languages descended, one might say, from the same IE dialect. Yet, they already had very ancient differences: the different fate of IE s after k, r, i, u (see this work, p. 81 ff.), different details of the satəm phenomenon (cf. Lithuanian klausyti and Slavic slyšati, see this work, p. 186 ff.), and so on. Later, when the IE linguistic community broke up, the Baltic and Slavic languages went through a certain period during which their territories were encircled by several new isoglosses. Some covered the entire territory of these languages, while others covered only part of it. Such isoglosses include the change of r, l, m, n into ir, il, im, in, and ur, ul, um, un (see this work, p. 76); the verb type with -auju: -ųja; and the nomina agentis Lithuanian artōjas, Old Prussian artoys, Old Church Slavonic ratajь, and so on. As long as, in Chr. Stang’s view, the possibility remained for new

¹² V. Pisani, Glottologia indoeuropea, Torino, 1949, p. 26.

¹³ M. Bartoli, Il carattere conservativo dei linguaggi baltici, StB, 3, 1933, p. 1—26.

¹⁴ G. Bonfante, „Arcaico“ e „conservativo“ nel gruppo baltico, StB, 5, 1935—1936, p. 30—37, plg. „Indogermanisches Jahrbuch“, 21, 1937, p. 326.

¹⁵ For more on this, see B. H. Топоров, Из истории изучения древнейших балто-славянских языковых отношений, p. 22—25.

¹⁶ Chr. Stang, Einige Bemerkungen über das Verhältnis zwischen den slavischen und baltischen Sprachen, NTS, 11, 1939, p. 85—98.

26

As far as the isoglosses are concerned, one may speak of a living isoglossal area (lebendiges Isoglossengebiet).

In 1963 Chr. Stang elaborated and specified his views on the earliest relations between the Baltic and Slavic languages¹⁷. He expressed the opinion that the differing fate of IE s after k, r, i, u, and certain differing reflexes of the satam phenomenon (klaušyti: sly-šāti) in Baltic and Slavic do not date back as far as the period of IE linguistic unity.

Having discussed some innovations in the Baltic and Slavic languages, Chr. Stang concludes that, after the breakup of IE linguistic unity, these languages formed a unified dialect area (Dialektgebiet), in which the innovations of these languages arose and which justifies the use of the expression “Baltic-Slavic languages” (“Baltisch-slavisch”). On the other hand, in his view, it must also be acknowledged that certain old isoglosses intersected in this dialect area, separating Baltic dialects from Slavic ones. The Baltic-Slavic dialect area split in stages, since there are isoglosses of which some encompass the Slavic–East Baltic area, while others cover the Slavic–Prussian area.

In his works, Chr. Stang frequently uses the expression “Baltic-Slavic languages.” The book “The Accentuation of Slavic Languages”¹⁸ notes that he did not use the term “Baltic-Slavic languages” (“Balto-Slavonic”) to mean an absolutely homogeneous language. By this term he understood “a dialect area sufficiently homogeneous for linguistic changes arising within it to affect the whole area or a part of it”¹⁹. Chr. Stang’s view on the earliest relations between the Baltic and Slavic languages was confirmed by the principal conclusions of his book “The Slavic and Baltic Verb”²⁰. In earlier times, the common features of the Baltic and Slavic verbal systems must have been much more extensive, and their present-tense systems must have been very similar after the breakup of IE linguistic unity²¹. Later, as a result of differing developmental tendencies, the verbal systems of these languages diverged.

Chr. Stang also points to major structural differences between the Baltic and Slavic verbs. Baltic, like Germanic, has no sigmatic aorist and apparently never had one. The Slavic sigmatic aorist closely resembles the Indo-Iranian one: both have a lengthened root vowel grade. In this connection, Chr. Stang observes that, with respect to the aorist, the Baltic and Slavic languages may have belonged to different geographical spheres. Slavic languages have no long-root preterite, which is found in Baltic, Germanic, Celtic, Italic, and Albanian. The sigmatic future tense is widespread in Baltic, but is absent from Slavic. In Slavic, the perfective present came to perform the function of the future tense, so the sigmatic future tense was no longer needed.

Chr. Stang’s analysis shows that some structural types and categories (for example, nasal and i-stem verbs, and the optative) are more likely the result of convergent or even parallel development.

A survey of research into the earliest relations between the Baltic and Slavic languages in the 1920s and 1930s shows that researchers of this period, with the exception of R. Trautmann and F. Specht, found no evidence permitting the recognition of an unbroken-

¹⁷ Chr. Stang, Über das Verhältnis zwischen slavischen und baltischen Sprachen, „Prace językoznawcze“, 5, 1963, p. 393–394.

¹⁸ Chr. S. Stang, Slavonic accentuation, Oslo, 1957.

¹⁹ Chr. S. Stang, op. cit., p. 174.

²⁰ Chr. S. Stang, Das slavische und baltische Verbum, Oslo, 1942.

²¹ Chr. S. Stang, loc. cit., p. 268, 271.

27

the Baltic-Slavic linguistic unity. On the contrary, the study of specific accentological and morphological phenomena made it possible to draw conclusions indicating that the earliest relations between the Baltic and Slavic languages were of a contact nature.

The study of the earliest relations between the Baltic and Slavic languages in the postwar years

Study of the problem of the earliest relations between the Baltic and Slavic languages intensified after the Second World War. Several articles devoted specifically to this problem were published. The earliest relations between these languages are also characterized in more than one major work on the history of the Baltic or Slavic languages. E. Fraenkel’s book The Baltic Languages, published in 1950¹, prompted the publication of an article by J. Endzelin, to which J. Otrębski, A. Zenas, and others responded in turn (see below). One might say that a certain discussion ensued. The earliest relations between the Baltic and Slavic languages were discussed especially actively in preparation for the Fourth International Congress of Slavists in Moscow.

In the postwar years, research on the earliest relations between the Baltic and Slavic languages was characterized by detailed analysis of known commonalities and differences and the identification of new ones, as well as by the raising of the question of their chronology. Moreover, researchers at this time avoided proposing hypotheses that offered new interpretations of the earliest relations between the Baltic and Slavic languages, being content instead with modifying theories advanced in 1908–1912, providing more detailed arguments for them, or criticizing them.

In 1946, two works by Polish linguists devoted to the problem of the earliest relations between the Baltic and Slavic languages appeared²—J. Safarewicz’s article “On the question of Baltic-Slavic unity”³ and T. Lehr-Splawiński’s book On the Origin and Homeland of the Slavs⁴.

In his article, Jan Safarewicz observed that, on the one hand, such innovations shared by the Baltic and Slavic languages as the formation of the present tense from aorist roots (cf. Lithuanian duomi, Old Church Slavonic damъ and Old Indic dādam, Greek ἔδομεν), and the derivation of the type minėti from the perfect root *men-, are very ancient and therefore could not have arisen when the Baltic and Slavic languages converged for a second time. In making this claim, J. Safarewicz undoubtedly had J. Endzelin’s convergence theory in mind. In his view, these and similar innovations indicated that an era of unity had probably existed, immediately after Indo-European linguistic unity, continuing it⁵. On the other hand, such ancient differences between the Baltic and Slavic languages as the use of the present-tense suffix *-do (cf. Old Church Slavonic idǫ and Lithuanian eimi, Lithuanian vėrda, which is absent in the Slavic languages), in J. Safarewicz’s view, show that the era of unity between these languages (okres wspólnoty) was not long-lasting. He believes that Baltic-Slavic unity disintegrated not because of geographical contacts, but as a result of an influx of Iranian tribes.

Later, J. Safarewicz’s remarks on Baltic and Slavic linguistic unity became even more critical⁶. In 1961 he wrote that “the existence of an era of Baltic-Slavic unity is by no means certain”⁷. J. Safarewicz further notes that the commonalities between the Baltic and Slavic languages

¹ E. Fraenkel, Die baltischen Sprachen, Heidelberg, 1950.

²During the war years, H. Pedersen’s article “Et baltoslavisk problem” (In memoriam Kr. Sandfeld, København, 1943) was published; unfortunately, the author of this work was unable to consult it.

³ J. Safarewicz, Przyczynki do zagadnienia wspólnoty bałto-słowiańskiej, „Sprawozdania z czynności i posiedzeń Polskiej Akademii Umiejętności“, 46, 1946, p. 199–202.

⁴ T. Lehr-Splawiński, O pochodzeniu i praojczyźnie słowian, Poznań, 1946.

⁵ J. Safarewicz, op. cit., p. 200.

⁶ J. Safarewicz, Bałto-słowiańskie stosunki językowie, „Słownik starożytności słowiańskich“, I, 1 (A–B), Wrocław–Warszawa–Kraków, 1961, p. 79–80.

⁷ J. Safarewicz, op. cit., p. 79.

28

are more similar to parallel, independently arisen innovations (cf. the singular genitive ending of o-stem nouns—Lith. -o, OCS -a and Pr. -as). He even regards such a conspicuous common feature of the Baltic and Slavic languages as adjective forms with pronominal endings as having developed independently in the two language groups. In J. Safarewicz’s view, only the beginnings of this phenomenon date back to the period of Baltic-Slavic unity. He believes that the Baltic and Slavic languages derive from a single dialect of the Indo-European linguistic community and that the period of Baltic-Slavic unity, if indeed it existed, may have occurred in the first half of the second millennium BCE.

The question of the earliest relations between the Baltic and Slavic languages was approached from a somewhat different perspective by another Polish linguist, Tadeusz Lehr-Splawiński (1891–1965). In his book “On the Origin and Homeland of the Slavs,” he succinctly set out his conception: “the ancestors of the Slavs and Balts once experienced a period of common linguistic development, which continued even after their ties with other groups of Indo-European dialects had already been severed⁹. T. Lehr-Splawiński follows the methodological principle that innovations provide much more reliable evidence of common linguistic development than Proto-Indo-European archaisms, which, in his view, could easily have survived in the independent development of those languages as well. To support his claim, T. Lehr-Splawiński lists several shared innovative changes in the phonetics, morphology, and lexicon of the Baltic and Slavic languages. Nevertheless, he considers lexical commonalities the most important argument; in his view, there are so many of them that they constitute a large (with as many as 1,600 entries) etymological dictionary of the Baltic and Slavic languages (the reference is to R. Trautmann’s dictionary).

In T. Lehr-Splawiński’s view, the old differences in morphology, and especially in vocabulary, do not refute the thesis of the original unity of the Baltic and Slavic languages. First, these differences show that the Indo-European dialects spoken by the ancestors of the Balts and Slavs during the period of the Indo-European linguistic community were closely related, but also differed considerably in both grammatical and lexical composition. This by no means prevented the formation of common developmental tendencies, which operated even after the breakup of the Indo-European linguistic community and later brought the two dialect groups closer together during their common period. Second, the old differences between the Baltic and Slavic languages show that the common Baltic-Slavic period (doba wspólności językowej bałto-słowiańskiej) existed in the distant past. It was followed by a fairly long period of independent development of the Proto-Baltic and Proto-Slavic dialects, during which tendencies toward differing grammatical structures arose and lexical differences increased, some of them inherited from the Indo-European linguistic community. At that time, contacts between the two dialect groups weakened considerably, but did not cease entirely. After some time, these contacts became active again, and their reflection was the borrowing of words by the Baltic languages from Slavic and by Slavic from Baltic.

This conception of T. Lehr-Splawiński¹⁰, set out in his 1946 book “On the Origin and Homeland of the Slavs,” was developed in many articles in subsequent years¹¹, but its essence remained the same.

⁸ J. Safarewicz, op. cit., p. 80.

⁹ T. Lehr-Splawiński, op. cit., p. 23.

¹⁰ Jos kvintesencija žr. ВЯ, I, 1958, p. 37–41 ir „Сборник ответов на вопросы по языкознанию (К IV Международному съезду славистов)“, М., 1958, p. 152–157.

¹¹ For a bibliography of T. Ler-Splavinskis’s works published in 1946–1955, see V. Falkenhahn, ZfSl, 2, 1956, p. 49–88; for a bibliography of offprints of his works (including those on the question of the earliest relations between the Baltic and Slavic languages) from 1951–1961, see “Studia linguistica in honorem Thaddaei Lehr-Splawiński”, Warszawa, 1963, p. 7–8.

29

T. Ler-Slavinskis’s article is worth mentioning, in which he attempts to characterize the Indo-European dialects from which the Baltic and Slavic languages arose¹².

In T. Ler-Slavinskis’s view, the Indo-European dialects from which the Baltic and Slavic languages later developed did not form a compact unity during the period of the Indo-European linguistic community; rather, they consisted of several closely related dialects that had quite numerous differences. To characterize the Indo-European dialectal basis of the Baltic and Slavic languages, T. Ler-Slavinskis presents the following features as evidence of its heterogeneity:

I. Phenomena characteristic of the Baltic languages and unknown in the Slavic languages:

a) the Baltic languages, like all Indo-European western languages (except Germanic), distinguish Indo-European ō and ā, whereas in the Slavic languages, as in Indo-Iranian and Germanic, the long Indo-European vowels merged (see this work, p. 61 ff.);

b) the Indo-European cluster sr- remained intact in Lithuanian (cf. Lith. sraujà and O.Ind. srá-vati), whereas in Latvian, Prussian, Slavic, and all Indo-European western languages it became str- (cf. OCS struja¹³);

c) different formation of the numeral “first,” cf. Lith. pìrmas, Pr. pirmas, O.Eng. forma, O.Sax. formo (suffix *-mo-) and OCS prъvъ (suffix *-uo-, as in Indo-Iranian, Tocharian, and probably Illyrian);

d) different formation of the numerals “eleven” and “twelve,” cf. the second element *-leig- (Lith. vienúolika, dvýlika, Goth. ainlif, twalif, O.H.G. einlif, zwalif) and the syntactic construction in Old Church Slavonic (jedinъ na desęte, dъva na desęte);

e) Prussian preserved the Indo-European medial first-person ending -mai (Pr. asmai “I am,” cf. Greek -μαι).

II. Phenomena characteristic of the Slavic languages and unknown in the Baltic languages:

a) In Slavic, as in Indo-Iranian, PIE s after k, r, i, u became (š>) ch, whereas in Lithuanian this happened only after r, k (cf. OCS mъchъ, ucho and Lithuanian músos, ausis; for a fuller discussion, see p. 81 ff.);

b) agent nouns with the suffix -telb (cf. Hittite -talla-);

c) the suffix *-es-, as in Hittite, in names of body parts (cf. OCS ucho ušese, oko očese);

d) the suffix of adjectival abstracts -ostb (cf. Hittite -ašti)¹⁴;

e) the formation of deverbal adjectives with participial function using the suffix *-lo- (cf. OCS bylъ and Hittite dalugmūla from dalugnu- ‘lengthen’)¹⁴;

f) the use of the suffixes *-no- || *-io to form intransitive and transitive verbs from the same root, a feature shared with the Germanic languages (cf. OCS užasnąti: užasiti and Gothic usgeisnan: usgaisjan);

g) the perfect 1st-person ending -é (OCS vedě=Latvian vidē) and the retention of the medial 3rd-person ending -tъ < *-tor.

These ancient differences between the Baltic and Slavic languages show that, at the time the PIE linguistic community broke up, Baltic and Slavic dialects differed in many grammatical features. Since the territories of the Baltic and Slavic languages bordered one another directly for many centuries—a circumstance that undoubtedly accelerated the process of leveling differences and the emergence of commonalities—the differences at the time of the breakup of the PIE linguistic community may, in T. Ler-Slavinskij’s view, have been much more numerous and more substantial.

¹² T. Lehr-Splawiński, Podstawy indoeuropejskie wspólnoty językowej bałto-słowiańskiej, „Z polskich studiów slawistycznych“, Warszawa, 1958, p. 125—136.

¹³ There are examples in Lithuanian with str- as well; cf. Lithuanian strañjas ‘fast (of water)’, strėbti, stragùs.

¹⁴ These two suffixes seem to occur in the Baltic languages as well; cf. Lithuanian bjaur-astis (: bjaurùs) and dęg-las (: dègti).

30

It follows from T. Ler-Splavinsky’s theory that (1) he derives the Baltic and Slavic languages from dialects of the Indo-European linguistic community that were not identical, and (2) the Baltic and Slavic dialects drew closer only later. Although at first glance T. Ler-Splavinsky’s theory recalls J. Endzelin’s concept of convergence, in essence the two differ considerably: the Polish linguist recognized not only the linguistic but also the ethnic kinship of the Balts and Slavs.

André Vaillant similarly addresses the question of the earliest relations between the Baltic and Slavic languages, readily acknowledging a period of Baltic-Slavic unity (une période d’unité balto-slave)¹⁵. In his view, the abundance of shared features in the Baltic and Slavic languages, especially striking ones such as adjectival forms with pronominal endings, clearly indicates that a period of Baltic-Slavic linguistic unity existed for quite a long time after the disintegration of the Indo-European linguistic community. On the one hand, the French Slavist acknowledges that the Baltic languages differ from Slavic no more than Swedish differs from German; on the other, he is compelled to observe that Slavs readily perceive the kinship of their languages, whereas Lithuanian is entirely foreign to them.

A. Vaillant is inclined to regard the Baltic Veneti (Vénèdes de la Baltique) as Balto-Slavs, whose languages still differed little around the beginning of our era. In his view, the common Balto-Slavic language should be conceived of as a cluster of dialects (un faisceau de parlers), whose central and peripheral dialects differed from one another to a greater or lesser degree. The Germanic tribes, especially the Goths from Scandinavia, the French Slavist further argues, divided the Balto-Slavic people into two groups around the beginning of our era and thereby finally broke apart the unity of the Balto-Slavic languages, which had already been dialectally fragmented.

A. Vaillant sets out the task of reconstructing the Balto-Slavic language. In that case, researchers of the Baltic and Slavic languages would have to work with three reconstructed linguistic models—the Indo-European, Balto-Slavic, and Proto-Baltic or Proto-Slavic.

In the article “Balto-Slavic Linguistic Unity”¹⁶, A. Vaillant analyzed the commonalities and differences between the Baltic and Slavic languages in considerable detail, particularly emphasizing the identity of those facts in which A. Meillet saw differences between these languages (for example, in the verb). At the end of the article, A. Vaillant sets out the task of finding a unified basis for the distinct phenomena and facts of the Baltic and Slavic languages¹⁷. The article shows that A. Vaillant does indeed attempt to find just such a common original basis for a number of Baltic and Slavic linguistic phenomena. Some Slavic linguistic phenomena that developed later do in fact seem interpretable on the basis of Baltic-language data. For instance, A. Vaillant quite reasonably points out that in the Slavic imperfect inflection one can recognize stems extended by the suffixes *-ējā- and *-ājā- (cf. OCS vidēa- and iškaa-), which are found only in the Baltic languages (cf. Lithuanian mylėjo ‘loved’ and ieškojo ‘searched for’). However, the common original basis that A. Vaillant proposes for other phenomena raises serious doubts. For example, A. Vaillant believes that Lithuanian neša ‘carries’ results from the merger of the singular third-person form *nešeti and the plural third-person form *nešanti. It is doubtful that the Baltic languages ever had thematic verb forms for the singular and plural

¹⁵ A. Vaillant, Grammaire comparée des langues slaves, I, Lyon—Paris, 1950, p. 13—15; Le problème de l’unité balto-slave, BSL, 51, 1955, p. XIII—XV.

¹⁶ A. Vaillant, L’unité linguistique balto-slave, „Filologija“, I, Zagreb, 1957, p. 23—35.

¹⁷ “The original unity must be recovered where it has ceased to be apparent,” – A. Vaillant, op. cit., p. 34.

31

third-person endings *-ti and *-nti. There is insufficient evidence to prove this. It goes without saying that A. V. Vaillant reconstructed the aforementioned singular and plural third-person forms on the basis solely of corresponding Slavic-language facts; cf. Old Russian nesetb and nesoutb.

In 1950, when A. V. Vaillant’s “Comparative Grammar of the Slavic Languages” appeared, a work on the history of the Baltic languages was also published—E. Fraenkel’s book “The Baltic Languages”¹⁸. The authors of both books devoted considerable attention to the question of the earliest relations between the Baltic and Slavic languages, but reached entirely opposite conclusions. A. V. Vaillant is an ardent supporter of the theory of Balto-Slavic linguistic unity, whereas E. Fraenkel holds that the common features and innovations of the Baltic and Slavic languages arose through long-standing neighborhood.

The distinguished German Baltist and author of a Lithuanian etymological dictionary¹⁹, Ernst Fraenkel (1881—1957), analyzed in detail in his book “The Baltic Languages” the shared and distinct phenomena of Baltic and Slavic phonetics, accentology, morphology, syntax, and lexicon. For many of the phenomena under discussion, he provided parallels from other Indo-European languages, showing that they are not characteristic only of the Baltic and Slavic languages. He also pointed out major structural differences between the Baltic and Slavic languages.

E. Fraenkel acknowledges that the Baltic and Slavic languages are linked by many isoglosses, which, in his view, arose because (1) the Baltic and Slavic languages, being distant from the center of active contact and exchange among Indo-European languages, retained an old form, and (2) they were neighbors for many centuries²⁰. In addition, these language groups often exhibit parallel changes that arose independently, but which in themselves do not indicate a closer relationship between the Baltic and Slavic languages. Significant differences also occur, especially in vocabulary and syntax. At times, the Slavic languages have fewer archaic features than the Baltic languages. The Baltic, Slavic, and Germanic languages form a Sprachbund in the sense of the Prague Linguistic Circle²¹.

J. Endzelin responded to E. Fraenkel’s book²² (see p. 19 of this work), engaging E. Fraenkel in polemics on particular questions; somewhat later, substantial articles by J. Otrębski and A. Zēns appeared. In their treatment of the question of the earliest relations between the Baltic and Slavic languages, these articles recalled works published in 1950: as then, one argued for Baltic and Slavic linguistic unity, while the other denied it.

In the article “Slavic-Baltic Linguistic Unity,” Jan Otrębski²³ revised his earlier position on the question of the earliest relations between the Baltic and Slavic languages. In 1947 he wrote that “a Balto-Slavic proto-language never existed”²⁴, since, in his view, the differing development of Indo-European s after k, r, i, u in Baltic and Slavic languages ruled out such a conclusion. By 1954, J. Otrębski held that the Baltic and Slavic languages had developed from dialects of a single Balto-Slavic linguistic group. This group should be regarded as a unity that developed from a single language, which in turn had branched off from the Indo-European linguistic group²⁵. The vocabulary of the Baltic and Slavic languages, according to J. Otrębski

¹⁸ E. Fraenkel, The Baltic Languages (Their relationships to one another and to the Indo-European sister idioms as an introduction to Baltic linguistics), Heidelberg, 1950.

¹⁹ E. Fraenkel, Litauisches etymologisches Wörterbuch, Heidelberg—Göttingen, 1955 tt.

²⁰ E. Fraenkel, op. cit., p. 75.

²¹ E. Fraenkel, op. cit., p. 75.

²² Я. М. Эндзелин, Древнейшие славяно-балтийские языковые связи, 1953, p. 67—82.

²³ Я. С. Отрембский, Славяно-балтийское языковое единство, ВЯ, 1954, 5, p. 27—42, 6, p. 28—46.

²⁴ J. Otrębski, „Slavia Occidentalis“, 18, 1947, p. 448.

²⁵ Я. С. Отрембский, ВЯ, 6, p. 43.

32

in his view, was so similar that Balts and Slavs, speaking their own languages, understood one another not only during the period of unity but also for a long time after its breakup. J. Otrębski sees identical phenomena and similar processes throughout the Baltic and Slavic languages: the Slavic languages have a sigmatic aorist—which the Baltic languages must also have had, where it later merged with the future tense; in the Slavic languages, the diphthongs oi and ei became ē and ī—a similar process took place in the Baltic languages, where the diphthongs oi and ei became ie, and so on. The weak point in J. Otrębski’s argumentation is that he treats phenomena from different chronological periods as having occurred simultaneously: the merger of the long vowels ō and ā, the reflexation of Indo-European syllabic sonants, the monophthongization of the diphthongs oi and ei, the assibilation of Indo-European palatal velar consonants, and the retraction of the accent according to the de Saussure–Fortunatov law (for more on some of these, see the second part of this work). Moreover, he analyzes phenomena in the Baltic and Slavic languages without relating them to similar phenomena in other Indo-European languages. For instance, the change of the cluster tt to st cannot be considered a phenomenon characteristic of the Baltic and Slavic languages and thus an argument for Baltic–Slavic linguistic unity, since this phenomenon is also known in other Indo-European languages (Iranian and Greek). Since the cluster tt changed to st quite late (after the separate Indo-European language groups had already formed), and therefore independently in Iranian and Greek, it could likewise have changed independently in the Baltic and Slavic languages.

To demonstrate the complete identity of nominal inflection in the Baltic and Slavic languages, J. Otrębski reconstructs the paradigm of the numeral “three” in the Baltic and Slavic languages, but fails to note that other Indo-European languages—Germanic, Old Indic, and others—have almost exactly the same case endings.

A. Vaillant, an ardent proponent of the theory of Baltic–Slavic linguistic unity, has repeatedly pointed out that the Baltic–Slavic language should be imagined as consisting of dialects, whereas in J. Otrębski’s view, the Baltic and Slavic languages derived from a single language that had separated from the Indo-European language group and only later split into dialects.

Alfredas Zenas²⁸ is categorically opposed to the theory of Baltic–Slavic linguistic unity defended by A. Vaillant, J. Otrębski, and others. He believes²⁹ that the ancient Balts (Urb alten) lived north of the Pripet Marshes (Pripet- oder Rokitno-Sümpfe). To the west, north, and east, the ancient Baltic tribes were surrounded by Finno-Ugric peoples, while to the southwest their immediate neighbors were the ancient Germans (Urgermanen). The impassable Pripet Marshes separated the Balts from the Slavs, hindered their contact, and finally isolated the Balts and their language. A. Zenas writes that because the Pripet Marshes were impassable, there could have been no direct contacts between the Balts and the Slavs³⁰.

The Baltic settlements north of the Pripet, A. Zenas further argues, were not original. The Proto-Balts (Vorb alten) lived west or southwest of the Pripet Marshes, where their neighbors were the Slavs, Germans and, possibly, the Illyrians. There they formed a unity of a state-like character (staatliche Einheit), but not an Illyrian-Germanic-Slavic-Baltic proto-language (Ursprache). For the independent linguistic development of Baltic—in other words, for its linguistic community (Urbaltisch) to take shape—the neighbors of the Balts, the Proto-Slavs, had to change

²⁶ W. Porzig, Die Gliederung des indogermanischen Sprachgebiets, Heidelberg, 1954, p. 76–78.

²⁷ Я. С. О тр ембский, ВЯ, 6, р. 28–29.

²⁸ A. Senn, On the Degree of Kinship between Slavic and Baltic, „The Slavonic and East European Review“, 20, 1941, p. 251–265; Die Beziehungen des Baltischen zum Slavischen und Germanischen, KZ, 71, 1954, p. 162–188.

²⁹ A. Senn, Die Beziehungen des Baltischen zum Slavischen und Germanischen, p. 162–188.

³⁰ A. Senn, op. cit., p. 164.

3. Baltų ir slavų kalbų ryšiai

33

and Proto-Germanic (Vorslaven und Vorgermanen) languages, while Proto-Baltic itself may still have remained “Proto-Indo-European” (“urindogermanisch”).

A. Zenas believes that the Balts and Slavs came into contact only after the Slavic expansion northward: the eastern Balts with the East Slavs, and the Prussians with the West Slavs.

A. Zenas lists the following common features of the Baltic and Slavic languages: a) the loss of postconsonantal į before front vowels, b) the transfer of nt participles to the io stem, c) the plural dative ending -mus, d) singular first-person forms of personal pronouns with *mur, e) the first-person plural genitive of personal pronouns *nō(s)sōm³¹.

According to A. Zenas, all the features shared by the Baltic and Slavic languages are either archaisms inherited from Indo-European times or innovations that arose independently in each language group.

The article “Relations of the Baltic Languages with the Slavic and Germanic Languages” seems to suggest that A. Zenas adheres to the theory of parallelism. However, in a later article, “Baltic and Slavic Relations³²,” A. Zenas says that he is not a proponent of A. Meillet’s theory of parallelism and adds that he does not believe in the parallel development of two neighboring languages whose speakers lived in close symbiosis for more than a thousand years. In A. Zenas’s view, the numerous correspondences and similarities between the Baltic and Slavic languages reflect the political and cultural influence that Russians and Poles exerted over Lithuanians and Latvians in the past.

Unlike J. Otrębski and A. Zenas, Walter Porzig approached the earliest relations between the Baltic and Slavic languages from an entirely different methodological perspective. In his book “The Division of the Indo-European Linguistic Area”³³, the author’s conclusions on the question of the earliest relations between the Baltic and Slavic languages are particularly important, not only because comparing the relations between these languages with those between other Indo-European language groups (for example, Indo-Iranian or Celtic-Italic) helps clarify what these earliest relations between the Baltic and Slavic languages were and which theory best explains their origin. These conclusions are especially important also because, in order to characterize the earliest relations between the Baltic and Slavic languages, W. Porzig uses an entirely different method. He writes: “Our knowledge allows us to conclude that the two language groups were neighbors throughout. This gave rise to many innovations peculiar to them, almost all of which are lexical or word-formational. The Baltic and Slavic languages appear to have no innovations in phonetics or inflection. It is possible that the dialectal areas from which the Baltic and Slavic languages inherited a large part of their systems were already neighbors. There is little direct evidence for this. The reflexation of syllabic sonants in the Baltic and Slavic languages, which took place in the separate languages, is the same, but in itself it has no evidential force. Our hypothesis that the original areas were neighboring would be confirmed or refuted if the two language groups had the same, or often different, relations with other languages”³⁴.

The quotation shows that, in characterizing the relations between the original Baltic and Slavic dialect areas—in other words, the earliest relations between the Baltic and Slavic languages—W. Porzig abandons the traditional argument based on the features shared by and the differences between the Baltic and Slavic languages: these languages

³¹ A. Senn, op. cit., p. 185.

³² A. Senn, The Relationships of Baltic and Slavic, „Ancient Indo-European Dialects“, Berkeley and Los Angeles, 1966, p. 139–151.

³³ W. Porzig, Die Gliederung des indogermanischen Sprachgebiets, Heidelberg, 1954.

³⁴ W. Porzig, op. cit., p. 140.

34

innovations which, in his opinion, arose as a result of the long-standing proximity of the Baltic and Slavic languages. For this reason, V. Porzig, wishing to characterize more adequately the oldest relations between the Baltic and Slavic languages, compares them with the relations between other Indo-European linguistic groups. Having reviewed the relations between Baltic and Slavic, on the one hand, and Indo-Iranian, on the other, and established that the area of these languages is encircled by twenty-one isoglosses, V. Porzig drew the following conclusion: the relationship between Indo-Iranian and Baltic is almost equivalent to that between Indo-Iranian and Slavic; consequently, in the first half of the second millennium BCE, the connections between the Baltic and Slavic languages must have been close³⁵.

Antanas Salys’s view is also close to that of V. Porzig. A. Salys writes: “The obvious and universally recognized similarity between the two linguistic groups is readily explained by their descent from two neighboring Indo-European dialects, which maintained contact even during the final stage of the disintegration of Indo-European linguistic unity. The migration of the Proto-Balts (pre-Balts) northward around 2000 BCE ended their contacts with the Proto-Slavs (pre-Slavs), and for this reason the two linguistic groups began to develop independently. As a result of Slavic expansion, new contacts arose in the north around the sixth century CE and somewhat earlier in the south”³⁶.

Manus Loiman writes about the very close relations between the Baltic and Slavic languages. In his article “Baltic and Slavic Languages”³⁷, he first noted that the theory of Balto-Slavic linguistic unity (die baltoslavische Theorie) has become the victim of a fundamental dispute: the disagreement concerns the principles on the basis of which the common features of the Baltic and Slavic languages should be interpreted. The author then listed several common features of the Baltic and Slavic languages and devoted the greater part of the article to examining the relations between their verbal systems. M. Loiman argues that the Baltic -ā, -ē preterite is essentially similar to the Slavic thematic and sigmatic aorist. In his view, the verbal features cited are without exception old innovations that point to a close connection between the Baltic and Slavic languages.

On the basis of these innovations, M. Loiman believes that after the dissolution of Indo-European linguistic community, the Baltic and Slavic languages were linked by a fairly large bundle of isoglosses that distinguished these two language groups from other languages. Therefore, in his view, a Balto-Slavic proto-stage of the same type as Indo-Iranian is the only suitable interpretation of that bundle of isoglosses³⁸.

M. Loiman based the claim of Balto-Slavic linguistic unity on common features of the verbal systems, whereas J. Kurylowicz relied mainly on prosodic relations.

Jerzy Kurylowicz considers the following phenomena to be the principal arguments for the theory of Balto-Slavic linguistic unity³⁹: a) the replacement of the ablaut “full grade: zero grade”; b) two types of oppositions, eR: iR and eR: uR (where R denotes a sonor

³⁵ W. Porzig, op. cit., p. 169.

³⁶ A. Salys, Baltic Languages, „Encyclopaedia Britannica“, 3, 1964, p. 53.

³⁷ M. Leumann, Baltisch und Slavisch, „Corolla linguistica, Festschrift F. Sommer“, Wiesbaden, 1955, p. 154–162.

³⁸ M. Leumann, op. cit., p. 162.

³⁹ J. Kuryłowicz, Balto-Slavic linguistic unity, “Dictionary of Slavic Antiquities (Trial Issue),” Warszawa, 1934, p. 4–7; L’apophonie en indo-européen, Wrocław, 1956, p. 209–243, 286–308; L’accentuation des langues indo-européennes, Wrocław–Kraków, 2nd ed., 1958, p. 162–356; On Balto-Slavic linguistic unity, BPTJ, 16, 1957, p. 71–113; On Balto-Slavic linguistic unity, ВСЯ, 3, 1958, p. 15–49.

3*

35

the emergence of l, r, m, n); c) quantitative changes (the emergence of new long vowels); d) changes in stress and intonation.

J. Kurylowicz considers the intonations of the Baltic and Slavic languages a particularly important innovation, calling them an essential and most characteristic feature of these languages⁴⁰. In his view, the intonation systems of Greek and of the Baltic and Slavic languages developed independently of one another, and therefore cannot be considered inherited from the Indo-European linguistic community. Kurylowicz believes that intonations in Greek word-final syllables arose in historical times, whereas the Baltic and Slavic acute accent is the result of stress retraction onto a long syllable.

Following prosodic principles, J. Kurylowicz thus periodizes the history of the Baltic and Slavic languages:

1) the Balto-Slavic period (from the emergence of intonations);

2) the period of the Lithuanian language (from the operation of de Saussure’s law);

3) the Common Slavic period (from the weakening of the yers and the emergence of the neo-acute accent)⁴¹.

Like M. Loimann and J. Kurylowicz, Peter Arumaa also firmly adheres to the theory of Balto-Slavic linguistic unity. He seeks to support this theory by presenting new facts. In an article devoted to the analysis of derivatives containing the element b(h)⁴², P. Arumaa regretted that word-formation relations between the Baltic and Slavic languages had been studied very little, and that suffixes generally tend to harden, thereby providing fertile ground for linguistic differences to arise. Nevertheless, in his view, “remarkable correspondences are found at every turn” in the derivation of the Baltic and Slavic languages⁴³. P. Arumaa considers the lengthening of short vowels, especially i and u, and the frequent use of the long vowels ū and ī, as well as the intonation system, to be highly significant facts for resolving the question of Balto-Slavic linguistic unity⁴⁴.

In the first volume of his “Proto-Slavic Grammar,” published in 1964⁴⁵, P. Arumaa argues that the Baltic and Slavic languages originated in a fairly unified Indo-European linguistic area, since, compared with other Indo-European languages, they share a very large number of isoglosses. After the dissolution of the Indo-European linguistic community, the Baltic and Slavic languages formed a unity (sprachliche Einheit), which may have split into two principal dialects. In P. Arumaa’s opinion, the Baltic and Slavic languages may once have been as closely related as the Indo-Iranian languages. This thesis is supposedly difficult to substantiate because written records of the Baltic and Slavic languages appeared very late, one and a half to two millennia after the appearance of records in the Indo-Aryan and Iranian languages. Numerous later studies (P. Arumaa has in mind works published after the 1908–1912 debate) demonstrate the theory of Balto-Slavic unity as a proven fact⁴⁶. P. Arumaa goes on to note that there is no evidence on which to determine when Balto-Slavic unity existed. Nevertheless, he is inclined to believe that by the fourth century BCE Balto-Slavic unity (in P. Arumaa’s terms, the language) had already split.

⁴⁰ J. Kuryłowicz, L’accentuation des langues indo-européennes, p. 163.

⁴¹ J. Kuryłowicz, op. cit., p. 163.

⁴² P. Arumaa, Die Verwandtschaftsverhältnisse zwischen Baltisch und Slavisch, ZfsIPh, 24, 1955, p. 9–28.

⁴³ P. Arumaa, op. cit., p. 28.

⁴⁴ P. Arumaa, De l’unité balto-slave, „Scando-Slavica“, 9, 1963, p. 86.

⁴⁵ P. Arumaa, Urslavische Grammatik (Einführung in das vergleichende Studium der slavischen Sprachen), I. Band (Einleitung·Lautlehre), Heidelberg, 1964.

⁴⁶ P. Arumaa, op. cit., p. 19–20.

36

Osvaldas Semerenjis also believes that a period of a common Balto-Slavic language once existed (a period of common language). In his extensive article “The Problem of Balto-Slavic Unity—A Critical Survey”⁴⁷, having reviewed many common features of the Baltic and Slavic languages and established the archaic nature of some of them (for example, the suffix *-mo- of present passive participles, cf. Lithuanian neša-mas, Old Church Slavonic neso-mъ and Hittite ašima-, “loved”), he concluded that the following shared features should be considered innovations of the Baltic and Slavic languages: 1) palatalization; 2) the appearance of the vowels i and (after velar consonants) u before Indo-European l, m, n, r; 3) the change of Indo-European s to š after k, r, i and u; 4) accentual innovations; 5) definite adjectives; 6) the jo declension of participles; 7) the singular genitive ending *-ā(t) of o-stem nominal forms; 8) new suffixes for degrees of comparison of adjectives; 9) the stem *men- (*mun-) of the singular of the first-person pronoun and the plural genitive *nōsōm; 10) the pronouns *tos/*tā instead of Indo-European *so/*sā; 11) the present-tense paradigm *dōmi, *dōsi, *dōst-, *dōdṇ-; 12) the past-tense stems -ē and -ā; 13) verbs with Lithuanian -āuju = and Old Church Slavonic -uja; 14) common vocabulary⁴⁸.

In O. Semerenjis’s view, these most important innovations of the Baltic and Slavic languages cannot be explained by chance or parallel development. Thus, they show that the Balts and Slavs shared a common period of language and life⁴⁹.

However, it should be added that some of the common features of the Baltic and Slavic languages listed by O. Semerenjis apparently cannot be regarded as innovations. For example, the jo declension of participles is found in Gothic; Indo-European s after k, r, i, u also became š in the Indo-Iranian languages, and, moreover, there is evidence to suggest that this process occurred independently in each language group (see this work, pp. 81ff.); the Indo-European pronoun *so/*sā was also replaced in the West Germanic languages (cf. Old Saxon thē, thia and Old High German der, diu); and Italic languages also have past-tense stems with -ē and -ā (cf. Latin dice-bam, amā-bam)⁵⁰.

Consonant palatalization does not appear to be a very old phenomenon dating back to Proto-Baltic times. The opposition, for example, p – pʹ (cf. taupau – taupʹau), most likely arose as j disappeared on Lithuanian linguistic soil, since its phonetic realization in Latvian was different—Latvian p – pl and p – pj. The claim that consonant palatalization is a recent phenomenon is also supported by the different development of the Baltic cluster sj, cf. Lithuanian siūti, Latvian šūt and Prussian schuwikis (= Lithuanian siuvikis). Thus, it may be assumed that the Baltic linguistic community had intact consonant clusters with j.

Nor do the other common features of these languages listed by O. Semerenjis demonstrate Balto-Slavic linguistic unity. The emergence of the pronoun stem *mun- is far too uncertain⁵¹ for it to be considered an innovation of the Baltic and Slavic languages that arose during the period of Balto-Slavic linguistic uni-

⁴⁷ O. Szemerényi, The Problem of Balto-Slav Unity—A Critical Survey, „Kratylos“, 2, 1957, p. 97—123. Jo 1948 m. išspausdintas straipsnis „Sur l’unité linguistique balto-slave“ (Études slaves et roumaines, I, 1948) šio darbo autoriui liko neprieinamas.

⁴⁸ For a more detailed description of some of these shared features, see the second part of this work.

⁴⁹ O. Szemerényi, op. cit., p. 120.

⁵⁰ Būtojo laiko kamienai su -ē ir -ā gali būti archaizmai, žr. A. Vaillant, L’origine des présents thématiques en -e/o-, BSL, 38, 1937, p. 96.

⁵¹ The pronoun munie may have arisen phonetically in the Samogitian dialect, since here a (Baltic o) between m and n changes to u; cf. Samogitian nūmas, kāmunos, kamunė instead of standard Lithuanian namas, kamanos, kamanė,—K. Būga, RR, II, p. 282; P. Jonikas, Pagramančio tarmė, Kaunas, 1939, p. 59; J. Aleksandravičius, Kretingos tarmės įvardis, „Kalbotyra“, 10, 1964, p. 121; Z. Zinkevičius, Lietuvių dialektologija, Vilnius, 1966, p. 299.

37

during the period. o-stem nominal singular genitive inflection -ā indicates that close relations existed only between the eastern Baltic and Slavic languages, since Prussian appears to have preserved here the old ending -as (cf. Hittite antuhšāš ‘of a human’). A. Vaillant’s view⁵², supported by O. Semerenjo, that the Prussian ending -as arose from the old *-ā (Lith. -o, Lat. -a), to which -s was added by analogy with the nominative, is highly hypothetical and therefore can hardly be relied upon.

O. Semerenjo’s argumentation, like that of other proponents of the theory of Baltic-Slavic linguistic unity, is characterized by a tendency to seek identity in every fact or phenomenon of these languages, even when they differ. For instance, O. Semerenjo derives the Slavic suffix *-dlo-, found in nomina instrumenti, from *-tlo (rather than from Indo-European *-dhlo-, as all scholars do), most likely because it also occurs in the Baltic languages (Lith. -kla-s). Yet in Slavic the consonant t did not, after all, become d.

The Bulgarian linguist Vladimir Georgiev is also a proponent of the theory of Baltic-Slavic linguistic unity.

V. Georgiev⁵³ raised the question of a criterion by which the degree of affinity between language groups could be determined. He notes that such a criterion might be provided by Indo-Iranian linguistic unity, whose existence is beyond doubt and which can be inferred from the evidence of the oldest written monuments of those languages. In the absence of monuments of the ancient Indo-Iranian languages (the oldest passages of the Rigveda are dated to the end of the second millennium BCE, while the Avesta dates no later than the eighth century BCE), V. Georgiev argues, it would scarcely be possible to establish a close relationship between the two language groups and thus arrive at the conclusion that Indo-Iranian linguistic unity once existed, since the present-day Indic and Iranian languages have changed greatly. Written monuments of the Slavic languages, he continues, began to appear only in the ninth–tenth centuries, and those of Lithuanian only from the mid-sixteenth century; hence the Slavic-language evidence was recorded almost 2,000 years later, and the Lithuanian evidence 2,500 years later, than the Indo-Iranian evidence. Nevertheless, the affinity between the Baltic and Slavic languages is supposedly evident, whereas such affinity is much harder to discern between the Indic and Iranian languages of the tenth–fifteenth centuries. V. Georgiev is therefore convinced that “the Baltic and Slavic languages once formed a unity that lasted for a long period and probably split apart later than the Indo-Iranian unity”⁵⁴.

Believing the affinity between Baltic and Slavic morphology to be so close that Proto-Slavic morphology is almost identical to Lithuanian morphology, V. Georgiev regards this morphological affinity as the most important argument for Baltic-Slavic linguistic unity: identical inflectional morphemes (nominal and verbal inflection) cannot be borrowed (precisely: borrowed!) or arise through languages drawing closer—they must have been inherited from a linguistic unity that once existed⁵⁵.

V. Georgiev gives the following stages in the development of relations between the Baltic and Slavic languages: the period of Baltic-Slavic linguistic unity—the third millennium BCE; the transitional period—the end of the third millennium BCE and the beginning of the second millennium BCE;

⁵² A. Vaillant, BSL, 44, 1948, p. 129.

⁵³ В. Георгиев, Към въпроса за балто-славянската езикова общност, „Български език“, 8, 1958, р. 105—110; Балто-славянский, германский и индо-иранский, СФ, 1, 1958, р. 7—26; Исследования по сравнительно-историческому языкознанию, М., 1958, р. 218—238.

⁵⁴ В. Георгиев, Балто-славянский, германский и индо-иранский, р. 10.

⁵⁵ В. Георгиев, ор. cit. р. 11.

38

the Proto-Slavic period — the second and first millennia BCE.⁵⁶

As can be seen, V. Georgiev places the Balto-Slavic linguistic unity in very remote times, in the third millennium BCE (in his view, the Indo-European linguistic community existed until the fourth millennium BCE). On this occasion, it is worth recalling how other proponents of the unity theory date Balto-Slavic linguistic unity.

A. Vaillant believes⁵⁷ that Balto-Slavic linguistic unity disintegrated around the beginning of the Common Era, when Germanic tribes swept in and separated the Balts from the Slavs. J. Otrębski⁵⁸ holds a similar view. P. Arumaa conjectures⁵⁹ that Balto-Slavic unity had already split in the fourth century BCE. T. Lehr-Spławiński⁶⁰ dates the breakup of the Balto-Slavic linguistic unity, which he calls the Balto-Slavic linguistic community, to 1500–1300 BCE.

In the postwar period, the theory of Balto-Slavic linguistic unity has also attracted a considerable number of critical remarks.

Anglų slavistas Viljamas Metjusas (? — 1958), straipsnyje „Baltų ir slavų kalbų tarpusavio santykiai“⁶¹ trumpai apžvelgęs ir kritiškai įvertinęs baltų ir slavų kalbų seniausiųjų santykių tyrinėjimo istoriją, konstatavo, kad šių kalbų bendrybių pilniausio inventoriaus sudarymas dar nepadeda spręsti jų seniausiųjų santykių problemos: bendrybės, kaip kalbos faktai, yra izoliuotos ir todėl argumentais negali būti laikomos, nes kalba yra ne izoliuotų duomenų inventorius, bet sistema, kurioje tie duomenys yra susiję ir vienas nuo kito priklauso. V. Metjusas mano, kad baltų ir slavų kalbų seniausiųjų santykių klausimas turi būti tyrinėjamas kita kryptimi: abi kalbų grupes reikėtų ištirti kaip sistemas. Detalūs sinchroniniai, pavyzdžiui, lietuvių ir rusų kalbų struktūrų aprašai esą būtini teisingam tyrinėjimui ir, galimas daiktas, seniausiųjų santykių problemos sprendimui. Jie leistų tyrinėtojui sulyginti visas abiejų sistemų detales, ir tuo būdu būtų galima prieiti labiau pagrįstų išvadų.

Researchers into the earliest relations between the Baltic and Slavic languages generally base their theories on reconstructed facts; however, everyone knows that these reconstructions are in every case based on comparisons of data from languages attested at different times. For example, the written monuments of Old Indian are said to be many centuries older than the monuments of the Slavic languages, while the latter are about half a millennium older than the monuments of Lithuanian. Moreover, in comparisons with ancient languages, it is generally not forms from old texts but forms of modern Lithuanian that are cited. Comparing Old Indian with Hindi shows how greatly the Indian languages have changed. The Baltic and Slavic languages may likewise have changed, but we know nothing about those changes. The reconstructed prehistoric forms are said to be unreal, since there is no evidence that they ever existed.

V. Matthews notes that subjectivism plays an important role in the theory of Balto-Slavic linguistic unity. He categorically rejects this theory as antihistorical: since the Baltic

⁵⁶ B. Георгиев, op. cit., p. 14.

⁵⁷ A. Vaillant, Comparative Grammar of the Slavic Languages, I, pp. 13–14.

⁵⁸ J. Otrębski, Development of Mutual Relations between the Baltic and Slavic Language Groups, „Z polskich studiów slawistycznych“, Warsaw, 1958, p. 146.

⁵⁹ P. Arumaa, Urslavische Grammatik, I, p. 19.

⁶⁰ T. Lehr-Spławiński, The Balto-Slavic Linguistic Community and the Problem of the Ethnogenesis of the Slavs, „Slavia Antiqua“, 4, 1954, p. 21.

⁶¹ W. K. Matthews, The Interrelations of Baltic and Slavonic, “The Slavonic and East European Review”, 35, 1957, p. 409—427; V. K. Mathews, On the Relationship between Slavic and Baltic Languages, SF, I, 1958, p. 27—44.

39

and the problem of the oldest relations between the Slavic languages is diachronic; therefore, we can formulate it as a problem of historical dialectology, thereby also “neutralizing” its synchronicity.

The theory of Baltic-Slavic linguistic unity is also rejected by the Czech linguist Adolf Erhart.

He believes⁶² that the similarity of the morphological systems of the Baltic and Slavic languages was determined by the conservative character of both language groups. A. Erhart noted that the common features (Übereinstimmungen) of these languages suffice to prove only the contiguity of the Indo-European dialects from which the Baltic and Slavic languages later arose. In his view, after the breakup of Indo-European linguistic unity, the Baltic and Slavic languages could have developed independently.

However, A. Erhart points out, we do not have a single fact with which we could definitively refute the theory of Baltic-Slavic linguistic unity.

A. Erhart then briefly described the differences and common features of the Baltic and Slavic languages. He identified fourteen differences and twenty-five common features.

A. Erhart listed the following common features of the Baltic and Slavic languages: the development of Indo-European ḱ, gʰ into š, ž (Slavic s, z), as in Indo-Iranian, Armenian, and other Indo-European dialects; the loss of voiced aspirated consonants, as in other Indo-European dialects; the merger of Indo-European a and o, as in other Indo-European languages; the development of Indo-European eu into jau; the change of Indo-European s to š in certain positions, as in Indo-Iranian languages; the development of Indo-European R into iR and uR (this common feature has no evidentiary force, since it is not a typical phonetic phenomenon); the genitive singular ending of o-stem nouns (here A. Erhart is inclined to recognize Slavic influence on the Baltic languages); the instrumental singular ending of ā-stem nouns (Lith. -a < *-ān, OCS -ą, -oją; Skt. -ā, -ayā without m); the transition of consonant stems to i-stems, as in Latin; pronominal adjectives (with a pronoun’s Indo-European root; cf. Iranian and Germanic evidence); the present participle suffix *-jo-, as in West Germanic languages; the pronoun *mun- (cf. Hittite ammu); the formation of the dative singular of first-person pronouns on the basis of the genitive singular form (A. Erhart notes that this is a parallel development); the genitive plural of pronouns *nōssom, *vōssom; the use of the pronoun root *i/j-o- as a personal pronoun, as in Latin and Germanic languages; the nominative forms of demonstrative pronouns *tos, *tā (instead of Indo-European *sos, *sā); the dative (and locative) of pronouns with the ending m (cf. Lithuanian tāmui, tamē), as in West Germanic languages; the numeral “nine” (Eastern Baltic d- may have arisen under the influence of Slavic languages); the numerals Lithuanian ašmas, OCS osmь; the formation of distributive numerals (cf. Lithuanian penkeri, OCS petorъ); word formation, cf. the Balto-Slavic suffixes -ējas: -ējė; -ovas: -avъ; -lis, -lys: -lb; -uolis: alb; -ininkas: -bnikъ; -ytis: -ištъ; -tukas: -tъkъ; -tojas: -tajь; -yba: -ьba; preterite formation with the suffixes *-ē, *-ā, as in Latin; the infinitive suffix *-ti (A. Erhart compares it with Vedic -tāye) and the present passive participle suffix *-mo-, as in Armenian, Luwian (and, it should be added, Hittite) languages.

As can be seen, among the common features of the Baltic and Slavic languages A. Erhart included not only innovations in the Baltic and Slavic languages (the development of Indo-European eu into jau; the transition of Indo-European R into iR and uR; the genitive singular ending of o-stem nouns; the ā-stem nouns’

⁶² A. Erhart, On the Problem of Baltic-Slavic Linguistic Unity, “Sborník prací filosofické fakulty Brněnské University”, Volume VII, Series on Linguistics (A), No. 6, 1958, pp. 123–130.

40

the singular instrumental ending; the pronoun *mun-; derivational morphemes, etc.), but also those common phenomena that are found in other Indo-European languages.

A. Erhart pointed out the following differences between Baltic and Slavic languages:

1. The first palatalization of the Slavic languages is quite old, and it is entirely possible that it is connected with the Indo-Iranian palatalization. This would be an important argument against the theory of Baltic-Slavic linguistic unity.

2. The different fate of Indo-European ō and ā (see this work, pp. 61 ff., for more detail).

3. The declension of adjectives in the Baltic languages, as in Germanic, differs from that of nouns (pronominal singular dative endings, etc.).

4. The comparative suffix of adjectives -esnis resembles the corresponding Germanic suffix (cf. Gothic -izan).

5. Word formation. Slavic nomina agentis are formed with the suffix -telb, which Lithuanian does not have. Lithuanian suffixes -ėjas, -tojas have fairly rare counterparts in Slavic languages. Lithuanian diminutive suffixes are dominated by the element -l-, which is alien to Slavic languages, whereas the very common Slavic diminutive element -k- is used rarely in Lithuanian. The Baltic suffix for nomina instrumenti was *-tlo-, while the Slavic one was *-dhlo. Lithuanian has a highly productive adjective suffix -ingas, which is absent from Slavic languages.

6. The replacement of the primary indeclinable numerals “five”–“nine”: in the Baltic languages, the adjectivizing element jo was added to them, whereas in Slavic, numeral abstracts are formed with the suffix *-ti-.

7. The way Lithuanian numerals “eleven”–“nineteen” are formed has parallels only in the Germanic languages.

8. Some present-tense formation types (nasal infix, sta stem) have no analogues at all in Slavic languages. The *ne/no verbs widespread in Slavic languages are almost unattested in Baltic languages. Baltic intransitive verbs have the suffix -i-, while Slavic ones have -i- (<ei?).

9. There are no traces of the sigmatic aorist in Baltic languages.

10. The Baltic languages’ first-person singular has retained the Indo-European inflection *-ō. The origin of the Slavic ending -ą is unclear.

11. The third-person plural in Baltic languages has no ending.

12. Baltic languages have no participles with the suffixes -l- and -n-.

13. The category of verbal aspect is less well preserved in Baltic languages.

14. Vocabulary.

Some of the differences between the Baltic and Slavic languages identified by A. Erhart may be very old, for example, the mismatch in numerals, the sigmatic aorist, and the third-person plural. These differences affect cardinal parts of the Baltic and Slavic language systems, showing that in the past the Baltic and Slavic languages were not identical. Others (the first palatalization of the Slavic languages, the different fate of Indo-European vowels ō and ā, sta-stem verbs, the first-person singular inflection, etc.) are likely to have arisen later. Consequently, they do not reflect the relations between the Baltic and Slavic languages after the breakup of Indo-European linguistic unity. In this way, A. Erhart pointed out major and ancient differences between the Baltic and Slavic languages.

41

One of the most consistent advocates of the convergence theory, Samuel Bernstein, also raised the question of why the theory of Baltic-Slavic linguistic unity cannot be accepted⁶⁴. In his view:

1. Many of the oldest processes in the Baltic and Slavic languages were different (for example, Indo-European o and a became a in the Baltic languages and o in Slavic; the Baltic languages preserved the distinction between Indo-European ē and ā, whereas the Slavic languages lost it; for more on this, see pp. 52 ff. of this work).

2. Many of the most important processes of Proto-Slavic, which took place in its early period, were wholly or partly unknown to the Baltic languages (for example, Indo-European s became ch in the Slavic languages after k, r, i, u, and became š in Lithuanian after k, r; for more on this, see pp. 81 ff. of this work).

3. The theory of Baltic-Slavic unity denies the ancient dichotomy of Proto-Baltic.

4. Advocates of the theory of Baltic-Slavic unity cannot reconstruct a Proto-Baltic-Slavic language; this is said to be obvious proof that the theory is unfounded.

5. The theory of Baltic-Slavic unity is not supported by archaeological evidence. Archaeologists, for example J. Kostrzewski, find no culture that could be considered the equivalent of Baltic-Slavic linguistic unity.

On this basis, S. Bernstein believes that neither Baltic-Slavic linguistic unity (a proto-language) nor a Baltic-Slavic proto-people ever existed.

S. Bernstein further noted that the fact that the Baltic and Slavic languages are closely related and share many features is universally acknowledged. This is not what scholars dispute. The usual subject of debate is the nature and origin of the common features of the Baltic and Slavic languages. S. Bernstein believes that the Baltic and Slavic languages are so closely related not because they descended from Indo-European linguistic unity, but because of the long-term contacts established between them, which began not immediately after the breakup of Indo-European linguistic unity and ended well before the Common Era. In other words, the closeness of the Baltic and Slavic languages and their common features arose from the linguistic community of the Baltic and Slavic languages (сообщность, Verkehrsgemeinschaft)⁶⁵. The assumption of long-term contacts is sufficient to explain the origin of the common features of the Baltic and Slavic languages.

Boris Gornung, too, rejects Baltic-Slavic linguistic unity and its dichotomous split. In his view, as Indo-European linguistic unity broke up, the integration of different dialect types (intensification of contact, interaction and merging of dialects) not only regularly accompanied processes of differentiation (the splitting of linguistic unity), but often predominated; as a result, differentiation itself was sometimes a consequence of integrative processes⁶⁶.

In B. Gornung’s view, the “northwestern” zone of Indo-European linguistic unity split at the end of the third millennium BCE and the beginning of the second millennium BCE into a “western” zone (Proto-Celtic-Illyrian-Italic-Germanic) and a “northeastern” zone (Proto-Baltic-Slavic-Tocharian). When the Proto-Tocharians migrated beyond the Volga, the remaining “northeastern” zone of Indo-European linguistic unity should be regarded as the common source from whose dialectal diversity the Proto-Baltic and Proto-Slavic linguistic communities took shape.

⁶⁴ S. B. Bernstein. Outline of the Comparative Grammar of the Slavic Languages (Prospectus), KSISl, 18, 1956, pp. 3–26; “Baltic-Slavic Linguistic Community,” SF, I, 1958, pp. 45–68; Outline of the Comparative Grammar of the Slavic Languages, M., 1961, pp. 27–37; see also VYa, I, 1958, pp. 48–49; “Collection of Answers to Questions on Linguistics (For the IV International Congress of Slavists),” M., 1958, pp. 167–169.

⁶⁵ S. B. Bernstein, “Baltic-Slavic Linguistic Community,” p. 45.

⁶⁶ B. V. Gornung, “On the Discussion of Baltic-Slavic Linguistic and Ethnic Unity,” VYa, 4, 1958, pp. 55–62; From the Prehistory of the Formation of Common Slavic Linguistic Unity, M., 1963.

42

The most questionable point in B. Gornung’s theory is the claim that the Proto-Prussian dialects had from ancient times belonged not to the Proto-Baltic, but to the Proto-Slavic linguistic community. The Baltic character of the Prussian language is beyond doubt. It does not seem that this Baltic character was any less pronounced several millennia ago than it is now. The Prussian-Slavic isoglosses⁶⁷ are clearly insufficient to support such a conclusion.

Around the second quarter of the second millennium BCE, the Proto-Slavic dialects—which, as already mentioned, included the Proto-Prussian dialects—and the Proto-Baltic dialects began to draw closer: the “Baltic-Slavic epoch”⁶⁸ was beginning (B. Gornung uses J. Endzelin’s term). However, in B. Gornung’s view, it did not last long, since around the middle of the second millennium BCE the Proto-Prussian dialects joined the Proto-Baltic dialects (that is, according to B. Gornung, the Proto-Lithuanian-Latvian dialects), and in this way the Baltic linguistic community was formed⁶⁹. B. Gornung believes that Baltic and Slavic contacts then ceased for a long time.

B. Gornung’s fundamental approach to the most ancient relations between the Baltic and Slavic languages recalls the conceptions of J. Endzelin, Chr. Stang, E. Fraenkel, and S. Bernštein.

B. Gornung’s idea that the genetic relations (or commonalities) between the Baltic and Slavic languages may date from chronologically distinct periods is especially noteworthy. B. Gornung derives them from five periods⁷⁰: 1) the period of the Globular Amphora culture; 2) the period of the expansion of the Battle Axe and Corded Ware cultures; 3) the period of contact between the Proto-Baltic and Proto-Slavic groups and the bearers of the Boat-Axe culture; 4) the period of contact with the Proto-Indo-Iranians, during which the satam isogloss was formed; and 5) only finally, the “Baltic-Slavic epoch.”

Projecting Baltic and Slavic linguistic commonalities onto distinct chronological planes is methodologically very important, since proponents of the theory of Baltic-Slavic linguistic unity usually disregard chronology: they tend to project all commonalities between the Baltic and Slavic languages onto a single chronological plane. This creates the impression that there are indeed many commonalities between the Baltic and Slavic languages and that these—most importantly—derive from a single period. In that case, the idea of Baltic-Slavic linguistic unity follows naturally. Analysis of the commonalities between the Baltic and Slavic languages shows that they are, nevertheless, most likely products of different chronological periods.

The review presented shows that in the postwar years, as before, linguists from various countries who took part in discussions of the most ancient relations between the Baltic and Slavic languages did not reach a consensus. Some scholars—T. Ler-Splavinskis, A. Vaillant, J. Otrębski, M. Loiman, J. Kurilowicz, P. Arumaa, O. Semerenyi, V. Georgiyev—adhere to the theory of Baltic-Slavic linguistic unity. Others—J. Safarewicz, E. Fraenkel, A. Zenas, A. Salys, V. Porzig, V. Matthews, A. Erhart, S. Bernštein, B. Gornung—reject this theory and tend to explain the affinity and commonalities of the Baltic and Slavic languages in terms of their community (J. Safarewicz), long-standing neighborhood (E. Fraenkel), close relations (V. Porzig), long-term contacts (S. Bernštein), or the neighborhood of the Indo-European dialects from which they derived (A. Erhart).

⁶⁷ J. Endzelīns, Par prūšu valodas radniecības sakariem, FBR, 11, 1931, p. 189—193; S. Stang, Einige Bemerkungen über das Verhältnis zwischen den slavischen und baltischen Sprachen, NTS, 11, 1939, p. 85—103.

⁶⁸ Б. В. Горнунг, Из предыстории образования общеславянского языкового единства, р. 96.

⁶⁹ Б. В. Горнунг, op. cit., p. 106, išnaša 9.

⁷⁰ Б. В. Горнунг, op. cit., p. 97.

43

The Question of the Earliest Relations between the Baltic and Slavic Languages and the Fourth and Fifth International Congresses of Slavists

Research into the earliest relations between the Baltic and Slavic languages was considerably stimulated by the Fourth and Fifth International Congresses of Slavists. The Fourth Congress of Slavists, held in Moscow from September 1 to 10, 1958, devoted particularly great attention to the question of the genetic relations between the Baltic and Slavic languages.

In preparation for the congress, the Soviet Committee of Slavists announced thirty questions for the congress to consider, covering the principal problems of Slavic linguistics. The twentieth question was: “Did a Baltic-Slavic linguistic and ethnic unity exist, and how should it be understood?”¹ The answers to these questions were published in a separate volume².

Four papers on the question of the earliest relations between the Baltic and Slavic languages were read at the congress³. These papers, and the problem of the earliest relations between the Baltic and Slavic languages in general, were vigorously debated during the congress⁴.

K. Janaček, V. Kiparsky, V. Ernits, E. Dikenman, L. Bulakhovsky, I. Lekov, and others answered the question: “Did a Baltic-Slavic linguistic and ethnic unity exist, and how should it be understood?”

K. Janaček noted⁵ that two major obstacles prevent an affirmative answer to this question: 1) monuments of the Baltic and Slavic languages appeared very late; 2) we do not know which features of the individual Indo-European languages may be considered features of the Indo-European protolanguage, nor how the Baltic-Slavic period should be distinguished from the period of the Indo-European protolanguage.

V. Kiparsky expressed the conviction that a Baltic-Slavic linguistic unity had existed⁶. In his view, the numerous lexical innovations in the so-called basic vocabulary could not otherwise be explained. V. Kiparsky tends to associate the Baltic-Slavic ethnic complex with Herodotus’s Neuri.

Like V. Kiparsky, V. Ernits recognizes Baltic-Slavic unity. He indicated⁷ that, after separating from the other representatives of the satem group, the Balto-Slavs retained their unity for a certain period of time. Later, as the population grew, the territory expanded, and dialectal features emerged, linguistic and territorial distinctions between the Baltic and Slavic languages began to become apparent

¹ ВЯ, 1, 1958, p. 36.

² Collection of Answers to Questions on Linguistics (For the Fourth International Congress of Slavists), Moscow, 1958.

³ R. Trost’s “The Current State of the Question of Baltic-Slavic Linguistic and Ethnic Relations,” V. Ivanov and V. Toporov’s “How to Pose the Question of the Earliest Relations between the Baltic and Slavic Languages,” J. Otrębski’s “The Development of Relations between the Baltic and Slavic Language Groups,” and V. Mažiulis’s “On the Earliest Relations between the Baltic and Slavic Languages”; see IV Международный съезд славистов (Материалы дискуссий), II, Moscow, 1962, p. 422—426.

⁴ In addition to the speakers, the following spoke about the earliest relations between the Baltic and Slavic languages: T. Ler-Splavinsky (Poland), M. Altbauer (Israel), K. O. Falk (Sweden), V. Georgiev (Bulgaria), J. Czekanowski (Poland), V. Illich-Svitych (USSR), M. Guchman (USSR), M. Vasmer (Federal Republic of Germany), V. Machek (Czechoslovakia), E. Seidel (German Democratic Republic), A. Vrăciu (Romania), B. Gornung (USSR), B. Larin (USSR), P. Ariste (USSR); see IV Международный съезд славистов (Материалы дискуссий), II, p. 429—450.

⁵ Collection of Answers to Questions on Linguistics (For the Fourth International Congress of Slavists), p. 162—166.

⁶ Op. cit., p. 170—171.

⁷ Op. cit., p. 173—174.

44

differences. But contacts continued to be maintained. Therefore, linguistic changes that arose in one group could easily have spread to the other as well. All this, in V. Ernits’s opinion, must have taken place before the contacts between the Balts and the Baltic Finnic peoples (around 2000 BCE), since these contacts were no longer reflected in the Slavic languages.

E. Dickenmann noted⁸ that there is insufficient evidence to answer affirmatively the question of whether there was ethnic unity between the Balts and Slavs, whereas their linguistic unity is beyond any doubt.

L. Bulachovskis (1888–1961), vienas stambiausių slavų akcentologijos specialistų, pabrėžė⁹, kad baltų kalbų kirtis iki šiol nepraradęs savo senosios morfologinės funkcijos (šokinėjimo paradigmoje). Tačiau to negalima esą pasakyti apie slavų kalbas. Svarbiausias baltų ir slavų kalbų ypatingo artimumo įrodymas — de Sosiūro-Fortunatovo dėsnis — gali būti paralelinės abiejų kalbų grupių raidos padarinys. Hirto dėsnio veikimas baltų ir slavų kalbose apskritai kelias abejonių. Ne įtikinamesni yra ir kiti šių kalbų bendri akcentologiniai reiškiniai (pavyzdžiui, metatonija supyno formose ir pan.).

The difference between the Baltic and Slavic languages is supposedly so obvious that it requires no proof. These are “closed” linguistic groups, with no transitional links between them. Nor does it seem that the state of relations between the Baltic and Slavic languages was fundamentally different even a thousand years ago. It may be assumed that the ancient connections between the Baltic and Slavic languages were eventually severed. This, however, did not prevent the Baltic and Slavic languages from not only retaining some traces of their former closeness, but also from independently developing new common features.

L. Bulakhovsky believes that most of the common features of the Baltic and Slavic languages arose when these languages still belonged to the satem dialect of the Indo-European linguistic community, within which Baltic-Slavic linguistic unity could have existed.

J. Lekov¹⁰ also disagreed with the theory of Baltic-Slavic linguistic unity. In his view, the features common to these languages consist of: 1) the legacy of the Indo-European linguistic community; 2) later phenomena arising from parallel development; 3) borrowings from various chronological periods. I. Lekov drew attention to the fact that terminological confusion makes it difficult to resolve the question of the earliest relations between the Baltic and Slavic languages, since researchers use a wide variety of terms—community, unity, common epoch or period, proto-language, parallel development, and so on.

As in the aforementioned replies, the papers presented at the congress likewise failed to reach a consensus on the earliest relations between the Baltic and Slavic languages.

In his paper, P. Trost¹¹ emphasized that the principal argument capable of proving Baltic-Slavic linguistic unity is specific morphological correspondences, namely, minor morphological features similar in both language groups. In his view, the numerous similarities between the two languages can be explained in two ways: by genetic relationships or by a linguistic union. However, the similarity of minor morphological features in the Baltic and Slavic languages points more to genetic relationships than to a linguistic union.

The demand made earlier by some researchers (for example, V. Matthews) that new approaches be sought for resolving the problem of the earliest relations between the Baltic and Slavic languages has already been mentioned

⁸ Op. cit., p. 175.

⁹ Op. cit., p. 157–162.

¹⁰ Op. cit., p. 172–173.

¹¹ P. Trost, K otázce baltoslovanských jazykových vztahů, „Československé přednášky pro IV. Mezinárodní sjezd slavistů v Moskvě“, Praha, 1958, p. 221–227.

45

in this way, using new methods, since compiling an inventory of the commonalities and differences between these languages is unlikely to help solve the problem.

From a methodological point of view, the paper by V. Ivanov and V. Toporov¹² was valuable: it approached the question of the earliest relations between the Baltic and Slavic languages from an entirely new perspective. The authors pointed out that the question of whether or not there was a Balto-Slavic protolanguage cannot, they argued, be posed. This is because research in recent decades indicates that the earliest state of the Baltic and Slavic, Indo-Iranian, and also Italic and Celtic language groups should be understood not as a traditional protolanguage, but as a spatiotemporal continuum of dialects. For this reason, V. Ivanov and V. Toporov pose the question of the earliest relations between the Baltic and Slavic languages as follows: what is the relationship between the earliest structural models of the Baltic and Slavic languages?

Having established the earliest phonological and morphological models for the Baltic and Slavic languages separately by means of internal reconstruction, and compared them, V. Ivanov and V. Toporov concluded that the earliest structural models of the Baltic and Slavic languages coincide and that they differ fundamentally from the Indo-European model. Moreover, the models of the Baltic and Slavic languages cannot, they argue, be placed on the same chronological plane: the Slavic model arose later and is the result of a change in the structural model of the Baltic languages¹³. V. Toporov characterized the essence of this solution to the problem of the genetic relationship between the Baltic and Slavic languages as follows: the Slavic protolanguage formed from peripheral dialects of the Baltic type¹⁴.

V. Ivanov and V. Toporov’s hypothesis accords well with the spatiotemporal regularity of linguistic facts observable in the actual life of languages: languages with little dialectal differentiation that occupy a large territory arose later and spread across that territory later. The Slavic languages are precisely of this kind. The Baltic languages, by contrast, are highly fragmented into dialects despite occupying a small territory. The relationship between the Slavic and Baltic languages has a good parallel in English, which has spread across various continents, and the highly dialectally differentiated West Germanic languages in Europe.

In his paper¹⁵, J. Otrebski emphasized that the Balts and Slavs once spoke a single language. In J. Otrebski’s view, the Balto-Slavic linguistic unity (łączność bałtycko-słowiańska), inherited from the Indo-European linguistic community, broke up very late, around the beginning of our era. The cause of the breakup was that some of the Balts, or Balto-Slavs, began speaking Finno-Ugric languages and became Finno-Ugric peoples.

V. Mažiulis¹⁶ focused mainly on the origin of the present stems Lithuanian duomi, demi, Old Church Slavonic damь. Drawing on Hittite data, the author derives these athematic forms from non-reduplicated Proto-Indo-European forms. The element -d- of the third person plural-

¹² В. В. Иванов и В. Н. Топоров, К постановке вопроса о древнейших отношениях балтийских и славянских языков, М., 1958.

¹³ В. В. Иванов и В. Н. Топоров, op. cit., p. 39.

¹⁴ В. Н. Топоров, К проблеме балто-славянских языковых отношений, КСИСл., 33—34, 1961, p. 213.

¹⁵ J. Otrębski, Rozwój wzajemnych stosunków między grupą językową bałtycką a słowiańską, p. 146—148.

¹⁶ В. Мажюлис, Заметки к вопросу о древнейших отношениях балтийских и славянских языков, Вильнюс, 1958. For the Lithuanian version of the paper, see Pastabos baltų ir slavų kalbų seniausių santykių klausimu, LKK, 2, 1959, p. 5—16.

46

could have arisen in that person, in V. Mažiulis’s view, as early as a result of the already formed Baltic and Slavic conjugational systems.

As language changes, numerals are often among the most stable elements. Therefore, they are of great importance in studying the relationships between languages. In this respect, the conclusions reached by V. Mažiulis in his study of the Baltic numeral system are of interest¹⁷: “In general, the derivational and morphological systems of Baltic numerals are in many respects similar to the Slavic ones. However, it should be noted that the oldest Baltic numerals, i.e. the common Indo-European ones, are direct relics of the Indo-European proto-language, whereas the later numerals, i.e. those not common to Indo-European, are correspondingly similar to numerals in Slavic, as well as Germanic and some other Indo-European languages”¹⁸. V. Mažiulis’s research shows that most of the common features of Baltic and Slavic numerals were inherited from the Indo-European linguistic community.

At the proceedings of the IV International Congress of Slavists, devoted to the problem of the oldest relations between the Baltic and Slavic languages, and in discussions of that question, two theories were represented—the theory of Baltic-Slavic linguistic unity and the theory of the convergent development of the Baltic and Slavic languages. As before, proponents of the theory of Baltic-Slavic linguistic unity emphasized innovations in the Baltic and Slavic languages, while representatives of the theory of the convergent development of the Baltic and Slavic languages, without denying these innovations, also stressed their no less numerous archaic differences. It goes without saying that, on the basis of commonalities (innovations) and differences whose chronology is unknown, it is hardly possible to resolve the problem of the oldest relations between the Baltic and Slavic languages. The congress discussion showed that resolving this problem requires the use of a range of linguistic methods, an important place among which should be given to the relations of the Baltic and Slavic languages with other languages and language groups; borrowings by the Baltic and Slavic languages from, and into, other languages; ethnonymy and toponymy; landscape terms (biological and geographical); and so forth. Solving these extralinguistic problems together with a deeper structural analysis of Baltic and Slavic linguistic phenomena will make it possible to reach conclusions of a geographical and chronological nature and to say whether Baltic and Slavic dialects existed together or separately¹⁹.

It is natural that, in the development of scholarship, a period of ebb follows feverish and intensive research. Such an ebb did in fact begin after the IV International Congress of Slavists.

For instance, apart from the outline of T. Ler-Splavinski’s sketch of Slavic ethnogenesis, where the genetic relations between the Baltic and Slavic languages were also briefly touched upon, not a single paper on the problem of the oldest relations between these languages was presented at the V International Congress of Slavists, held in Sofia on September 17–23, 1963²⁰. True, here too there was

¹⁷ V. Mažiulis, Lithuanian Numerals and the Relationship of Baltic Numerals to Numerals in Other Indo-European Languages (manuscript), M., 1955; The Indo-European Decimal System of Numerals, VY, 4, 1956, pp. 53–59.

¹⁸ V. Mažiulis, Remarks on the Question of the Oldest Relations between the Baltic and Slavic Languages, p. 14.

¹⁹ V. M. Illich-Svitych, The Balto-Slavic Problem at the IV International Congress of Slavists, VY, 1, 1959, pp. 139–141.

²⁰ Slavic Philology, VI (proceedings), Sofia, 1965.

47

The discussion concerned the relations between the Baltic and Slavic languages, but only their later, historical-period relations. Papers by V. Falkenhan, J. Otrębski, T. Zdancewicz, and E. Smulkova were devoted to these questions²¹.

The Problem of the Earliest Relations between the Baltic and Slavic Languages in Recent Years

It has already been mentioned that, after the Fourth International Congress of Slavists, interest in the problem of the earliest relations between the Baltic and Slavic languages declined. The reason for this may be the complexity of the problem, the lack of chronologically differentiated comparative phenomena, and the absence of objective criteria for determining the degree of kinship between languages.

Nevertheless, individual articles or book chapters that outline the earliest relations between the Baltic and Slavic languages continue to appear.

A r i t o n a s V r a č i u s believes¹ that the Baltic and Slavic languages derive from a unified Indo-European dialectal area. J a n i s L o j a holds that “the Baltic and Slavic languages formed independently within the Indo-European linguistic community, but, owing to long-lasting and fairly close Baltic-Slavic contacts, one or another shared phenomenon later emerged in their languages”². In his view, there can be no question of Baltic-Slavic linguistic unity, all the more so since the great differences between the Baltic languages themselves (Prussian on the one hand, and Lithuanian-Latvian on the other) do not allow J. Endzelin to recognize a Proto-Baltic language with certainty. J. Loja regards such shared phenomena in the Baltic and Slavic languages as the influence of the io-stem on nt-stem participles, the plural dative and instrumental endings with -m- (cf. corresponding Germanic-language facts), the shift of consonant stems to i-stems (cf. corresponding Latin-language facts), and pronominal adjectives as innovations that arose independently in the Baltic and Slavic languages. Fedot Filin³ is also categorically opposed to the theory of Baltic-Slavic linguistic unity. In his view, this theory has no firm scientific basis: the arguments of its proponents are characterized by apriorism, antihistoricism, and tendentiousness in the selection and interpretation of facts. As an example of such antihistoricism and disregard for the chronology of phenomena, F. Filin cites V. Georgiev’s claim that Lithuanian, which is characterized by conservatism in phonetics and morphology, can almost serve as a substitute for the unattested Slavic proto-language. F. Filin also disagrees with the opposite view, that the Baltic and Slavic languages separated independently from the Indo-European linguistic community and that there was never any closer contact between them. He believes that a group of dialects formed within the Indo-European linguistic community itself, and that their speakers had fairly close and long-lasting contacts. Related dialects, not isolated from other Indo-European…

²¹ V. Falkenhan’s “Lithuanian-Polish Linguistic Union,” J. Otrębski’s “The Slavicization of Lithuanian Hydronyms and Toponyms,” T. Zdancewicz’s “The Influence of the Lithuanian and Russian Vocabulary in the Białystok Region,” and E. Smulkova’s “On Belarusian g ega „tитnagас“, see J. Marvan, Vzajemné vlivy mezi slovanskými a baltskými jazyky (V. Mezinárodní sjezd slavistů v Sofii), „Slavia“ 33, 1964, p. 252–256.

¹ A. V r a c i u, Problema comunității lingvistice balto-slave, „Romanoslavica“, 4, 1960, p. 87–106; see review LKK, 6, 1963, p. 325–330.

² J. Loja, Valodniecības vēsture, Rīga, 1961, p. 151–154.

³ Ф. М. Филин, Образование языка восточных славян, М.–Л., 1962, p. 123–130; žr. recenziją LKK, 7, 1964, p. 240–243.

48

dialects, became enriched with distinctive innovations, and underwent many shared changes. The ancestors of Proto-Baltic and Proto-Slavic were among this group of Indo-European dialects. As the Indo-European linguistic community broke up, Proto-Baltic and Proto-Slavic separated; their systems, though not identical, remained similar down to our own day. From this, F. Filin concludes that contacts between the Baltic and Slavic languages existed, with only brief interruptions, throughout the history of those languages.

An important contribution to the study of accentual relations between the Baltic and Slavic languages is Vladislav Illich-Svitych’s book on the history of nominal paradigms.⁴ The nominal paradigm systems of these languages are similar. V. Illich-Svitych’s findings suggest that the accentuation systems of Baltic and Slavic nouns continue the Indo-European accentuation system. Lithuanian nouns with a baritone paradigm (except for the group with a long, non-apophonic vowel in the root) correspond to nouns of the same paradigm in Greek, Old Indian, and Germanic. Likewise, Lithuanian nouns with a mobile-accent paradigm correspond to nouns of the oxytone paradigm in those languages. The Indo-European opposition “baritone accentuation paradigm: mobile-accent or oxytone paradigm” has essentially been preserved in the Slavic languages as well.⁵

Thus, the accentuation of nouns in the Baltic and Slavic languages is very similar, and was even more similar in the past. As V. Illich-Svitych’s study has shown, the similarity in nominal accentuation between the Baltic and Slavic languages is an archaism inherited from the period of the Indo-European linguistic community, and therefore does not indicate that Baltic-Slavic linguistic unity existed after that community broke up.

In a recent work devoted to questions of Baltic ethnogenesis, Vytautas Mažiulis⁶ uses the term “earliest contacts” for contacts between related dialects (idioms) characterized by the following: 1) the tribes maintaining them understood one another fairly well; 2) the territory in which the earliest contacts existed must have been more or less limited, depending on population density, since compact linguistic contacts over a large territory would sooner or later have ceased. On the basis of the boat-axe culture attributed to the Baltic tribes, which spread across large areas of northeastern Europe at the beginning of the second millennium BC, V. Mažiulis is inclined to think that the disintegration of the earliest Baltic contacts (in other words, of the Baltic linguistic community) began precisely at that time. And since, in his view, the earliest Baltic contacts and the earliest Baltic-Slavic contacts began to disintegrate simultaneously, he dates the beginning of the disintegration of the earliest Baltic-Slavic contacts to the first centuries of the second millennium BC. V. Mažiulis believes that, moving further back in time from the beginning of the second millennium BC, contacts between the Baltic and Slavic languages must have become increasingly close. He concludes that a Slavic proto-language existed, but a Baltic one did not; the Baltic proto-language (Ursprache) existed within the framework of Baltic-Slavic linguistic unity (within the framework of the Balto-Slavonic „Ursprache“).⁷

The book by Anatolij Nepokupno devoted chiefly to present-day areal relations between the Baltic and Slavic languages is entitled “Areal Relations between the Baltic and Slavic Languages

⁴ В. М. Иллич-Свитыч, Именная акцентуация в балтийском и славянском (Судьба акцентуационных парадигм), М., 1963; see review LKK, 7, 1964, p. 231–236.

⁵ В. М. Иллич-Свитыч, op. cit., p. 83.

⁶ V. Mažiulis, Linguistic Notes on Baltic Ethnogenesis, Moscow, 1964.

⁷ V. Mažiulis, op. cit., p. 3.

4. Baltų ir slavų kalbų ryšiai

49

“aspects”⁸. However, both in its methodology and in the geographical distribution of the semantic and syntactic phenomena examined, this book may also prove useful for research into the earliest relations between the Baltic and Slavic languages.

Georg Y. Shevelov, who addressed the earliest relations between the Baltic and Slavic languages, points out⁹ that these languages have many striking commonalities, but no fewer striking differences. It is these differences above all that make it difficult to determine the date of Baltic-Slavic unity, if such unity ever existed at all.

In G. Shevelov’s view, the comparison of sound changes provides much important information, especially when attention is paid to their chronology. G. Shevelov considers it a simple fact that “of the five periods in the history of Proto-Slavic, the phonetic and phonological changes of the first period (approximately 2000–1500 B.C.) were common to the Baltic languages as well, whereas those of the later periods were not”¹⁰. During this period, the Baltic and Slavic languages underwent the following changes: the Baltic languages, like the Slavic languages, lost aspirated plosives and labiovelar consonants, simplified consonant and spirant clusters, changed syllabic sonants R into iR and uR, acquired two intonations of long syllables—the acute and the circumflex—and lost final r. Some of the changes listed here also occurred in other languages, but the twofold vocalization of syllabic sonants and the emergence of phonemic intonations were supposedly phenomena specific to the Baltic and Slavic languages. It can therefore be asserted with confidence that during the first period of Proto-Slavic, the development of the Baltic and Slavic languages was shared.

In the later periods of Proto-Slavic development, the Baltic languages were not completely isolated from the Slavic languages, but the connection was very tenuous. The fact that relations between the Baltic and Slavic languages had greatly weakened is shown by two important Slavic sound changes that were only partly reflected in the Baltic languages: in the Baltic languages, s after k, r, i, u did not change to ch, as it did in the Slavic languages; in the Baltic languages, s changed only after k, r (see this work, pp. 81 ff.). The second characteristic Slavic change—the merger of Indo-European o and a—affected only the short vowels of the Baltic languages.

Thus, on the basis of phonetic changes, we can, in G. Shevelov’s view, establish four periods in the development of the kinship relations between the Baltic and Slavic languages: 1) a period of shared development, approximately 2000–1500 B.C.; 2) a period of weakened contacts, whose final centuries were the sixth and fifth centuries B.C.; 3) a period of independent development (period of no established contacts), approximately from the fifth century B.C. to the sixth or seventh century A.D.; 4) a period of contact between the East Slavs and Latvians (in the Pskov region).

It should be said that G. Shevelov’s picture of the earliest relations between the Baltic and Slavic languages is relatively objective. There is no doubt that the vocalization of syllabic sonants and the formation of accentual oppositions are innovations of the Baltic and Slavic languages. They suggest that in the first centuries after the breakup of the Indo-European linguistic community, the Baltic and Slavic languages developed together in the sense that a wave of innovation originating in the Slavic-language area could spread unhindered across the entire Baltic-language area or a part of it,

⁸ А. П. Непокупный, Ареальные аспекты балто-славянских языковых отношений, Киев, 1964.

⁹ G. Y. Shevelov, A Prehistory of Slavic, Heidelberg, 1964, p. 613—614.

¹⁰ G. Y. Shevelov, op. cit., p. 613.

50

and vice versa. In other words, they suggest that the Baltic and Slavic languages formed a “living isogloss territory” (Chr. Stang’s term).

After analyzing the terminology of Slavic crafts (weaving, woodworking, pottery, and smithing) and comparing it with the corresponding terminology of the Baltic languages, Oleg Trubachev¹¹ concluded that: 1) the lexicon in this field differs substantially between the Baltic and Slavic languages¹²; 2) correspondences between the Baltic and Slavic languages occur in later, newly formed layers of the lexicon, which suggests that the affinity between the Baltic and Slavic languages in this field is secondary¹³. Summarizing the results of his research, O. Trubachev writes: “…the material itself indicates that the ancient Slavs oriented themselves not toward contacts with the Balts, but toward contacts with Indo-Europeans living farther to the west, whose interaction with the Slavs in the creation of terminology was so extensive and so significant that we are compelled to acknowledge the former existence of a Central European cultural area encompassing the dialects (or some of them) of the ancient Germanic, Italic, and Slavic peoples, but not affecting the Baltic dialects, with which contact may have begun later”¹⁴.

* * *

Using the diagrams of M. Weingart and N. van Wijk¹⁵, all the theories of the earliest relations between the Baltic and Slavic languages proposed to date can be represented graphically as follows:

To determine which of the existing theories should be preferred, and which should be considered the most reliable and the most objective reflection of the earliest relations between the Baltic and Slavic languages, it is necessary to review the most important commonalities between these languages, especially those regarded as the principal arguments for the theory of Baltic-Slavic linguistic unity.

¹¹ О. Н. Трубачев, Ремесленная терминология в славянских языках, М., 1966.

¹² О. Н. Трубачев,, op. cit., p. 209.

¹³ О. Н. Трубачев, op. cit., p. 275.

¹⁴ О. Н. Трубачев, op. cit., p. 393.

¹⁵ See N. van Wijk, Die baltischen und slavischen Akzent- und Intonationssysteme, p. 12.

4*

51

A CHARACTERIZATION OF THE MOST IMPORTANT COMMON FEATURES OF THE BALTIC AND SLAVIC LANGUAGES

In addressing the problem of the earliest relations between the Baltic and Slavic languages, scholars operate chiefly with phonetic, morphological, and lexical features shared by these languages, which may be said to constitute the argumentation of every theory. Proponents of the theory of Baltic-Slavic linguistic unity place particular emphasis on innovations in the Baltic and Slavic languages.

In the linguistic literature, the following features shared by the Baltic and Slavic languages are particularly often cited: the identical fate of the Indo-European short vowels o and a and the diphthong eu; the identical reflexation of Indo-European syllabic sonants R; the change of Indo-European consonant s after k, r, i, u; the affrication of palatal velar consonants k’ and g’; the emergence of pitch accents; the laws of Hirt and de Saussure-Fortunatov; the formation of suffixes *-ėjo-, *-tėjo-, etc.; the emergence of the pronominal root *mun-; the formation of pronominal adjectives; the identical genitive singular inflection of o-stem nouns; the origin of the second-person verb ending *-sēi, etc.

Proponents of the theory of Baltic-Slavic linguistic unity project all these shared features, as well as others, especially innovations, onto a single chronological plane. This creates the impression that the Baltic and Slavic languages share a great many features, especially innovations, and that these—as is most important—arose in a single period. From this it is concluded that Baltic-Slavic linguistic unity undoubtedly existed. However, clarifying the conditions and details under which shared phenomena and processes in the Baltic and Slavic languages arose, and shedding light on their chronology, may substantially alter the picture of how these languages’ shared features, especially innovations, formed and developed.

This gives rise to the task of examining in detail the most important features shared by the Baltic and Slavic languages, especially those considered the principal arguments for the theory of Baltic-Slavic linguistic unity. In this work, for the reasons stated in the preface, the analysis is limited to phonetic phenomena¹.

Indo-European short vowels o and a

The reflex of the Indo-European short vowels o and a is a in the Baltic languages, and o in the Slavic languages. Cf.: a) Indo-European o: Lithuanian naktis, Latvian nakts, Old Prussian naktin (acc. sing.), Old Church Slavonic noštь (from *nakti-): Latin nox noctis; Lithuanian akis, Latvian acs, Old Prussian ackis, Old Church Slavonic oko: Latin oc-ulus, Greek ὄσσε (nom. du.); b) Indo-European a: Lithuanian ašis, Latvian ass, Old Prussian assis, Old Church Slavonic osь: Latin axis, Greek ἄξων, Old High German ahsa, Old Indian ákṣaḥ; Lithuanian marios, Latvian mare, Old Prussian mary, Old Church Slavonic morje: Latin mare, Gothic marei. Latin and Greek have best preserved the Indo-European short vowels o and a. Their examples therefore clearly show where the vowel o and where a originally occurred. In addressing the question of the earliest relations between the Baltic and Slavic languages, it is important to establish which vowel the Baltic linguistic community and the Proto-Slavic language had in place of present-day Baltic a and Slavic o. Researchers’ opinions on this question differ.

¹ For a brief characterization of shared accentological and morphological phenomena in the Baltic and Slavic languages, see S. Karaliūnas, Baltų ir slavų kalbų seniausiųjų santykių klausimu, „Lietuvos TSR Mokslų akademijos darbai“, Serija A, 1(20), 1966, p. 167—177; С. Каралюнас, Древнейшие отношения балтийских и славянских языков (автореферат), Вильнюс, 1967.

52

M. Much¹, P. Kretschmer², J. Endzelīns³, J. Mikola⁴, W. Vondrák⁵, E. Schwarz⁶, T. Lehr-Splawiński⁷, A. Vaillant⁸, J. Otrębski⁹, K. Horálek¹⁰, P. Arumaa¹¹, V. Georgiev¹², and others believed or continue to believe that, both in the Baltic linguistic community and in Proto-Slavic, the Indo-European short vowels o and a became a. M. Bartoli¹³, V. Porzeziński¹⁴, H. Jakobson¹⁵, F. Fortunatov¹⁶, and others held the view that during the Balto-Slavic linguistic unity there existed a vowel ă (a with a low degree of labialization), from which Baltic a and Slavic o later developed.

All these scholars were proponents of theories of Balto-Slavic linguistic unity or of a common Balto-Slavic period (J. Endzelīns), and this was most likely what determined their view of the development of vocalism in the Baltic and Slavic languages: the same changes were thought to have taken place in the Baltic languages as in the Slavic languages.

Since in all Baltic languages Indo-European o and a were represented by a, these scholars therefore considered that a must likewise have existed in the Baltic linguistic community. The conclusion that the Baltic languages had formerly had the vowel a was not adequately supported by evidence from the Baltic languages themselves. For example, J. Endzelīns¹⁷ was inclined to think that the vowel a must have existed in the Balto-Slavic period for two reasons: 1) the wave of the change of Indo-European o into a, which had swept across the Indo-Iranian, Albanian, and Germanic linguistic areas, must also have reached the neighboring Baltic and Slavic languages; 2) in Greek and Finno-Ugric loanwords of Slavic origin, a occurs in place of the current Slavic-language vowel o. These loanwords were also regarded by the other scholars mentioned above as the principal argument that the short vowel a existed in Proto-Slavic.

Another view had also been expressed quite some time ago. K. Brugmann¹⁸ believed that Indo-European o and a became the vowel a during the period of Balto-Slavic linguistic unity; it was preserved in the Slavic languages, but changed to a in the Baltic languages. However, his view that the Baltic languages had formerly possessed

¹ M. Much, Abhandlungen zur germanischen Philologie, „Festgabe für R. Heinzel“, Halle, 1898, p. 213.

² P. Kretschmer, Die slavische Vertretung von indogermanischen o, AfslPh, 27, 1905, p. 228–240.

³ Я. Эндзелин, Славяно-балтийские этюды, p. 106–107.

⁴ J. J. Mikkola, Urslavische Grammatik, I, Heidelberg, 1913, p. 48.

⁵ W. Vondrák, Vergleichende slavische Grammatik, I, 2. Aufl., Göttingen, 1924, p. 106.

⁶ E. Schwarz, Zur Chronologie von asl. a > o, AfslPh, 41, 1927, p. 124–136.

⁷ T. Lehr-Splawiński, O pochodzeniu i praojczyźnie słowian, p. 24.

⁸ A. Vaillant, Grammaire comparée des langues slaves, I, p. 107.

⁹ Я. С. Отрембский, Славяно-балтийское языковое единство, ВЯ, 5, 1954, p. 28.

¹⁰ K. Horálek, Úvod do studia slovanských jazyků, Praha, 1962, p. 99.

¹¹ P. Arumaa, Urslavische Grammatik, I, p. 74–77.

¹² В. Георгиев, Вокалната система в развоя на славянските езици, София, 1964; Нови теории и традиционални заблуди, „Български език“, 15, 1965, p. 297–307.

¹³ M. G. Bartoli, Riflessi slavi di vocali labiali romane e romanze, greche e germaniche, „Zbornik u slavu V. Jagića“, Berlin, 1908, p. 30–60.

¹⁴ W. Porzeziński, Die baltisch-slavische Sprachgemeinschaft, p. 16.

¹⁵ H. Jacobsohn, Arrier und Ugrofinnen, Göttingen, 1922, p. 75.

¹⁶ Ф. Ф. Фортунатов, Лекции по фонетике старославянского (церковно-славянского) языка, Петроград, 1919, p. 26.

¹⁷ Я. Эндзелин, op. cit., p. 106–107.

¹⁸ K. Brugmann, Grundriss der vergleichenden Grammatik der indogermanischen Sprachen, I, 1, p. 146.

53

but it was not well substantiated. V. Thomsen¹⁹ and H. Jacobsohn²⁰ based their claims that Baltic languages once had the vowel o on Baltic-origin borrowings in the Finnic languages, where the short vowel o occurs instead of the present-day Baltic a.

On the basis of Baltic-origin borrowings in the Finnic languages of the Baltic region, K. Būga²¹ believed that during the earliest period of contact between Baltic and Finnic, the Balts had preserved Indo-European o intact only in unstressed position, while changing it to a in stressed position.

Recently, an attempt has been made to support the claim that a short vowel o existed in the Baltic linguistic community with new evidence. J. Kazlauskas²² believes that the Eastern Aukštaitian dialects’ in, un and the Samogitian dialects’ ẹn, ọn in Lithuanian may have developed from ẹn, ọn. J. Kazlauskas further argues that the change of Latvian ẹn, ọn sequences into ē, ō and then (after the ē: ō correlation had emerged) into ie, uo is simpler and more comprehensible. The sequences en, an need not necessarily have yielded ē, ō, since nasal ẹ, ą are also possible.

V. Mažiulis²³ believes that in words of the Prussian type doalgis ‘scythe’, the grapheme oa denotes a long vowel ō, which arose through circumflex lengthening from short o; in his view, this vowel should be regarded as Common Baltic. The claim that the Baltic linguistic community had a short vowel o may also be supported by two further pieces of evidence.

1. In studying the development of Baltic-language vocalism, the principle of polarization in the development of vowel systems had already been identified in scholarly literature²⁴. It is clearly evident in the development of the vowel systems of the Samogitian dialects. Here, not only the vowels a, e, but also the vowels ɔ, ẹ are broad and open. The fact that ɔ, ẹ developed from the close vowels u, i is shown by the sequences um, un, im, in, in which the vowels u, i mostly remained unchanged. The fact that the vowels u, i retained their closeness in all other Lithuanian dialects, but became more open in Samogitian, should most likely be explained by the strong labialization and considerable raising of the long vowels; compare Lithuanian uo, ie → Samogitian ū, ī, ọu, ẹi, Lithuanian o, ė → Samogitian uo, ie.

Thus, the system of long vowels in the Samogitian dialects changed in the direction of raising and labialization, whereas the system of short vowels changed in the opposite direction—toward opening and delabialization. In that case, the only conclusion is that not only the vowels ɔ, ẹ, but also a, e developed from close and rounded short vowels (u, i and o, ẹ, respectively). Given this tendency for short vowels to open, their closeness in some positions may be regarded as archaic. For instance, in Samogitian sequences, as already mentioned, um, un, im, in the vowels u, i are always close; in other words, under the influence of nasal consonants they will have preserved their original character. Likewise, the vowels ɔ, ẹ in the Samogitian sequences ɔm, ɔn, ẹm, ẹn should be considered older than those in open syllables, a, e. This has already been discussed in scholarly literature²⁵.

In the position of the ancient accent, the Northern Dounininkai dialects have close vowels ẹ/ie, in some cases diphthongized by considerable raising, if the following syllable contains a close—

¹⁹ V. Thomsen, Beröringer mellem de finske og de baltiske (litauisk-lettiske)Sprog, Köbenhavn, 1890, p. 89.

²⁰ H. Jacobsohn, op. cit., p. 75.

²¹ K. Būga, RR, I, p. 457.

²² И. Казлаускас, К развитию общебалтийской системы гласных, ВЯ, 4, 1962, p. 20–24.

²³ В. Мажюлис, Заметки по прусскому вокализму, „Вопросы теории и истории языка (Сборник в честь профессора Б. А. Ларина)“, Ленинград, 1963, p. 192–193; V. Mažiulis, Remarques sur le vocalisme du vieux prussien, „Acta baltico-slavica“, 2, 1965, p. 55.

²⁴ И. Казлаускас, op. cit., p. 21–22.

²⁵ I. Kazlauskas, op. cit., p. 23.

54

The narrow vowels u, i (short or long). Before other vowels in the following syllable, there is a broad e here. The behavior of the vowels u, i shows in which position the older character of the vowel e has been preserved. The narrow vowels u, i remain unchanged before the dialect’s narrow vowels u, ū, uo, i, ī, ie in most of the area of the dounininkai dialects. Before other dialect vowels, they have become more open, turning into ọ, ẹ²⁶. If the vowels u, i preserved their original character before narrow vowels, then the vowel e in this position could likewise have preserved its original character (narrowness).

Thus, the facts described here show that the original short vowels of the Samogitian dialects were most likely narrow o, ẹ.

In the East Aukštaitian dialects (the so-called rotininkai dialects), o is pronounced in place of the literary language’s stressed, non-final a (in dialectological transcription, written ə and ă); cf. the Anykščiai and Kupiškis forms rotai, gražū (acc. sing.), vokaras, dorbas. In unstressed position and at the stressed word-final position, which is phonologically equivalent to an unstressed position, a occurs here; cf. vakarai, gražūs, giria²⁷.

The question arises whether rotininkavimas is a newly emerging phenomenon or has been inherited from ancient times (in other words, which position’s vocalism—stressed or unstressed—is older).

Rotininkavimas is spread across a relatively small area between Papilys and Ukmergė²⁸. Since vowel assimilation does not operate in the southern part of the rotininkavimas area, and sound relationships are therefore clearer, the discussion will initially be confined to examining the vowel systems of the dialects in this part. According to Z. Zinkevičius’s data²⁹, for example, the vowel systems of the Anykščiai dialect are as follows:

short vowels long vowels in stressed position u. i. ū ī ə. ẹ. uo ie in unstressed position u i u. i. a e a. e.

Before the emergence of the diphthongs uo and ie (which, from a synchronic perspective, may perhaps be treated as tense phonemes), the system of long vowels in stressed position could have been as follows:

ū ī ō ē ā ē

The position of each phoneme in the system could have been characterized by the following distinctive features:

the phonemes /ū/ and /ī/—diffuse (opposed to the phonemes /ō/ and /ē/) and non-compact (opposed to the phonemes /ā/ and /ē/);

²⁶ Z. Zinkevičius, Lietuvių dialektologija, Vilnius, 1966, p. 61–62.

²⁷ For a fuller discussion, see Z. Zinkevičius, op. cit., p. 50.

²⁸ Z. Zinkevičius, op. cit., pp. 50–51 and map no. 6.

²⁹ Z. Zinkevičius, op. cit., p. 50 ff.

55

The phonemes /ō/ and /ē/ are non-diffuse (opposed to the phonemes /ū/ and /ī/) and non-compact (opposed to the phonemes /ā/ and /ē/); the phonemes /ā/ and /ē/ are compact (opposed to the phonemes /ō/ and /ǭ/) and non-diffuse (opposed to the phonemes /ū/ and /ī/). In addition, the phonemes /ū/, /ō/, /ā/ and /ī/, /ē/, /ė/ seem to have had the differential features “flatness—sharpness.” The phonemes /ū/, /ō/, /ā/ are characterized as flat because, at the phonetic level, not only the vowels ū and ō were strongly labialized, but ā too seems to have had a certain degree of labialization (a more adequate notation would be ā or even ā). That the vowel ā was labialized is shown by its labialization to the diphthong uo in Samogitian and some Aukštaitian dialects (around Zietela, Lazūnai, Ramaškonys, and around Adutiškis and Gerdasiai near Druskininkai ³⁰). This diphthongization in itself presupposes a certain initial degree of labialization in the vowel ā. The fact that the vowel ā was the manifestation of a compact, non-diffuse, and also flat phoneme is seemingly also confirmed by the rendering of the Slavic vowel a as Lithuanian long o (cf. Lith. striōkas and Old Rus. strachъ, Polish strach). The vowel most likely reconstructable for the Eastern Baltic linguistic community is one of the kind now found among Eastern Aukštaitians (around Dusetos, Svėdasai, Alanta, Utena) and Western Samogitians (around Karklė, Žardė). The archaism of this vowel (i.e., ā) is also indicated by geography: according to the principles of areal linguistics, a linguistic phenomenon or fact occurring in marginal or mutually distant areas is older than a linguistic phenomenon or fact widespread in a central or continuous area³¹. In other words, there is reason to believe that the ancient Eastern Baltic linguistic-community ā has survived in the above-mentioned Eastern Aukštaitian and coastal Samogitian dialects.

Comparing the long-vowel system of the Eastern Baltic linguistic community with the long-vowel system of the Anykščiai dialect, we see that: 1) at the phonemic level, the relative position and relations of its elements remained unchanged: the phonemes are characterized by the same differential features as before; 2) at the phonetic level, its elements changed in the direction of slight labialization and narrowing: ō, ē→uo, ie and ā, ē→ɔ, ė.

The question arises as to the direction in which the short-vowel system of the Anykščiai dialect changed—toward narrowing or widening—and how its development related to the development of the long-vowel system. In answering this question, it is important that in the Eastern Aukštaitian dialects (at least those within the rotininkas area) a correlation has formed between short vowels, lengthened here to half-long vowels, and long vowels. Numerous facts confirm the existence of this correlation. Around Ukmergė, not only the length–shortness opposition but also the localized (timbre) opposition has been neutralized; cf. rōtas ‘wheel’: prōtas ‘mind’. In the northern part of the rotininkas area, where vowel assimilation operates, both short e and long ē underwent the same development. Before flat vowels and in absolute word-final position, Kupiškėnai dialect has a and ā, respectively, in place of literary-language e and ē, while before sharp vowels and etymological j it has e and ē, respectively. In the position following k, g, š, ž, č, j, most Kupiškėnai speakers have e and ē, respectively, irrespective of the quality of the following syllable; around Papilys, however, before flat vowels and in absolute word-final position, in the aforementioned position, the vowels are æ and ǣ, respectively³².

³⁰ Z. Zinkevičius, op. cit., p. 71.

³¹ В. Пизани, Общее и индоевропейское языкознание, „Общее и индоевропейское языкознание (Обзор литературы)“, Москва, 1956, p. 112–114.

³² Z. Zinkevičius, op. cit., p. 54, 74.

56

Once a correlation had formed between short and long vowels (chiefly as a result of the lengthening of short vowels), the two vocalism systems had to change in one and the same direction. Since, as we have seen, the long-vowel system became labialized and narrowed, the short-vowel system, too, must have changed in the direction of labialization and narrowing. Thus, the narrow vowels o, ẹ of Eastern Aukštaitian are most likely new developments from earlier open vowels a, e.

In unstressed position, the long-vowel system in rotininkas dialects changed in the opposite direction—it shortened and widened. For example, in the Anykščiai dialect ō, ē became delabialized and widened, thereby merging with ā, ē, while the latter, through delabialization and widening, became a., æ.. In unstressed position, rotininkas dialects everywhere have open short vowels. Thus, the uniform development of the vocalism systems of rotininkas dialects in unstressed position also supports the claim that their development in stressed position was uniform as well.

The existence of a correlation between short and long vowels in Eastern Aukštaitian dialects (that is, the lengthening, narrowing, and labialization of short vowels in stressed position) is also shown by the fact that across a larger part of the rotininkas area, alongside the long vowels ō, ō, we find the short (here lengthened) ɔ, while in the eastern part of this area, alongside ā, we find the short vowel ā. In border dialects these relations may be confused: alongside ɔ there may be ā (in Trakiniai village between Kavarskas and Alanta), while alongside ā there may be ɔ (in the Švedasai dialect)³³. The existence of the correlation itself, however, is beyond doubt here as well.

The alternation of vowels in the Kupiškėnai dialect depending on position (assimilation) will be understandable if we assume that open vowels e and ē, respectively, existed here earlier; before flat vowels and in absolute word-final position they widened further and became a and ā, respectively, while before sharp vowels and etymological j they narrowed and became ẹ and ē, respectively, i.e.

e and ē, respectively → a and ā, respectively; e and ē, respectively → ẹ and ē, respectively

The narrowing of the open vowel a in rotininkas dialects is also shown by the fact that the acute diphthong au contains the vowel a (but not o). It apparently did not narrow because of the nature of the stress and because of the u, before which a narrow vowel (e or o) seemingly cannot occur.

A labialized vowel a also occurs in some Samogitian dialects (around Laukuva, Seda, etc.)³⁴. Here, too, however, it is most likely a recent development, since the strongly labialized and narrowed long vowels in Samogitian dialects must regularly have caused short vowels to become open and wide.

The aforementioned changes in the vocalism of Lithuanian dialects (the polarized development of short and long vowels in the Samogitian dialect, the formation of a correlation, and the phenomena of assimilation in Eastern Aukštaitian dialects) may indicate that the short vowels o, e became open a, e relatively long ago.

2. In Latvian, the eastern dialects of the Aukštzemnieki have the short vowel o in place of the literary-language a. The vowel o here may occur both in stressed and in unstressed position (cf. Latv. vosara and vosora ‘summer’), both before a back vowel and before

³³ Z. Zinkevičius, op. cit., p. 50.

³⁴ Z. Zinkevičius, op. cit., p. 51.

57

before front vowels (cf. Latvian lobs ‘good’ and obi ‘both’)³⁵. Without investigating the development of vowel systems in Latvian dialects, it is difficult to say anything about the antiquity of this vowel.

The short vowel o also occurs in Prussian. In the Elbing Vocabulary of the Prussian language, a is quite often replaced by o³⁶; cf. Prussian wobse ‘wasp’, sompisinis ‘coarse bread, Grobbrot’ (Prussian som- corresponds to Lithuanian sam-). This o may be a remnant of the short vowel o, which was formerly more widely used³⁷.

3. The existence of the short vowel o in the Baltic linguistic community may perhaps also be indicated by loanwords of Baltic origin in the Baltic Finnic languages, which in certain cases have o; cf. Finnish morsian and Lithuanian marti, Finnish oinas and Lithuanian avinas, Finnish lohi and Lithuanian lašiš, etc.³⁸ Recently, V. Steinitz expressed the well-founded view that words with o belong to the oldest layer of loans of Baltic origin in the Finno-Ugric languages³⁹.

Some researchers, beginning with K. Brugmann⁴⁰, tend to believe that Proto-Slavic also had the short vowel o. M. Vasmer⁴¹, A. Meillet⁴², S. Bernstein⁴³, F. Filin⁴⁴, and others held or continue to hold this view.

M. Vasmer⁴⁵ explained the fact that the Slavic vowel o is represented by the vowel a in Greek loanwords and place names of Slavic origin as follows: the close Greek vowel o was unsuitable for rendering the Slavic vowel o, since the latter was open. In Proto-Slavic, in A. Meillet’s view⁴⁶, there had been no great difference in timbre between short o and long a. The difference was chiefly quantitative. S. Bernstein argues⁴⁷ that the Slavic vowel o formerly had a low degree of labialization.

Thus, some researchers claim that Proto-Slavic had a short vowel a or ă, while others tend to believe that Proto-Slavic had the vowel o. In fact, these two views do not differ as greatly as may appear at first glance. For those who claim that Proto-Slavic had the vowel a also believe that this short vowel was slightly labialized and more close than the vowel a. And those,

³⁵ J. Endzelins, Latviešu valodas gramatika, Rīgā, 1951, p. 110–123; M. Rudzīte, Latviešu dialektoloģija, Rīgā, 1964, p. 267–271.

³⁶ E. Berneker, Die preussische Sprache, I. Lautlehre, Strassburg, 1895, p. 27; J. Endzelin, Zum o für a im Elbinger Vokabular, StB, 5, 1935–1936, p. 94–98; Senprūšu valoda, Rīgā, 1943, p. 24.

³⁷ Plg. И. Казлаускас, op. cit., p. 21.

³⁸ For literature from recent years, see E. Nieminen, Über einige Eigenschaften der baltischen Sprache, die sich in den ältesten baltischen Lehnwörtern der ostseefinnischen Sprachen abspiegelt, „Sonderabdruck aus den Sitzungsberichten der Finnischen Akademie der Wissenschaften 1956“, Helsinki, 1957, p. 199–200; A. Sabaliauskas, Relations between the Baltic and Baltic Finnish languages, LKK, 6, 1963, p. 109–136.

⁳⁹ W. Steinitz, Zur Periodisierung der alten baltischen Lehnwörter im Ostsee-Finnischen, „Symbolae linguisticae in honorem Georgii Kurylowicz“, Wrocław–Warszawa–Kraków, 1965, p. 297–303.

⁴⁰ K. Brugmann, op. cit., p. 146.

⁴¹ M. Vasmer, Zwei kleine Abhandlungen. II. Zur slavischen Vertretung von arioeurop. o, KZ, 41, 1907, p. 157–164.

⁴² А. Мейе, Общеславянский язык, Москва, 1951, p. 42–44.

⁴³ С. Б. Бернштейн, Очерк сравнительной грамматики славянских языков, p. 144–146.

⁴⁴ Ф. П. Филин, Образование языка восточных славян, p. 101.

⁴⁵ M. Vasmer, op. cit., p. 157–164.

⁴⁶ A. Meje, op. cit., p. 44.

⁴⁷ С. Б. Бернштейн, op. cit., p. 143.

58

who consider Proto-Slavic to have had a short vowel o also emphasize its weak degree of labialization and openness.

The following facts seem to allow a more precise definition of the quality of the short vowel in Proto-Slavic:

1. That the Proto-Slavic vowel originating from PIE o and a may phonetically have been closer to a than to o is shown by loans and place names of Slavic origin in Finno-Ugric, Baltic, Germanic, and Greek languages, all of which have a instead of the present-day Slavic vowel o; cf. Finnish papu ‘bean’ from Russian боб, Prussian maddla ‘prayer, Bitte, Gebet’ from Polish modła, Old Bavarian Rasa from Slavic ražb, Greek βάβεω ‘old woman’ from Slavic *baba, and so on. In other languages, Slavic a is everywhere rendered by the vowel o; cf. Old Church Slavonic kotblъ ‘kettle’ from Gothic *katils (cf. katile gen. pl.). This shows that at the time of borrowing, Slavic had the vowel a.

2. Belarusian and some southern Russian dialects, as well as some very archaic Bulgarian dialects⁴⁸, have the vowel a in unstressed position (the so-called “akan’e”); cf. Belarusian каза́, Russian каза́ (spelled коза), and Bulgarian kazý-te (nom. pl.).

In unstressed position, these dialects will have preserved the Proto-Slavic short vowel a⁴⁹. In the recent discussion⁵⁰ of the origin of “akan’e,” the prevailing view was that this feature of Belarusian and southern Russian dialects arose relatively late (around the 13th–14th centuries)⁵¹, either spontaneously or under substrate influence. Nevertheless, the first interpretation of the origin of “akan’e” seems more reliable: the a in unstressed position in these dialects should be regarded as a relic of Proto-Slavic a, because: a) this is required by the geographical evidence (cf. the analogous phenomenon in archaic Bulgarian dialects); b) phonological studies confirm the existence of a short a in Proto-Slavic.

3. All researchers to date have reconstructed for Proto-Slavic the correlation a: ā (resp. ă: ǟ)⁵². The change of the PIE short vowel o to a and its merger with the reflex of PIE a were connected with the delabialization of the other Proto-Slavic vowels. In Proto-Slavic, alongside the palatalization of pre-vocalic consonants, there occurred labialization of pre-vocalic consonants (or, respectively, delabialization of vowels)

⁴⁸ Ст. Стойков, Акане в говора на с. Триград (Девинско), „Български език“, 13, 1963, p. 8–21.

⁴⁹ A. Vaillant, Grammaire comparée des langues slaves, I, p. 107; G. Y. Shevelov, A Prehistory of Slavic, p. 386–387; V. I. Georgiev, Russian akanye and its relation to the phoneme system of Proto-Slavic, VYa, 2, 1963, p. 20–29; The pan-Slavic significance of the problem of akanye, VYa, 4, 1964, p. 151–152.

⁵⁰ В. Г. Руделев, К фонологической интерпретации русского аканья, ВЯ, 2, 1963, p. 30–38; П. С. Кузнецов, К вопросу о происхождении аканья, ВЯ, 1, 1964, p. 30–41; В. В. Колесов, О некоторых особенностях фонологической модели, развивающей аканье, ВЯ, 4, 1964, p. 66–79; Я. Риглер, К проблеме аканья, ВЯ, 5, 1964, p. 36–45; В. И. Лыткин, Ещё к вопросу о происхождении русского аканья, ВЯ, 4, 1965, p. 44–52; Г. А. Хабургаев, О фонологических условиях развития русского аканья, ВЯ, 6, 1965, p. 55–63. Cf. also M. Halle, Akan’e (The Treatment of Unstressed Nondiffuse Vowels in Southern Great Russian Dialects), „Symbolae linguisticae in honorem Georgii Kuryłowicz“, p. 103–109.

⁵¹ Cf. В. И. Борковский, П. С. Кузнецов, Историческая грамматика русского языка, М., 1963, p. 142.

⁵² R. Jakobson, Remarks on the phonological evolution of Russian compared with that of the other Slavic languages, TCLP, 2, 1929, p. 20–23; F. V. Mareš, The emergence of the Slavic phonological system and its development up to the end of the period of Slavic linguistic unity, “Slavia”, 25, 1956, p. 445; G. Y. Shevelov, A Prehistory of Slavic, p. 150–162; H. van Wijk, On the history of the phonological system in late Common Slavic, “Slavia”, 19, 1950, p. 293.

59

process. Recent phonological studies have established⁵³ that the short vowels o and a (like ō and ā) first merged in the position after labial consonants. Here the differential feature of the vowel o (like ō), “rounding,” became irrelevant, since that feature was a constitutive element of labial consonants (for more details, see pp. 67 ff. of this work).

The claim that Proto-Slavic had a short vowel a (but not o)¹ is contradicted by the singular nominative and accusative ending -ō of masculine o-stem nouns—it could have arisen phonetically only from *-os and *-om through the intermediate forms *-u(s) and *-u(m).

G. Shevelov is quite reasonably inclined to explain the emergence of the singular nominative ending -ō instead of *-as (cf. OCS vьlkъ) by morphological factors⁵⁴. The ending *-am could easily have become -ō, since vowel raising before nasal consonants is reportedly a common phenomenon, found in many languages. Following the phonetic changes at the ends of words in Proto-Slavic, an analogy between o- and u-stems produced the following proportion: *-u (acc. sing., u-stem): *-u (nom. sing., u-stem)= *-u (acc. sing., o-stem): x (nom. sing., o-stem), where x must have been equal to *-u.

Thus, nothing prevents us from assuming that Proto-Slavic had a short a, perhaps somewhat labialized, since in Romanian loanwords of Slavic origin, Proto-Slavic a is represented by oa; cf. Romanian groază “horror” from Slavic *graza⁵⁵.

In the opinion of researchers, the Slavic short vowel a became o quite late (around the 8th–9th centuries), just before the appearance of the earliest written monuments of the Slavic languages⁵⁶. After the 9th century, loanwords of Slavic origin in other languages quite often already have o; cf. Greek Σϑλοβενοί from Slavic *slověne.

Almost all researchers agree that in both Baltic and Slavic languages the original short-vowel system was triangular:

But the alternative—that the original short-vowel system of these languages was quadrangular—is equally possible:

in which there was no opposition a: o.

Besides Baltic and Slavic, the Proto-Indo-European short vowels o and a also merged in Germanic, Albanian, Old Indic, and Hittite. In these languages, they are reflected as a.

53 B. K. Zhuravlev, The Formation of Group Singarmonism in the Proto-Slavic Language, ВЯ, 4, 1961, p. 35–36.

54 G. Y. Shevelov, op. cit., p. 156–157.

55 G. Y. Shevelov, op. cit., p. 162.

56 E. Schwarz, op. cit., p. 136; H. ван Вейк, op. cit., p. 293; F. V. Mareš, op. cit., p. 445; G. Y. Shevelov, op. cit., p. 151; B. K. Журавлев, Из истории вокализма в праславянском языке позднего периода, ВЯ, 2, 1963, p. 13, 19; В. И. Георгиев, ВЯ, 2, 1963, p. 29; Нови теории и традиционални заблудки, „Български език“, 15, 1965, p. 297–307.

60

The short vowel a in the Indo-European linguistic community was apparently rare. It occurred most often in expressive, technical, and everyday vocabulary, and participated hardly at all in ablaut⁵⁷.

The vocabulary of European languages distinguishing the phonemes o and a, which contains the short vowel a, is limited mainly to the sphere of agriculture; for this reason, this vocabulary, and thus the phoneme a, may be of later origin⁵⁸. Some researchers tend to think that the o: a opposition in the southern Indo-European languages (Armenian, Greek, Italic, and Celtic) may have arisen later as a result of the loss of laryngeals⁵⁹.

It is therefore entirely possible that those Indo-European languages which lack the o: a opposition never had it. The Proto-Baltic short vowel o and the Proto-Slavic a may correspond to the o and a of the southern Indo-European languages.

In addressing the question of the most ancient relations between the Baltic and Slavic languages, it is important to determine which manifestations of unstressed compact phonemes, that is, low-back short vowels, the Baltic and Slavic languages had in the first centuries after separating from the Indo-European linguistic community. This cannot be stated for many reasons, the most important of which are:

a) we have no data on which to base a reconstruction of the vocalism of such remote epochs; that is, we do not know whether the Baltic dialects at that time really had the short vowel o, and the Slavic dialects the short vowel a;

b) it is unclear which short-vowel system (triangular or quadrangular) was original in the Baltic and Slavic languages;

c) we do not know the exact absolute chronology of when the reflexes of the Indo-European short vowels o and a (assuming the triangular short-vowel system to be original) coincided in the Baltic and Slavic languages.

In any case, whichever short-vowel system we consider original, one thing is clear: other Indo-European dialects (e.g., Germanic) also had an original short-vowel system exactly like that of the Baltic and Slavic languages. Thus, reconstructing an identical original short-vowel system for the Baltic and Slavic languages in no way indicates a common stage of Baltic-Slavic development distinct from that of other Indo-European dialects, or a Baltic-Slavic linguistic unity.

Indo-European long vowels ō and ā

Indo-European long vowels ō and ā merged in the Slavic languages. The Baltic languages retained their distinction. Thus, the Baltic and Slavic languages differ greatly in the development of Indo-European long vowels. In addressing the problem of the earliest relations between the Baltic and Slavic languages, it is important to determine when the Slavic languages ceased to distinguish Indo-European long vowels.

Indo-European ā is preserved intact in Latvian (ā); in Lithuanian it became o, while the Slavic languages have a here; compare:

Lith. brólis, broter-ėlis, Latv. brãlis, brāter-ītis, OCS bratrъ: OInd. bhrātā, Gk. φράτηρ “member of a brotherhood,” Lat. frāter, Goth. broþar (with o from Indo-European ā), Toch. A pracar;

⁵⁷ A. Meillet, Introduction à l’étude comparative des langues indo-européennes, 7 éd., Paris, 1934, p. 99, 166–168. Compare: “La voyelle a se trouve notamment dans des mots de caractère populaire, technique ou affectif” (p. 99).

⁵⁸ H. Kuhn, Ablaut, in: Altertumskunde, KZ, 71, 1954, p. 129–161.

⁵⁹ J. Kuryłowicz, L’apophonie en indo-européen, Wrocław, 1956, p. 167–168.

61

Lith. moté, móteris, Latv. māte, OCS mati: OInd. matā, Gk. (Doric) μάτρη, Lat. māter, Toch. A mācer.

In the Slavic languages, Indo-European ō became a. In the Eastern Baltic languages, it had two different reflexes. In some varieties Indo-European ō became uo; compare:

Lith. belt, Lat. belt, OCS po-jasъ ‘belt’: Gr. ζωστός ‘girded’ (= Lith. belt), Avest. vāh ‘belt’ (ā from Indo-European ō);

Lith. puo-tà: Gk. πώ-νω “I drink,” Lat. pō-tus “drunk,” OInd. pā-ti “(he) drinks” (ā from Indo-European ō);

Lith. uoga, Latv. uōga: OCS jag-oda;

Lith. uolektis (and uolektis), Latv. uōlekts: Gk. ὠλένη “elbow”: *ōl- is the lengthened grade of the root *el-/*ol- (cf. Lith. elcūnė, alkūnė and Latv. elkuons);

Lith. ūosis, Latv. uōsis: Russ. яс-ень, Serbo-Croatian jas-ēn;

Lith. dūo-ti, duo-snùs, duo-slùs, dūo-klė, Lat. duō-t, OCS da-ti, da-nъ ‘tribute’ (= Lith. duo-nis ‘gift’¹), da-rъ ‘gift’: Gr. δῶ-μι ‘I give’, δῶ-ρον ‘gift’, Lat. dō-num ‘gift’, Skt. dānám ‘gift’ (ā from Indo-European ō); but with o Lith. do-snùs (Lat. dā-sns), do-slùs, do-vanà (Lat. dō-vana), dō-vamas ‘given as a gift’, dō-vis ‘giving’, dō-sai ‘gifts from the bride’s family’, dō-tas ‘giving, gift’, dō-talas (Lat. dā-tali ‘gifts’) ‘giving, tribute’, do-tilas ‘ditto’: cf. OPr. dā-t, dā-twei ‘to give’.

Indo-European ō, which alternated with Indo-European ē, also became uo in certain cases; cf. brėžti: bruožas; plėšti, Latv. plēst: plūostas, Latv. plūōsti “to tear to pieces,” pluosa “frenzy”; rėžti, OCS rěz-ati “to strike”: ruožas, OCS razъ “blow; stroke”; sėsti: sūodžiai, Latv. suōd-rēji “soot,” OCS sazъda; ėsti: ūodas, Latv. uōds.

Elsewhere, Indo-European ō became o in Lithuanian and ā in Latvian; compare:

Lith. nor-ėti: Gr. νωρ-εῖν “to act, do,” νώρ-εμενος “large, broad”;

Lith. krōk-ti (ir kriōk-ti), kriog-uoti “to speak or shout in a hoarse voice,” Latv. krākt, krākt, O. Russ. krak-ati “to hawk”: Lat. crōc-āre, crōc-īre “to croak (of ravens),” Gr. κρώζω “I hawk” (-ζ- from *-gj-), OHG hruoh “raven” (uo from Germ. ō);

Lith. žmōnės, žmona: Lat. acc. sing. homōnem (nom. sing. homō) and hemōnem; root *gʰm-/*gʰem-/*gʰom-; in the words žmōnės, žmona we have the zero-grade root, while the full-grade root is Lith. žem-ė; but O. Lith. žmuo (from *gʰm-ōn);

Lith. prō, pró- (cf. pró-plaiša): Lat. prō, prō(d)-, Gr. πρω-; but Latv. pruō-jām, pruō-m “away, off”;

The genitive singular of o-stem nouns: Lith. vilk-o, Latv. vilk-a, O. Slav. vlъk-a: O. Lat. (personal name) Gnaiv-ōd, Lat. lup-ō “from the wolf,” O. Ind. vṛk-āt, vṛk-ād (-āt -ād from *-ōd), Gr. τω (from *pōd) “from where.”

Instead of the regular vowel alternation (Lith. ē: uo, Latv. ē: uo from PIE ē: ō), Lithuanian exhibits the alternation ē: o, while Latvian has ē: ā; cf.:

Lith. bėgti, Latv. bēgt: Lith. bog-inti “to carry, transport, haul, convey something quickly or with difficulty,” but Latv. buōk-stīt “to hide,” buōk-stīties “to wander, roam;”

Lith. glėbti, Latv. glēbt: Lith. glōbti “to wrap, swaddle, embrace,” Latv. glābt “to protect, guard,” Lith. globà (acc. sing. glōbą) and glōba; the full grade of the vowel alternation is Lith. glab-ōti “to embrace, take, grasp,” Latv. glab-āt “to protect, guard; bury” (from Baltic *glo-b-);

¹ Lith. donis should be regarded as a Slavic borrowing.

62

Lith. korýs (acc. sing. kõrį), Latv. kãris: Gk. χηρός “wax,” Lat. cēra; cf. kor-ýti, kor-ýtas;

Lith. mólis, mõlė “earth by clay,” mol-iūgas “clay-colored, clayey,” Latv. mãls (usually plural mãli): OCS mẽlъ “lime”; E. Fraenkel also includes here mėlas, mėlynas²;

Lith. nókti, Latv. nãkti “to come”: Goth. nehv(a) “to approach,” nehv-jan “to draw near,” OHG nãh “near” (ā from ē < PIE ē);

Lith. plėšti: plótas, alongside them the roots *plet- in the normal grade (cf. plėčia, plet-óti) and *plot- (cf. platus);

Lith. slėgti (ir slėgti), Latv. slėgt: Lith. slog-inti “to press down, compress,” sloga “such an illness; oppression, trouble,” slõgas “a weight used to press down flax or hemp,” Latv. slāga, slāgs “burden, weight, suffering, illness”; but Lith. šliuogas (ir šliuogas) “a weight used to press down,” Latv. šluõga, šluõgs “the same,” šluõdz-it, šluõdz-et “to press down, compress”;

Lith. sėsti, sėd-ėti: sod-inti, sõdas (ir sõdnas), OCS sadъ, Lith. sodà (acc. sing. sõdą), sõdžius (Latv. sādzā, sādzius are presumably Lithuanian loanwords); Samogitian svad-inti “to plant” may derive from *suod-inti;

Lith. tvėrė, Latv. tvēru “I groped, seized”: Lith. tvorà; the normal grade is found in tvēr-ia, cf. Old Rus. tvor-iti “to do, create”;

Lith. vókas (ir vokà) “cover, lid; blinking covering of the eye,” Latv. vāks “cover”: OCS vėko “(eyelid), cover”;

Lith. vóras “spider,” Latv. vāris: Lith. vėrė, Latv. vēru “I threaded”: the normal grade is vēr-ia, var-tai;

Lith. žolė, Latv. zāle: Lith. žėlė; the normal-grade root *g’hel-/*g’hol- occurs in the words žėl-ia, žãl-ias.

The reflexes of PIE ō and ā in Prussian are very unclear.

In the Elbing Vocabulary (a monument of the Pomesanian dialect), PIE ā is represented by the graphemes o and oa (cf. Pr. mothe and Lith. mótė, Latv. māte; Pr. moazo “aunt” and Lith. mõša, Latv. māsa “sister”). PIE ō is likewise represented here (cf. Pr. podalis and Lith. Samogitian puo-dãlis; Pr. woasis and Lith. úosis, Latv. uosis). The use of the same graphic notation for the reflexes of PIE ō and ā might suggest that PIE ō and ā had merged in the Elbing Vocabulary dialect of Prussian³. However, J. Kazlauskas believes⁴ that the identical graphic notation does not indicate neutralization of the opposition between the reflexes of PIE ō and ā. After all, in old Lithuanian texts the letter o apparently represents both o and uo. In J. Kazlauskas’s view, Prussian o (from PIE ō) may have been higher and more strongly rounded than o (from PIE ā). The latter may have been a lower, weakly rounded sound.

In the Prussian catechisms (written monuments of the Sambian dialect), the reflex of PIE ā is represented by ā or a (cf. Pr. brāti and Lith. brólis, Latv. brālis). The reflex of PIE ō is represented here mainly by ō or o (cf. Pr. per-ōn-i “community” and Lith., Latv. -uon- in words such as gel-uon-is ir dzel-uon-s). In certain cases the reflex of PIE ō is represented here by ā or a (cf. Pr. dāt, dātwei and daton and Lith. duóti, Latv. duôt; Pr. sālin and Lith. žolė, Latv. zāle).

Thus, in the dialect of the catechisms, the distinction between the reflexes of PIE ā and ō is preserved, except in certain cases where PIE ā and ō have merged, as in the East Baltic languages.

^² E. Fraenkel, LEW, p. 430.

^³ Cf. K. Būga, RR, III, p. 104.

^⁴ I. Kazlauskas, On the Development of the Common Baltic Vowel System, p. 21.

63

The twofold reflex of the Proto-Indo-European long vowel ō in the Baltic languages puzzled researchers for a long time. Some of them (J. Hanuš⁵, O. Wiedemann⁶, J. Mikkola⁷, and others) considered Lithuanian o and Latvian ā to be the sole reflexes of Proto-Indo-European ō in the Baltic languages, while deriving Lithuanian and Latvian uo from Proto-Indo-European *ōu (or *ōn and *ōns). This view is contradicted by the fact that the regular development of the Proto-Indo-European diphthong *ōu in Lithuanian was not uo, but áu, cf. Lithuanian gáuti, káuti, etc. That their roots derive from *gōu-, *kōu- is shown by the forms gōvė, kōvė. The derivation of káuti from *kōu- is also confirmed by data from other Indo-European languages (cf. Old High German houwan ‘to strike’ and Latin cūd ere ‘to strike, beat’ < *kou-/*kōu-).

Other researchers (G. Malov⁸, H. Pedersen⁹, V. Porzeziński¹⁰, A. Vaillant¹¹, and others) believed that Proto-Indo-European ō became only uo in Lithuanian and Latvian. The most renowned specialists in the Baltic languages, K. Būga¹², J. Endzelīns¹³, P. Skardžius¹⁴, E. Fraenkel¹⁵, and J. Otrębski¹⁶, also held this view. All these researchers regarded the Lithuanian and Latvian vowels ō and ā, instead of the expected uo, as having arisen through ablaut: alongside the regular alternation a: uo (from Proto-Indo-European o: ō), Lithuanian ablaut a: o (from Proto-Indo-European a: ā) had also given rise to an alternation a: o (instead of Proto-Indo-European o: ō). An important role here was played by the circumstance that the Proto-Indo-European short vowels a and o had merged as a in the Baltic languages.

The shortcoming of this hypothesis is that it does not explain the ablaut in words such as bog-inti, which lack corresponding forms with the short vowel a (for example, *bag- from Proto-Indo-European *bog-).

Moreover, there are cases whose explanation has compelled researchers to devise rather complicated schemes. For example, in order to explain the ā vowel in the Old Prussian word dāt (cf. Lithuanian dúoti and Latvian duôt), J. Endzelīns developed a complicated theory of analogy¹⁷. Judging by Latvian dakstīt ‘to give’, Greek δόσις ‘gift’, and Latin datio ‘giving’, the Baltic languages must have had an infinitive *dat(e)i. The long ā (cf. Old Prussian dāt) arose here, he proposed, because the vowel immediately before *-t(e)i in infinitives was always long. The Lithuanian and Latvian words dúoti and duôt, however, acquired their diphthong uo from the future-tense form (cf. Lithuanian dúosiu and Sanskrit dāsyā-mi).

For words with a root vowel o (Latvian ā) alternating with ē, instead of the expected uo, J. Endzelīns attempted to find an Indo-European basis, proposing a—

⁵ J. Hanusz, O dobie litewsko-słowiańskiej w stosunku do prajęzyka indoeuropejskiego, p. 254.

⁶ O. Wiedemann, Das litauische Präteritum, Strassburg, 1891, p. 45–51.

⁷ J. Mikkola, RSI, 1, 1908, p. 8; Urslavische Grammatik, p. 50–51.

⁸ G. Mahlow, Die langen Vocal e A, E, O in den europäischen Sprachen, Berlin, 1879, p. 82.

⁹ H. Pedersen, Wieviel Laute gab es im Indogermanischen?, KZ, 36, 1900, p. 101–103.

¹⁰ B. K. Поржезинский, К истории форм спряжения в балтийских языках, Москва, 1901, p. 7.

¹¹ A. Vaillant, Grammaire comparée des langues slaves, I, p. 112.

¹² K. Būga, Priesagos -ūnas ir dvibalsio uo kilmė, RR, II, p. 340–366.

¹³ J. Endzelin, Le. nākt: li. nók̇ti: got. nēhʷ and the alternation of ē and ā, KZ, 62, 1934, p. 23–28; Le. buokstīt and related forms, APh, 5, 1935, p. 10–13; What became of Indo-European ō (and ā) in Old Prussian, StB, 4, 1935, p. 133–143; Lettische Grammatik, Riga, 1922, p. 39; Latviešu valodas gramatika, p. 58–59.

¹⁴ P. Skardžius, Lietuvių kalbos žodžių daryba, Vilnius, 1941, p. 47–48, 544.

¹⁵ E. Fraenkel, Die baltischen Sprachen, p. 76.

¹⁶ Я. С. Отрембский, Славяно-балтийское языковое единство, ВЯ, 5, 1964, p. 29; however, in his “Gramatyka języka litewskiego” (I, Warszawa, 1958, p. 180–181), J. Otrębski also acknowledged Proto-Indo-European ō > Lithuanian o.

¹⁷ J. Endzelin, Was ist im Altpreussischen aus ide. ō (und ā) geworden? p. 136.

64

to identify the Indo-European vowel alternation ā: ē¹⁸. However, such apophony, with few exceptions, is not characteristic of Indo-European languages¹⁹, and therefore J. Endzelin’s idea was viewed critically²⁰.

Other scholars (J. Zubatas²¹, K. Brugmanas²², H. Hirtas²³, A. Meje²⁴, N. van Veikas²⁵, Chr. Stangas²⁶, J. Safarevičius²⁷, H. Krahe²⁸, G. Ševeliovas²⁹, and others) recognized a dual outcome of Indo-European ō in the Baltic languages: in some instances it became uo, in others ā (Lithuanian o, Latvian ā).

On the basis of the dual development of Indo-European ō in the Baltic languages, and thereby overemphasizing the archaic character of Lithuanian, J. Zubatas³⁰ advanced the idea that the Indo-European linguistic community had two ō sounds—one that became uo in Lithuanian, and another that yielded o in Lithuanian. Later, when it was established that the dual development of Indo-European ō depended on specific conditions in Lithuanian, this theory was abandoned.

Once a dual development of Indo-European ō in the Baltic languages had been recognized, the question naturally arose as to which instances of Indo-European ō became uo and which became ā (Lithuanian o, Latvian ā, Old Prussian ā or a). H. Hirtas and A. Meje tended to argue³¹ that Indo-European ō alternating with Indo-European ē became ā in the Baltic languages, whereas Indo-European ō not participating in vowel alternation became uo. The alternation in forms of the type brėžti: bruožas (see p. 62 of this work), as well as the fact that the Indo-European non-apophonic vowel ō is quite often represented by the vowel ā in the Baltic languages (see p. 62 of this work), shows that this view is unfounded.

In recent years, the development of Indo-European ō (and of Baltic-language vocalism in general) has tended to be linked (J. Kazlauskas, V. Mažiulis³²) to the accentual system: in stressed position, Indo-European ō became uo in the East Baltic languages, was probably preserved unchanged in the Prussian dialects, and in unstressed position merged with Indo-European ā in the Baltic languages. The claim that the development of Indo-European ō (and vocalism in general) depends on the accentual system is supported by evidence from living Lithuanian speech. For example, in many East Aukštaitian dialects the opposition between the reflexes of Indo-European ō and ā has been eliminated in unstressed position: some East Aukštaitian dialects have a (half-long or short) in place of Indo-European ō and ā, others (e.g., around Kupiškis and elsewhere) have o (half-long or short), and still others (e.g., around Raguva and elsewhere) have u (usually reduced and short). The opposition between the reflexes of Indo-European ō and ā is also

¹⁸ J. Endzelin, Le. nākt: li. nókti: got. nēh und der Wechsel von ē und ā, p. 27.

¹⁹ A. Meillet, Introduction à l’étude comparative des langues indo-européennes, p. 153–168.

²⁰ E. Fraenkel, Die baltischen Sprachen, p. 78.

²¹ J. Zubatý, Zum baltischen ū, BB, 18, 1892, p. 241–266.

²² K. Brugmann, Grundriss der vergleichenden Grammatik der indogermanischen Sprachen, I, 1, p. 151–152, 158.

²³ H. Hirt, Indogermanische Grammatik, II, Heidelberg, 1921, p. 33, 180.

²⁴ A. Meillet, op. cit., p. 100–104.

²⁵ N. van Wijk, Les voyelles uo, ie en letto-lituanien et en russe, StB, 4, 1935, p. 144.

²⁶ Chr. S. Stang, Das slavische und baltische Verbum, p. 196 (footnote).

²⁷ J. Safarewicz, Przyczynki do zagadnienia wspólnoty bałto-słowiańskiej, p. 199–202.

²⁸ H. Krahe, Indogermanische Sprachwissenschaft, I, Berlin, 1958, p. 57.

²⁹ G. Y. Shevelov, A Prehistory of Slavic, p. 150.

³⁰ J. Zubatý, op. cit., p. 241. Cf. J. Otrebski, Przyczynki słowiańsko-litewskie, I, Wilno, 1930, p. 53–54. K. Brugman and H. Hirt also incorporated J. Zubatas’s theory into their comparative grammars of the Indo-European languages; see K. Brugmann, op. cit., p. 151–152, 158 and H. Hirt, op. cit., p. 33, 180.

³¹ H. Hirt, op. cit., p. 33; A. Meillet, op. cit., p. 100–104; for critical remarks on H. Hirt’s opinion, see K. Būga, RR, II, p. 242.

³² I. Kazlauskas, On the development of the Common Baltic vowel system, p. 24; V. Mažiulis, Notes on Old Prussian vocalism, p. 195–196; V. Mažiulis, Remarques sur le vocalisme du vieux prussien, p. 53–59.

5. Baltų ir slavų kalbų ryšiai

65

is absent in the Western Aukštaitian dialects around Smalininkai (here there is a long o with very slight diphthongization, and sometimes even a clear diphthong ou) and in the Western Samogitian dialects around Priekulė (here there is a long o³³. Thus, in unstressed position, these dialects had a labialized sound ā (a more adequate graphic notation would be ā) instead of Proto-Indo-European ō and ā; in some dialects it became a, in others o, u, or even the diphthong ou.

In stressed position, the distinction between the reflexes of Proto-Indo-European ō and ā is preserved in these dialects. The Baltic-language evidence concerning the fate of Proto-Indo-European ō in the Baltic languages also appears not to contradict the aforementioned idea (that the development of Proto-Indo-European ō depended on the accentual system): in Lithuanian, uo occurs in words with a barytone paradigm in many cases (cf. brúožas, júosta, plúoštas, rúožas³⁴, súodžiai, úodas, úoga, úolektis, úosis, etc.). In words with mobile-accent paradigms (e.g., puõta), the vocalism of barytone case forms (cf. acc. sing. puõtą) may have been generalized.

The Lithuanian o corresponding to ō in other Indo-European languages is quite often unstressed (cf. nor-ėti, bog-inti, kriog-úoti, kor-ỹs, kor-ýtas, slog-à, slog-inti, sod-inti, tvorà, vilk-o, vorai, žmonių, žolė, etc.). Here one must also bear in mind factors such as analogy and generalization: the vocalism of forms with unstressed ō may have been generalized to forms with stressed ō as well. This may have given rise to o in stressed position. The reverse may also have occurred: the vocalism of forms with stressed ō, which regularly became uo, may have been transferred to the unstressed position in forms of the same root. This may have given rise to uo in unstressed position. In this way, the existence of doublets of roots becomes understandable; cf. Lithuanian šúo-lis and šõ-k-ti, as well as Lithuanian dúo-ti, Latvian duõ-t, and Old Prussian dā-t.

The circumflex accentuation of the Latvian words juõsta, úoga, úoleks, uõsis, etc., shows that these words formerly belonged to a mobile-accent paradigm in Latvian. Here, too, the vocalism of barytone case forms may have been generalized.

Analysis of the Baltic-language evidence shows that Proto-Indo-European ō and ā, as already mentioned, merged in the Baltic languages in unstressed position. The result of the merger in the Baltic languages, as in Slavic, appears to have been ā; cf. Latvian ā and Old Prussian (in the catechism dialect) ā or a. That Lithuanian o derives from an earlier ā (probably with a slight degree of labialization) is shown by Lithuanian dialects and old written sources (cf. M. Mažvydas’s catechism a[ō]); for further discussion, see p. 56 of this work.

Thus, the merger of Proto-Indo-European long vowels had begun in the Baltic languages as well. The reason that prevented the further merger of Proto-Indo-European long vowels ō and ā in the Baltic languages was apparently that Baltic ō (Proto-Indo-European ō) entered into a correlation with Baltic ē₂, which arose in stressed position from ē and, apparently, from ā³⁵ (cf., for example, Lithuanian piẽmõ and Finnish paimen from Baltic *pai-men, cf. Greek ποιμήν). The correlated development of Baltic ō and ē₂ is shown by their diphthongization (cf. Aukštaitian uo and ie, Samogitian ộu and ẹi) and subsequent raising (cf. Samogitian ū and ī) in Lithuanian dialects. Once this correlation had arisen, the long vowel of the Eastern Baltic languages vo-

³³ Z. Zinkevičius, Lietuvių dialektologija, p. 69, 86—87 and maps Nos. 29 and 59.

³⁴ DLKŽ (p. 687) assigns the word rúožas to the third accentual paradigm.

³⁵ H. Hirt, On the sliding and struck tone in the Indo-European languages, IF, 1, 1892, p. 32—41; J. Endzelin, Lithuanian-Latvian miscellanea. I. On Lithuanian-Latvian ie, „Lietuvių tauta“, II, Vilnius, 1911, p. 284—291; FBR, 12, 1932, p. 180—181; K. Būga, RR, II, p. 90, 277; Chr. S. Stang, A contribution to the problem of the diphthong ie in Lithuanian, „Die Welt der Slaven“, I, 2. Heft, 1956, p. 137—139; J. Kuryłowicz, Latvian-Lithuanian monophthongization ai, ei>ie, BPTJ, 15, 1965; P. Skardžius, On the change ai, ei>ie in Lithuanian, ZfslPh, 26, 1958, p. 375—382; И. ЭнДзелин, On the origin of Lithuanian-Latvian ie, ИОРЯС, 12, I, 1907, p. 40—66.

56

The vocalism system became symmetrical, with no empty slots left in it, and the shift of long vowel ō toward ā was halted.

However, the ō: ē₂ correlation arose only within the linguistic community of the East Balts (after the 4th–3rd centuries BCE). It seems there were no conditions for its emergence in Prussian, since here the diphthong ei (and ai) did not undergo monophthongization (cf. Prussian deiws, eit, etc.)³⁶.

For this reason, the merger of the Proto-Indo-European long vowels ō and ā in Prussian may have continued, that is, it may have extended from unstressed to stressed position as well. Admittedly, the fragmentary nature of the written monuments of Prussian does not allow us to say what the actual fate of Proto-Indo-European ō and ā was in Prussian. The following circumstances would suggest that the opposition between ō and ā may also have been eliminated here:

1) the identical graphic representation of the Proto-Indo-European ō and ā reflexes in the Elbing Vocabulary (graphemes o and oa);

2) the Proto-Indo-European ō and ā reflexes becoming ū or u after labials and, apparently, also after velar consonants in the catechism dialect.

All the changes to Proto-Indo-European ō and ā examined here took place in later stages of the development of the Baltic languages.

In addressing the question of the earliest relations between the Baltic and Slavic languages, it is important that the Baltic languages (and certainly the East Baltic languages) continued to distinguish Proto-Indo-European ō and ā. There can likewise be no doubt that the Baltic languages distinguished these vowels in the earliest periods of their history.

That the Baltic languages distinguished two long vowels in the earliest periods of their development is evident from loans of Baltic origin in the Finnic languages of the Baltic region, where Baltic ō is represented in many cases by Finnish uu (cf. Finnish huuta and Lithuanian šluota; Finnish kuuro and Latvian kuorns ‘deaf’), while Baltic ā is represented by Finnish uo (cf. Finnish vuota and Lithuanian óda, Finnish tuohi and Lithuanian tošis, Finnish luoma and Lithuanian lomà). Moreover, the loans show that Baltic ā had a certain degree of labialization (Finnish uo <*oo).

As already mentioned, in the Slavic languages the opposition between the reflexes of Proto-Indo-European long vowels ō and ā disappeared.

Phonological studies of recent years (A. Martinet, V. Zhuravlyov, et al.) have shown that, alongside the Proto-Indo-European short vowels o and a, the merger of the Proto-Indo-European long vowels ō and ā was one of the earliest changes in the Slavic vowel system. Phonological studies have also established that, alongside the process of consonant palatalization in the Slavic languages, another process took place—the labialization of consonants preceding labialized vowels. The phonological essence of these processes was that vowels transferred their distinctive features “acute—grave” to the preceding consonants³⁷. The merger of Proto-Indo-European long vowels ō and ā in the Slavic languages is indeed connected with one of these processes—with vowel delabialization and consonant labialization. The beginning of the process of vowel delabialization is considered to be the neutralization of o and a after labial consonants; that is, the sequences po: pa, bo: ba, me: ma, uo: ua, pō: pā, bō: bā... could merge and cease to be distinguished. “Grave” is a constitutive feature of labial consonants, and therefore became irrelevant for vowels immediately following labial consonants. The result of the merger of vowels o and a

³⁶ E. Berneker, Die preussische Sprache, I, Lautlehre, p. 16—18; R. Trautmann, Die altpreussischen Sprachdenkmäler, Göttingen, 1910, p. 137—141; J. Endzelīns, Senprūšu valoda, p. 29.

³⁷ V. K. Zhuravlev, The Formation of Group Singarmonism in Proto-Slavic, VYa, 4, 1961, p. 35.

5*

67

As a result of dissimilatory neutralization, it must have been a. Later, when the distinctive feature “non-bemolization” of the vowels o and ō shifted to the consonants preceding them, the vowels o and a coincided in other positions as well (after dental consonants, etc.)³⁸.

The exact chronology of the merger of ō and ā is unknown. It is clear only that during contacts between Slavic and Germanic and between Slavic and Greek (approximately from the first centuries CE), Slavic languages no longer had ō: Germanic and Greek long close ō is rendered in Slavic languages by the vowel ū; cf. OCS bukъvi ‘letters’ from Gothic bōkōs ‘books’, OCS Solomunъ from Greek Σολομών.

It may be assumed that before the neutralization of the opposition between the reflexes of IE ō and ā, Slavic languages retained these vowels as distinct. This may also be indicated by certain differences between their reflexes in endings. For example, the singular dative of the IE o-stem had the ending *-ōi, while that of the ā-stem had *-āi. If these endings had become phonetically identical as a result of the neutralization of the ō–ā opposition, both stems would have the same ending. This is not the case, however, since in Old Church Slavonic the o-stem has -u here (cf. rab-u), while the ā-stem has -ē (cf. žen-ē).

The original system of long vowels in Slavic, as in Baltic, was probably as follows:

ī ū ē ō ā

Both Baltic and Slavic inherited this triangular system from the northern-area dialects of the IE linguistic community (cf., for example, the entirely identical original system of long vowels in Germanic languages) and probably retained it unchanged during the first centuries of their independent existence.

Since the problem of the earliest relations between Baltic and Slavic languages concerns the relations between Baltic and Slavic dialects that existed during the first centuries after their separation from the IE linguistic community, it is self-evident that the preservation of the IE ō–ā opposition in Baltic languages and its elimination in Slavic languages can be regarded neither as an argument for nor as a counterargument to the theory of Baltic–Slavic linguistic unity. The fate of the vowels ō and ā characterizes the later stages and tendencies in the development of Baltic and Slavic languages.

A feature common to the subsequent development of IE ō and ā in Baltic and Slavic languages was the tendency of long vowels to merge. Through this feature, Baltic and Slavic languages belong to the same area as Germanic and Indo-Iranian languages.

However, Baltic and Slavic languages also differ considerably in the fate of IE ō and ā. In Slavic languages, the IE long vowels merged everywhere, whereas in Baltic languages they merged only in unstressed positions. Baltic languages (Eastern Baltic, without doubt) preserved the distinction between long vowels inherited from the IE linguistic community right up to the present day (Lith. uo: o, Latv. uo: ā).

Baltic and Slavic languages differ not only in the outcomes of the development of IE long vowels, but also in the developmental tendencies of their long-vowel systems. The long-vowel system of Baltic languages tended to become narrower and more labialized; as a result, Baltic ō and ē₂ became uo and ie in Lithuanian, while Baltic ā and ē changed to o and ė. In Slavic languages, by contrast, the long-vowel system showed a marked tendency toward expansion and delabialization. As a result, Slavic ō and ū became the vowels a and y.

³⁸ Б. К. Журавлев, op. cit., p. 35–36.

68

The development of the IE diphthong eu

The reflex of the IE diphthong eu in Baltic and Slavic languages has not yet been definitively clarified.

In some cases, the IE diphthong eu (ēu) is reflected by Baltic (Eastern) jau and Slavic ju (from Slavic *jau), cf.:

Lith. liaukà and liauka (Karsakiškis, Panevėžys), plural liaukos ‘a certain disease of pigs’, liaūkti ‘to run, flow abundantly’, Latv. ļauka and laukas¹ ‘pig disease’: Gr. λεύκη ‘eruption’ (this word is thought to have been borrowed from Baltic by the Finns, cf. Finnish leuka ‘chin’²; this would indicate that at the time of borrowing the Balts had only eu);

Lith. liaupsė, acc. sing. liaūpse (from *leub-s-), OCS ljubъ ‘dear’: Goth. liufs, OHG liob ‘dear’ (Goth. iu, OHG io from IE eu);

Lith. liaudis and liaudžia, acc. sing. liaudžią (Valkininkai, Rodūnia), Latv. plural ļaudis and dialectal ļauži, OCS ljudьje ‘people’, Old Rus. ljudi ‘the people’: OHG liut ‘people’ (iu from IE eu);

Lith. siāuras, Latv. sāurs: Gr. εὔριπος ‘strait’ (Gr. εὐ- may derive from *seu-);

Lith. šiāurė (from *ḱeu-r-): Old Rus. séverъ (from *k’ēu-er-).

There are quite a number of cases in which the eu diphthong of other IE languages corresponds to au in Baltic languages and u in Slavic (from Slavic *au), cf.:

Lith. baūsti: OCS bljusti ‘to observe, perceive’, Gr. πεύθομαι ‘I perceive, experience’, Goth. ana-biudan ‘to command, oblige’ (Goth. iu from IE eu);

Lith. naūjas, Pruss. nauns, Lith. *navas (a form reconstructed by linguists from the village name Nav-ikai and the surname Nav-ikas): Goth. niujis, Gr. νέος ‘new, young’ (from *νέFος), Hitt. neua-; however, OCS novo, Lat. novus, Gaulish (place-name) Novio-dūnum (cf. Lith. Naujamiestis) indicate the existence of IE *nou-;

Lith. laūkas and láukas ‘an animal with a white muzzle or white forehead’, Latv. lauks: Gr. λευκός ‘white’ (the Finns borrowed this word from Baltic, cf. Finnish laukki ‘white spot on an animal’, which has the diphthong au in its root);

Lith. raūdas, Latv. raūds, OCS ruda ‘metal’, Rus. румяный (Rus. rum- from *raud-m-): Old English réod ‘reddish’, Gr. ἐρεύθω ‘I blush’.

The material presented so far shows that it is not easy to state straightforwardly what the reflex of IE eu was in Baltic and Slavic languages—Lith. jau and Slavic ju, or Lith. au and Slavic u.

Some researchers (J. Schmidt³, G. Ul’janov⁴, E. Berneker⁵, V. Por že zin-

¹ K. Širvydas’s form laukos (C. Szyrwid, Dictionarium trium linguarum, Quinta editio, Vilnae, 1713, p. 136), which entered the dictionaries of K. Mülenbach–J. Endzelin and E. Fraenkel (K. Mülenbach, Latviešu valodas vārdnīca. Redigējis, papildinājis, turpinājis J. Endzelīns, II, Rīgā, 1925–1927, p. 426; E. Fraenkel, LEW, p. 361), should most likely be emended to liaukos, because: 1) the form laukos (with hard l-) is not known from elsewhere; 2) the third edition of K. Širvydas’s dictionary (C. Szyrwid, Dictionarium trium linguarum, Tertia editio, Vilnae, 1642, p. 157, 528) has Laukos (= liaukos). See K. Pakalka, Konstantino Širvydo žodyno „Dictionarium trium linguarum“ lietuviškų žodžių registras (LTSR MA Lietuvių kalbos ir literatūros institutas, 1964, typescript).

² Some researchers (E. Nieminen, Y. Toivonen) deny the Baltic character of Finnish leuka; see A. Sabaliauskas, *Baltų ir Pabaltijo suomių kalbų santykiai*, LKK, 6, 1963, p. 121.

³ J. Schmidt, Was beweist das e der europäischen Sprachen für die Annahme einer einheitlichen europäischen Grundsprache?, KZ, 23, 1877, p. 352–353.

⁴ Г. Ульянов, Основы настоящего времени в старославянском и литовском языках, Варшава, 1888, p. 36, 64, 100.

⁵ E. Berneker, Von der Vertretung des eu im baltisch-slavischen Sprachzweig, IF, 10, 1899, p. 145–167; plg. AfslPh, 25, 1903, p. 487.

69

ski⁶, E. Cupica⁷, J. Endzelynas⁸, H. Pedersen⁹ and others) believed that Indo-European eu regularly became jau in the Baltic languages, and ju in the Slavic languages. H. Pedersen added that the single word liaukà (cf. Greek λεύκη) was supposedly sufficient to prove that Indo-European eu became jau in Lithuanian¹⁰.

Since the change of Indo-European eu to Lithuanian jau was considered regular by the researchers mentioned above, all cases in which Lithuanian (and other Baltic) au corresponded to eu in other Indo-European languages had, in their view, to be exceptions. It was therefore necessary to find ways of explaining them.

Having investigated this question, V. Porzeziński concluded¹¹ that, in the Baltic linguistic community, soft consonants before eu, under the influence of consonants that had stood before u (where they were hard), had been replaced by hard consonants in some words. In this way, the group “soft consonant + diphthong eu” became “soft consonant + diphthong au” in Lithuanian and Latvian, and “soft consonant + diphthong eu” in Prussian. The group “hard consonant + diphthong eu” became “hard consonant + diphthong au.” E. Berneker¹² agreed with V. Porzeziński’s view, pointing out that this could explain many exceptions in which au occurs instead of the expected jau.

H. Pedersen¹³ tried to explain the origin of au quite differently. In the word tautà, in his view, t…t could have changed to t…t through the assimilatory influence of the second t. The word laũkas could have acquired its vowel o from another word of the same root. And naũjas might have arisen in place of *niaujas through dissimilation or by some other means.

E. Cupica¹⁴ had already explained the emergence of the au diphthong in tautà in the same way in 1907. These explanations for the emergence of au instead of the expected jau are not entirely successful. For instance, V. Porzeziński did not prove why soft consonants were replaced by hard ones in some cases but not in others. His theory does not explain why, for example, Prussian peuse did not become *pause, although a word in which p- before u was hard existed alongside it in the Baltic linguistic community (cf. Lithuanian pušis ‘pine’). V. Porzeziński’s theory would lead us precisely to expect such a change. One may, of course, suppose that assimilation occurred in the word tautà, as E. Cupica and H. Pedersen claim, but it is difficult to substantiate this with evidence.

O. Wiedemann¹⁵, K. Johansson¹⁶, J. Mikola¹⁷, A. Meillet¹⁸ and others believed that jau arose in the Baltic languages only from Indo-European ēu, whereas the diphthong au reflected Indo-European eu. In their view, in the Slavic languages ju could have arisen from both Indo-European ēu and Indo-European eu. Many Lithuanian examples in which jau derives from the short Indo-European diphthong eu are consistent with these researchers’ view.

J. Zubat¹⁹ tried to establish a criterion for determining where Indo-European eu became jau and where it became au. In his view, Indo-European eu became jau and ju, respectively, in the Baltic and Slavic languages before front vowels, whereas before back vowels it changed to au and u, respectively.

⁶ В. Поржезинский, К истории форм спряжения в балтийских языках, Москва, 1901, p. 79–80.

⁷ E. Zupitza, Lit. naũjas, KZ, 40, 1907, p. 250–251.

⁸ И. Эндзелин, Славяно-балтийские этюды, p. 79.

⁹ H. Pedersen, Lit. IAU, StB, 4, 1934–1935, p. 150–154.

¹⁰ H. Pedersen, op. cit., p. 152.

¹¹ В. Поржезинский, op. cit., p. 80.

¹² E. Berneker, AfslPh, 25, 1903, p. 489.

¹³ H. Pedersen, op. cit., p. 150–154.

¹⁴ E. Zupitza, op. cit., p. 254–255.

¹⁵ O. Wiedemann, Das litauische Präteritum, p. 17–18, 32, 184.

¹⁶ K. Johansson, Indische Etymologien, IF, 8, 1897, p. 172 (išnaša).

¹⁷ J. Mikkola, Urslavische Grammatik, I, p. 57–61.

¹⁸ A. Meillet, Le traitement de *-eu- en letto-lituanien, BSL, 25, 1924, p. 174–175.

¹⁹ J. Zubatý, AfslPh, 20, 1898, p. 396–397.

70

J. Zubat’s explanation, with a slight modification, was endorsed by J. Endzelin²⁰. In his view, the Baltic languages’ au developed from Indo-European eu probably not before all back vowels, but only before stressed ones; cf. the examples given by J. Endzelin: Lithuanian tautā and Gothic þiuda (the -d- from -t- shows that these words had been stressed on the ending since ancient times), Lithuanian laũkas and Greek λευκός.

The hypothesis of J. Zubat and J. Endzelin²¹ does not explain Lithuanian forms in which au already occurs before back vowels; cf., for example, liaukà, siaũras, kiáuras, kiáutas, etc. Moreover, the diphthong au is already quite often found in words with fixed root stress; cf. kiaũlė, kiaũšis, liáudis, šiáurė, etc.

E. Berneker²², E. Zupitza²³, K. Būga²⁴, G. Shevelov²⁵ and others tended to believe that the Baltic diphthong au derived from an Indo-European diphthong with a different ablaut grade from that of the diphthong eu. J. Endzelin²⁶ did not rule out the possibility of ablaut either. Thus, for example, Lithuanian tautā should be derived not from Indo-European *teutā, but from Indo-European *toutā or *tautā.

J. Otrębski’s view²⁷ needs to be mentioned separately. According to it, the regular reflex of Indo-European eu in (Eastern) Baltic languages is au preceded by a hard consonant. In the Slavic languages, Indo-European eu became *jau (from which ju later developed) as a result of early consonant palatalization before front vowels. Since Indo-European eu changed in the Baltic languages before consonant palatalization (that is, consonants before front vowels and before j must still have been hard), the diphthong into which eu developed must have been au preceded by a hard consonant. Here J. Otrębski supports one hypothetical proposition (eu developing into au) with another equally hypothetical one (the hardness of consonants and their palatalization). His hypothesis is contradicted by many Lithuanian and Latvian words whose diphthong au clearly derives from Indo-European eu.

In the Baltic and Slavic languages, cases are fairly common in which the same root has two variants—one with jau, the other with au (Slavic ju and u). The Lithuanian kiáunė and Latvian caune (c- from *kj-) have jau in the root, whereas Prussian caune and Russian куна́ have au. Some Slavic languages have ju (cf. Old Russian плюща “lungs” and Old Czech plučé), while others have u (cf. Old Church Slavonic plušta and Polish płuca). Baltic examples show that the root vocalism here is of the o-grade; cf. Lithuanian plaučiai, Latvian plāušas, and Prussian plauti. If we accept that Indo-European eu regularly developed into au and u, respectively, in certain cases in the Baltic and Slavic languages, these facts compel us to conclude that this development took place very late, in the individual languages. It goes without saying that this could not have been the case. The Finnish loanword laukk(i) shows that, at the time of Baltic–Finnic contact, the Baltic languages already had au in place of eu in other Indo-European languages.

²⁰ И. Эндзелин, Славяно-балтийские этюды, p. 83; J. Endzelin, Lettische Grammatik, p. 40—41.

²¹ Šią koncepciją pripažino V. Vondrakas ir A. Vajanas, W. Vondrák, Vergleichende slavische Grammatik, I, p. 41; A. Vaillant, Grammaire comparée des langues slaves, I, p. 122—124.

²² “The few cases where au and u seem to correspond can be dismissed by assuming a different ablaut,” — E. Berneker, IF, 10, 1899, p. 167.

²³ E. Zupitza, op. cit., p. 250—251.

²⁴ K. Būga, RR, II, p. 348.

²⁵ G. Y. Shevelov, A Prehistory of Slavic, p. 272—275.

²⁶ И. Эндзелин, Славяно-балтийские этюды, p. 82; J. Endzelin, Lettische Grammatik, p. 41.

²⁷ J. Otrębski, LP, 4, 1953, p. 308—312; Я. С. Отрембский, Славяно-балтийское языковое единство, ВЯ, 5, 1954, p. 31—32.

71

If we were to assume that the change of Indo-European eu in the Baltic and Slavic languages was a regular process and took place within Baltic and Slavic linguistic communities, then there is no way that some languages could have already developed respectively ju, while others had au.

Such variation shows that either the o-grade vocalism was inherited from the Indo-European linguistic community, or it arose in individual languages through structural analogy. Since, for example, alongside Prussian caune there was no form with a different ablaut grade, and structural analogy was therefore impossible, the Prussian word could only have developed from the root *kou-.

In the Baltic (and Slavic) languages, Indo-European eu had already changed into jau, and an ablaut series jau: au: u (Indo-European eu: ou: u) had formed. In many cases, a soft consonant displaced the hard consonant that had stood before the u-grade. For example, in Lithuanian griūti, instead of the expected *grūti, the form arose by analogy with griáuti < *grēu-. The soft consonant was generalized also in forms in which o had arisen as a result of ablaut; cf. griovė instead of the expected *grovė < *grōu-.²⁸

The reverse process can also be imagined: the hard consonant that had stood before the u-grade could have displaced the soft consonant. For instance, one might suppose that Latvian laukas (with hard l-) arose instead of *liaukas by analogy with a Baltic word of this kind (cf. Prussian luckis “splinter”), in which hard l- was regular. However, such an assumption is precluded by words such as Prussian peuse and Lithuanian pušis, Prussian caune and Lithuanian kiáunė. Although pušis also existed alongside peuse in the Baltic linguistic community, peuse did not thereby become *pause. Likewise, by analogy Prussian caune could not have arisen from *kiaune, since there is no word in which the consonant would have stood before the u-grade.

All this shows that the alternations with au already found in words of the type Lithuanian kiáunė: Prussian caune must be explained by phonological and morphological factors.

Finnish loanwords of Baltic origin may also support this conclusion. One may suppose that the Finns borrowed the words leuka and laukki at the same time, since both are associated with names for the characteristics of living beings. These loans show that, at the time of borrowing, 1) Baltic languages already had au in place of the eu of other Indo-European languages, and 2) the diphthong eu had not yet changed into jau. The diphthong eu changed into jau much later than eu changed into au. Thus, the processes eu > au and eu > jau were most likely unrelated both chronologically and structurally.

The facts mentioned above seem to make it possible to derive some Baltic and Slavic words having au or u in place of the eu found in other Indo-European languages from an o-grade root; cf.:

*kēu-g-: Lithuanian kiáugė and *kōu-g-: Lithuanian káugė, Latvian kaudze, Old Icelandic haugr “hill”; *kēu-n-: Lithuanian kiáunė, Latvian caūne and *kōu-n-: Prussian caune, Russian куча; *leu-k-: Greek λευκός, Gothic liuha “light” and *lou-k-: Lithuanian laukas, Latvian lāuks, Latin lūc-idus “bright,” Greek λούσσον (-σσ from *-kj-) “white spruce heartwood”; *leu-k-: Lithuanian liaukà, liaūkti, Latvian lauku, Greek λεύκη and *lou-k-: Latvian laukas; *leu-|*lēu-: Lithuanian liáutis, liav-ōnas, Latvian laut “to let,” Czech leviti “to release,” Gothic lēwjan “to leave to fate” and *lou-|*lōu-: Lithuanian lavōnas, Prussian au-lāut “to die”; *pleu-: Old Russian pljuća “lungs,” Old Czech pljuće, Greek πλεύμων and *plou-: Lithuanian plaučiai, Latvian plaušas, Prussian plauti, Old Church Slavonic plušta, Polish płuca; *reu-d-: Old English réod “reddish,” Greek ἐρεύθω “I blush” and *rou-d-: Lithuanian raūdas, raud-ōnas, Latvian rauds, Old Church Slavonic ruda “ore,” Russian румяный (rum- from *raud-m-); *sreu-: Lithuanian sriaumuō, Greek ῥεῦμα “stream” (ῥ- from *sr-) and *srou-: Lithuanian srav-ėti, sraumuo, Latvian strāume, Old Russian strumenь “brook” (stru- from *srau-);

²⁸ Perhaps for similar and other, less clear reasons, the r is soft in some Lithuanian dialects (in certain forms), but hard in others; see Z. Zinkevičius, Lietuvių dialektologija, pp. 153–156.

72

*teut-: Gothic piuda ‘people’, Illyrian (personal name) Teuta and *tou-t-: Lithuanian tautà, Latvian tāuta, Old Prussian tauto, Oscan touto ‘state’, Welsh tūd ‘country’ (tū- from *tou-), Old Irish tuath ‘people’ (tua- from *tou-).

It should be added that only some Baltic and Slavic words with root vocalism o may go back to the period of Indo-European linguistic unity. This applies, for example, to the Lithuanian word tautà and its cognates: some Indo-European dialects may have had the *teu- grade, while others had the *tou- grade. However, variants of most words with the e- and o-vocalism grades of ablaut may have arisen already in the Baltic and Slavic languages, since ablaut in the Baltic languages (for example, Lithuanian) is still productive today. For instance, some Baltic dialects had the word kiáunę with the e-grade of ablaut, while others had the o-grade (cf. Old Prussian caune).

In some Lithuanian words, the diphthong au may in fact stand for Indo-European eu. Present-tense forms such as keičia, sėka (pasaka), etc., whose roots have e-grade vocalism, show that forms such as baudžia, plaukia, smaukia must likewise have e-grade vocalism, since iterative-causative verbs generally have o-grade vocalism; cf. kaityti, sakyti and baũdinti, plaũkyti, smaũkyti. That the old root diphthong in these present-tense forms was eu is also shown by corresponding words in other Indo-European languages; cf. Lithuanian baudžia and Old Church Slavonic bljūsti, Lithuanian plaukia and Old High German fliogen ‘to fly’ (io from eu), Lithuanian smaukia and Old Scandinavian smjūga ‘to force one’s way in, slip’ (ju from eu). In the old forms *beudja, *pleukja and *smeukja, the root diphthong eu was replaced by au not because Indo-European eu regularly became au in the Baltic languages, but apparently because the vocalism of these present-tense forms was assimilated to that of other forms of the same root with o-vocalism.

Thus, it may be supposed that, apart from a few verbs in which ou most likely arose by analogy in place of the older eu, the Indo-European diphthong eu did not become au in the Baltic (and Slavic) languages. All Baltic and Slavic words whose roots have au or u where other Indo-European languages have eu most likely derive from forms with o-grade vocalism.

The Indo-European diphthong eu regularly became jau or ju in the Baltic and Slavic languages. This phonetic development is regarded as a Baltic-Slavic innovation, indicating close relations between these languages and supporting the theory of Baltic-Slavic linguistic unity.

The development of Indo-European eu into jau or ju can be considered an argument for Baltic-Slavic linguistic unity only if Indo-European eu became jau or ju during the period of Baltic-Slavic unity. However, there seems to be no basis for treating this phonetic development as so ancient.

It cannot be stated with certainty that the development of Indo-European eu into jau was common to all the Baltic languages. It is difficult to determine from the Old Prussian monuments what the reflex of the Indo-European diphthong eu was in Old Prussian. Spellings of the type bleusky in the Elbing Vocabulary are ambiguous. Some scholars believe²⁹ that the spelling eu denotes the diphthong au after a palatalized consonant. Thus, Indo-European eu also became jau in Old Prussian. R. Trautmann also draws attention to similar spellings of jau in old Lithuanian texts, e.g. gereus instead of geriaus in the writings of M. Daukša³⁰. Others hold the view³¹ that Old Prussian eu actually represents the diphthong eu, not jau. According to them, Indo-European eu was preserved unchanged in Old Prussian.

²⁹ R. Trautmann, Die altpreussischen Sprachdenkmäler, p. 148—149; J. Endzelīns, Prūsų tekstu grafika, FBR, 15, 1935, p. 97—98; Senprūšu valoda, p. 30; V. Mažiulis, Prūsų kalbos paminklai, Vilnius, 1966, p. 54—55.

³⁰ R. Trautmann, op. cit., p. 148.

³¹ J. Otrębski, LP, 4, 1953, p. 308—312.

73

All researchers cite as an argument Prussian iaukint (= Lithuanian jaukinti), which supposedly shows that in Prussian the Indo-European eu likewise became iau. This argument, however, is unreliable, since the word may also have had the root *ouk- (not *euk-). The o-vocalism of the root is precisely appropriate to the causative meaning of the verb Prussian iaukint and Lithuanian jaukinti. That j- here may be prothetic is shown by Old Church Slavonic učiti (u- from *au-) and Sanskrit ókah ‘house’.

The words written with the digraph eu (ew) in the Elbing Vocabulary are as follows: bleusky 286, geauris 757 (cf. Lithuanian giaurÿs LKŽ, III, 282), gleuptene 247, keutaris 762, keuto 156, 497 (cf. Lithuanian kiáutas), peuse 597, skewre 685 (cf. Lithuanian kiáglė). What strikes one is that the graphemes i, y, or g—which could be interpreted as signs of consonant softness—never occur before this digraph. We would expect precisely this kind of graphic marking if Prussian had the sequence iau in these words. It is true that the mentioned graphemes sometimes also occur in the Elbing Vocabulary before front vowels; cf. plieynis 38 alongside pleynis 75. In this connection, attention should be drawn to the spelling of keuto: on both occasions it is recorded identically, that is, with the digraph eu. All this would seem to indicate that the words in question in the Elbing Vocabulary had the diphthong eu.

On the assumption that the Prussian sequence *sj became š (cf. schuwikis 496, schu-meno 507, schuwan 471), it is generally assumed that in Prussian j was lost after other consonants as well (dentals, velars), palatalizing those consonants³². Certain spellings, however, especially the use of the grapheme g, seem to indicate that a sound j is nevertheless represented here; for example, cf. angurgis 565, ansalgis 506, kargis 410, saligan 468, wargien 525 (cf. cugis 518 = Lithuanian kūjis, where the grapheme g represents the sound j). It is highly doubtful that the grapheme g here denotes consonant softness³³. As Lithuanian and Latvian show, j need not have disappeared simultaneously in all positions: in some places (e.g. after t, d) it disappeared earlier, in others (e.g. after labial consonants) later; in some places the outcome of its loss was one thing, elsewhere another. Therefore, the change of the sequence *sj does not at all show that other consonant sequences with j also underwent a change. Indeed, even proponents of the “jotated” consonant palatalization theory do not entirely rule out the possibility that j still existed in Prussian after labial consonants³⁴. In this case, the Elbing Vocabulary has the following spellings: piuclan 547 (word-initially), kalabian 424, tarbio 325, knapios 268, wupyan 9, wormyan 465, samyen 235, clawio 375, perwios 281. In determining whether Prussian still had the diphthong eu or whether its dialects had already developed jau in its place, it is not so important what the spellings bia, bio, piu, pio, pya, mya, mye, and wio represent: a labial consonant followed by j or a palatalized labial consonant before a back vowel. What matters is that the graphic representation of this position differs from the graphic representation of the analogous position in peuse: if the Prussian diphthong eu had become jau, then here too (that is, in peuse) we would expect the same graphic representation of a labial consonant (with j or palatalized) before a back vowel (*pjau, *pyau or *pyeu). Does the different graphic representation reflect only a different phonetic complex (pja or p’a... in the examples above, but pe in the word peuse)? When considering the phonetic interpretation of the spelling in Prussian monuments, it is also worth recalling how J. Endzelynas explains the spelling pvienkts in the Second Catechism. He conjectures that

³² J. Endzelins, Prūšu tekstu grafika, pp. 96–98; T. Milewski, Przyczynki do gramatyki pruskiej, „Slavia Occidentalis“, 15, 1936, pp. 102–108; V. Mažiulis, op. cit., pp. 54–55.

³³ Cf. J. Endzelynas’s tentative interpretation of the word wargien: “je te vēl iesprausts aiz g varbūt tāpēc, lai g-burtu te kas nelasītu ka vēlāro balsīgo šļēdeni”, — J. Endzelins, op. cit., p. 97.

³⁴ J. Endzelins, op. cit., p. 98; V. Mažiulis, op. cit., p. 55.

74

here, after a labial consonant, j could have arisenj³⁵. Thus j could be represented even before front vowels.

In exactly the same way, other consonants (with j or palatalized) are represented before back vowels in the Elbing Vocabulary; cf. maldian 438, median 586, medione 699, brunyos 419, dragios 386, and so on. They should likewise be represented in the words keutaris, keuto, skевre, geauris if the consonants there were before back vowels. In addition, attention should be drawn to the fact that in the Elbing Vocabulary the graphemes i, y are generally not written after consonants before front vowels (cf. deywis 1, deynayno 5, metis 166, etc.), with, of course, a few exceptions (such as plienis 38 alongside pleynis 75, geyте 339, etc.). Thus these words with the digraph eu (ew) most likely also had a front vowel and, consequently, the diphthong eu.

At the same time, it becomes clear that there is no reliable evidence showing that the change of Indo-European eu to the diphthong jau was already a common Baltic development.

At the time of the change of the Indo-European diphthong eu, entirely different phonological situations had developed in Baltic (Eastern Baltic) and Slavic: in the Baltic languages the change was associated with one set of phonological processes, while in the Slavic languages it was associated with entirely different ones.

The phonetic change of eu to jau presupposes the widening and backing of the vowel e in the position before u (apparently first where that vowel bore the acute accent), i.e. e→ā. That such a phonetic change is possible is shown by Lithuanian, where there are instances of the vowel e widening and backing before u or v; cf. the kliāvas and klēvas of many dialects. Before v (u), Samogitians in many areas pronounce a with a following palatalized consonant. Here, even in place of ē, which developed from ei, a is pronounced; cf. kr’āus ‘crooked’³⁶, š’āvā ‘askew’. The a that developed from e before u may phonetically have been very close to the vowel a, which likewise occurred before u (from the ancient diphthongs āu and ōu); after j, the vowel a may even have merged completely with ā. When the vowel o became a, the vowels a and ā became variants of the phoneme /a/. In this way, the variant ā could become phonemicized; in other words, instead of a manifestation of the phoneme /e/ in the segment CāuC (C denotes consonants), a manifestation of the phoneme /a/ arose, i.e. CāuC → C’auC.

Consonants occurring before front vowels must already then have differed from consonants occurring before back vowels. The former were palatal, the latter velar. The position before j was probably neutral. Following the changes mentioned above, palatal consonants came to stand before the back vowel a. This situation could not have persisted for long, since it ran counter to the entire system, in which palatal consonants occurred only before front vowels, while back vowels occurred only after velar consonants. This anomaly could have been eliminated in two ways: 1) C’auC had to become CauC (only a velar consonant is possible before a back vowel); 2) C’auC had to become CjauC (before a back vowel, a palatal consonant is possible only before j). Realizing the first possibility involved greater difficulties than the second: consonants in this position would have had to shift from the class of palatal consonants to the class of velar consonants. Although the classes of palatal and velar consonants stood in a relation of complementary distribution, realizing the first possibility would nevertheless have constituted a radical shift in the system. This is apparently why the second possibility was realized; moreover, its realization was a compromise, since the position before j was neutral, and thus

³⁵ J. Endzelīns, op. cit., p. 96.

³⁶ Z. Zinkevičius, Lithuanian Dialectology, p. 51.

75

Here the consonants were neither palatal nor velar (or rather, both palatal and velar). The insertion of a yod eliminated the contradiction that had arisen in the system.

Incidentally, in some cases, especially where a root variant with the diphthong au (and ou) existed alongside it, a root of the type C’auC- could easily have been replaced by a root of the type CauC-. This is apparently what happened in the verbs; cf. daužia instead of the regular *djaūžia, following daužyti.

In Proto-Slavic, the monophthongization of the diphthong eu in the presence of a preceding palatal consonant was associated with the formation of complex group phonemes. As complex group phonemes formed, sonorants (including u) joined the phonemes preceding them and became syllabic, since the syllable-forming element was usually the final element of the group phoneme. Later, when the differential features “openness: closeness” were transphonologized as the differential features “length: shortness,” the vowels, by then no longer syllabic, became redundant indicators of palatalization and therefore readily disappeared.

When the group phoneme C’eu became C’+u, a unique case arose in the Proto-Slavic phonological system: a palatalized consonant occurred before a non-palatalized vowel. The contradiction between the palatalized consonant and the following non-palatalized vowel could be eliminated through the formation of neutral group phonemes. The resulting neutral consonants (neither palatalized nor non-palatalized), for example b, must have merged with the neutral consonants already present in the system (for example, b from bjV) and produced the same reflexes. This is apparently why the reflexes of consonants occurring both before j and before the diphthong eu are identical in all Slavic languages³⁷.

Thus, the phonological interpretation of the transformation of IE eu into jaū in the Baltic languages and ’u in the Slavic languages finally confirms the intuitive conjecture that this change was an entirely independent development in the Baltic and Slavic languages. The IE diphthong eu changed in the Baltic and Slavic languages not only under different systemic conditions and under the influence of different systemic factors, but the phonetic nature of the change also differed: in the (eastern) Baltic languages eu became jaū, whereas in the Slavic languages it became simply ’u. The fate of the IE diphthong eu is a clear example of the parallel development of the Baltic and Slavic languages. This example shows that the general developmental tendencies shared by the Baltic and Slavic languages, sustained by geographical contact, continued to exist in later stages of their history as well.

The fate of IE syllabic sonorants

IE syllabic sonorants l, r, m, n (hereafter denoted R) became il, ir, im, in or, less often, ul, ur, um, un (hereafter denoted iR and uR, respectively). Their identical fate brings the Baltic and Slavic languages especially close.

IE short sonorants R became iR or uR (with a circumflex); cf.:

a) Lithuanian vilkas, Latvian vilks, Old Prussian wilkis, Old Church Slavonic vlъkъ, Russian волк, Serbo-Croatian vuk (from Slavic *vьlkъ): Sanskrit vṛ́kaḥ, Gothic wulfs, Greek λύκος, Latin lupus;

Old Prussian kirsnan “black” (cf. Lithuanian Kirsna, a tributary of the Šešupė), Old Church Slavonic čьrnъ, Russian чёрный (from Slavic *čьrnъ): Sanskrit kṛṣṇáḥ “black”;

Lithuanian šimtas, Latvian simts, Old Church Slavonic sъto (probably a loanword from Iranian languages): Sanskrit śatám (śa- from IE *k’ṃ-), Greek ἑκατόν (-κα- from IE *k’ṃ-), Tocharian kānt(e), Latin centum, Gothic hund;

³⁷ В. К. Журавлев, Развитие группового сингармонизма в праславянском языке, Минск, 1963, р. 29–41.

76

Lith. mintis (acc. sing. minti), at-mintis, Lat. mit “to think, guess,” Pr. po-minisnan “memory,” OCS pa-mętъ (ę from bn): OInd. matíh (-a- from n) “intention, desire; prayer,” Lat. mens mentis “spirit, mind,” Goth. munš “thought”;

b) Lith. kulnas (and kulnis, acc. sing. kūlni), Lat. kulna “heel,” Pr. kulnis “ankle,” OCS klъka “hollow of the knee,” Bulg. kъlka “thigh,” Serbo-Croatian kūk (from Sl. *kъlk-): Lat. calx “heel,” Toch. A kolye “thigh”; Pr. curwis “bull”: Lith. kárvė, OCS krava, Russ. корова, Gk. κέρας “horn”: the Lithuanian, Slavic, and Greek words have a full-grade root *kerH-/*korH-; Lat. cornu “horn,” Goth. haúrn “horn” from *kṛH-; Lith. ungurys (gen. sing. ūngario): Pr. angurgis, Russ. угорь (у- from *ang-), root *ng-/*ong-.

PIE long syllabic sonants R̄ (from PIE *RH) developed into iR or uR (with acute accent), cf.:

a) Lith. vilna, Latv. vīna, Pr. wilnis and wilna “skirt; coat,” OCS vъlna, Serbo-Croatian vūna (from Sl. *vъlna): OInd. ūrṇā, Av. varəna, Gk. λῆνος, Lat. lāna, Goth. wulla (from Germ. *wulna); Lith. žirnis, Latv. zirnis, Pr. syrne “grain,” OCS zrъno, Russ. зерно (from Sl. *zьrno): Lat. grānum (rā from PIE *ṛH), Goth. kaúrn “grain”; Lith. pintì, Latv. pīt, OCS peti (ę from bn, but may also be from en, cf. OCS mrěti, pres. mьrǫ), pres. pьnǫ: Arm. henum “I spin, weave,” Lat. pendere “to weigh, assess”; also cf. Lith. pántis; b) Lith. gurklỹs (gen. sing. gūrkliō), Pr. gurcle “throat” (from Baltic *gur-tl-), O. Russ. gъrlo (Russ. горло), Pol. gardło (from Sl. *gъr-dl-): OInd. gūrtáh “expected, pleasant,” Lat. gurges “whirlpool,” gurgulio “windpipe”; cf. also Lith. gūrkšnis, gerklė, gargaliuoti; Lith. mùlkis, Latv. muļķis: OInd. mūrkháh “foolish.”

PIE sonants R in Baltic and Slavic developed into iR and uR not only in preconsonantal position, but also in prevocalic position, cf.:

Lith. pilìs, Latv. pils, Pr. (place names) Pill-aw, Pille-kaym: OInd. pūḥ (acc. sing. pūr-am) “defensive wall,” Gk. πόλις “city”; Lith. girìà, Latv. dzira (and dzire) “forest”: OInd. giríh “mountain”; Pr. garian, garrin “tree” and OCS gora “mountain” have a different ablaut grade (*gor-); Lith. gun-ioti, gūn-yti “to drive about, chase,” Pr. gun-twei, OCS gъn-ati “to drive,” Lith. giñti, Latv. dzīt (i from in), OCS žęti (ę from ьn or en), pres. žьnǫ “I reap (crops)”: Lith. gan-ýti, Latv. gan-īt, OCS gon-iti.

The material presented shows that PIE syllabic sonants R developed in Baltic and Slavic into iR in some cases, and uR in others. The question arises as to what determined the distribution of the anaptyctic vowels i and u. There is no uniform answer to this question. J. Endzelīns¹, J. Mikola², K. Būga³, R. Trautmann⁴, and others tended to explain the emergence of i or u by ablaut. According to this conception, iR arose from eR,

¹ I. Endzelin, Slavic-Baltic Studies, pp. 3—24.

² J. J. Mikkola, Urslavische Grammatik, I, pp. 68—71; cf. Woher lit. jau und slav. ju?, IF, 16, 1904, p. 99.

³ K. Būga, RR, II, pp. 128—129, 260—264.

⁴ R. Trautmann, Viennes baltų slavėnų gramatikos bruožas, „Švietimo darbas“, 1922, 3—6, Kaunas, p. 271—272; Ein Kapitel aus der Lautlehre der baltisch-slavischen Sprachen, „Slavia“, 2, 1923.

77

from the zero grade of the diphthong eR, and uR—from the zero grade of oR, the diphthong oR. Thus, in the Baltic and Slavic languages, the reduced e and o became the full vowels i and u.

In the Baltic languages, in many cases iR alternates with eR, and uR with aR; cf. girtas: gérti (but cf. gurklys), gilti: gėlti, kilti: kėlti; kūlti: kūlti, kūmpas: kampas, Old Prussian curwis: Lithuanian karvė. However, there are cases where iR may occur alongside aR, and uR alongside eR; cf. miltai: málti, Lithuanian minia: Gothic manags ‘many’, Lithuanian giria: Old Prussian garian, Old Church Slavonic gora, Old Prussian gulsennin ‘pain’: Lithuanian gėlti. Forms of the same root may exhibit both iR and uR, cf. tilpti: telpa: talpinti: tūlpinti.

Those cases in which iR occurs alongside aR, and uR alongside eR, do not contradict the conception of the authors mentioned, since, for example, alongside *m̥n- and *mon- (cf. Lithuanian minia and Gothic manags), *men- may also have existed earlier, whose zero grade may have been Lithuanian minia.

Consequently, it is hardly possible either to refute or to substantiate the vowel ablaut theory with facts.

It is useful to raise the question of why precisely the vowels i and u, but not, say, the vowels e or a, as in other Indo-European languages, appeared before syllabic sonants in the Baltic and Slavic languages. Diphthongs eR and oR, inherited from the Indo-European period, existed in the Baltic linguistic community. For this reason, apparently, e and o (Lithuanian a) could not have appeared before syllabic sonants.

In the four-vowel system of short vowels in Baltic

i u e o,

when compared with the system of the first components of diphthongs

e(R) o(R),

the unmarked members were the vowels i and u, since only the vowels e and o occurred as the first components of diphthongs. Therefore, the first components of diphthongs arising from syllabic sonants could only be the unmarked vowels i and u.

From what has been said, it is clear that the distribution of iR and uR may have depended on forms with the full grade of vowel ablaut, in other words, on the diphthongs eR and oR: iR (from R) could have arisen in forms alongside which there existed forms of the same root with the diphthong eR (cf. kil-ti: kel-ti), while uR (from R) could have arisen in those forms alongside which there were forms with the diphthong oR (cf. kūl-ti: kūl-ti). If the root had the ablaut eR/oR, then its forms could have developed both iR and uR (cf. Lithuanian tém-ti: Latvian tim-sa: Lithuanian tamša: Latvian tūm-sa).

K. Brugmann, A. Vaillant, J. Kurilowicz, G. Shevelov, and others believed that the occurrence of the anaptyctic vowel i or u depended on the phonetic environment. K. Brugmann⁵ noted that the vowel u appeared after labial and velar consonants. This claim was not adequately supported by evidence from the Baltic and Slavic languages. K. Brugmann’s explanation was based on data from Old Indian and Greek, where ul and ur occur after labial and velar consonants.

A. Vaillant⁶ believes that in the Baltic and Slavic languages the anaptyctic vowel u derives from the vocalic element of the Indo-European labiovelar consonants kʷ, gʷ, gʷʰ. For example, Baltic and Slavic *gen-: *gun- (cf. Lithuanian gėna: gunióti, gūnyti and Old Church Slavonic ženą: gʷona-ti) developed from Indo-European *gʷʰen-: *gʷʰn- (that is, *gʷʰn- > *gʷʰn- > *gun-). Thus, according to A. Vaillant, Baltic and

⁵ K. Brugmann, Kurze vergleichende Grammatik der indogermanischen Sprachen, Berlin und Leipzig, 1922, p. 123.

⁶ A. Vaillant, Comparative Grammar of the Slavic Languages, I, p. 171.

78

In the Slavic languages, the prothetic vowel u appeared after labiovelar consonants, while the vowel i appeared after all the others.

J. Kurilowicz⁷ holds a similar view. According to him, syllabic sonants R developed into uR in the Baltic and Slavic languages not only after labiovelars but also after plain velar consonants (IE k, g, gh), whereas after all other consonants they developed into iR. J. Kurilowicz bases this conception on the assumption that the palatalization of consonants in the Baltic and Slavic languages is an older phenomenon than the vocalization of syllabic sonants. J. Kurilowicz believes that R shifted to iR in the clusters kR and gR because k and g had changed into k’ and g’. However, since velar consonants could not be palatalized before R and were therefore hard, the vocalic coloring of the syllabic sonant could only have been back (i.e., u).

The statistical data presented by J. Kurilowicz⁸ do not entirely support his claims, since there are no fewer examples of uR after other consonants than after velar consonants. For instance, uR occurs in 35.6% of examples after k and g, in 34.8% after m, and in 33.3% after n. The proportion of examples is the same after j as after n. The largest share (44.4%) of examples have the diphthong uR after the consonant s. All the other examples have iR.

The original distributional relationship between iR and uR may have become confused in the course of language development. For instance, the vowel alternation eR: R after hard velar consonants regularly became eR: uR (cf. Lithuanian gerklė: gurklys from Baltic *ger-tl-: *gur-tl-). Since the alternation eR: iR existed after all other consonants, and since it was more frequent and could therefore have been generalized, forms with iR could easily have arisen alongside forms with uR after hard velar consonants (cf. Lithuanian girtas).

Having examined the question of the distribution of iR and uR, G. Shevelov reached the following conclusions⁹:

a) after palatal palatals, R developed into iR;

b) between velar palatals (including labiovelars), R developed into uR;

c) after velar palatals and before non-velars, the phonetic reflex of R was uR;

d) after labials and before velar palatals, the phonetic reflex of R was uR;

e) after labials before non-velar palatals, and after dentals (including s) before any consonant, R developed into iR.

A strong argument for the theory of phonetic environment is the fact that in the Baltic and Slavic languages uR is virtually unattested after the consonants š and ž (from IE k’ and g’). The few examples found (cf. Lithuanian šultis, žulsnūs) are probably of late origin. This fact suggests that, when syllabic sonants R developed into iR and uR, the Baltic and Slavic languages still had k’ and g’, or had already changed them into š and ž (perhaps š and ž). Both k’, g’, rš, ž must have been very soft.

Consequently, on the basis of the theory of phonetic environment, it must be acknowledged that in the Baltic and Slavic languages the regular development of IE syllabic sonants was R > iR. The development R > uR, which depended on the phonetic environment (after k and g), was anomalous.

⁷ J. Kuryłowicz, L’apophonie en indo-européen, Wrocław, 1956, p. 227—243; O jedności językowej bałto-słowiańskiej, BPTJ, 16, 1957, p. 81—94; O балто-славянском языковом единстве, ВСЯ, 3, 1958, p. 23—24.

⁸ J. Kuryłowicz, L’apophonie en indo-européen, p. 229 (footnote).

⁹ G. Y. Shevelov, A Prehistory of Slavic, p. 90.

79

The vowel u after certain consonants before syllabic sonants has also arisen in other Indo-European languages. In V. Georgiev’s view¹⁰, Indo-European syllabic sonants after velar consonants became uR in Tocharian (after all others, āR). In Old Indian, after labial and, apparently, old labiovelar consonants, r before vowels became ur¹¹. In the same phonetic environment, l and r became ul and ur in Greek as well.

However, there is no genetic relationship here between the Baltic and Slavic, Tocharian, Indo-Iranian, and Greek languages, since there can be no doubt that after certain consonants the vowel u before l and r arose independently in each language.

In other Indo-European languages, too, the vowel i has also arisen before sonants. For example, in Albanian, before vowels Indo-European R became not only iR, but also Ri¹². In the Celtic languages, before consonants the vowel i arose after the sonants l and r¹³. Nor can there be any question of a genetic relationship between the Baltic, Slavic, Albanian, and Celtic languages, since the circumstances in which the vowel i arose alongside a sonant differed in each language: in Baltic and Slavic, iR arose both before vowels and before consonants, whereas in Albanian it arose only before vowels; in Baltic and Slavic the vowel i arose before sonants, whereas in Celtic it arose after sonants.

The closest reflexation of Indo-European syllabic sonants is found in Pelasgian: here, as in Baltic and Slavic, syllabic sonants became iR or uR¹⁴. However, a genetic relationship between this phenomenon in Baltic and Slavic, on the one hand, and Pelasgian, on the other, cannot be demonstrated, since the distribution of Pelasgian iR and uR is unknown. The fact that in certain cases Pelasgian also developed lu, ru and om, on most likely indicates that the reflexation of Indo-European syllabic sonants in Baltic and Slavic and in Pelasgian proceeded independently.

All that has been said suggests that the change of Indo-European syllabic sonants R into iR or uR is an innovation of the Baltic and Slavic languages.

No other Indo-European language has the same reflexation of syllabic sonants as the Baltic and Slavic languages. In this reflexation, Baltic and Slavic are entirely in agreement: Slavic has iR or uR wherever Baltic does (with one or another exception¹⁵); in both Baltic and Slavic, Indo-European syllabic sonants R became iR or uR not only before consonants but also before vowels. All this indicates that when Indo-European syllabic sonants R changed into iR or uR, the Baltic and Slavic languages maintained close relations: an innovation wave that arose in the territory of one language group could have spread into the territory of the other.

The absence in every other Indo-European language of the same reflexation of syllabic sonants as in Baltic and Slavic also suggests that Indo-European syllabic sonants were vocalized (i.e., became iR or uR) in the Baltic and Slavic languages after they had diverged from the dialects of the northern area of the Indo-European linguistic community.

Indo-European syllabic sonants R in Baltic and Slavic most likely became iR or uR before the formation of the intonational opposition “circumflex: acute,” since iR

¹⁰ В. Георгиев, Балто-славянский и тохарский языки, ВЯ, 6, 1958, p. 8—15.

¹¹ T. Burrow, The Sanskrit Language, II impression, London, 1959, p. 86.

¹² N. Jokl, Die Verwandtschaftsverhältnisse des Albanischen zu den übrigen idg. Sprachen, „Die Sprache“, 9, 1963, p. 120—122.

¹³ Г. Льюис и Х. Педерсен, Краткая сравнительная грамматика кельтских языков, Москва, 1954, p. 29.

¹⁴ В. И. Георгиев, Исследования по сравнительно-историческому языкознанию, p. 92.

¹⁵ Cf. Lithuanian sùrbti and Russian сра́ба́то, Polish sarbáć (but OCS srobati), Lithuanian būrlas and Russian берлога. Similar discrepancies are not significant Baltic and Slavic differences, since they also occur within Baltic itself (cf. Latvian tūmsa in the Central dialect and tīmsa in the High Latvian dialect).

80

and uR were treated as diphthongs: iR and uR respectively received the circumflex, while iR and ūR respectively (from IE *RH) received the acute. Since Baltic-origin Finno-Ugric loanwords have iR and uR (cf. Finnish villa and Lithuanian vilna, Finnish kurpponen and Lithuanian kūrpė), the reflexation of IE syllabic sonants in the Baltic languages must therefore have taken place before contact between the Baltic and Finno-Ugric languages.

The fate of Indo-European s

In Slavic and Indo-Iranian, the IE consonant s changed after k, r, i, u: in Slavic it became ch, while in Indo-Iranian it became š (Skt. ṣ, Av. š). In Lithuanian, IE s became š after k and r. Whether IE s also changed after i and u in Lithuanian has not yet been conclusively established. Cf.:

Lith. aukštas, Latv. augsts: Lat. augustus “elevated” (: Lith. augti ‘to grow’);

Lith. makštis, Latv. maksts: OCS mošьna “box” (from *maks-ina) (: Lith. makas “leather pouch (for tobacco, money),” Latv. maks “pouch,” Pruss. danti-maks “Zahnfleisch, gums,” OHG mago “belly”);

Lith. veršis, Latv. vērsis “bull,” Pruss. werstian “Kalb, calf”: Skt. vṛṣaḥ, Av. varšni- “ram,” Lat. verrēs “boar, hog” (Lat. -rr- from *rs);

Lith. viršus, Latv. virsus: OCS vrъchъ, ORus. vrъchъ; Skt. vārṣmā “summit,” Lat. verrūca “protuberance; wart,” Ir. ferr “better” (Lat. -rr- and Ir. -rr from *rs);

Lith. maišas, Latv. maiss, Pruss. moasis “Blasebalg, bellows”: OCS měchъ “sack,” Skt. meṣáḥ “ram,” Av. maeša- “sheep”;

Lith. paišyti, Latv. pāisīt “to break flax”: ORus. pъchati “to pound,” Skt. piṣṭáḥ “ground,” Av. pišant- “crushed”;

Lith. jūšė (also jušė) “fish soup,” Pruss. iuse “Juche, a kind of soup”: OCS jucha “soup,” Skt. yúḥ (yūṣám), Lat. iūs;

Lith. ausis, Latv. auss, Pruss. ausins (acc. pl.): OCS ucho, Av. uši, Goth. ausō, Gk. οὖς, Lat. auris (Lat. -r- from -s-).

As the preceding examples show, in Lithuanian š occurs after i and u in some cases (cf. maišas and jūšė), and s in others (cf. paišyti and ausis). This naturally raises the question: did IE s change after i and u in Lithuanian, or not? Researchers’ opinions on this question differ.

J. Endzelin¹, K. Būga², J. Šeftelovičius³, E. Frenkelis⁴, J. Kurilovičius⁵, S. Bernšteinas⁶, G. Ševel iovas⁷, and others hold that IE s did not change after i and u in Lithuanian. They derive the consonant š in many examples of the maišas type from sk’ or k’. However, in explaining the origin of this š, proponents of this theory encounter considerable difficul-

¹ I. Endzelin, Slavic-Baltic Studies, p. 51—60; J. Endzelin, Latvian Grammar, p. 110; On the Slavic-Baltic Reflex of Indo-European sk’, ZfslPh, 16, 1939, p. 107—115.

² K. Būga, RR, III, p. 589—590, 823; cf. K. Буга, Судьба индоевропейского s в литовском языке (Vilniaus Valstybinio universiteto rankraštynas, E 180).

³ J. Scheftelowitz, The verbal and nominal sk’- and sk-stems in Balto-Slavic and Albanian, KZ, 56, 1929, p. 156—170.

⁴ E. Fraenkel, The Baltic Languages, p. 113; LEW, p. 397 (s. v. maišas).

⁵ J. Kuryłowicz, Apophony in Indo-European, p. 374.

⁶ S. B. Bernstein, Outline of the Comparative Grammar of the Slavic Languages, p. 161.

⁷ G. Y. Shevelov, A Prehistory of Slavic, p. 129.

6. Baltų ir slavų kalbų ryšiai

81

of the bag. This is how variously and unreliably the origin of the sound -š- is explained, for example, in the word laiškas.

H. Peterson⁸ derived the Lithuanian word laiškas from *loik’sko-. To explain the origin of the -šk- combination in words of the laiškas type, K. Būga⁹ expressed the view that IE sk’ became šk’¹⁰ in Lithuanian. Accordingly, he held that Lithuanian laiškas derives from *loisk’-. J. Scheftelowitz¹¹ connected Lithuanian laiškas and Old Church Slavonic listъ with Old Indic līs- ‘to tear’ and derived them from *lik’-ko- and *lik’-to-.

Without assuming that IE s after i and u changed into š, it is hardly possible to explain the consonant š in the suffix -iškas. Not knowing how to explain that š, J. Scheftelowitz¹² considered the suffix -iškas to have been borrowed from the Slavic languages. J. Endzelin thought¹³ that alongside *-isko- there had formerly also been a suffix *-isk’o (from which Lithuanian *-iša-). J. Endzelin’s claim that the suffix -iškas arose through contamination of these suffixes is not convincing.

H. Pedersen¹⁴, K. Brugmann¹⁵, H. Arntz¹⁶, A. Vaillant¹⁷, J. Otrębski¹⁸, and O. Semerenyi¹⁹ believed that in the Balto-Slavic linguistic community, IE s changed not only after k and r, but also after i, u. This conclusion was not supported by Baltic-language evidence. It rested on the assumption that the Baltic and Slavic languages had once formed a linguistic unity: since IE s after i and u had changed in the Slavic languages, it must also have changed in the Baltic languages, which had formerly formed a unity with the Slavic languages.

To explain why Lithuanian has š after i and u in some cases and s in others, H. Pedersen formulated the following law²⁰: š was retained after an acute-accented vowel (cf. jūšė), whereas after a circumflex-accented vowel š again became s (cf. saũšas). However, Lithuanian facts contradict this law: in many cases s occurs after an acute-accented vowel (cf. ãusas ‘gold’, lýsė, viẽšulas, kláusė, etc.), while š occurs after a circumflex-accented vowel (cf. maišas, kiũši, etc.).

A. Meillet²¹, R. Trautmann²², V. Pisani²³, and V. Porzig²⁴ argued that in Lithuanian, IE s after i and u corresponds in some instances to the consonant š and in others to the consonant s.

⁸ H. Petersson, Arische und armenische Studien, Lund—Leipzig, 1920, p. 91.

⁹ K. Būga, RR, II, pp. 285—286.

¹⁰ IE sk’ probably became š in Lithuanian; cf. J. Scheftelowitz, op. cit., pp. 165—170; J. Endzelin, Über den slavisch-baltischen Reflex von idg. sk’, pp. 107—115; M. Leumann, Idg. sk’ im Altindischen und im Litauischen, IF, 58, 1941, pp. 1—26, 113—130.

¹¹ J. Scheftelowitz, op. cit., p. 196.

¹² J. Scheftelowitz, op. cit., p. 197.

¹³ J. Endzelin, Über den slavisch-baltischen Reflex von idg. sk’, pp. 112—113.

¹⁴ H. Pedersen, Indo-European s in Slavic, IF, 5, 1895, pp. 33—87.

¹⁵ K. Brugmann, Grundriss der vergleichenden Grammatik der indogermanischen Sprachen, I, 2, p. 781.

¹⁶ H. Arntz, Sprachliche Beziehungen zwischen Arisch und Balto-Slawisch, Heidelberg, 1933 p. 9—10.

¹⁷ A. Vaillant, Grammaire comparée des langues slaves, I, p. 32.

¹⁸ Я. Отрембский, Славяно-балтийское языковое единство, ВЯ, 5, 1954, p. 32—34.

¹⁹ O. Szemerényi, The Problem of Balto-Slav Unity — A Critical Survey, p. 105—107.

²⁰ H. Pedersen, op. cit., p. 77—78.

²¹ A. Meillet, Les dialectes indo-européens, p. 85.

²² R. Trautmann, Die altpreussischen Sprachdenkmäler, p. 169; Baltisch-Slavisches Wörterbuch, p. 165 and 250 (s. v. maiša- and sausa-).

²³ V. Pisani, Balto e slavo, StB, 2, 1932, p. 16.

²⁴ W. Porzig, Die Gliederung des indogermanischen Sprachgebiets, p. 165.

82

An analysis of Lithuanian-language material helps answer the question of whether PIE s became š in Lithuanian after i and u.

Here are examples in which Lithuanian has š instead of s in other PIE languages: maišas, riešas (riešutas), višta, glušas, jušė (jušė), kriaušė, kruša, kūšys, rūšė, pušė (puškas), aiškus (eiškus), ieškoti, kišką, laiškas (laiskas), miškas, reiškia; -uš- (ėpušė, kermušė, vėtušas), *-ištja- (eriščias), and -išk- (moteriškas). The verbs rišti, aušti, and traušti are uncertain, since their -š- may also have developed from -s-sk’ or -sk’.

Here, by contrast, are examples in which PIE s after i and u did not become š: ausis, mušė, mušas, pusė; ausas (auksas), blusà, lỹsė, usnis, iesmė, viesulas, visas, pliauskà (plauska); paisyti, tiešti (teisti), veisti, daūsos, klausyti, kniaustis, niausti, praušti, saūsas, tausytis, triušti, the past participle suffix -us- (cf. mirusį²⁵). Comparison with the corresponding facts in other PIE languages shows that the element -s- in these words is old and was undoubtedly present before PIE s changed to š. Hence š, rather than s, is expected here after i and u.

To explain why PIE s did not become š in these words, we must bear in mind that the consonant remained unchanged when a morpheme boundary lay between it and the vowels i and u²⁶. In some of the words listed here, the element s likewise belonged to one morpheme, while the vowels i and u belonged to another; cf.:

plau-sk-à, pliau-sk-à (: plaũ-t-as, pliaũ-t-as, Latv. plàu-kt-s, plàu-t-s),

vė-s-ulas, vė-s-ulas (: vė-jas, vė-tra),

pai-s-yti (: pai-l-yti, pai-v-yti, peĩ-k-ė, peĩ-k-enà),

kniaũ-s-tis, kniaũ-s-ėsi (: kniaũ-t-esi)²⁷.

The same applies to Lithuanian niaũ-s-ti ‘to lower, bow one’s head’, apsi-niaũ-s-ti ‘to be gloomy’, and niau-s-óti ‘to be sad, gloomy’. Neither the sounds nor the meaning contradicts their connection with Lithuanian niaũ-k-tis (niáu-k-tis) ‘to become overcast’, niau-k-ùs ‘overcast, dark’, and niau-r-ùs ‘gloomy, angry’.

At the time when PIE s changed to š, the s may have been perceived as a suffix—and thus as belonging to a different morpheme from i and u—because forms of similar meaning with different root-final consonants existed alongside it.

The explanation that the element -s- did not change to -š- because it belonged to a different morpheme from i and u does not apply to all words that retained unchanged s. It does not apply to words that apparently had no asigmatic form alongside them and in which s was probably not a formative; cf. Lithuanian tausytis ‘to calm down (of the wind)’, Latvian tusēt ‘to breathe with difficulty, suffocate’, Old Prussian tusnan ‘calm, quiet’, Old Church Slavonic po-tuchnǫti ‘to go out’, and Sanskrit tuṣyati ‘(he) becomes calm’. These words therefore require a different explanation.

It is quite possible that PIE s after i and u did not become š in Lithuanian before certain consonants (for example, before j)²⁸.

The material reviewed from this perspective shows that most of the words have -s- before j(i); cf. the element *-sja, identifiable in *peisja (from paisyti), tiešia, teisčia, veisia, kláusia (klausyti, with s before i), kniausiasi, praušia, *teusjasi (from tausytis, with s before i), and triūšia. The past participle suffix -us- was likewise always present

²⁵ All words and instances in which š and s correspond to s in other PIE languages are catalogued in the article “К вопросу об н.-е. *s после i, u в литовском языке”, “Baltistica”, I(2), 1966, pp. 113–126.

²⁶ Op. cit., p. 122.

²⁷ For meanings and etymologies, see E. Fraenkel, LEW; K. Būga, RR, I–III (Indices).

²⁸ “Baltistica,” I(2), 1966, p. 125.

83

before j; cf. mirusio and mirusios from *mirusiō and *mirusjās. Such i-stem nouns as ausis, musis (cf. also jā stem musià, where s was before j) and pùšis, derived from old consonant-stem words, appear to be much older than their ē-stem forms; cf. mùsė, pùsė.

In other words, s occurred before v(u) (cf. lýsvė, viesulas) m and n (cf. iesmė and usnis) in other words.

Thus, most words in which s after i and u remained unchanged have that s before the sonants j(i), v(u), m, and n. Apparently, such a position of s must also be sought in those words in which s is not now followed by a sonant. For example, in the words aũsas, blusà, aũsos, dùsas, mùsas or mùsos, saũsas, visàs, the morphemes aus-, blus-, daus-|dus-, mūs-|mus-, saus-, vis- were much more often followed by sonantal consonants and vowels than by the morpheme -a- (-o-); cf. auksinis, auksinti, auksýti, auksius (stress unknown), auksnas, áuksinas, auksioti (-k- here is secondary; cf. Old Prussian ausis and Latin aurum); blusia (stress unknown), blùsinas, blusýnas, blusinėti, blusingas, blusinis, blusinti, blùsius; dūsúoti, dūsinti (Latvian dusinat), dūsýti, dùsnà, dusulýs (Latvian dusulis), duslùs; mūsýti, musýti; saũsis, sausmė, sausra, sausumà, susnà; vismet, visur, Latvian visur (cf. also Old Church Slavonic vьsь and Old Indic viṣú-, which may indicate that the Baltic word, too, could formerly have had i or u).

All the material presented here seems to allow the conclusion that Proto-Indo-European s after i and u remained unchanged in the Baltic languages, possibly when it immediately preceded sonorants.

Thus, nothing contradicts the assumption that Proto-Indo-European s after i and u also became š in the Baltic languages, except, of course, in cases where that s belonged to a different morpheme from i and u and immediately preceded sonorants.

Thus, in the Baltic languages, as in Slavic and Indo-Iranian, Proto-Indo-European s also changed in the position after k, r, i, u. That this phenomenon was common Baltic is shown by borrowings of Baltic origin in the Baltic Finnic languages, where the Baltic consonants š (and ž) correspond to the sound h, which could have arisen only from š (cf. Finnish halla, Estonian hall and Lithuanian šalnà, Latvian salna; Finnish hihna, Estonian ihn and Lithuanian šìkšnà, Latvian siksna). Proto-Baltic š, which arose from Proto-Indo-European s after k, r, i, u and from Proto-Indo-European ḱ, as well as its correlate ž, which arose from Proto-Indo-European gʰ and gʰh, became s and z in Latvian and Prussian dialects because of the particularities of their phonological systems. They remained unchanged only in Lithuanian.

Some researchers (A. Meje⁽²⁹⁾, S. Bernštein⁽³⁰⁾, and others) tend to discern traces of the change of Proto-Indo-European s into š in Prussian, where (in the Third Catechism) -rsch- does indeed occur alongside -rs- (cf. Prussian kirscha, kerscha alongside kirsa, pirschau alongside pirsdai). This Prussian phenomenon should most likely be regarded not as a remnant of the Proto-Baltic state but as a new development in Prussian (possibly due to the influence of German⁽³¹⁾), associated with the change of *sj into š and the emergence of the sound š in the clusters šp-, šk-, šl-.

The change of Proto-Indo-European s after k, r, i, u into š is a phenomenon shared by the Baltic, Slavic, and Indo-Iranian languages. Therefore, it cannot be an innovation of the Baltic and Slavic languages. What is important for characterizing the earliest relations between the Baltic and Slavic languages is that the circumstances under which Proto-Indo-European s became š differed in each language group.

A. Meillet, Les dialectes indo-européens, p. 85.²⁹

³⁰ S. B. Bernštein, Очерк сравнительной грамматики славянских языков, p. 161.

³¹ I. Endzelin, Славяно-балтийские этюды, pp. 60—68; J. Endzelīns, Senprūšu valoda, p. 37.

84

In the Slavic languages, IE *s became ch only before vowels, whereas when followed by a sonorant or a consonant it remained unchanged³² (cf. OCS pracho from *parsas and OCS prъstъ from *pṛstis). In the Baltic and Indo-Iranian languages, unlike in the Slavic languages, IE *s became š not only before vowels but also before consonants. In the Baltic languages, as in the Slavic languages, IE *s remained unchanged after i and u before sonorants.

In Old Indic, unlike in the Baltic and Slavic languages, IE *s became š not only after k, r, i, u, but also after k’, l, and i derived from IE *H. In the Indo-Iranian languages this phenomenon was so intensive that z, too, became ž after the aforementioned sounds (cf. Avestan mūždam and OCS mbzda). In Old Indic, the r that followed IE *s prevented its change to š (cf. OInd. tisrāḥ, feminine, “three”), and often so did anusvara³³.

All this suggests that IE *s after k, r, i, u most likely changed independently in the Baltic, Slavic, and Indo-Iranian languages³⁴. The fact that in the three language groups mentioned, the development of IE *s proceeded in parallel may also be indicated by the fact that in the modern Indic languages (Sindhi, Hindi, etc.), OInd. š has become kh and h³⁵. There is no doubt that this Indic development is genetically unrelated to the similar Slavic development (Slavic ch most likely comes from an earlier š). The fact that in the Kafir languages IE *s after u remained unchanged³⁶ suggests that IE *s changed quite late, in dialects of the Indo-Iranian group that had already taken shape.

Finally, the conclusion that IE *s after k, r, i, u changed independently in the Baltic and in the Slavic languages is also supported by evidence from the Baltic languages themselves. That by the time IE *s after k, r, i, u changed to š the Baltic and Slavic languages were already two independent linguistic units is shown by the fact that in the Baltic languages a consonant s belonging to a morpheme other than i and u remained unchanged, whereas in the Slavic languages it also became ch in this position (cf. the plural locatives Lithuanian avi-sù and OCS gostb-chъ, Lithuanian tri-sù and OCS trb-chъ). If the Baltic and Slavic languages had formed a unity at that time, one would expect that such a discrepancy would not have occurred. At the time IE *s changed to š, or immediately afterward, the Baltic and Slavic languages displayed clearly opposing tendencies: in Baltic inflection, s was generalized throughout, whereas in Slavic, conversely, inflectional ch also entered positions where it was not phonologically motivated (cf. OCS synъ-chъ and žena-chъ instead of the expected form *žena-sъ). In the Baltic languages, IE *s after i and u did not become š before sonorants, whereas in the Slavic languages, by analogy, ch was transferred to suffixes and word roots where it could not have arisen phonetically (cf. OCS spě-chъ alongside spě-ti and Russian ma-x-ámъ alongside má-ámъ, i.e. má-j-atъ).

Although, as mentioned, IE *s became š quite late, in dialects that had already taken shape—for example, the Indo-Iranian dialects—on the other hand, it changed much earlier than many phonetic changes in the Baltic and Slavic languages took place. For instance, in the Baltic languages, IE *s seems to have become š before t and d were lost before s (cf. Lithuanian garšas from *gord-s-os). This suggestion is prompted by the fact that the change of IE *s to š after k, r also affected suffixes (cf. garšas, garšvà).

³² J. Kurylowicz believes (L’apophonie en indo-européen, p. 374) that in the Slavic languages IE *s became š even before consonants (> Slavic ch), and only later changed to s.

³³ A. Thumb, R. Hauschild, Handbuch des Sanskrit, I, Heidelberg, 1958, p. 306; T. Burrow, The Sanskrit Language, p. 79–80.

³⁴ C. C. Uhlenbeck, Miscellen. 1. Zur Geschichte des s im Slavischen, KZ, 39, p. 599–600; A. Meillet, Les dialectes indo-européens, p. 84–85; И. Эндзелин, Славяно-балтийские этюды, p. 74.

³⁵ J. V. Lazicius, Zur Wandlung idg. s>sl. ch, IF, 51, 1933, p. 196–199.

³⁶ G. Morgenstierne, Indo-European k’ in Kafiri, NTS, 13, 1945, p. 225–238.

85

The change of Slavic s to ch preceded the monophthongization of the diphthong ai and the metathesis of er and or in South Slavic or pleophony in East Slavic, since after the diphthong ai and er and or, Proto-Indo-European s changed.

The fact that in Baltic-origin loans in the Finnic languages of the Baltic region, Baltic š (and ž) is reflected by the sound h, which may have developed from š, indicates that contacts between Baltic and Baltic-region Finnic began at a time when Proto-Indo-European s after k, r, i, u and Proto-Indo-European k’ had already changed to š in the Baltic languages.

The change of Proto-Indo-European s after k, r, i, u to š in the Baltic languages was structurally connected with the assibilation of Proto-Indo-European k’. As early as 1908, A. Meillet noted that palatal velar consonants had been assibilated in those Proto-Indo-European dialects in which Proto-Indo-European s after k, r, i, u changed to š³⁷. That the change of Proto-Indo-European s after k, r, i, u and the assibilation of Proto-Indo-European k’ were synchronically and structurally connected processes is clearly shown by their identical outcome: s after k, r, i, u and Proto-Indo-European k’ both changed to š in Lithuanian. This also clarifies the phonological process by which Proto-Indo-European s after k, r, i, u changed to š.

After the stop consonant k, after the apical r, and after the high vowels i and u, the articulation of consonant s was probably higher, and the sound itself was likely pronounced more softly and less distinctly. Acoustically, this variant of consonant s may have been an intermediate sound between present-day Lithuanian s and š, that is, it may have been š. The change of the “high variant” of consonant s into an independent phoneme depended on the merger, in the course of language development, of the sequences *-rś- and *-rḱ-, *-iś- and *-iḱ-, etc. Their merger resulted in -rš- and -iš-. However, the sequence *-oḱ- also changed to -aš-. But since the sequence *-os- contained precisely the phoneme /s/ (and not its “high variant” š), š and s were phonologically distinct sounds. Thus the phoneme /s/ split into two distinct units—i s and š³⁸.

On the basis of the data presented in this section, the following conclusions can be drawn concerning the relations among the dialects of the Proto-Indo-European linguistic community:

1) by the time Proto-Indo-European s after k, r, i, u changed to š, the Baltic, Slavic, and Indo-Iranian dialects were already independent linguistic units (cf. the different details of this change noted above);

2) at the time of this change, the Baltic, Slavic, and Indo-Iranian dialects maintained certain contacts: in all three language groups, Proto-Indo-European s changed in the same position, and the result of this change was also the same.

The development of Proto-Indo-European palatal velar consonants

In the Baltic and Slavic languages, the palatal velars k’, g’, g’h of the Proto-Indo-European linguistic community underwent assibilation. In Lithuanian they became š, ž; in Latvian, Old Prussian, and Slavic, they became s, z; cf.:

Lith. šimtas, Latv. simts, OCS sъto, Skt. śatám, Av. satəm: Gk. ἑκατόν, Lat. centum, OHG hund, Toch. kante;

Lith. širdis, Latv. sirds, OPr. seyr (from *šer(d)), OCS srъdъ-ce, Old Russian sьrd-ce: Gk. καρδία, Lat. cor cordis, Hitt. kard-;

Lith. paršas, OPr. parstian (attested as prastian in Voc. 686), Old Russian porosja, Serbo-Croatian prâse: OHG farah, Lat. porcus “pig, boar,” Gk. πόρκος;

Lith. žiemà, Latv. ziema, OPr. semo (s- is to be read as z-), OCS zima, Skt. hēman “in winter” (h- from Proto-Indo-European g’h-): Gk. χεῖμα, Lat. hiems, Hitt. gimmant-;

³⁷ “The line of the š/s treatment thus broadly coincides with that of the treatment of the gutturals,” A. Meillet, Les dialectes indo-européens, p. 86.

³⁸ A. Martinet, Concerning some Slavic and Aryan Reflexes of IE s, “Word,” 7, 1951, p. 91–95.

86

Lith. žvėris, Latv. zvērs, Pr. swīrins (acc. pl.), OCS zvěrъ: Gk. θήρ, Lat. ferus (f- from g’h);

Lith. bėržas, Latv. bērzs, Pr. berse, OCS brěza, Serb.-Croat. breza, Skt. bhūrjāh: OHG birihha (hh from Germ. k < IE g’).

The examples given show that some IE languages have spirants, while others have stops in the same words. Formerly, great importance was attached to this phonetic feature of the IE languages: on its basis, the IE languages were divided into two dialects—the centum and satem dialects (from the Latin and Avestan words for “hundred,” centum and satam). The first to propose such a dialectal division of the IE linguistic area was P. von Bradke in 1888¹. Thanks to the support of K. Brugmann², the division of IE languages into centum and satem dialects became generally accepted. According to this division, besides the Balto-Slavic languages mentioned above, the Indo-Iranian, Armenian, and Albanian languages were assigned to the satem group, while Greek, Italic, Celtic, Germanic, and Illyrian belonged to the centum group. The geographical distribution of the IE languages corresponded to this linguistic classification: all satem languages lay in the east of the IE linguistic area, and the centum languages in the west.

Over time, however, doubts arose as to whether the isogloss of affrication of palatal consonants could serve as the basis for a dialectal division of IE languages. Study of the recently discovered Tocharian and Hittite languages showed that, although geographically situated among the satem languages, they were nevertheless centum languages. Thus the centum and satem theory lost its geographical argument. Moreover, centum-language elements are also found in the satem languages themselves, especially those situated farther to the west (for example, Baltic and Slavic), while elements of the affrication of palatal velar consonants occur in centum languages (cf. some affrication phenomena in Hittite). In Baltic and Slavic, a velar consonant and a spirant often alternate within the same root, resulting in the alternations Lith. š: k, ž: g and Slavic s: k, z: g; cf. Lith. šaukti, Latv. saukt: Lith. kaukti; Lith. šeima: kiemas, kaimas; Lith. šleivas: klevas, Latv. klēvs; Lith. šliubas: kliubas, Latv. kluhurs; Lith. šliaukti: kliaukti; Lith. ašmuõ, Skt. aśmā “stone, rock, sky”: Lith. akmuõ, Latv. akmens, Gk. ἄκμων “anvil”; Lith. aštrùs: akūotas; Lith. krėšti, krešėti: krėkti, krekėti, Latv. krecēt; Lith. šùnpis: kùmpis; Lith. šuõ, Latv. suns, Pr. sunis: Latv. kuņa “bitch”; Lith. švañkšti: kvañkšti; Lith. šviẽti, OCS světěti: Latv. kvitēt; Lith. ker̃džius, Pr. kerdan (acc. sing.) “time,” OCS črěda “herd; cattle”: Skt. śárdhah “herd”; Skt. rúsant-: Lith. laũkas; Lith. smãkras (also smakrà), Albanian mjekër “chin; beard” (from *smekrā-): Skt. śmáśru “beard” (ś- from s-); Lith. pėkus “cattle,” Pr. pecku, Lat. pecu: Skt. paśúh, Av. pasu-; Lith. žárdis “enclosed place for animals”: gárdas, OCS gradъ “town”; Lith. žėltas “yellowish”: geltãs; Lith. žentas, OCS zetъ: Lith. gentis “relatives”; Lith. žnybti, žnaibyti: gnybti, gnaibyti; Lith. aržóti “to shout, call out; quarrel, fight”: argóti “to beat”; Pr. gunsix “Beule, lump”: Lith. gūnglė “lump”; Lith. įstrižas: įstrigas; OCS brěgъ “shore”: Av. barǝzah- “mountain” (cf. OHG berg); OCS kotora “dispute”: Skt. śátruh “enemy”; OCS kloniti: Old Russian Sl. sloniti se, Skt. śrāyate “(he) leans”; Old Russian světo: květo.

¹ P. von Bradke, Beiträge zur Kenntnis der vorhistorischen Entwicklung unsres Sprachstamms, Giessen, 1888.

² K. Brugmann, Grundriss der vergleichenden Grammatik der indogermanischen Sprachen, I, 1, p. 542—544.

87

There are cases in which Baltic has a spirant while Slavic has a velar, or Baltic has a velar while Slavic has a spirant; cf. Lith. klausyti, Latv. klāusīt, Pr. klausīton “to hear, listen to”: OCS slušati (but Baltic also has a spirant, cf. Lith. šlovė, šlovinti, Latv. slava, slavět); Lith. šešūras, Skt. śváśuraḥ: OCS svekrъ; Lith. šokti: OCS skokъ; Lith. žąsis, Latv. žūose, Pr. sansy: Slavic *gǫsь (Russian гусь, Polish gęś); Lith. žvaigždė, Latv. zvaigzne, Pr. swāigstan (acc. sing.) “shine, light”: OCS zvězda (zv- from gv-), Polish gwiazda; Pr. massī “I can, I am able”: OCS moga, etc.

No great significance should be attached to these different reflexes of IE palatal consonants in Baltic and Slavic, since such differing reflexes also occur within the Baltic languages themselves.

The examples given show that centum elements are found in Baltic and Slavic. Thus, they are not purely satem languages.

Baltic and Slavic are not the only satem languages in which centum elements are found. A similar alternation of spirants and velar consonants is also evident in Albanian, Thracian, and Phrygian. In Hittite and some other Anatolian languages, IE k’ before u became s (cf. ašuwa “horse”), although in all other positions velar and palatal consonants merged³.

All those languages in which spirants and velar consonants alternate in the same root form a group of transitional dialects between the IE languages in the east and west. For this reason, V. Porzig quite rightly distinguished two areas within the zone of palatal-consonant affrication: the first, comprising the Indo-Iranian and Armenian languages, where palatal consonants—with a few minor exceptions—have consistently become spirants; and the second, comprising the transitional dialects mentioned above⁴. The feature of the so-called satem languages whereby spirants alternate with velar consonants was noted as early as 1904 by E. Cupica, who observed that forms with velar consonants made up a larger proportion in the west of the satem area than in the east⁵. Since Slavic languages lie farther east within the IE linguistic area, forms with velar consonants in place of spirants are correspondingly fewer⁶.

There is no consensus as to why k, g arose in place of the expected š, ž, or s, z in Slavic, in Baltic and also in other IE languages.

Most examples of the type Lith. pėkus, Slavic *gąsь were considered loans of Germanic origin⁷ or were thought to have undergone phonetic alteration under the influence of centum languages⁸.

Since there are quite a number of forms with k, g in place of the expected š, ž (Slavic s, z), it seems inappropriate to regard them as loanwords. If pėkus is treated as a loanword, consistency would require

³ Вяч. В. Иванов, Проблема языков centum и satem, ВЯ, 4, 1958, p. 19; Хеттский язык, Москва, 1963, p. 90. However, some scholars (e.g., H. Kronaser, see his Vergleichende Laut- und Formenlehre des Hethitischen, Heidelberg, 1956, p. 229) are inclined to regard the words containing spirants, Hittite ašuwa “horse,” šuwana (acc. sing.) “dog,” etc., as loans from Indo-Iranian languages.

⁴ W. Porzig, Die Gliederung des indogermanischen Sprachgebiets, p. 74.

⁵ E. Zupitza, Zur Gutturalfrage, K.Z. 37, 1904, p. 400.

⁶ G. Shevelov thinks otherwise; see G. Y. Shevelov, A Prehistory of Slavic, p. 141–145.

⁷ For Lithuanian pėkus, see K. Būga, RR, II, p. 85–86; for Slavic *gąsь, see K. Brugmann, Kurze vergleichende Grammatik der indogermanischen Sprachen, Berlin und Leipzig, 1922, p. 158.

⁸ For example, M. Fasmer (REW, I, p. 324) thinks that the sound z- in Slavic *zǫsь (cf. Lithuanian žąsis) was replaced by g- under the influence of Germanic languages, cf. Old High German gans, while E. Fraenkel (Die baltischen Sprachen, p. 15) argues that the consonant k in Lithuanian pėkus, Old Prussian pecku arose when the old Baltic word for ‘cattle’, with š (cf. Lithuanian pešti), crossed with Gothic faihu ‘money’ (from *peku).

88

to regard as a loanword and kleivas. It would seem implausible that the Balts, having their own word šleivas, borrowed kleivas as well, with exactly the same meaning.

Other researchers were inclined to explain the emergence of the velar consonants k and g in place of the expected š and ž (Latvian, Old Prussian, and Slavic s and z) by the dissimilatory effect of neighboring consonants. A. Meillet argued that k’ and g’ became k and g in Slavic under the influence of s⁹. E. Fraenkel¹⁰ also agreed with Meillet’s view. A. Vaillant¹¹ held a similar opinion, saying that k’, g’, g’h became k, g in Baltic and Slavic when they followed s or š. According to A. Vaillant, Lithuanian klausyti acquired its k from the construction iš-klausyti. The interpretations of A. Meillet and A. Vaillant¹² leave unexplained the fairly frequent alternation of k, g with š, ž.

In the opinion of S. Agrell and K. Būga¹³, Proto-Indo-European k’ and g’ before r did not become š and ž, or s and z respectively, in Baltic and Slavic, cf. Lithuanian smákras and Sanskrit śmáśru. K. Būga did not consider words of the type šleivas: kleivas, in which velar consonants do not occur next to r, to be derivable from a single root¹⁴. However, Lithuanian forms can be found in which the consonants k and g, although not before r, occur in place of the expected š and ž, cf. Lithuanian akmuo and Sanskrit aśmá.

V. Machek is inclined to think¹⁵ that palatal velar consonants could have been depalatalized before the liquid sonorants r and l and before back vowels (vor dunklen Vokalen).

Other researchers (A. Brückner¹⁶, G. Shevelov¹⁷, V. Lehmann¹⁸, and others) attempted to explain the Baltic and Slavic alternation k, g: š, ž, or s, z respectively, as an alternation between Proto-Indo-European velar and palatal consonants (k, g, gh: k’, g’, g’h). In examining the origin of this consonant alternation, V. Lehmann established the following facts: 1) in Indo-European languages, the articulation of palatal and velar consonants depends on the articulation of the vowels that follow them: before e, a velar consonant has a front-of-the-tongue articulation; before o, a back-of-the-tongue articulation; before a, it is neutral, that is, non-palatal; 2) the sატəm phenomenon occurred in those Indo-European dialects that simplified the vowel system; 3) k’, g’, g’h never occur before a and r¹⁹. On the basis of these facts, V. Lehmann concludes that the original opposition among Indo-European velar consonants was k: kʷ (or g: gʷ and gh: gʷh), and that the palatal consonant k, which arose before e in satem languages, was a positional variant of the phoneme k. V. Lehmann further argues that, as a result of fundamental changes in the vowel system (cf. Sanskrit a from Proto-Indo-European e, o, a; Lithuanian a from Proto-Indo-European o and a; Old Church Slavonic o from Proto-Indo-European o and a), a phonemic contrast arose between velar and palatal velar allophones, that is,

⁹ “Slavic uses the velar when the word contains a sibilant: gostь: hostis, gažb: ža-šis,” — A. Meillet, Indo-iranica. II. — Trois notes sur la phonétique des gutturales, MSL, 9, 1896, p. 347. See also A. Meillet, Varia, II. — A propos de v. sl. gažb, MSL, 13, 1905–1906, p. 243–245.

¹⁰ E. Fraenkel, The Baltic Languages, p. 15.

¹¹ A. Vaillant, Comparative Grammar of the Slavic Languages, I, p. 37–38.

¹² For critical remarks on A. Meillet’s conception, see M. Vasmer, Beiträge zur slavischen Grammatik, I. Meillet’s urslavisches Sibilantendissimilationgesetz, ZfslPh, 2, 1925, p. 54–58.

¹³ S. Agrell, Baltoslavische Lautstudien, Lund, 1919, p. 27; K. Būga, RR, II, p. 217, 223.

¹⁴ “Such word pairs cannot be derived from a single proto-root. Consequently, šleivas and kleivas must have arisen from separate roots, š(e)l-ei- and k(e)l-ei-,” — K. Būga, RR, II, p. 221.

¹⁵ V. Machek, Zur Vertretung der indogermanischen Palatale, IF, 53, 1935, p. 89.

¹⁶ A. Brückner, Die lituslavische Spracheinheit, KZ, 46, 1914, p. 233.

¹⁷ G. Y. Shevelov, A Prehistory of Slavic, p. 141–142.

¹⁸ W. P. Lehmann, Proto-Indo-European Phonology, Austin, 1964, p. 100–101.

¹⁹ W. Lehmann, op. cit., p. 101.

89

k and k’ had become phonemes. In this way, W. Lehmann attempts to answer the fundamental question: for what reasons did the allophones k and k’ acquire phonemic status²⁰? However, it cannot be said that W. Lehmann’s answer to this question is entirely unassailable: the explanation he sets out is more applicable to Indo-Iranian than to Baltic and Slavic, where the opposition between front and back vowels (between e i a and o, respectively) was strictly maintained.

Because e and o alternate in forms of the same root (cf. the alternation e/o in verb stems), in W. Lehmann’s view, an alternation of palatal and velar dorsals also arose within one and the same morpheme. Since consonant alternation was not possible in the Indo-European linguistic community, after k’ and k became phonemes, either the palatal or the velar consonant must have been generalized. However, it may also have been that the palatal consonant was generalized in some forms and the velar consonant in others.

This explanation by W. Lehmann of the origin of the alternation between palatal and velar consonants seems to apply only to alternation among root-final consonants (type kre-š-ėti: kre-k-ėti). Alternation among word-initial consonants (type šleivas: kleivas) should be explained differently.

Examples in which dorsal and spirant consonants alternate can be divided into three groups:

1. In the words of the first group, the alternating dorsal and spirant consonants may be regarded as formatives: Old Indian rūśant-: Lithuanian laũkas, Lithuanian krešėti: krekėti, aržóti: argóti, Avestan barəzah-: Old Church Slavonic bregъ; įstrižas: įstrigas, Prussian massi: Old Church Slavonic moga, šliaužti: sliaugti.

2. The second group consists of words whose dorsal consonants occur before sonorants, but are separated from them by another sound: kaũkti: šaukti, kiẽmas: šeima, kermušė: šermūkšnis, Lithuanian ker̃džius: Old Indian śárdhah, Lithuanian katãryti, Old Church Slavonic kotora: Old Indian śátruḥ, Lithuanian gãrdas: žãrdis, Lithuanian žąsis: Slavic *gaśb, Lithuanian geltas: želtas, gentis: žentas.

3. The third group includes words in which dorsal consonants occur immediately before sonorants:

kl, gl: klausýti: šlovė, kleivas: šleivas, kliaũkti: šliaũkti, kliùrti: šliùrti, Old Church Slavonic kloniti: Old Russian sloniti, klúbas: šlúbas, Lithuanian gùnglė: Prussian gunsix;

kr: Lithuanian smãkras: Old Indian śmáśru, Old Church Slavonic svekrъ: Lithuanian šešuras;

km: akmuõ: ašmuõ;

kn, gn: gnýbti: žnýbti; agnùs: ažnas;

kj (kj) —;

kv (ku), gv: kvaĩkšti: švañkšti, Old Russian květъ: svěť, Latin kvitēt: Lithuanian švitėti, Polish gwiazda: Lithuanian žvaigždė, kùmpis: šùmpis, Lithuanian šuo: Latvian kuņa, Lithuanian pėkus: Old Indian páśuh.

On the basis of the words assigned to the last group, one might suppose that in Baltic (and perhaps Slavic) the palatal consonants k’, g’ were depalatalized before sonorants, i.e. k’l(g’l), k’r, k’m, k’n and k’v(k’u) became kl(gl), kr, km, kn and kv(ku)²¹ They must certainly have been depalatalized before r, since Lithuanian has no clusters *šr-, *žr- (not-

²⁰ J. Kurylowicz, On the Methods of Internal Reconstruction, „Proceedings of the Ninth International Congress of Linguists“, London, 1964, p. 9—36.

²¹ Cf. С. Каралюнас, Древнейшие отношения балтийских и славянских языков, p. 8.

90

including, of course, loanwords). It would be logical to assume that palatal velar consonants hardened also before other sonorants—before l, m, n, and also before j, v. A structural analogy supporting this conclusion would be the fate of s after k, r, i, u: the consonant s did not become š, apparently likewise in the position before sonorants (see p. 84 of this work). However, the assumption that palatal velar consonants were depalatalized before sonorants is contradicted by many Lithuanian words beginning with šl-, žl-, šm-, žm-, šn-, žn-, šv-, žv-. It is important, however, not to lose sight of the fact that most of these words may be suspected of having an onomatopoeic origin; compare, for example, those beginning with the clusters šn-, žn-: šnabždėti/šnibždėti, šnairas, šnarėti/šnėrėti, šnarkšti/šnėrkšti, šnar̃pšti/šnėr̃pšti, šnekėti, šnervės, šniaukšti, šniaukti, šniokšti, šnioti, šnypšti, šnir̃kšti/šnir̃kšti, šniukštinėti, šniurti, šnopsėti; žnaibyti, žnairùs, žneksėti, žniaugti, žniaukti, žnybti, žniugti. Consequently, the alternation itself (šl-: kl-, šm-: km-, etc.) may likewise have arisen because many words exhibiting it are onomatopoeic in origin. Furthermore, K. Būga’s view may also be correct: not all variants containing a velar and a sibilant consonant can be derived from a single root; the variant with a velar consonant may have originated from one root, while the variant with a sibilant consonant may have originated from another.

If the assumption that palatal consonants were depalatalized before sonorants (should it prove correct in the future) is accepted, certain conclusions could be drawn. In the Balto-Slavic linguistic community, even before the assibilation of palatal consonants, alongside forms with the normal grade of ablaut (e.g., *k’eu-/*k’ou- and *g’er-/g’or-), in which the development of palatal consonants immediately preceding e and o must have been regular (that is, they must have become š and ž; cf. šau-kti and žar-dis), zero-grade forms existed as well (e.g., *k’u- and *g’r-). In the latter forms, palatal consonants immediately preceding u and r could have been depalatalized, that is, *ku- and *gr-→ku- and *gr- (cf. kù-kti, kù-kti; there are no examples with *gr-). Thus, alternation between palatal (> sibilant) and velar consonants could have arisen in forms of a single root. Later, through generalization, k and g could have come to occur before e and o, and š and ž before i and u. This could have led to doublets with identical or similar meanings (cf. šaikti: kaukti, šukti: kūkti), and sometimes with a certain semantic differentiation as well (cf. žardis: gardas). Consequently, the velar consonants in the words akmuõ and agnùs could have arisen through depalatalization of palatal consonants before m, n, whereas the words ašmuõ and ãžnas must have acquired their sibilants from a position before a vowel, where palatal consonants could regularly undergo assibilation.

Since the origin of the alternation between š and k (respectively ž and g) is very unclear, no conclusions characterizing the earliest relations between the Baltic and Slavic languages can be drawn. The assibilation of palatal velar consonants itself took place in the Baltic, Slavic, and Indo-Iranian languages. This circumstance also neutralizes the relevance of this phenomenon to the question of the earliest relations between the Baltic and Slavic languages. However, the assibilation of palatal consonants, which was structurally connected with the change of consonant s after k, r, i, u (see p. 86 of this work), suggests that, at the time of these changes, the Baltic, Slavic, and Indo-Iranian dialects were in linguistic and territorial contact. On the other hand, the differing details of these changes show that by then the aforementioned dialect groups were already autonomous linguistic units. The assibilation of palatal consonants in the Baltic languages took place before contacts between the Baltic and Finno-Ugric languages (in loanwords of Baltic origin, h arose from š and ž).

91

CHRONOLOGICAL QUESTIONS CONCERNING THE EARLIEST RELATIONS BETWEEN THE BALTIC AND SLAVIC LANGUAGES

Some features shared by the Baltic and Slavic languages are archaisms, others are innovations of the Baltic and Slavic languages, and still others are innovations also familiar to other Indo-European languages. Shared Baltic and Slavic morphological phenomena, such as nominal inflection, the suffixes *-esti-/*-osti-, *-lo-, *-er-/*-or-, the inflection of athematic and thematic verbs (*-mi, *-si, *-ti, *-ō, *-ēi,-), and so on, were inherited from the Indo-European linguistic community and should therefore be regarded as archaisms. Adjectives with pronominal inflection could also be archaisms, in part. The change of Indo-European s after k, r, i, u (Baltic š, Slavic ch) and the transformation of Indo-European palatal consonants k’, g’, g’h into Baltic š, ž and Slavic s, z are innovations also found in other Indo-European groups (Indo-Iranian). Moreover, there is evidence that these phenomena occurred independently in each language group. The baritone accentuation of long-root nouns (Hirt’s law) and the singular genitive ending *-ō in o-stem nouns likewise belong among phenomena found in other Indo-European languages.

Finally, the innovations of the Baltic and Slavic languages themselves are not all equivalent. The change of the Indo-European diphthong eu and de Saussure–Fortunatov’s law are undoubtedly innovations of the Baltic and Slavic languages. However, they arose in parallel in Baltic and Slavic, and, moreover, rather late and under different circumstances.

The innovations of the Baltic and Slavic languages whose conditions of emergence and final outcome were completely identical are as follows:

1) the transformation of the Indo-European syllabic sonant R into iR and uR;

2) the emergence of the intonational opposition “acute: circumflex”;

3) the emergence of certain deverbal formants (for example, *-ējo-, *-tājo-).

These innovations of the Baltic and Slavic languages require us to acknowledge an epoch of certain active contacts and communication between the Baltic and Slavic languages, in other words, an epoch of community. The question arises as to when this Baltic–Slavic community (or period of contact) may have existed.

Among archaeologists, the prevailing opinion is that the representatives of the Battle Axe and Corded Ware cultures may have been Baltic, Slavic, and Germanic tribes¹. Linguistic evidence appears to support this view as well. Such innovations shared by the Baltic, Slavic, and Germanic languages as the common ending -m- in some plural cases (cf. Lithuanian rankoms, Old Church Slavonic rąkamъ, Gothic gibom), the word for “thousand” (cf. Lithuanian tūkstantis, Prussian tū-simtons (acc. pl.), Old Church Slavonic tysęšti, tysąšti, Gothic þūsundi), the shared word for gold (cf. Latvian zèlts, Old Church Slavonic zloto, Gothic gulþ), the shared verb denoting rule (cf. Lithuanian valdyti, Old Church Slavonic vladъ, Gothic waldan), and numerous other shared lexical parallels² suggest that past relations among the Baltic, Slavic, and Germanic languages were exceptionally close, and that these languages experienced an epoch of community (or contact). Contacts among the Baltic, Slavic, and Germanic languages must have existed at the time when the inflectional system of these languages was taking shape, the concept of a thousand emerged, the first signs of authority appeared, and there were

¹ M. Gimbutas, The Balts, London, 1963, p. 43; А. Я. Брюсов, Об экспансии „культур с боевыми топорами“ в конце III тысячелетия до н. э., СА, 3, 1961, p. 14–33, Восточная Европа в III тысячелетии до н. э. (Этногенетический очерк), СА, 2, 1965, p. 47–56; П. Н. Третьяков, Финно-угры, балты и славяне на Днепре и Волге, М.—Л., 1966, p. 69–73.

² W. Porzig, The Structure of the Indo-European Linguistic Area, p. 140–143.

92

to that involving metals (cf. also the name for ‘silver’, borrowed from a common source but separately by each language group: Lithuanian sidãbras, Russian серебро, Gothic silubr). All this attests to the archaic nature of contacts among the Baltic, Slavic, and Germanic languages and is consistent with dating them to the third millennium BCE.

The dialects of the northern Indo-European area of that period, from which the Baltic, Slavic, Germanic, and other languages later arose, may have had a vocalic system consisting of five short (i, u, e, o, a) and long (ī, ū, ē, ō, ā) vowels, intact syllabic sonants R, a three-series consonant system (bh – b – p, dh – d – t, etc.), oppositions between the velar consonants kʷ – k, gʷ – g, and laryngeal consonants that had not yet disappeared in certain positions. It must be noted that it remains unclear whether an opposition between the short vowels o and a existed, since the Baltic, Slavic, and Germanic languages never distinguished these two vowels. The dialectal community of Baltic, Slavic, and Germanic must also have had free phonological stress.

Although the dialects of the northern Indo-European area shared a unified phonological model, old structural differences already existed and new ones were developing among the Baltic, Slavic, and Germanic dialects (especially in morphology). This claim may perhaps be supported by the further circumstance that the degree of affinity of the Baltic and Slavic languages with Germanic differs: the Baltic and Germanic languages are linked by more isoglosses than are the Slavic and Germanic languages.

Innovations shared by the Baltic and Germanic languages, such as the formation of the numerals “first” (cf. Lithuanian pirmas, Gothic fruma, the Old English form with the suffix *-mo-; Slavic has the suffix *-uo- here, cf. Old Church Slavonic пръвъ, Sanskrit pūrvah), “eleven,” and “twelve” (cf. Lithuanian vienúo-lika, dvy-lika, Gothic ain-lif, twa-lif, Old High German ein-lif, zwei-lif, whose second component is related to the Indo-European root *leigʷ- ‘remain’), the identical first- and second-person pronouns in the dual (cf. Lithuanian Samogitian vēdu, Gothic wit < *we-duo; Lithuanian jü-du, Gothic jut < *ju-duo), shared constructions for denoting natural phenomena (cf. Lithuanian sniegas drimba, Latvian sniegs drēb, Old Icelandic drífr snær), and numerous shared lexical items (cf., for example, the verb meaning ‘walk’, Lithuanian žengia and Old English gengān), show that the Baltic and Germanic dialects directly bordered one another and that their contacts were active and long-lasting. These striking innovations shared by the Baltic and Germanic languages suggest that there may even have been a Baltic–Germanic linguistic community within the framework of the dialects of the northern Indo-European area.

The numerals “first,” “eleven,” and “twelve” took shape in the Baltic and Germanic languages; the common Baltic and Germanic vocabulary denotes primitive conditions of everyday life; these languages lack social terminology⁴—all this shows that the similarities between the Baltic and Germanic languages are highly archaic and that their contacts may possibly predate contacts among the Baltic, Slavic, and Germanic languages. This assumption is reinforced by the fact that the innovations in the Baltic and Germanic languages arose before the palatalization and assibilation of velar consonants (cf. even the Lithuanian žengia and Old English gengān already mentioned).

Relations between the Slavic and Germanic languages are not characterized by such an abundance of innovations: these languages share chiefly lexical neologisms⁵.

³ For more recent works, see W. Porzig, Die Gliederung des indogermanischen Sprachgebiets, pp. 145–147; H. C. Чемоданов, К вопросу о германо-балтийских языковых связях, „Вопросы германского языкознания“, 1961, pp. 68–87; Сравнительная грамматика германских языков, I, pp. 70–79.

⁴ Сравнительная грамматика германских языков, I, pp. 78–79.

⁵ W. Porzig, Die Gliederung des indogermanischen Sprachgebiets, pp. 143–145; Сравнительная грамматика германских языков, I, pp. 79–81; О. Н. Трубачев, Ремесленная терминология в славянских языках, М., 1966.

93

An analysis of the Balto-Germanic, Slavo-Germanic, and Balto-Slavo-Germanic isoglosses seems to permit the following conclusion: the Baltic and Slavic languages most likely arose from dialects of the northern area of different Indo-European linguistic communities. Initially, the “Baltic” dialects may have maintained closer contacts with the “Germanic” dialects. If the Baltic and Slavic languages had arisen from a single dialect of the northern area of the Indo-European linguistic community, the Balto-Germanic isoglosses would probably not have appeared: they would have extended to include the territory of the Slavic languages as well.

The convergence of the “Baltic” and “Slavic” dialects may have begun during the Balto-Slavo-Germanic community (or, alternatively, period of contact). In addition to isoglosses encompassing the territories of all three dialect groups, separate Balto-Slavic isoglosses may also have arisen through immediate proximity and contact. It is entirely possible that during the Balto-Slavo-Germanic community (or, alternatively, period of contact), the anaphoric-demonstrative pronoun began to be used more frequently alongside adjectives in Baltic and Slavic dialects; the transformation of certain elements of the Indo-European morphological system (“Balticization” and “Slavicization”) began; and the first shared Balto-Slavic vocabulary appeared. At that time, certain distinct features of the Baltic and Slavic languages may have taken shape, for example, the sigmatic future tense in the Baltic languages and the sigmatic aorist in Slavic.

In connection with the migration of representatives of the battle-axe and Corded Ware cultures at the end of the third millennium BCE and the beginning of the second millennium BCE, the linguistic community of the Baltic, Slavic, and Germanic peoples that had existed in the northern area of the Indo-European linguistic community disintegrated.

A new epoch in the earliest relations between the Baltic and Slavic languages begins.

In the first half of the second millennium BCE, the Baltic and Slavic dialects seem to have experienced a period of particularly close contact. Although, according to archaeological evidence, the Balts and Slavs occupied vast areas of northeastern Europe at that time, it must be acknowledged that their dialects drew considerably closer. The particularly close contacts between the Baltic and Slavic dialects in the first half of the second millennium BCE are indicated not only by the aforementioned innovations in the Baltic and Slavic languages, which most likely arose at that time, but also by the relations of the Baltic and Slavic languages with the Indo-Iranian languages. Such isoglosses shared by Baltic, Slavic, and Indo-Iranian, as the genitive singular of personal pronouns (cf. Lithuanian dialectal mane, Old Church Slavonic mene, Avestan mana), the use of the pronoun *jos with nouns, the change of Indo-European consonant s after k, r, i, u, the affrication of palatal consonants, and numerous shared vocabulary items⁸, suggest that in the past the territories of the Baltic, Slavic, and Indo-Iranian dialects were adjacent and that these dialects maintained certain contacts. These contacts are dated to the first half of the second millennium BCE⁷. There is no doubt that contacts between the Baltic, Slavic, and Indo-Iranian languages occurred several centuries later than those between the Baltic, Slavic, and Germanic languages.

In addressing the question of the earliest relations between the Baltic and Slavic languages, the fact that the affinity between Baltic and Indo-Iranian is almost as great as that between Slavic and Indo-Iranian is important; in other words, the Baltic and Slavic languages are equally distant from the Indo-Iranian languages⁸. This fact seems to indicate that in the first half of the second millennium BCE (that is, approximately between 2000 and 1500 BCE) the Bal-

⁶ H. Arntz, Linguistic Relations between Aryan and Balto-Slavic; W. Porzig, The Structure of the Indo-European Linguistic Area, p. 164—169.

⁷ W. Porzig, op. cit., p. 169.

⁸ W. Porzig, op. cit., p. 169.

94

These dialects drew closer to the Slavic dialects. The following changes may have occurred in the systems of Baltic and Slavic dialects of that period:

the laryngeals were lost;

the opposition between voiced and voiced aspirated consonants (b: bh>b, etc.) was eliminated;

syllabic sonants R became iR and uR;

an opposition between intonations (“acute: circumflex”) was formed;

some new suffixes appeared (for example, *-ėjo- and *-tājo-);

palatal consonants k’, g’ were assibilated (Lith. š, ž; Sl. s, z);

PIE s after k, r, i, u changed (Lith. š, Sl. ch).

Some of these phenomena (the assibilation of velar consonants, the change of PIE s, and the loss of laryngeals) were shared by the Baltic, Slavic, and Indo-Iranian languages, and thus must have arisen at the time when the three aforementioned dialect groups were in contact, i.e. in the first half of the second millennium BCE. The change of PIE s after k, r, i, u and of palatal consonants, which most likely occurred independently in the Baltic, Slavic, and Indo-Iranian dialects, confirms the idea implicit here that dialects of PIE with a certain degree of autonomy already existed at that time.

In addition to the shared Baltic, Slavic, and Indo-Iranian isoglosses, it is self-evident that separate isoglosses of the Baltic and Slavic languages must also have arisen at that time (in the first half of the second millennium BCE). These were the aforementioned common features of the Baltic and Slavic languages—the change in syllabic sonants, the emergence of new derivational morphemes and lexemes, and, apparently, the intonational opposition that arose as a result of the disappearance of the laryngeals.

The date when these Baltic and Slavic innovations arose is unknown. We have no data that would help determine precisely when, for example, such Baltic and Slavic suffixes as *-ėjo-, *-tājo- arose. This raises the question of why the emergence of these Baltic and Slavic innovations is nevertheless dated to the first half of the second millennium BCE.

If, for example, one assumes that the opposition between voiced and voiced aspirated consonants was eliminated and that intonations arose as a result of the loss of laryngeals, the simultaneity of the elimination of the aforementioned consonant opposition, the emergence of intonations, and the disappearance of the laryngeals becomes particularly evident.

The iR and uR that developed from syllabic sonants R were already treated as diphthongs when the intonations were forming; therefore, it must be assumed that the syllabic sonants changed before the emergence of the intonations. It follows that all these phenomena arose at approximately the same time or immediately one after another.

Since the change in syllabic sonants, the origin of intonations, and the formation of the derivational morphemes *-ėjo-, *-tājo- were all innovations of the Baltic and Slavic languages, it may be surmised that they arose at approximately the same time. Since the causes and conditions behind the emergence of the aforementioned Baltic and Slavic innovations, as well as their final outcomes, were identical in the Baltic and Slavic languages, it must be assumed that these innovations arose when contacts and communication between the Baltic and Slavic languages were most active. The fact that, during the change of PIE s after k, r, i, u and the assibilation of palatal consonants, the Baltic, Slavic, and Indo-Iranian dialects maintained certain contacts, together with the nearly equal degree of closeness between the Baltic and Slavic languages and the Indo-Iranian languages, indicates that the most active and closest relations between the Baltic and Slavic dialects existed in the first half of the second millennium BCE. It follows that the emergence of Baltic and Slavic innovations may

95

dateable to the first half of the 2nd millennium BCE (i.e., approximately 2000–1500 BCE).

It seems that contacts between Baltic and Finnic languages, in which Slavic dialects did not participate, preclude regarding those innovations as having arisen later. Thus, during contacts between Baltic and Finnic languages, the relations between Baltic and Slavic languages must, perhaps because of the great distances involved, have become considerably weaker. Contacts between Baltic and Finnic languages must have begun after the syllabic sonorants R had already changed into iR and uR, and the palatal consonants k’, g’ had already changed into š, ž; compare Finnish villa and Lithuanian vilna; Finnish kurpponen and Lithuanian kùrpė; Finnish halla, Estonian hall and Lithuanian šalnà; Finnish hihna, Estonian ihn and Lithuanian šikšnà (Finnish and Estonian h from Baltic š and ž). If we date the aforementioned changes in the Baltic languages to the first half of the 2nd millennium BCE (i.e., approximately 2000–1500 BCE), then contacts between Baltic and Finnic languages must have begun around the third quarter of the 2nd millennium BCE or later (but by no means earlier). Since relations between Baltic and Slavic languages seem to have been rather weak during contacts between Baltic and Finnic languages, that weakening must likewise have begun in the third quarter of the 2nd millennium BCE or later. This by no means implies, however, that relations, neighborly contact, and communication between Baltic and Slavic dialects had been entirely broken off from then on. From that time, the parallel development of Baltic and Slavic languages may have begun. The independent emergence of parallel, entirely analogous individual phenomena in Baltic and Slavic languages (the change of the IE diphthong eu, Saussure–Fortunatov’s law and, perhaps, the 2nd person singular ending *-sēi of verbs) may have depended on these languages’ descent from the IE linguistic community, inherited shared developmental tendencies, their identical structural type, and geographical contiguity.

One circumstance may seem to contradict dating the emergence of innovations in Baltic and Slavic languages and the active contacts between those languages to the first half of the 2nd millennium BCE. Archaeologists believe that, at the end of the 3rd millennium BCE and the beginning of the 2nd millennium BCE, in connection with the expansion of the Battle Axe and Corded Ware cultures, the ancient Balts, or Proto-Balts, occupied a territory extending from the Baltic Sea to the upper reaches of the Volga and Oka⁹, while the Slavs occupied roughly the area between the upper reaches of the Vistula and the middle Dnieper. A quite reasonable question arises: could active and close contacts between Baltic and Slavic languages have existed across so vast an area, and could identical innovations affecting all Baltic and Slavic dialects have emerged? Numerous innovative isoglosses generally arise in a limited area as a result of active and close contacts. It should be said, first of all, that a date such as “at the beginning of the 2nd millennium BCE or in its first centuries” may include (and probably does include) the beginning or even the middle of the second quarter of the 2nd millennium BCE. Naturally, ancient tribes could not have occupied so vast a territory all at once and simultaneously—the process must have been complex and prolonged. The scholar of the Baltic past M. Gimbutienė believes¹⁰ that the Balts settled this vast territory gradually, spreading eastward from the Sambia Peninsula and the shores of the Baltic Sea. The final settlement of this territory and the Balts’ coming into contact with the Finns may have been processes that took place much later than the beginning of the 2nd millennium BCE.

⁹ X. A. Moopa, O древней территории расселения балтийских племен, СА, 2, 1958, p. 9–33; А. Я. Брюсов, op. cit.; П. Н. Третьяков, op. cit.; Gimbutas, op. cit.

¹⁰ M. Gimbutienė, Rytprūsiai ir Vakarų Lietuvos priešistorinės kultūros apžvalga (Atspaudas iš „Mažosios Lietuvos“), New York, 1958.

96

Thus, it may be assumed that in the first half of the 2nd millennium BCE (approximately 2000–1500 BCE), the Baltic and Slavic dialects, which had previously differed considerably more, drew closer and underwent an epoch of active and close contacts, neighborly relations, and communication—in other words, an epoch of community (or contacts), during which a bundle of innovative isoglosses came to encircle the Baltic and Slavic dialects. For this reason, the epoch of community (or contacts) between Baltic and Slavic languages may also be called the epoch of isoglossic community.

In the dynamics of the development of the earliest relations between Baltic and Slavic languages, one can observe moments of independent, divergent, and parallel development. The epoch of isoglossic community encompasses only one of the periods in the development of the earliest relations between Baltic and Slavic languages.

The understanding of the earliest relations between Baltic and Slavic languages set out here approaches J. Endzelin’s conception (славяно-балтийская эпоха) and the views held by Chr. Stang (Isoglossengebiet), E. Fraenkel (Nachbarschaft), S. Bernstein (сообщность), B. Gornung („балто-славянская эпоха“), G. Shevelov (a period of common development), J. Safarewicz, V. Porzig, A. Salys, A. Erhart, and other scholars.

Thus, the earliest relations between Baltic and Slavic languages, and hence the origin of their shared innovations, can be interpreted in terms of the existence of a Baltic-Slavic isoglossic community (Isoglossengemeinschaft), that is, the existence of a spatiotemporal continuum of Baltic and Slavic dialects, which most likely arose through the convergence of dialects that had initially differed.

Conclusions

1. In the 3rd millennium BCE, within the framework of linguistic community among the dialects of the northern area, the “Baltic” dialects had become distinct from the “Slavic” dialects. At that time, the “Baltic” dialects had maintained closer contacts with the “Germanic” dialects.

2. In the first half of the 2nd millennium BCE (approximately 2000–1500 BCE), the Baltic dialects had drawn closer to the Slavic dialects and undergone an era of active contacts, or isoglossal community, with them. The Baltic and Slavic dialects of that period had been encircled by the following bundle of isoglosses:

the laryngeals disappeared;

the opposition between voiced and voiced aspirated consonants was eliminated (b: bh > b, etc.);

syllabic sonorants Ṛ became iR and uR;

an opposition of intonations (“acute: circumflex”) developed;

some new suffixes appeared (for example, *-ėjo- and *-tājo-);

the palatal consonants k’, g’ were assibilated (Lith. š, ž; Sl. s, z);

PIE s changed after k, r, i, u (Lith. š, Sl. ch).

The change in syllabic sonorants, the development of the intonational opposition, and the emergence of specific Baltic and Slavic language morphemes and lexemes were common innovations of the Baltic and Slavic languages.

3. Around the middle of the 2nd millennium BCE or later, the convergent development of the Baltic and Slavic dialects had ceased, and their parallel development had begun. From then on, an era of contacts between the Baltic and Finnic languages may have begun, in which the Slavic languages did not participate. Such shared phenomena of the Baltic and Slavic languages as the PIE diphthong eu becoming

7. Baltų ir slavų kalbų ryšiai

97

Baltic jau and Slavic ju, de Saussure–Fortunatov’s law, and perhaps the formation (at a later stage of development) of the singular second-person endings *-sēi of pronominal adjectives and verbs were most likely consequences of the parallel development of the Baltic and Slavic languages.

4. The earliest relations between the Baltic and Slavic languages, and thus the origin of their shared innovations, can be interpreted in terms of the existence of a Baltic–Slavic isogloss community (Isoglossengemeinschaft), that is, a spatiotemporal continuum of Baltic and Slavic dialects that most likely arose through the secondary convergence of dialects that had initially been distinct.

Over the long history of the Baltic and Slavic languages, their relations can be observed to have been independent, convergent, and parallel. The era of the Baltic–Slavic isogloss community was only one period in the development of the earliest relations between these languages.

SOME QUESTIONS CONCERNING THE EARLIEST RELATIONS BETWEEN THE BALTIC AND SLAVIC LANGUAGES

С. КАРАЛЮНАС

Summary

This paper examines selected questions concerning the problem of the earliest Baltic–Slavic linguistic relations. The first chapter presents the history of research on this problem. The account begins with a brief overview of views on the relations among the Baltic, Slavic, and other Indo-European languages in the works of the founders of comparative-historical linguistics, and concludes with a presentation of the results of recent research on the problem.

The second chapter is devoted to an examination of certain shared phonetic phenomena in the Baltic and Slavic languages. A short phoneme-class a is reconstructed for the Proto-Baltic linguistic community. Its existence in this community is indicated by the development of the short- and long-vowel systems according to the principle of polarization in the Samogitian dialect, the preservation of the more archaic character of the short vowel before m, n in that same dialect, possibly forms such as Old Prussian wobse (with o instead of the expected short vowel) in the Elbing Vocabulary, and Baltic loanwords in the Finno-Ugric languages (cf. Finnish morsian and Lithuanian marti). In East Aukštaitian dialects, the lengthening of short vowels in stressed position created a correlation between short and long vowels; therefore, both systems must have undergone labialization and raising. Thus, the existence of short o in these dialects is most likely an innovation. Phonological research in recent years (especially that of V. Zhuravlev) has established that in Proto-Slavic, through the process of consonant labialization, resp. vowel delabialization, the opposition o: a, resp. ō: ā, was neutralized first after labial consonants (since the distinctive feature of bemolization was a constitutive element of these consonants), and later after the other consonants as well. The result of the merger through dissimilative neutralization was o, resp. ā. Therefore, akanye in Belarusian and the southern Russian dialects, as well as in some Bulgarian dialects, may be considered a relic of the Proto-Slavic state. Indo-European ō in the (East) Baltic languages merged with Indo-European ā in unstressed position (hence Lithuanian o, Latvian ā), while in stressed position it became uo; that is, the opposition ō: ā was apparently preserved in this position because Baltic ō entered into a correlation with ē (from ei and ai).

The regular development of the Indo-European diphthong eu in the Baltic and Slavic languages was its change to Lithuanian-Latvian jau and Slavic *u. There are some grounds for supposing that this diphthong was preserved in Old Prussian. The diphthong au (Slavic u) in those words where it occurs in place of eu in other Indo-European languages is most plausibly considered to have developed from the Indo-European diphthong oi (or *Hu).

98

The transformation of Proto-Indo-European syllabic sonants R into iR and uR is a shared Balto-Slavic innovation. The distribution of i and u is explained by two theories—the ablaut theory and the phonetic-environment theory—which are equally plausible, since, on the basis of the relevant material from the Baltic and Slavic languages, neither can be proved nor disproved.

Proto-Indo-European s changed to i and then to š also in the Baltic languages (in Latvian and Old Prussian, in accordance with the specific development of the phonological systems of these languages, Common Baltic š reverted to s). This change of Proto-Indo-European s did not occur in two cases: 1) if i, u belonged to one (root) morpheme, while the consonant s belonged to another (suffixal) morpheme, i.e. if a morpheme boundary occurred between i, u and s; 2) if a sonant (r, l, m, n, j, v) followed s. The change of Proto-Indo-European s after k, r, i, u was structurally connected with the assimilation of Proto-Indo-European palatal k’ in the Baltic languages (cf. the identical reflex š of these changes). In these same languages we find many cases in which the consonant k (or g) occurs instead of the expected š (or ž). Since in most cases this k (or g) occurs before the sonants r, l, m, n, j, v, it might be assumed that depalatalization of palatal k’ (or ǵ) took place precisely in this position. This assumption is valid in the case of r, since Lithuanian has no clusters *šr-, *žr-. However, numerous examples containing šl-, žl-, šm-, žm-, etc., contradict the assumption that k’ (or ǵ) was depalatalized before other sonants as well. It should nevertheless be borne in mind that most of these examples may be suspected of having an onomatopoeic origin. Consequently, the fairly frequent alternation š/k, ž/g in Lithuanian may in most cases have the same origin. Moreover, some doublets with š/k and ž/g may derive from different roots (the form with š or ž from one root, and the form with k or g from another).

Some details of the change of Proto-Indo-European s after k, r, i, u and the assimilation of Proto-Indo-European k’ (cf., for example, the change of this s after k, l, i < H and the depalatalization of Proto-Indo-European k’ before s> Old Indian ś in the Indo-Iranian dialects, etc.) indicate, on the one hand, that these changes took place independently in the Baltic, Slavic, and Indo-Iranian language groups. In other words, by the time these changes took place, the Baltic, Slavic, and Indo-Iranian dialects already possessed a certain degree of autonomy. On the other hand, during these changes, certain territorial and linguistic contacts apparently existed between these dialects.

Chapter Three attempts to address some questions concerning the earliest relations between the Baltic, Slavic, and other Indo-European (and non-Indo-European) languages from a chronological perspective. The linguistic evidence does not appear to contradict the conclusion drawn from archaeological data that the bearers of the Battle Axe and Corded Ware cultures were probably Baltic, Slavic, and Germanic tribes: the known Balto-Slavic-Germanic isoglosses point to the existence of Balto-Slavic-Germanic linguistic contacts and apparently make it possible to date them to the third millennium BCE. There are, however, also separate Balto-Germanic isoglosses whose archaic character, as well as the absence of a shared Balto-Germanic social terminology, inevitably suggests that Balto-Germanic linguistic contacts are extremely ancient and apparently predate Slavo-Germanic contacts. Consequently, the Baltic and Slavic languages most likely developed from different dialects of the northern area of the Proto-Indo-European speech community. Initially, the Baltic dialects may have maintained closer contacts with Germanic dialects. The rapprochement of the Baltic and Slavic dialects apparently began during the period of Balto-Slavic-Germanic contacts. These dialects apparently had particularly close contacts in the first half of the second millennium BCE, that is, during the period of Balto-Slavic-Indo-Iranian linguistic contacts. At that time, the following changes may have taken place in the systems of the Baltic and Slavic dialects: the loss of laryngeals in certain positions and, as a consequence of this loss, the elimination of the opposition between voiced and voiced aspirated consonants and the formation of an intonational opposition; the reflexation of Proto-Indo-European syllabic sonants; the formation of new morphemes; the change of Proto-Indo-European s after k, r, i, u; and the assimilation of Proto-Indo-European palatal consonants. All these phenomena, except the last two, were shared Balto-Slavic

7*

99

innovations. Close Balto-Slavic linguistic contacts apparently existed before the Balto-Finno-Ugric linguistic contacts, which began approximately in the second half of the second millennium BCE or later, since Slavic dialects did not participate in the latter.

Thus, the earliest Balto-Slavic linguistic relations are interpreted as the existence of a Balto-Slavic isogloss community (Isoglossengemeinschaft), that is, a spatiotemporal continuum of Baltic and Slavic dialects which apparently arose, however, through the secondary rapprochement of dialects that were originally distinct.

ON THE EARLIEST RELATIONS BETWEEN THE BALTIC AND SLAVIC LANGUAGES

S. KARALIŪNAS

Summary

In the 3rd millennium BCE, “Baltic” and “Slavic” dialects probably differed within the northern area of the Indo-European linguistic community. At that time, the “Baltic” dialects seem to have maintained fairly close relations with the “Germanic” dialects. Well-known Baltic-Germanic commonalities probably point to this.

In the first half of the 2nd millennium BCE (approximately in the years 2000–1500 BCE), Baltic and Slavic dialects may have drawn closer to one another and lived in an era of active contacts. The following bundle of isoglosses probably arose in the Baltic and Slavic dialects of that time: loss of the laryngeals, elimination of the opposition between voiced stops and voiced aspirated stops (b: bh>b, etc.), change of syllabic sonants R to iR and uR, emergence of phonological intonations (acute: circumflex), formation of some common suffixes (e.g. *-ēio-, *-tāio-), assimilation of IE k’, g’ (Lith. š, ž, Slav. s, z), and change of IE s after k, r, i, u (Lith. š, Slav. ch).

The change of syllabic sonants, emergence of phonological intonations, and formation of specifically Balto-Slavic morphemes were shared innovations of the Baltic and Slavic languages.

One may assume that around the middle of the 2nd millennium BCE or later, the convergent development of the Baltic and Slavic dialects came to an end and their parallel development began. From then on, the period of Baltic-Finnic contacts may have begun, in which the Slavic languages did not participate. Such common features of the Baltic and Slavic languages as the change of IE eu to Lith.-Latv. jau, Slav. ju, de Saussure–Fortunatov’s law, the formation of definite adjectives and the verbal ending *-sei of the 2nd person singular present, among others, probably arose as a result of the parallel development of the two branches.

The earliest relations between the Baltic and Slavic languages, and thus the emergence of common Balto-Slavic innovations, could be conceived as involving a community between the Baltic and Slavic dialects, that is, a spatiotemporal continuum of these dialects, which apparently arose, however, because of the secondary convergence of originally distinct Baltic and Slavic dialects. The period of the Balto-Slavic linguistic community was therefore only one phase in the development of the earliest relations between the Baltic and Slavic languages.