LIETUVIŲ KALBOTYROS KLAUSIMAI, II LIETUVOS TSR MOKSLŲ AKADEMIJA

NOTES ON THE QUESTION OF THE EARLIEST RELATIONS BETWEEN THE BALTIC AND SLAVIC LANGUAGES

V. MAŽIULIS

If nineteenth-century Indo-European linguistics (A. Schleicher, the Neogrammarians, etc.) adhered fairly unanimously to the hypothesis of a Balto-Slavic protolanguage [which was born of nothing other than A. Schleicher’s “family tree” theory of languages (Stammbaumtheorie)], later other opinions began to emerge on this matter. One of the most prominent Russian linguists of that period, J. Baudouin de Courtenay, in opposition to the Neogrammarians, believed that there was no basis for speaking of a Balto-Slavic protolanguage, since there were supposedly at least as many arguments “against” as there were “for”; however, J. Baudouin de Courtenay never developed his claims in greater detail. The first to try to substantiate with arguments the nonexistence of a Balto-Slavic protolanguage was the well-known French linguist A. Meillet. In his study “Les dialectes indoeuropéens” (Paris, 1908), criticizing K. Brugmann’s conception of a Balto-Slavic protolanguage, he concluded that the genetic relations between the Baltic and Slavic languages should be explained by the principle of parallel development. Understandably, this innovation by A. Meillet seemed heretical to the traditionally minded Neogrammarians. Yet it was (and remains) far from acceptable to all other linguists as well, especially those of more recent times—not only to proponents of a Balto-Slavic protolanguage. The celebrated researcher of the Baltic languages J. Endzelīns, like A. Meillet, was disinclined to recognize a Balto-Slavic protolanguage, but approached the problem differently from A. Meillet. In his work “Славяно-балтийские этюды” (Харьков, 1911), having reviewed the similarities and differences between Baltic and Slavic with extraordinary care, J. Endzelīns concluded that the Balts and Slavs descended from separate Indo-European tribes, speaking separate (though quite similar) dialects, which later experienced an epoch of coexistence—the “Balto-Slavic epoch” (l. c., pp. 200–201)¹. Our great linguist K. Būga held precisely the same opinion²—the foremost scholar of this problem from the perspective of Baltic toponymy, who in the last years of his life viewed the hypothesis of a Balto-Slavic protolanguage with particular skepticism. “Other linguists,” K. Būga wrote, “believe that the common features shared by the languages of the Aestii and the Slavs,”

¹ Cf. also J. Endzelīns, Lekcijas par baltu valodu salīdzināmo gramatiku, 6, Rīgā, 1927; J. Endzelīns, Latviešu valodas gramatika, 18, Rīgā, 1951; Я. Эндзелин, Древнейшие славяно-балтийские языковые связи. — „Труды Академии наук Латвийской ССР“, 2, 67—82, Рига, 1953; cf. Вопросы языкознания, 4, 125, Москва (1952)

² K. Būga, Kalbų mokslas bei mūsų senovė. — „Draugija“, 20, 9—10, Kaunas (1913). K. Būga, Rinktiniai raštai, 1, 407—408, Vilnius, 1958.

5

show that the Aestii and Slavs also had a separate protolanguage after the Indo-European one. I do not believe this, for I find not a single linguistic fact that compels us to recognize the alleged Aestii-Slavic protolanguage. All the common features of the Aestii and Slavs show only that the Aestii have lived as neighbors of the Slavs from time immemorial, that is, since the Indo-European period³. Some⁴ believe that K. Būga was a complete supporter of A. Meillet’s conception, but this cannot be accepted. K. Būga’s quite clearly negative view of a Balto-Slavic protolanguage does not entitle us to say that K. Būga opposed J. Endzelīns’s conception, since K. Būga, while always emphasizing the particular closeness of the Baltic and Slavic languages, supported the principle of contact-driven development of these languages (not without the influence of J. Schmidt’s “wave theory”⁶); cf. “the common features of the languages show that the Aestii were neighbors of the Slavs for much longer than the Italic peoples were neighbors of the Celts”⁵. Pr. Skardžius, it seems, held an opinion similar to Endzelīns’s⁶. A considerable number of rational arguments in favor of Endzelīns’s theory (or, rather, against a Balto-Slavic protolanguage) have recently been put forward by the prominent Soviet Slavist Prof. Dr. S. Bernstein (С. Б. Бернштейн) in his article “Балто-славянская языковая сообщность”⁷, whose central idea is succinctly expressed in the author’s following words: “Our article (that is, S. Bernstein’s—V. M.) will deal precisely with Verkehrsgemeinschaft (a term the author subsequently translates as сообщность—V. M.), that is, with the commonality of the Baltic and Slavic languages, which arose not from their common descent from the Indo-European protolanguage, but from prolonged contact that began considerably later than the Indo-European period and ended well before our era”⁸.

Without examining in greater detail the history of the question of the genetic relationship between the Baltic and Slavic languages (many works have already been devoted to it), it should be added that, despite the major studies of J. Endzelīns, K. Būga, E. Fraenkel, Chr. Stang, M. Vasmer, L. Bulakhovsky, and other scholars (Balticists, Slavists, and Indo-Europeanists, as well as archaeologists, etc.), resolving once and for all the problem of the earliest relations between the Baltic and Slavic languages undoubtedly still requires considerable work⁹. In any case, it should be said that nowadays typology is becoming increasingly important in the study of both individual languages and related languages; V. Ivanov and V. Toporov, for example, make fairly fruitful use of it in investigating the problem of the earliest relations between the Baltic and Slavic languages¹⁰.

This article has examined the origins of the earliest relations between the Baltic and Slavic languages

³ K. Būga, Šis-tas iš lietuvių ir indoeuropiečių senovės. — „Tauta ir žodis“, 2, 106, Kaunas, 1924; cf. also op. cit., 1, 431.

⁴ Cf., for example, „Славянская филология (Philologie slave)“ (hereafter — СФ) 1, 8, Москва (1958).

⁵ K. Būga, op. cit., 106.

⁶ Pr. Skardžius, Lietuvių kalba, jos susidarymas ir raida (cited according to СФ, 1, 31. — V. M.).

⁷ See СФ, 1, 45—67.

⁸ СФ, 1, 45.

⁹ Some linguists (especially proponents of a Balto-Slavic protolanguage) maintain that the fundamental question in this problem (that is, the existence of a Balto-Slavic protolanguage, etc.) has now been definitively resolved, a view by no means shared by everyone.

¹⁰ See В. В. Иванов и В. Н. Топоров, К постановке вопроса о древнейших отношениях балтийских и славянских языков (IV Международный съезд славистов), Москва, 1958.

6

we do not intend to address the problem, but limit ourselves to a few specific remarks on certain questions that have some significance for investigating the oldest relations between these language groups¹¹.

I Lith. duomi, demi, OCS. damb stem

§ 1. For more than a hundred years, linguists have been studying the present-tense stem of Lith. duomi, duosti, duo-dąs, OCS damb, dastō, dadetō, dadų, and Lith. demi, desti, dedąs. The whole difficulty is that many of the reconstructed reduplicated proto-forms *dōd(a)mi resp. *dedh(a)mi¹² do not correspond in their vocalism to the Proto-Indo-European *dedōmi resp. *dedhēmi (usually reconstructed on the basis of Sanskrit dadāmi, dadhāmi, Greek δίδωμι, δίθημι¹³, etc.). There are numerous explanations for this discrepancy (see footnote 12 on p. 7), generally based on contamination of the root *dō- resp. *dhē- with the corresponding reduplicated present-tense stems (for an additional point, cf. footnote 17 on p. 8); whatever the case, however, this leaves unexplained the disappearance of the long-vowel stem ending in the singular (cf. Sanskrit dadāmi, Greek δίδωμι, etc.).

2. J. M. Kořinek explains Lith. duomi and OCS damb differently¹⁴. In his view, Baltic and Slavic *dōmi, *dōme derive from earlier *dōmi resp. *dōmes, which are nothing other than Proto-Indo-European forms, except that they have lost the reduplicative syllable in all numbers¹⁵. However, J. M. Kořinek’s explanations concerning the reduplicated pre-

¹¹ The same ideas and material are presented here (though in some places revised slightly differently) as were set out in my pamphlet: Заметки к вопросу о древнейших отношениях балтийских и славянских языков (IV Международный съезд славистов), Вильнюс, 1958.

¹² The forms Lith. duomi, OCS damb, and Lith. demi, with a similar reduplicated stem in Proto-Baltic resp. Proto-Baltic and Proto-Slavic, are reconstructed directly or as compromise forms by: A. Schleicher, Litauische grammatik, 253—254, Prag, 1856; G. Mahlow, Die langen Vocale AEO in den europaeischen Sprachen, 84—85, Berlin, 1879; O. Wiedemann, Das litauische Präteritum, 54—55, Strassburg, 1891; O. Wiedemann, Handbuch der litauischen Sprache, 114, Strassburg, 1897; K. Brugmann, Grundriss... 2 Band, Teil 3, 1 Lieferung, 110 ir kt., Strassburg, 1913; W. Vondrák, Vergleichende slavische Grammatik, 1, 506 ir kt., Göttingen, 1906; A. Leskien, Grammatik der altbulgarischen Sprache, 188—189, Heidelberg, 1909; A. Meillet, Le slave commun, 207—208, Paris, 1934; A. Meillet, Introduction... 204, Paris, 1934; R. Trautmann, Baltisch-slavisches Wörterbuch, 47, 56—57, Göttingen, 1923; Chr. Stang, Das slavische und baltische Verbum, 21—22, 268, Oslo, 1942; A. Vaillant see Bulletin de la société de linguistique de Paris, 40, 22 (although for a different earlier view, see Revue des études slaves, 14, 30, etc.); A. Баян, Руководство по старославянскому языку, § 222, Москва, 1952; Я. С. Отрембский, Славяно-балтийское языковое единство. — „Вопросы языкознания“ (hereafter ВЯ), 6, 35—36, (1954); J. Otrębski, Gramatyka języka litewskiego, 3, § 430, Warszawa, 1956; J. Endzelīns, Latviešu valodas gramatika, 723 ir kt., Rīgā, 1951; M. Vasmer, Russisches etymologisches Wörterbuch, 1, 327, 347, Heidelberg, 1953; В. Георгиев, Към въпроса за балтославянската езикова общност. — „Български език“, год VIII, 2, 107, (1958).

¹³ „The Greek distribution (ε for aorist and perfect, ε for present) is secondary“ — E. Schwyzer, Griechische Grammatik, II Lief., 648, München, 1939.

¹⁴ J. M. Kořinek, Present forms of the root *dō-, ‘to give’, in the Slavic and Baltic languages. — „Listy filologické“, volume 65, Praha, 445—454, 1938.

¹⁵ E. Fraenkel supports J. M. Kořinek’s explanations: E. Fraenkel, Baltische Sprachwissenschaft in den Jahren 1938—1940, 11—12, Helsinki, 1941; E. Fraenkel, Die baltischen Sprachen, 92 et seq., Heidelberg, 1950; E. Fraenkel, Litauisches etymo-

7

The aorist stem—OChurchSl. dadetъ, dady, Lith. duodąs (cf. Skt. dadat, Gk. διδοὺς)—and likewise the 2 pl., 2 du., and 3 sg. forms are, in my opinion, quite artificial¹⁶.

§ 3. First, it should be said that there are two possible approaches, or two alternatives, for resolving this question. According to one, the traditional approach, reduplication in the present stems of these verbs goes back to the Proto-Indo-European period (see § 4–§ 8); according to the other, it could have arisen later as an isoglossal innovation among the Indo-European dialects (see § 11).

§ 4. Under the first alternative, it would seem most reliable to assume that Lith. duomi, demi and OChurchSl. dambъ lost the reduplicated syllable only in the singular forms: that is, in the singular forms these two present stems had no consonant *-d- before the ending; cf., for example, the Sanskrit and Greek paradigms: dádāmi (dá-dāsi, dádāti: dadmās, etc.), δίδωμι (δίδως, δίδωσι: *δίδομεν > δίδομεν, etc.), τίθημι (τίθης, τίθησι: *τίθεμεν > τίθεμεν, etc.), and so on.

§ 5. In the Baltic languages (cf. Slavic and others), the singular forms of athematic verbs and, more generally, of all verbs never exceed the plural and dual forms in the number of stem segments; cf., however, Skt. dádāmi: dadmās, dadvás, Gk. δίδωμι: *δίδομεν > δίδομεν, etc. Consequently, in Baltic (and Slavic), the reduplicated stems of the singular forms *dedōmi, *dedēmi became phonologically and morphologically isolated in the singular (they were lost):

sg. 1. *dōm(e)i pl. 1. *dedmē du. 1. *dedvā part. praes. act. *dedan(t)s

2. *dōs(e)i 2. *destē 2. *destā

3. *dōti 3. *dedint(i), -ant(i).

sg. 1. *dēm(e)i pl. 1. *dedmē du. 1. *dedvā part. praes. act. *dedan(t)s

2. *dēs(e)i 2. *destē 2. *destā

3. *dēti 3. *dedint(i), -ant(i).

§ 6. Owing to the qualitative (and quantitative) opposition between the stem vowels, i.e. *-ō-: *-ē-, the stem vowel *-ō- of the singular forms *dōmi, *dōsi, *dōti could have been generalized throughout the entire present stem¹⁷ (cf. Lith. ēmi: *imē ≈ Skt. emi: imās, Lith. esmi: *smē ≈ Skt. asmi: smas, etc.):

sg. 1. *dōm(e)i pl. 1. *dōdmē du. 1. *dōdvā part. praes. act. *dōdan(t)s

2. *dōs(e)i 2. *dōstē 2. *dōstā

3. *dōti 3. *dōdint(i), -ant(i).

§ 7. The third-person singular form *dōti became *dōstē by analogy with the 2 pl. *dōstē, 2 du. *dōstā, 3 pl. *dōdint(i), -ant(i), and the pres. act. participle *dōdan(t)s (cf. 3 sg. ēsti: 2 pl. ēste, 3 pl. *ēdint(i), -ant(i), pres. act. participle *ēdan(t)s, etc.). In the first-person plural, *-d- regularly disappeared before -m-, and by analogy it also disappeared in the corresponding dual form¹⁸, no doubt with a corresponding influence on the singular as well

logisches Wörterbuch, 112, Heidelberg, 1955; V. Machek, Etymologický slovník jazyka českého a slovenského, 81 (however, cf. “jen 3. plur. stč. dadie z dā-d-redupl.”), Praha, 1957. Similarly, cf. O. Szemerényi, Zwei Fragen des urslav. Verbums. — Études slaves et roumaines, 1, 7—14, (1948); however, this work, like that of J. Safarewicz [J. Safarewicz, Przyczynki do zagadnienia wspólnoty bałto-słowiańskiej. — Sprawozdania z czynności i posiedzeń P.A.U., 46, Nr. 8, 199—202, (1945)], was unavailable to the author.

¹⁶ Cf. also Endzelīns’s critical remarks concerning 3 sg.—J. Endzelīns, op. cit., 724.

¹⁷ In this case, the root -ō-, e.g., *dōt(e)i (infinitive), can hardly have had any influence; cf. *dēt(e)i > Lith. dėti, but demi, desti, OCS dĕti, but dežǫ (see below).

8

of the first-person form of the count¹⁸ (in which *-d- was absent; see above). Similarly arose OCS damo, dasi, dastõ, damž, daste, dadetõ (part. praes. act. dady, -ąšta), dabē, daste, daste, -ta stem forms.

§ 8. In the paradigm of the verb demi, through the quantitative opposition of the stem vowels, i.e. -ē- (sg.): -ē (pl., du. and in the corresponding participles), the vowel -ē- becomes generalized (there are also more present-tense forms with this stem-vowel quantity). In Lithuanian, *dēm(e)i developed similarly to *dōm(e)i, but the stem vowel in the present tense is ē-¹⁹, while in the *-d- forms *dedmē, *dedvā disappeared, apparently by analogy (not phonologically at all!) with the forms *dōmē, *dōvā (see above), *ēdmē > *ēmē, *ēdvā > *ēvā, etc., all the more since in the singular *dēm(e)i (instead of the older *dēm(e)i) had no *-d-; thus arose Lith. demi, desi, destis, etc. The Slavic present-tense paradigm of this verb developed similarly, becoming thematic at a fairly early date—OCS deždą < *dedjǫ < *demi (but forms analogous in terms of the paradigm are OCS damb, dasi, etc.); cf. OCS viždǫ (cf. the athematic imperative viždь): Lith. veimu ‘I seek’ (Dieveniškės—Lingua Posnaniensis 5, 72–73) < *veimi < *veidmei; cf. A. Meillet, Le slave commun, II éd., p. 206.

§ 9. Dar keletas pastabų — dėl šių veiksmažodžių daugiskaitos trečiojo asmens (3pl.) lyties. Pirmiausia pasakytina, kad 3 pl. lyties tam tikras formantinis santykis — koreliacija su praesentinio veikiamosios rūšies dalyvio lytimi gana aiškiai matomas indoeuropiečių kalbose (formantas *-nt-²⁰). Jeigu baltų prokalbėje iš tikrųjų egzistavo (tuo esame įsitikinę visi) 3 pl. lytis, atitinkanti (šiuo atveju) ssl. dadetъ, dadǫtъ, skr. dadati, dadhati ir pan., tai ja, t. y. *dōdint(i), *dōdant(i), *dedint(i), *dedant(i), aplinkiniu keliu, gal būt, rodo dalyvinė lytis tokiuose išsireiškimuose: nėr kas duodą, nėr kas dedą (Dusetos, Antazavė ir kt.) ²¹. [Savo ankstyvesnį hipotetinį spėjimą (bei prie jo esančią išnašą)²², kad D. Kleino gramatikoje pateikiamosios 3 sg. pl. du. lytys dūd, (pra)ded netiesiogiai, gal būt, rodančios prabaltiškąsias 3 pl. lytis *dōdint(i), *dedint(i), dabar laikau esant niekuo nepagrįstu]. Kadangi praesentinės šių dviejų, kaip ir kitų atematinių veiksmažodžių veikiamosios rūšies dalyvis jau, rodos, iš labai senų laikų turi formantą -ant- < ide. *-ont-²³ (ne -int-!), tai vietoje senesnės 3 pl. lyties *dōdint(i), *dedint(i), (plg. ssl. dadetъ) galėjo atsirasti (anksčiau minėtos 3 pl.

¹⁸ Cf. “The dual davē, vēvē, ēvē (javē) has analogical loss of d after 1st person singular and plural.”—N. van Wijk, Geschichte der altkirchenslavischen Sprache, 1, 53, Berlin u. Leipzig, 1931.

¹⁹ On Lith. demi, and not *démi, cf. K. Būga.—“Tauta ir žodis,” 1, 402; K. Būga, Kalba ir senovė, 1, 158, 213; K. Būga.—Archivum Philologicum, 1, 62.

²⁰ Perhaps these two forms did originate (?) from a single form, i.e. a verbal-nominal or participial form that eventually entered the verb paradigm?

²¹ For a fuller discussion, see J. Endzelin IF 33, 124–125, (1914). J. Endzelynas, Baltų kalbų garsai ir formos, 161–162, Vilnius, 1957; J. Endzelīns, Latviešu valodas gramatika, 708–712; Chr. S. Stang, Das slavische und baltische Verbum, 233–234, Oslo, 1942; cf. F. Specht KZ, 56, 121, (1929).

²² See B. Mažiulis, Notes on the Question of the Earliest Relations between the Baltic and Slavic Languages, 9 (lines 18–24 from the top and footnote no. 4), Vilnius, 1958.

²³ “There was, however, undoubtedly an Indo-European ablaut grade in the participle of most athematic verbs” (A. Thumb, Handbuch des Sanskrit, 1, 223, Heidelberg, 1905). The Baltic (Slavic, etc.) participial formant -ant- (Slavic *-ont-) is generalized here from the stem of direct (oblique—*-nt-) case forms; cf., for example, the old noun of participial origin Lith. dantis, Prussian dantis, Sanskrit dantam (acc. sg.); datās (acc. pl.), Greek ὀδόντα, (Aeolic) ἔδοντες (pl.); Latin dentem, Old Irish dét, Gothic tunþus, etc.

9

through the correlation with the participle praes. act. and the example of the corresponding thematic form) a new form with the formant -ant-, i.e. *dōdant(i), *dedant(i), cf. OCS dadǫtъ, Russian dadym²⁴, and so on.

§ 10. It is reasonable to suppose that, among other factors²⁵, the correlation between the stem of the praesens verbi finiti and that of the praesens verbi infiniti²⁶ may also have influenced the thematization of the Lithuanian duomi paradigm (admittedly, at a fairly late date). Much the same can be said about the development of the Slavic *demi paradigm’s present stem, which became thematic quite early.

§ 11. The alternative mentioned earlier (§ 4—§ 8), namely, dating the present reduplication of the verbs examined here back to Proto-Indo-European, seems open to question, since this projection of reduplication into the Indo-European proto-language is ultimately based only on the relevant Indo-Iranian, Greek and, in part, Armenian language evidence. It is known, however, that Indo-Iranian, Greek and Armenian form a distinct dialect group [cf. the works of Bonfantes (Bonfante), Porcigo (Porzig), Kraės (Krahe), and others], which has a number of features distinguishing it from the other Indo-European languages (cf. the verbal tense system, the augment, and so on). The evidence of other Indo-European languages also seems to suggest that the Lithuanian duomi, demi present stems may never have had reduplication, including, therefore, in the Indo-European proto-language; cf., for example, Hittite tahhi ‘I take’ (probably from PIE *dō- ‘to give’), temi ‘I speak’ (cf. the semantic development—OCS děti ‘to put; to say’), tehhi ‘I put’ [both forms seem to derive from PIE *dhē- ‘to put’ (reduplication is not alien to Hittite, but it is not a formative element as, for example, in Old Indian or Greek)]. Accordingly, the reduplication in Sanskrit dadāmi, dadhāmi, Greek δίδωμι, τίθημι, and so on, could be regarded as an innovation of this dialect group. In this connection, the present forms of these two verbs in Baltic (and Slavic) would have to be reconstructed with an archaic pattern (structurally corresponding to the Proto-Indo-European one):

Sg. 1. *dōm(e)i pl. 1. *dōmē

2. *dōs(e)i 2. *dōtē

3. *dōti²⁷ 3. *dōnti, and so on.

Sg. 1. *dem(e)i pl. 1. *demē

2. *des(e)i 2. *detē

3. *deti 3. *denti, and so on.

Cf. Old High German tôm, tôs, tót, tômes, tót, tónt, and so on (note the relationship between the root vocalism—*de-mi > *de-mei > Lithuanian de-mi: dē-ti, *de-mi > OCS de-ždǫ: dē-ti; cf. Lithuanian ėmi, OCS jamb: Sanskrit admī, and so on). Thus, one may speak of a 3 pl. form *dōnti, *denti belonging to a certain period in the history of Baltic and Slavic, which was later, by analogy with thematic 3 pl. forms such as *vedant(i), *nārint(i) and so on, interpreted as *d-ōnt(i), *d-entį²⁸. Since this position was the weakest in the pa-

²⁴ Cf. P. S. Kuznetsov, Historical Grammar of the Russian Language, 280, Moscow, 1953.

²⁵ Cf. J. Endzelins, op. cit. 725 (from the 4th line from the top).

²⁶ The root duoda cited by Endzelīns, correlated with 3 pl. *dōdant(i) (see J. Endzelīns, Latviešu valodas gramatika, 724), in Baltic and especially Lithuanian, is perhaps not actually incompatible with this: in old Lithuanian writings, 3 sg. was generalized to the third person plural and dual as well, while the thematic present forms duomi (and demi) seem to have appeared later (when 3 pl. had long since disappeared).

²⁷ Cf. Vedic dāti..., see A. A. Barannikov, Elements of the Comparative-Historical Method in Indology, VYa, 2, 49, (1952).

10

in the paradigm [a monosyllabic root alongside the simple *dō-, *dē-,²⁸, and also for other reasons: for example, in the great majority of athematic verbs in Baltic and Slavic, the 3 pl. forms have more syllables than the other (athematic) forms, especially those of the plural, cf. Lith. ēmi, OCS jamb: OCS jadętъ, Balt. *smē: OCS sątъ, and so on], consequently -d- could have arisen in the 3 pl. form²⁸ (i.e. *dō-nti > *dō-d-nti > *dō-d-inti, *de-nti > *de-d-nti > *de-d-inti), perhaps in a manner similar to the following forms, which did not originally have the consonant -d-²⁹: Lith. aumi ‘I weave’: austi ‘he/she weaves’, audžas (cf. Skt. otum ‘to weave’, etc.); Lith. giemi ‘I sing’: gėsti ‘he/she sings’, giedàs (cf. O.Rus. azamu ‘to crow (of a rooster)’, Skt. gāyati ‘he/she sings’, gī-tas ‘sung’, etc.); Lith. čiaumi ‘I sneeze’: čiausti ‘he/she sneezes’, čiaudas (cf. Latv. škaunt, Skt. kṣauti ‘he/she sneezes’, etc.);³⁰ cf. Lith. eimi ‘I go’, eiti ‘he/she goes’: OCS idǫtъ, Russ. idym (OCS idǫ, etc. = Russ. idy, etc.), and so forth. If -d- in the forms *dōdant-, *dedant- were indeed of this origin, the formation of Lith. duosti, desti, OCS dastъ, and other forms, and the emergence of -s-, respectively, would be self-evident.

§ 12. Alongside the athematic Lith. duomi, demi, OCS damb, *demb (≥dežd), we find in Slavic and Baltic: OCS dajo (dajati), dějo (dějati), Latv. dēj ‘(a hen) lays’ [dēt ‘to lay (eggs)’]. These parallel forms were apparently associated with aspectual meanings³¹. Cf. also Lith. joja: Ved. yāmi, OCS jadǫtъ, jadǫ (cf. OCS idǫtъ: Lith. eimi); Lith. stoju, OCS stajǫ: Gk. ἵστημι < ἵσταμι; OCS vějo, Goth. waia ‘I blow’: Skt. vāti, Gk. ἄησι, and so forth.

II. On Baltic and Slavic Enclitic Personal Pronouns

§ 1. For all three persons of the personal pronoun, Baltic and Slavic have atonic, or enclitic, forms: Lith. mi, ti, si (-i, presumably from *-ei), OCS mi, ti, si (-i, most likely from *-oi). The other Indo-European languages are usually said to have corresponding forms only for the first and second persons, e.g. Skt. me, te, Gk. μοι, σοι (Homeric τοι). It is claimed that such patterns are shared only by Baltic and Slavic³². However, similar claims do not prove that the Baltic and Slavic feature in question is a phenomenon exclusive to these two idioms; it must be assumed that it is ultimately an old Indo-European relic, cf. Gk. μοι, τοι, οἱ, and so forth.³³ Corresponding enclitics from the same root for all three persons are also found, for example, in the cuneiform Hittite language: -mu ‘to/for me, me’, -ta ‘to/for you, you’, -ši ‘to/for him, her, oneself’³⁴ (these Hittite enclitics also have plural-

²⁸ My Moscow colleague V. Toporov communicated this idea to me (May 20, 1958).

²⁹ In similar cases, might the consonant -d- (?) be derived from an earlier *-n- (later changed by dissimilation to -d- in the present 3 pl. form and the corresponding participle praes. act.)? Cf., in one respect, the Baltic present suffix -n-a?

³⁰ Perhaps -d- in the causative verb vė-d-inti is of similar(?) origin (cf. ges-inti, and so on).

³¹ Plg., pvz., A. Meillet, Le slave commun, II éd., 208, 223. A. Meillet, Introduction... VIII éd., 219.

³² See, for example, Я. С. Отрембский, Славяно-балтийское языковое единство ВЯ, 6, 35 (1954); J. Otrębski, Gramatyka języka litewskiego, 3, 137, Warszawa, 1956; others, too—for example, J. Endzelynas, Baltų kalbų garsai ir formos, 148—do not present corresponding evidence for the third person from other Indo-European languages.

³³ See K. Brugmann, Grundriss... 2 Band, 2 Teil, 1 Lief., 407, Strassburg, 1909.

³⁴ See И. Фридрих, op. cit., 73.

11

these forms, cf. the corresponding facts in Slavic, Indo-Iranian, etc. ³⁵, which are probably derived from the older *me, *te, *soi³⁶; moreover, compare these enclitic pronouns with the (Hittite) possessive enclitic pronouns, e.g., dat. sg. -mi, -ti, -ši³⁷ (these possessives occur especially in texts from the Hittite “Old Kingdom”). If -mu, -ta (and their plurals) have dative and accusative meanings, then -ši (and its plural) has only a dative meaning³⁸. The meanings of the Hittite first- and second-person enclitics seem to have a certain connection with the tonic personal pronouns uk ‘I’, zik ‘you’ (singular and plural), whose accusative, dative, and even nominative can be expressed by one and the same form.

2. Dative and accusative³⁹ meanings are quite clearly observable in the Lithuanian enclitics mi, ti, e.g., nemipatika Br. Pred. 2. 17, pamijakik Br. I Mos. 37. 16, Tob. 5. 17, Isz to mi auga džiaug/mas K. 13, Pamijāke T. Ap. G. 11. 9, Neti dave KS 176, dekawojameti KM. 11; Anas mi weda K. 18, kodel mi apžiedei T. Mat. 27.46, kad ti gärbintumim K. 209, Kad ti pažintu žmones K. 218, etc.⁴⁰ Similar meanings were also held by the enclitic si (which became a reflexive particle quite early), traces of which can indeed still be observed in the present-day language, e.g., auñuosi ‘I put on my shoes’: kelìuosi ‘I get out of bed’, cf. Latvian aunies ‘put on your shoes!’: celies agri! ‘get up early!’ etc. Moreover, there seem to be facts⁴¹ indicating that mi, ti, si also had a genitive meaning⁴², which should be old; cf. Sanskrit me, te (dat.-gen.). Slavic mi, ti, si apparently mean only the dative⁴³. If the earlier meanings of these three Baltic enclitics were indeed dative-genitive, then the accusative meaning in the East Baltic languages would have to be explained not only by the assumptions indicated by J. Endzelīns⁴⁴, but also by the influence⁴⁵ of the relationship between the genitive and accusative of the tonic personal pronoun⁴⁶ (cf. Hittite, see above). On the other hand, the case meanings of these three Indo-European enclitics—old atonic particles—

³⁵ It is assumed that this was formerly also the case in the Baltic languages; see J. Endzelīns, Latviešu valodas gramatika, 915 (bibliography there).

³⁶ See H. Kronasser, Vergleichende Laut-und Formenlehre des Hethitischen, 142, 144, Heidelberg, 1956.

³⁷ Cf. “Possessifs -mi-: -ti: -ši- = μοι, τοι, οἱ etc.” — J. Kuryłowicz, “Le hittite”. — Reports for the Eighth International Congress of Linguists, Vol. II, 295, Oslo, 5—9 August 1957.

³⁸ See И. Фридрих, op. cit., 73; H. Kronasser, op. cit., 142—144.

³⁹ The dative uses of mi, ti in old Lithuanian texts occur much more frequently than their accusative uses — E. Hermann, Litauische Studien, 69, Berlin, 1926.

⁴⁰ The examples are taken from A. Bezzenberger, Beiträge zur Geschichte der litauischen Sprache, 163—164 (author and abbreviations provided), Göttingen, 1877.

⁴¹ See E. Hermann, op. cit., 74 et seq.

⁴² Its loss must have begun quite early.

⁴³ For remnants of a similar (i.e., exclusively dative) meaning in the Baltic languages, cf. E. Hermann, op. cit., 73.

⁴⁴ J. Endzelīns, l. c.

⁴⁵ See J. Endzelīns, op. cit. 505 et seq.; J. Endzelynas, Baltų kalbų garsai ir formos, 147—148 (§ 300, § 302).

⁴⁶ In a similar way, the old Baltic tonic forms *mēn, *tēn, *sēn may also have been displaced (having acquired enclitic properties from a certain time onward; cf. for Slavic A. Байан, op. cit., 411; Н. ван-Бейк, История старославянского языка, 280, Москва, 1957), as well as the atonic (enclitic only) forms *mē, *tē, *sē (which may perhaps be indicated by Old Prussian si, if it indeed derives from *sē; cf. Sanskrit mā, tvā, Greek με, σε < *tFe, etc.; see J. Endzelynas, l. c.).

12

The meanings, it seems, may not have been strictly differentiated morphologically since ancient times⁴⁷.

§ 3. OCS sę, Prussian sien(sin) < *sēn, si, < *sē (presumably; see footnote no. 46, p. 12), could, it seems, correspond to the Lithuanian reflexive particle -se. We find it (alongside -si) in our old writings, e.g., Mažvydas⁴⁸: raschose 13₁₄, kiesiese 21₁₆, makase 38₁₀, ₁₁, skaitose 13₁₄, pradestyse 17₅, 40₁, prasidėase 21₉, biakiese 32₆–₇, etc. (alongside: pradestysy 14₁, dabaisey 34₁₅, giestyśy 54₂, etc.); Bretkūnas, Biblia: dabaiase Luke 6, 7, piktinasse Matthew 13, 21, biotumse Jeremiah 32, 39, budawiose Acts 9, 31; Bretkūnas, Postilla: mokintuse I 33₁₂, randasse I 86₂; Vilentas, Enchiridion: testojese 28; “Summa (‘…Ewangeliv… Kiedaynise… 1653)”: isedoutumey 21₂₁₉. The particle -se in old writings seems also to be confirmed by present-day dialects, namely, those of the southern Samogitians [where the -e in the particle -se cannot be a Samogitian pronunciation of the vowel -i] and the Western Aukštaitians adjoining them: šakose alongside šakos, pasidarė (Pagramantis)⁴⁹, pėštise, vaipytise, šukiotise, pėšase, vaipose alongside pėštis, vaipytis, pasišukuok (Girdžiai, Skirsnemunė, Šimkaičiai), sukūose, sukėse (Agilas, Tovė, Lūja)⁵⁰, etc. Lithuanian -se (which does not differ semantically from -si) seems to have developed from East Baltic *sēn or *sē (acc. sg.; for the ending, cf. savę, tavę, manę). On the basis of this form, one may suppose that East Baltic also had accusative forms of other pronominal enclitics — *mēn or *mē, *tēn or *tē⁵¹, cf. Prussian mien, tien, sien, si, OCS mę, tę, sę, Sanskrit mām, tvām, mā, tvā.

§ 4. It should also be said that Baltic and Slavic use only the third-person enclitic of personal pronouns—the reflexive—for all three persons of reflexive verbs; however, a very similar fact is also found in Hittite reflexive verb forms, where the reflexive particle for all three persons of the verb is -za, possibly derived from *sē⁵², cf. Prussian si < *sē (see above), Latin sē.

III. On Baltic and Slavic Numerals⁵³

§ 1. In terms of its root, the Baltic cardinal numeral “1” has parallels not only in the Slavic languages. The Indo-European numerals “2”–“4” (compare the relationship between Baltic, or Lithuanian, abù and Slavic, or OCS, oba with that between Greek ἀμφω and Latin ambo) developed in the Baltic and Slavic languages in accordance with their declensional systems. The morphological fate of the indeclinable Indo-European numerals “5”–“9” differs in Baltic and Slavic. The Baltic numeral “100” is a direct Proto-Indo-European relic,

⁴⁷ Cf. E. Hermann, op. cit., 78 et seq.

⁴⁸ Quoted from the 1922 edition by J. Gerulis.

⁴⁹ See P. Jonikas, Pagramančio tarmė, 66, Kaunas, 1939.

⁵⁰ See J. Gerullis and Chr. Stangas, Lietuvių žvejų tarmė Prūsuose, 47 et seq., Kaunas, 1933.

⁵¹ In the dialects of Eastern Lithuania (including all the Dzūkai), the two old forms *sei and *sen (> *sin > -si) may have phonetically merged into a single si; similarly, in Latvian, *sei and *sēn (> *sie > si) may have become a single -si-.

⁵² See H. Kronasser, op. cit., 145.

⁵³ For more on numerals, see B. Мажюлис, Литовские числительные и соотношение балтийских числительных с числительными других индоевропейских языков (диссертация и автореферат), Москва, 1955; В. Мажюлис, Индоевропейская десятичная система числительных ВЯ, 4, 53—59 (1956); Some Questions of Lithuanian Grammar, 40—64, Vilnius, 1957.

13

whereas the Baltic “1000” has parallels in Slavic and Germanic languages. In the cardinal numerals “11”—“19” of Baltic, Slavic, and Germanic, we find certain similarities which are, most likely, parallel developments. The Baltic, Slavic, and also Germanic cardinal numerals “20”—“90,” as word combinations (not compound forms!), are Proto-Indo-European relics, not innovations⁵⁴. The multiplicative numerals of Baltic and Slavic are typologically quite close, but compare the same numerals in other Indo-European languages.

§ 2. If the East Baltic languages and all Slavic languages have initial d- instead of the older n- in the numeral “nine, ninth” [cf. Lithuanian devyni, devintas (but Old Prussian newīnts), Old Church Slavonic devętь, devętь], this is, perhaps, a parallel feature of the Baltic (eastern) and Slavic languages; compare Old Church Slavonic nebo: Lithuanian debesìs, Latvian debess, and probably also Old Church Slavonic domъ: Lithuanian nãmas (where old Baltic *d- may have changed to East Baltic n-⁵⁵); in addition, compare Hieroglyphic Hittite tapas ‘sky’: Cuneiform Hittite nepiš ‘sky’, Hieroglyphic Hittite adamai|azamai ‘name’: Greek ὄνομα⁵⁶.

§ 3. The formation of Baltic collective numerals probably begins with the cardinal numeral “2”: Lithuanian dveji, dvėjos, dvėja, Latvian divēji, divejas (: Baltic *d(u)vō, *dvei and stem *dvei-), Old Prussian *dwaien (:dwai) ≈ abbaien (:dwai); compare Old Church Slavonic dъvoji, dъvoje (=dъva, dvė and stem *dvoi-), Sanskrit dvayas ‘double’, Greek δοιοί ‘etc.’, and so on. By analogy with dveji we have treji [and? (to some extent) ketveri (: *ketveres⁵⁷)], cf. dъvoji, troji [and četvorъ, četverъ (: *ketveres⁵⁷)]. The form dveji in Baltic (and also Slavic) languages was probably conditioned also by the fact that when the numeral *dvō, *dvei was used with pluralia tantum nouns, a mismatch could arise (*d(v)ō, *dvei—dual forms—with plural forms of pluralia tantum nouns), and this was eliminated by the form dveji. In general, the derivational and morphological system of Baltic numerals is similar to the Slavic one in many respects. However, it should be noted that the oldest Baltic numerals, that is, the common Indo-European ones, are direct relics of the Proto-Indo-European language, whereas the later, that is, non-common-Indo-European numerals, are correspondingly similar to numerals in Slavic, as well as Germanic and certain other Indo-European languages.

⁵⁴ This is what A. Meillet, for example, thinks, Introduction... VII éd., 413.

⁵⁵ In my pamphlet Заметки к вопросу о древнейших отношениях..., 18, an editorial error has crept in: this parenthetical addition is missing.

⁵⁶ On Hittite n-: d- in these words (the appearance of d- is unclear), see J. Gelb, The Contribution of the new Cilician Bilinguals to the Decipherment of Hieroglyphic Hittite. — „Bibliotheca orientalis“, vol. VII, Nr. 5, 140 (1950); S. Alp, Bemerkungen zu den Hieroglyphen des hethitischen Monuments von Imamkulu. — „Archiv Orientální“, vol. 18, Nr. 1—2, 8 (1950); H. T. Bossert, Die phönizisch-hethitischen Bilinguen vom Karatepe. — „Oriens“, vol. II, Nr. 1, 110 (1949); H. T. Bossert, Die phönizisch-hethitischen Bilinguen vom Karatepe. — „Jahrbuch für kleinasiatische Forschung, Band II, H. 2, 187; Goetze, The linguistic Continuity of Anatolia as shown by its proper names. — „Journal of Cuneiform Studies“, vol. VIII, Nr. 2, 77 (and note 59) (1954); J. Friedrich, Hethitisches Wörterbuch, Lief. 4, 334, Heidelberg, 1954; H. Kronasser, op. cit., 228—229.

⁵⁷ Baltic *ketveres, Slavic *ketveres, perhaps from older Baltic *ketvares and respectively Slavic *ketvores (nom. sing.) (Baltic *-a- > -e-, Slavic *-o- > -e-, perhaps due to “assimilation”; compare Greek (Ionic) τέσσερες, where “-e- prob. par assim. vocalique” — E. Boisacq, Dictionnaire étymologique de langue grecque, 960, Heidelberg—Paris, 1916), cf. Sanskrit catvāras, Greek (Doric) τέτορες, and so on (the -o- in the Slavic form četvoro is hardly old).

14

IV. Other brief remarks

§ 1. In the genitive of ō-stem nouns in East Baltic and all Slavic languages, we find the ending -ā; compare Lithuanian vilko (dialectal vilkā), Latvian vilka, Old Church Slavonic vlьka, whereas Old Prussian has -as [compare the endings: Old Prussian nom. sing. deiw's, gen. sing. deiwaś — Hittite nom. sing. attaš ‘father’, gen. sing. attaś; Old Prussian nom. sing. gijwan ‘life’ (neut.), gen. sing. gīwas — Hittite nom. sing. pedan ‘place’ (neut.), gen. sing. pēdaś]. This genitive type (i.e. with the ending -ā) can hardly be particularly ancient in either Baltic or Slavic: this is indicated not only by Old Prussian, but also, in general, by the forms of the aforesaid stem in all Indo-European languages, which developed differently in the various Indo-European branches. It is assumed that the East Baltic and Slavic genitive ending -ā is genetically related to the ablative (which at a certain time may have become a paradigmatic form in individual Indo-European branches⁵⁸). J. Otrębski’s arguments⁵⁹ do not seem to overturn this traditional thesis.

* * *

In Lieu of Conclusions

Without attempting to resolve the problem of the origin of the ancient relationship between the Baltic and Slavic languages, I am nevertheless inclined to think that the ancient (as well as innovative) similarities between Baltic and Slavic languages mentioned in these notes—the “parallels”—do not call for the assistance of the traditional hypothesis of a Proto-Balto-Slavic language (or, rather, of the genetic unity of the Baltic and Slavic languages): for now, the principles of linguistic differentiation and integration indicated by J. Endzelīns suffice here.

NOTES ON THE QUESTION OF THE EARLIEST RELATIONS BETWEEN THE BALTIC AND SLAVIC LANGUAGES

В. П. МАЖЮЛИС

Abstract

At the beginning of the article, some historical information is presented concerning attempts to solve this Balto-Slavic problem, with particular emphasis on the theory of the distinguished scholar J. Endzelin. The author of the article by no means claims to offer a specialized solution to the question of the earliest relations between the Baltic and Slavic languages, but merely sets out a number of specific observations which are nevertheless relevant to the problem. Examining two alternatives concerning the present stem of the Balto-Slavic verbs “to give” and “to put,” the author questions their traditionally reconstructed archetype *dōdmi or *dedmi, respectively. In the following three sub-

⁵⁸ Cf. Fr. Specht, Der Ursprung der indogermanischen Deklination, 371, Göttingen, 1947; Fr. Specht, Zum Ersatz des Ablativs, 69, 127 (1948).

⁵⁹ See Я. Отрембский, op. cit., 31.

15

The sections of the article discuss enclitic personal pronouns, numerals, and the genitive (such as Lithuanian vilko, Russian волка) in the Baltic and Slavic languages.

A more detailed summary of this article would be unnecessary, since it contains the same facts and ideas that were presented in the same author's pamphlet “Notes on the Question of the Earliest Relations between the Baltic and Slavic Languages (IV International Congress of Slavists),” Vilnius, 1958.

21