the end would be Latvian biete ‘ein Kalk-, Lehmschlägel’ ME I 306. The relationship between the vowels in the root ju would be like that found, for example, in Old Prussian playnis: Lithuanian plienas, Latvian pliens, while the linking semantic step would perhaps be Latvian bīte ‘der Proviantsack, die Speisepaudel’ ib. 304 (cf. further the verbs bit ‘schieben’, bī-d-īd ‘schieben, rücken, schiebend stossen’).

S. KARALIŪNAS

ON THE INTERPRETATION OF A PAIR OF PHONETIC PHENOMENA

1. The Lithuanian literary language diphthongs ie, uo and their equivalents in the Dzūkian Žemaitian dialect

That the Dzūkian Žemaitian i˙ ~ y and i˙ ~ ie had already merged in pronunciation by the 16th century is clearly shown by the evidence in J. Sprogys’s dictionary¹. Here, place names from the Dzūkian area, written in Russian, are often given both with e, ie (=Žem. i˙, lith. ie) alongside forms with u, and with e, ie (=lith. ie), for example: Сетова (p. 289), Ситова (p. 290) ‘a river connecting Paršėžeris and Lūkstas’; Скабя пиева (p. 290) (Pavandenė district), Кявлес пивас мишкас (p. 159) (Karšuva district), Жардинветис (p. 112) ‘a field in Kražiai district’, Жардинвиты (p. 112) ‘a field in Karšuva district’, and many others. Forms with e, ie (=lith. ie) are most likely the result of adapting place names to the Aukštaitian dialect. That this is probably indeed the case is shown by the hypercorrect spellings of place-name forms found in this dictionary, that is, writing e (=ie) instead of i˙ ~ y, for example: Ежня (p. 108) = Ÿžnė ‘a river in Karšuva district’, alongside Пожижни (p. 246) = Payžnys ‘an estate in Karšuva district’; Пекойти (p. 225) = Pykaičiai ‘an estate in Pajūris district’, alongside Пикоити (the same), and so forth. Thus, already in the 16th century, speakers of the Dzūkian Žemaitian dialects could not distinguish the literary-language y and ie (as well as ū and uo), which had merged in those dialects.

The first references to the reflexes of the diphthongs ie, uo as į, ū in Žemaitian dialects are already found in A. Schleicher’s Lithuanian grammar². J. Juška³ later made this claim more specific and detailed. It is also clearly attested in the works of all the linguists who studied the Dzūkian Žemaitian dialects—K. Jaunius, A. Salys, J. Gerulis, P. Jonikas, V. Grinaveckis, Z. Zinkevičius, A. Balasaitis, V. Vitkauskas, and others—as well as in the texts of these dialects recorded with relative expertise in the publication “Lithuanian Dialects”⁴. However, linguist A. Girdenis does not accept this merger. In his article “What the Dzūkian Žemaitians have instead of the literary-language ie, uo”⁵, he seeks to test whether “the equivalents of standard Lithuanian ie, uo in the Dzūkian Žemaitian dialect are long monophthongs ī, ū, indistinguishable from long į, ū = y, ū” (p. 143). To this end, as the article’s author states, “word pairs were selected which in the standard language and the Aukštaitian dialects differ only by the oppositions ie: ī, uo: ū: grūodas: grūdas, kuodas: kūdas ‘thin’, lieti:

¹ See Географический словарь древней жомойтской земли XVI столетия, Составленный Иваном Яковлевичем Спрогисом, Вильна, 1888, 363 стр.

² Žr. Handbuch der litauischen Sprache von August Schleicher, I, Grammatik, Prag, 1856, p. 5.

³ See Jon’s Juszka, Kalbos lėtuviszko lėžuvo’ ir lėtuviszkas statrazimas arba ortograpija, Peterburge, 1861, p. 12, 14.

⁴ See LKT, p. 93–129.

⁵ See A. Girdenis, Ką turi džūnininkai vietoj literatūrinės kalbos ie, uo, “Baltistica” VI(2), 1970, p. 143–146.

14. Užs. Nr. 1026

209

lyti, rietas: rytas, žiedukas: žvydūkas, and an attempt was made to determine whether they are distinguished in dialect speech by people originating from the following localities in the dünininkai area: Karklėnai, Laukuva […], Užventis, Varniai […], Batakiai […], Eržvilkas, Lauksargiai […], Skaudvilė, Šaukenai, Tauragė […], Vainutas […], Viduklė […]. Informants were asked to translate these pairs into their own dialect and say whether they differ in the dialect and (if it turned out that they do) how they differ” (pp. 143—144). “The results of the study,” the author states, “proved remarkably clear and unambiguous: in all the aforementioned localities in the dünininkai area (including such peripheral dialects as Eržvilkas and Viduklė), these pairs are distinguished without any significant variation—only people who had become detached from their dialect long ago and no longer actively use it are somewhat uncertain. Instead of the standard-language ie, uo, sounds with somewhat less tension and distinctness are pronounced everywhere than instead of y, ū […].” (p. 144). Accordingly, “the dünininkai dialect reflexes of ie, uo would apparently be most appropriately interpreted and written as diphthongs iį, uu” (p. 144). However, from a scientific point of view, it is difficult to regard this experiment and its method as successful. The first pair of Lithuanian words selected for the experiment, grūodas and grūdas, cannot even be translated into the dünininkai dialect (and thus cannot be contrasted or tested), since the dünininkai do not have the word grūodas⁶. Thus, an experiment based on an artificially translated, nonexistent Samogitian form is erroneous in this respect. As the quotation shows, the experiment’s informants were not actual local speakers of the dünininkai dialects (a scientific experiment—and especially one intended to revise an earlier view—would seemingly have to rely only on the speech of the latter), but merely people originating from the dünininkai area⁷. But even the latter, being detached from their native dialect and influenced by Lithuanian, naturally cannot fully represent the phonetics of the dialects of that area in their Samogitian pronunciation. And simply translating the corresponding pairs of Lithuanian words into a dialect and attempting to make them sound Samogitian—as A. Girdenis points out—would hardly enable anyone to pronounce the sounds in these pairs exactly as they are pronounced in the dialect. After all, in this case the experiment’s informants played only the role of imitators. And no equals sign can be placed between imitation and reality. Thus, the objective truth does not in fact become any clearer from these informants’ pronunciation, and it cannot tell us what the dünininkai actually have in place of Lithuanian ie, uo. The experiment’s transcription method also complicates matters. As A. Girdenis points out, “After a little practice (4—5 trials each), they (= the auditors, i.e. the people listening to the pronunciation of the sounds being tested; our note—E. G.) could easily determine which sound—i or į, ū or ų—was present even in words presented in isolation, without context” (p. 145). The object under experiment is usually transcribed in this way, but if the auditors recognized what they heard not immediately, but only after 4—5 practice trials, this is the clearest evidence that the results of this study, too, cannot be as “remarkably clear and unambiguous” as A. Girdenis considers them. As we can see, the experiment’s conclusions are considerably strained. It is therefore truly difficult to accept them and rely on them. Thus, A. Girdenis’s experiment is not yet capable of revising the old law concerning the dünininkai dialect reflexes of Lithuanian ie, uo.

⁶ Instead, plikšalai are actively used in these dialects. Moreover, the word grūuc ‘grūodas’ is not mentioned anywhere even by V. Grinaveckis, who raised this issue and whose article A. Girdenis directs readers to consult for examples of this type (p. 143).

⁷ It is worth noting that A. Girdenis was not always exacting in selecting experimental informants: an informant from the Lauksargiai dialect, probably through oversight, represents the dünininkai in the experiment, although the Lauksargiai dialect actually belongs to the northwestern Aukštaitian dialects; see LKT, p. 173 and the map.

210

2. Reflexes of Indo-European *ā in the Baltic languages

The Lithuanian linguist V. Mažiulis, having gained an in-depth knowledge of Prussian sources and the Prussian language, has advanced a new conception in his significant linguistic studies of recent years, according to which he argues that already in (Common) Baltic antiquity Indo-European *ā had developed into *ō⁸. On this basis, he offers novel explanations for the origin and development of many morphological phenomena in Lithuanian and, more broadly, the Baltic languages, as well as their relationship to corresponding morphological features in other Indo-European languages. Since this conception remains the subject of lively discussion⁹, without claiming to resolve the issue definitively, we will examine the development of Indo-European *ā in the Baltic languages from the perspective of a dialectologist.

V. Mažiulis adduces the vocalism of sixteenth-century Prussian words (the Sambian dialect)¹⁰, such as m-ū-ti III 67₃ ‘mother’, br-ā-ti ‘brother’, and the thirteenth- to fifteenth-century place names of this region By-o-ten, Bev-o-ten (cf. Lithuanian Bij-o-tai), Grind-o-s (cf. Lithuanian grìnd-o-s), Layp-o (cf. Lithuanian liep-a)¹¹, on the one hand, and the vocalism of fourteenth- to fifteenth-century forms (apparently representing the Prussian Pamedian dialect¹²), such as m-o-the E 170 ‘mother’, br-o-te E 170 ‘brother’, mens-o E 154, 374 ‘meat’, on the other. Taking Prussian data as reference points for the history of vocalism in all Baltic languages¹³, he argues for and substantiates two hypotheses: a) “the Prussian *ū in the word m-ū-ti III 67₃ (= Lithuanian m-ó-tė, Latvian m-ā-te) cannot have developed from *ā[…]. This Prussian *ū probably derives from a long, rounded, low back vowel, denoted Prussian *ō”; b) “the Sambian vowel *ā (br-ā-ti III) derives from an earlier Sambian *ō”¹⁴. He also asserts that “Such a Prussian *ō […] is also evidenced by the oldest written monument of the Prussian language—the Elbing Vocabulary”¹⁵. On this basis, he concludes that “the development of the Indo-European *ā vowel ‘in the direction of the vowel *ō’ (Prussian E oa or o=Lithuanian o, i.e. ō, Latvian dialectal ō, etc.) should be regarded as a feature of (Common) Baltic antiquity”¹⁶, and consistently adheres to this view.

In this connection, it should be noted that, when discussing the Prussian language, K. Būga clearly identifies its two most distinct dialects: northern and southern, which differ considerably from one another¹⁷. The southern (Pamedian) dialect is represented by the first source of the Prussian language, the Elbing (E) Vocabulary, written around 1400; the northern (Sambian) dialect, by the Prussian catechisms written about 150 years after the appearance of this vocabulary¹⁸. Thus, when discussing Indo-European *ā

⁸ See B. Мажюлис, Заметки по прусскому вокализму. — Вопросы теории и истории языка, Ленинград, 1963, стр. 192—194; V. Mažiulis, Baltų ir kitų indoeuropiečių santykiai, Vilnius, 1970 (hereafter — V. Mažiulis, Baltų sant.), p. 11—25 ff. This question is also discussed in Й. Казлаускас, К развитию общебалтийской системы гласных, „Вопросы языкознания“, 1962, № 4, стр. 20—24.

⁹ See З. Зинкявичюс, О развитии балтийского вокализма, — Балто-славянский сборник, Москва, 1972 (hereafter — Z. Zinkevičius, BV), стр. 5—14.

¹⁰ See V. Mažiulis, Baltų sant., pp. 11—12.

¹¹ Cf. also K. Būga, Rinktiniai raštai, III, Vilnius, 1961 (hereafter — K. Būga, RR, III), p. 106, where the Sambian place names Dywone—lauken, Kopte (Kupte), Aucopte, Pobeti, alongside Pu-bethyn, and the words pōstam, pōmien, perōniskan, encops, encopts, etc., are mentioned.

¹² See V. Mažiulis, Baltų sant., p. 13.

¹³ See V. Mažiulis, ibid., p. 21.

¹⁴ See V. Mažiulis, ibid., p. 12.

¹⁵ See V. Mažiulis, ibid., p. 13.

¹⁶ See V. Mažiulis, ibid., p. 24.

¹⁷ See K. Būga, RR, III, p. 104—110.

¹⁸ Žr. V. Mažiulis, Prūsų kalbos paminklai, Vilnius, 1966, p. 27—29.

14*

211

When considering the correspondences in the vocalism of the Old Prussian written sources and seeking to draw conclusions from them about the development of Indo-European *ā in all the Baltic languages, one must take into account the vocalic features and differences of not one but two independent phonological systems—the southern and northern Prussian dialects. Among other things, the difference between these two Prussian dialects is also reflected in the vocalism of Prussian words and Prussian place names cited by V. Mažiulis (different reflexes of Indo-European *ā). However, both of V. Mažiulis’s hypotheses derive solely from the ū and ā reflexes of this origin in the Sambia dialect (the northern Prussian dialect). But how, then, are we to explain the simultaneous presence, in the same Sambia-dialect texts, before the same labials (or velars), alongside ū (m-ū-ti, poglab-ū), of ō and o as well (pōstam, Layp-o)? Thus the vocalism with ō (o) in pōstam, Layp-o, and analogous examples, deriving from Indo-European *ā, would indicate that this ō reflex objectively existed in the Sambia dialect alongside its ū(u). Its actual coexistence with ū does not yet show which of the two is older, especially since K. Būga gives the Sambian place name Kopte with its variant Kupte¹⁹. The presence of Sambian ō(o), like ū(u), alongside ā (brāti) also provides no basis for deriving their ā from Sambian *ō (i.e. from the Sambian archetypal *ō). Since in the southern Prussian dialect V. Mažiulis mentions only the o and oa reflexes of Indo-European *ā (in the case of oa: s-oa-lis E 293 ‘grass’, V. Mažiulis’s view is that this too undoubtedly represents Old Prussian *ō̄²⁰), its vowel a (< Indo-European *ā) remains unexplained in words such as pon-a-sse E 90 ‘philtrum’ (cf. n-o-zy E 85 ‘nose’), pom-a-tre E 180 ‘stepmother’ (cf. m-o-the), which exists alongside reflexes of the same origin, o and oa (m-o-the, lubbo E 206 ‘Brett’²¹, m-oa-zo E 178 ‘Mumė’). Thus, if Indo-European *ā is represented by three reflexes in the southern dialect, all of them must be taken into account here, and their mutual relationship must be established. All this would indicate that the actual (not archetypal!) reflex of Indo-European *ā in the written sources of both Prussian dialects was ō. It existed alongside other reflexes, which most often occurred in other word positions. This, however, does not warrant the claim that all the other Indo-European *ā reflexes in both dialects could have developed from it, or that it was their precursor—the very first Indo-European *ā reflex in Proto-Baltic antiquity—or that it had any precedence over them. Thus, the attested situation of the different reflexes of Indo-European *ā does not support the view that its development “in the direction of the vowel *ō” began with *ō, which should be regarded as a feature of Proto-Baltic antiquity. The different reflexes of Indo-European *ā in these Prussian written sources show that its development was not the same in the two Prussian dialects. First and foremost, they indicate that the a reflex in the southern dialect could (and probably does) represent the original Indo-European *ā equivalent there, supporting the traditional view that Indo-European *ā in Proto-Baltic antiquity was also *ā. The fact that it is short in this case could not alter the situation. However, the very fact that the Indo-European *ā reflex a was actually present even in the oldest Prussian source—the Elbing Vocabulary—again seems not to support the conclusion that the development of Indo-European *ā in Proto-Baltic antiquity began with *ō. The other Indo-European *ā reflexes in the written sources of this dialect (oa, o) would indicate that Baltic *ā (not *ō) did nevertheless become narrower and somewhat closer. As a result, it became labialized and only then turned into (s-oa-lis), or developed into o (m-o-the, gald-o- E 365 ‘trough’). In the northern Prussian dialect, the development of this *ā must have differed clearly; it followed a different course. The Prussian written sources of this dialect likewise attest three kinds of Indo-European *ā reflexes. After labials and velars in the word stem, the usual reflex is ū, although sometimes ō (o) also occurs in place of it in the same contexts. In all other cases Indo-European *ā has remained unchanged here

¹⁹ Cf. K. Būga, RR, III, p. 106.

²⁰ See V. Mažiulis, Baltų sant., pp. 13–14.

²¹ Žr. J. Endzelīns, Senprūšu valoda, Rīgā, 1943, p. 205.

212

unchanged, while alongside u it is also common in place-name stems and after m, v, j, k, s, p²². At the end of a word, after labials and velars, ū is regularly found (only here does u also occur alongside it: laiku III 37₁₈ ‘šie halten’). In all remaining cases, word-final Indo-European *ā corresponds to the reflex o. Thus, the original stage of Indo-European *ā here, too, must have been *ā. And *ā changed in the direction of o only after labials and velars, and at the end of a word (in place names also after j, v, k, s). In other cases it remained unchanged (br-ā-ti). But this change in the specified contexts must again have differed, apparently, from that in the southern dialect. As V. Mažiulis has correctly noted²³, Baltic *ā could not have arisen phonetically directly from *ā in the form ā(u) found here instead. Thus this *ā’s change into ū must, in any case, have passed (and did pass) through the ō stage²⁴. Words with this reflex are precisely what attest to that stage. Here, however, ō must have become markedly narrower and closer. It could therefore have come to resemble ū, turned into it, or even changed into it altogether (e.g., pōstam, -m-ū-ti, deiv-ū-tai III 115₁₄ ‘sėlig’, poglab-ū, laik-u). A similar development of unstressed Baltic *ā has also occurred in part of the Lithuanian Veliuona dialect and in nearby areas²⁵.

The Indo-European *ā reflexes discussed here in the Old Prussian written sources allow the following conclusions:

1. Indo-European *ā was probably known to all Baltic languages and their dialects as a broad, open *ā.

2. In the southern Prussian dialect, that is, the dialect in which the Elbing Vocabulary was written, this Baltic *ā (whose reflex is a) must have become narrower and labialized, approached it (oa), or changed completely into o.

In the northern Prussian dialect—the dialect of the catechisms—Baltic *ā (whose reflex is ā) must have become narrower and labialized only after labials and velars, and at the end of a word (in place names also before v, j, k, s). In other cases it remained unchanged here. Its development here must have differed somewhat from that in the southern dialect. As a result of greater narrowing, it must have become increasingly tense and close, turning into ō, whose quality would have been closer to ū, or changing into it completely. The wave of Baltic *ā narrowing and labialization, spreading across the historical Baltic territory from southwest to northeast, must first have reached the Lithuanian-speaking area, giving a reflex ō to a large part of its dialects and turning it into uo in Samogitian (except in the Karklė dialect). It must have reached the East Aukštaitian and eastern Dzūkian dialects last. That is why ā has remained intact here, both everywhere in some places and only in part in others. The wave of narrowing and labialization of Baltic *ā did not reach the Latvian language (except Latgalian), and there it remained unchanged.

3. The reflex o of Indo-European *ā in the Baltic languages and their dialects is probably secondary.

E. GRINAVECKIENĖ

M. Kondratiuk, Place Names of the Southeastern Part of Białystok Region, Wrocław – Warszawa – Kraków – Gdańsk, 1974, 304 p. (with map).

Published in issue 29 of the Polish People’s Republic Academy of Sciences serial publication “Monografie Slawistyczne,” the work “Place Names of the Southeastern Part of the Białystok Voivodeship” by Dr. Michał Kondratiuk, a research fellow at the Institute of Slavonic Studies, is interesting to Balticists in many respects.

²² See K. Būga, RR, III, p. 106.

²³ See V. Mažiulis, Baltų sant., p. 12.

²⁴ Cf. Z. Zinkevičius, BV, p. 12.

²⁵ Plg. A. Salys, Kelios pastabos tarmių istorijai, APh IV, Kaunas, 1933, p. 28; Z. Zinkevičius, Lietuvių dialektologija, Vilnius, 1965, p. 68 ir žemėl. Nr. 32, 33.

213