IŠ LIETUVIŲ KALBOTYROS ISTORIJOS LIETUVIŲ KALBOTYROS KLAUSIMAI XX (1980)

E. GRINAVECKIENĖ

PHONETICS OF THE LITHUANIAN DIALECTS AROUND DAUGPILS

I. INTRODUCTION

a) During the 1977 dialectological expedition conducted by the J. Kolas Institute of Linguistics of the Academy of Sciences of the Byelorussian SSR together with the Institute of Language and Literature of the Academy of Sciences of the Latvian SSR and the Institute of the Lithuanian Language and Literature of the Academy of Sciences of the Lithuanian SSR, the Belarusian dialects of the Daugpils (Latvian Daugavpils), Kraslava (Latvian Krāslava), Ludza (Latvian Ludza), and Rezekne (Latvian Rēzekne) districts of the Latvian SSR were investigated. Our task was to collect Lithuanianisms used in the Belarusian dialect of the southern part of Daugpils District—the left bank of the Daugava (Latvian Daugava). Most of our work was carried out in the vicinity of Skrudaliena (Latvian Skrudaliena¹) (18 km southeast of Daugpils), especially in Raguliškiai (Latvian Raguliški) and neighboring villages. Among the Belarusian speakers of those places, we recorded quite a few words of Lithuanian origin, place names, surnames, and fragments of Lithuanian expressions². In the Daugpils suburb of Gryva (Latvian Grīva, now the part of Daugpils on the left bank of the Daugava), we learned that until quite recently several local people who spoke Lithuanian had lived there³. In nearby Kalkūnai (Latvian Kalkūne; older residents more often pronounce it Kálkune.⁴) The local Lithuanian pronunciation of Daugpils is Daũgpilis, Daugpilỹs, Daũgavpilis, Daũguvpilis.

¹ Tai, be abejo, lietuviškos kilmės vietovardis. Būga K. (Rinktiniai raštai. V., 1961, t. 3, p. 565) visai pamatuotai jį laiko kilusiu iš liet. Skrudėlynė (plg. liet. dial. skrudėlynė ‘skruzdėlynas’).

² Among the place names of Lithuanian origin, the names of villages with the suffix -iški should be mentioned first, e.g., Бойцушки, Сӯрышки, По̄ришки, Ёнишки, Басерушки, Кирлишки, Руманишки, Лынто̄пишки, Микалишки, Дзирванишки, etc. In the opinion of K. Būga (Būga K. Rinktiniai raštai, vol. 3, pp. 565–566) and J. Safarewicz (Safarewicz J. Litewskie nazwy miejscowe na -iszki. – Onomastica, 1956, no. 2, p. 15), wherever place names with the suffix -iški are widespread, there were historically Lithuanian settlements. Other elements of Lithuanian-origin toponymy and anthroponymy are also abundant here, e.g., names of bogs and forests—Шлапаки, Грьнишки, Кимсы, etc.; and some surnames of the old local inhabitants—Сервуц, Мажейка, Каралён, Лынку́н, etc. In the village of Raguliškiai we also recorded fragments of four Lithuanian expressions, e.g., šakė, pasakė. || vėrna šáka, tilvikis paplikis nimáka || pelupe. dėseru. ‘nulupo dešrą’ || visi šáka, visi krūta, be.larūst bū.lbes skūta || The people who uttered these expressions could no longer explain their meaning very well. They only said that their grandparents and parents used to speak that way, especially when they wanted children not to understand something secret. Translations of some Lithuanian idioms are also commonly used here, e.g., баю́та іх (cf. Lithuanian balà jų); каб ты скісла (cf. Lithuanian kad tu surūgtum), etc. It was also observed that when addressed in Latvian, residents of these places (interviewed by the Latvian linguist, Candidate of Philological Sciences B. Reidzāne) used quite a few words, forms, and calques characteristic of Lithuanian in their Latvian, e.g., es no pieciem būnu (cf. Latvian 1st person singular esmu, Lithuanian būnu) mājās; bērni atiet (cf. Latvian 1st person singular atnāk, Lithuanian ateina) un palidz; vilciens atier (cf. Latvian 1st person singular pienāk, Lithuanian ateina) uz staciju ap četriem, etc.

³ One local 75-year-old resident of Gryva, Jānis Jadzēvičius, had by then moved to live in the neighboring Kūpčiškiai area (6 km south of Gryva). He spoke Lithuanian quite well with us. Here is a fragment of his autobiographical account: “I was born in Grīva || I am seventy-five years old || my father is Lithuanian and my mother Lithuanian || my mother was born in Zvirbliškiai, my father in Grīva || with Maskalāts from Kalkūnai, when we get together, we speak only Lithuanian” || The other old local Lithuanians of Gryva have all died.

⁴ Būga K. Rinktiniai raštai, vol. 3, p. 567, note 2, indicates that in Lithuanian it should be called Kálkumone.

223

In the area (7 km south of Gryva), many old local residents speak Lithuanian at home and consider themselves Lithuanians⁵. A few old local Lithuanians also live in Mėdumas (Latvian Medumi; older residents more often pronounce it Mėdinas; cf. also the name of the lake there, Mėdinas⁶), and in the vicinity of Smiltynė (Latvian Smiltene). Accordingly, during the expedition the dialect of the Lithuanians in the Daugpils area we visited, collectively known here as daugupieši, was also investigated. This article is based on those studies.

b) According to the information provided by K. Būga and A. Bielenstein, Lithuanians have lived since ancient times on the left bank of the Daugava in the area around Daugpils⁷. The dialect of the Lithuanians in these places has, one might say, remained unstudied to this day. However, K. Būga published some information about it, particularly its phonetics. In 1911, together with E. Volteris, he conducted some research on the Lithuanian dialect of the Ciskoda parish in Rezekne District (Latvian Tiskādi) and published the findings. In them, he compares some data on the Ciskoda dialect with data on the Lithuanian dialects of Gryva, Kalkūnai, and Laukesa⁸. Our study merely confirms and partly supplements the linguistic data published by K. Būga, broadens our knowledge, and presents new facts about the phonetics of the Lithuanian dialects in these places.

c) With minor exceptions⁹, dialect examples in this article are presented using conventional Lithuanian orthography. The dialect has vowels of three different lengths—long, short, and half-long.

⁵ The following people speak Lithuanian here: Augustas Maskoliūnas (83), Jūlija Maskoliūnienė (81), Povilas Kergalvys (also says Kiergalvys, Kėragalvis, Kėrgelis, Kiergelis) (81), Monika Kergalvienė (76), Antanas Katkevičius (67), Katkevičienė (about 65), Benediktas Igliškà (about 60), Apolionija Igliškienė (about 55), and their two school-age children; Gravelỹs (about 80), Gravelienė (about 78), Lapkauskienė (about 80), Plečkėnas (about 75), and Ana Jonėnas (also says Jānina) (about 80), among others. They all know Latvian literary language well, as well as Russian. We interviewed in Lithuanian only Kergalvys, Kergalvienė, Katkevičius, Maskoliūnienė, and Igliškienė. It should be noted that Kalkūnai belongs to the Laukesa (Latvian Laucēsa) Catholic parish, whose church, according to local accounts, held Lithuanian singing every third Sunday until 1977. In the cemetery of Kalkūnai village we found many Lithuanian inscriptions on monuments, e.g., Cia Ilsys Ona Lauriene Mire 1958 || Laurenas Antanas || Barėys Antanas || Cia ilsys Barcis Antaluk as ir Julitia || Barėyene Julja || Cia Ilsis Grawelania Jieva || Gravelis Ignatas || Rozalia Isadiene ||

⁶ Būga K. (Rinktiniai raštai, vol. 3, p. 567) reconstructs the source form Meddum as Medina. Local residents speak Lithuanian here: Stasys Masėnas (74), Smilginys, Mikelis Liskinys (69), Petrauskienė, Baimasovičienė, and others. In the neighboring Smiltynė area (6 km northwest of Medumas), we interviewed a local Lithuanian woman, Verdanienė (pronounced Vierdaniene). Anelė (70) and her brother Pranas Liskinys (57). For comparison, two school-age children from Igliškiai were also interviewed.

⁷ See Būga K. Rinktiniai raštai, vol. 3, pp. 563–568, and also pp. 99–100; see also Bielenstein A. Die Grenzen des lettischen Volksstammes und der lettischen Sprache in der Gegenwart und im 13. Jahrhundert. St. Petersburg, 1882, pp. 5–10, 172, 390. Of course, this should be verified against the findings of the latest research in linguistics, archaeology, and ethnography.

⁸ See Būga K. Rinktiniai raštai, vol. 3, pp. 570–572; see also ibid., pp. 88, 96, 97. Particularly significant and relevant to us is K. Būga’s observation here (p. 571) that “the Lithuanian dialect of Ciskoda differs in almost no respect from the dialects of Gryva, Kálkunė (Kalkuhen), and the Laukesa area”; this provides some of the very first general data on the Lithuanian dialect of the Daugpils area that we are investigating.

⁹ A broad unstressed vowel e, intermediate between e and a, is represented by æ; a weakly labialized unstressed o is written ɔ; e retracted after a hardened l is represented by ę; a mid vowel i, developed from e after a hard l, is represented by ь; consonant palatalization is marked by a small hook written high on the right side of the letter representing the consonant; variants of examples are separated by a vertical bar; two parallel bars are placed at the end of a dialect sentence; punctuation between examples—the small hook and the semicolon—is as in standard Lithuanian.

224

Since pitch accents here seem to be clearly distinguished, the length of a stressed vowel is indicated first and foremost by the letter representing it (e.g., e, ė, ę, y, i, ū, ų) and the corresponding pitch-accent mark (e.g., ų́ža.las ‘oak’, bū́va. ‘was’, skiẽda ‘splinter’); only the half-long quality of the first component of a stressed acute i, u + l, m, n, r diphthong (as well as the half-long quality of every long unstressed vowel) is indicated by a dot placed low at the right of the letter representing it (e.g., kí.lpa, bū.rtai, gi.vì.nte.). The length of an unstressed vowel, except ē, y, ī, ū, ų (which are written as such), is indicated by a bar written above the letter representing it (e.g., kny-gėlė ‘booklet’). The meaning of unclear dialectal examples is explained in quotation marks (e.g., už-aukl.a.t ‘to raise’).

II. CONSONANTS

1. Pronunciation of the older *dj, *tj clusters

The Lithuanian dialect of the Daugavpils area belongs to the eastern Dzūkian dialects. K. Būga was the first to notice and record this characteristic of it¹⁰. Although linguistically the dialect is more or less homogeneous, certain differences and fluctuations in pronunciation do occur. They are more pronounced only between the speech of the oldest dialect speakers and that of the young¹¹. Thus, Dzūkian affrication here is not entirely consistent either. It is more consistent and more pronounced in the speech of the older generation and in older words more familiar to the dialect, e.g., kniūpsčia, meldžias, glūdžia.s ‘smooth, straight’, pū.nciu., pū.ncia.te. || In the speech of the young, especially in newer words, it is usually entirely absent, e.g., dviračia., trupučiūku., va.ke.čiais, pamečiau || It also happens in the speech of the elderly that the same form may be pronounced by the same speaker both with and without Dzūkian affrication, e.g., čia.nai / čia.nai; plačiai, kiečiai / medžiai; jaunučiam / mažūčiam, mažūčiai; stačiau, arčiau / skaičiau; atáverčia, šaláldžia / paláldžia, pripučia || In general, we have observed Dzūkian affrication somewhat more regularly before fronted a, especially in first-person singular conditional forms, e.g., matýcia, na.rėcia, negalėcia; išvercia, nupláudžia ‘washes off’; trečia; duńcia ‘of a tooth’, žódzia.; šalaldžia.s, gárdzia.s; dzia.vėna ‘dried’; dzia.vi.ntas, but only grińčia ‘cottage’, grińčia.s || *tj, *dj > c, dz before fronted ai, au, o, uo are much rarer, e.g., kve.čiai, džiaugias, áudžiau, pradžios, pačioj, kve.ciuõs, but va.ke.-čiai, apsivedžiau ‘I married’, pakraščiūos; compare also paukščiukus, viščiõkus ‘chicks’, tičiojas ‘mocks’, suvaiškė.jit ‘you have gone a bit mad’. Dzūkian affrication is rarest before fronted u, e.g., paciùs, peciùs, girdžiu, sudžiùvus, kve.cių, but ma.čiùte., mergyčiu., baldagudžiu ‘Belarusians’, perinciau ‘having passed’, prieža.dziu., švińčiu., minučiù., vėrdunčiu., skaičius, grečiam ‘alongside’. Dzūkian affrication is barely perceptible in the shortened infinitive ending, e.g., nuskūsic, nėšč (< nešč due to progressive assimilation of š), but only gi.vi.nt ‘to live’, pridižia.vi.nt’, ge.dót’, palaist’, nusirámninte ‘to calm down’, aite ‘to go’, etc. Nowadays Dzūkian affrication is almost entirely absent before every i, y, ie, in of the dialect, e.g., gaŕdi, paspáudi, dýkas ‘empty’, gaidýs, mótina, de.dùtis ‘grandfather’, sėdi, ažudì.rpt’, ù.ntis ‘duck, duck’s’, gimdyšena (compare Latvian dzemdēšana) ‘childbirth; farrowing’, vidutinis, šimtinė.s ‘hundreds’, stótis, netýras (compare Latvian netīrs) ‘dirty’, nepatińka, atlaidit ‘release’, pridig s ‘will sprout’, artie ‘near’, dievàs, netiešà, diela.s

¹⁰ See Būga K. Rinktiniai raštai, vol. 3, p. 571.

¹¹ Differences in pronunciation are highlighted by examining each phonetic feature of the dialect separately.

225

‘intestinal worms, roundworms; a childhood disease of this kind’, daytime, although nevertheless it was also possible to hear cį.nsta ‘swells’, cikras alongside the more frequent tikras (an example with c is also recorded on tape No. VIII¹, held at the Institute of the Lithuanian Language and Literature). However, even in this case, the fact that džūzavimas must formerly have been entirely regular here¹² is shown especially well, first of all, by the name of the lake in these localities, Skirscinys (some also pronounce Skirščinis) < Skirstinys (possibly also the village name Džipaniūnkai in the Skrūdaliena area), and words used in the dialect such as išči.nsta, viščiena, arkliarūgščis ‘horse sorrel’ (< iščinšta, viščiena, arkliarūgščis owing to the progressive assimilation of š), as well as the hypercorrect forms found here, va.ke.ečys (< va.ke.ečys) ‘German’, tūkstuncis (< tūkstuncys) ‘thousands’. The leveling of džūkavimas, especially before the vowel i, was most likely influenced above all by the Latvian literary language, which everyone here knows well and in which a similar phonetic phenomenon is not characteristic¹³. It should also be noted that no cases of tv, dv changing into cv, dzv have been observed in the dialect at all, e.g., pusdvi.likis ‘half past twelve’, tvirtus ‘firm’. Very rarely, a palatalized t is pronounced instead of dialectal c or standard č, e.g., less often / less often / less often (dial.), more firmly / more firmly / more firmly (dial.). This is possibly most likely an analogy with other forms of the same stem. On the other hand, perhaps this could also be some atypical hypercorrection.

The č of onomatopoeic origin and in loanwords is generally preserved in the dialect, e.g., čiū.lba, čiupūlis ‘one who handles things’; pėtmį.čia ‘Friday’; punčiaka.s ‘socks’, vorčiukas ‘long-legged’, telvčia.s, skle.mčia.s, či.staciėlas (cf. Belarusian чытачел) ‘celandine’, although pėčius also occurs (alongside the more frequent and usual pėčius).

2. Pronunciation of other consonant clusters with j

In the dialect, consonant clusters with j have generally disappeared, softening the consonant before them; only word-initial bj, pj always remain intact, e.g., bjaurūs, spjaut’; bu.lbū., kaltybin.; laibiau, patepiau, iškepiau; kelių, negaliu, apipilia, nemi.lia ‘nemylį’; žėmiu..

¹² K. Būga, claiming that the Ciskodas dialect differs very little from the Lithuanian dialect of Daugpilis, notes that here dzūkavimas occurs according to only one feature: čia, svečias, marčios, but marti, tikras (Būga K. Rinktiniai raštai, vol. 3, p. 571). Zinkevičius Z. repeats K. Būga’s view in Lietuvių dialektologija. V., 1966, p. 139. The instances of dzūkavimas in the Daugpilis Lithuanian dialect that we have provided refine this claim and allow us to assert that here, as in most of the eastern Dzūkian dialects in Lithuania, both features have traditionally been present, but the second of them had apparently already become very indistinct and leveled out by the time of K. Būga’s research, and in recent times only a few traces of it have survived.

¹³ It should be noted that in the area of the Lithuanian dialect under study, we did not find local residents who knew the Latvian High Latvian (Latv. augšzemnieku) dialect (according to the classification of Latvian dialects, these places belong to the High Latvian dialect). None of the old local residents remembers, nor have they heard from their ancestors, that the High Latvian dialect was ever spoken here. Before the war, the official language of the people of this region was literary Latvian. It was used in schools, and so all local residents (including non-Balts) learned that language well (rather than the Latvian dialect) and use it when necessary. That there are in fact no speakers of the local Latvian dialect here is evident from the work of Latvian dialectologists. For example, when mapping the phonetic, grammatical, and lexical features of Latvian dialects, the maps most often show only blank areas for data from Laukesa (point no. 456), Demene (point no. 460), Skrudaliena (point no. 457), and neighboring points (see Рудзите М. К. Латышская диалектология (Фонетика и морфология). Автореф. дис. на соискание ученой степени доктора филологических наук. Рига, 1969. Карты № 4,6–9, 13–16, 23, 26, 32, 33; Лаумане Б. Э. Лексический материал диалектологического атласа латышского языка, отражающий латышско-русско-белорусско-польские контакты. — В кн.: Контакты латышского языка. Рига, 1977. Карты № 1, 8, 10, 11, 16.

226

latviām; šiūndie ‘šiandien’, gelažnia.; kirviù, turiù, kētūria.s; miršiu, nuskindusiu, unksti.vōsias ‘ankstyvasias’, pirmasiai ‘pirmasis’; apsigrīšiu; grīžiam ‘gręžiam’, žiu.rėk, nuvažiuosim || There are instances in which clusters bj, pj are also pronounced intact at the end of a word, e.g., kvėpja ‘kvėpi’, šlapja ‘šlapia’, griebja ‘griebia’, grėbja ‘grėbia’.

Recently, a tendency has been observed in the dialect to harden stem-final l before the deleted v, as in some Lithuanian border dialects and neighboring Latvian dialects¹⁴, e.g., prōte.vai ‘protėviai’, le.tuvų ‘lietuvių’, su le.tuvų ‘su lietuvių’. Forms with a hard stem-final v are more prominent in the speech of young people or people who have somewhat forgotten the Lithuanian language, whereas the older generation of the dialect does not usually pronounce it this way.

It should be noted that in the dialect, the vowel a and diphthong ai following j or softened consonants, when fronted, do not turn into e(ei), e.g., sėjai, grėbja, grīžiam ||

3. Pronunciation of consonant l

In the dialect, l is usually hardened before e, ė and ei. The vowel ė (which occurs only under stress) becomes a mid-tongue e in all word positions after a hardened l, e.g.: plėšia, plėniu., gulėte., vabuolėliai ‘vabalėliai’, paslėgia, tilėk ‘tylėk’, glėbi., mi.lėja., našlė, ža.lėj, pu.s-lės, ža.lėm || When, under the influence of Latvian lk, stress is shifted from stressed lė to the first syllable of the word, the now unstressed lė does not turn into la or shorten, e.g., gailėčia ‘gailėčiau’. This shows that Latvian-like retraction of stress to the beginning of the word is an entirely new phenomenon in the dialect. The dialect’s stressed ė also becomes e when it occurs after l followed by another consonant, e.g., papilvė, gelbėk, kalbėk.

Every e (and ei in a diphthong) in the dialect becomes a after a hardened l, e.g., laša, vaŕ-lās ‘varles’, pu.slāla.s ‘pušelės’, ku.lālis ‘kūlelis’; išlaidžiau ‘išleidau’, palaide. ‘palei-do’, kōlai ‘kol’, atlaidit ‘atleiskite, paleiskite’; avēla. ‘avis’, kipēla. ‘milžtuvėlė’, kīela. ‘kielė’, pirmāde.la., vā́rla.s, kakēla.s ‘speneliai’, skidēla.s ‘skiedrelės’, vadēla.s ‘va-džios’, sūpo.kla.s ‘sūpuoklės’, kīrme.la.s, vaisla.s ‘veislės’, āgla.m, ka.iēla.m, brō-la.nas ‘brolėnas’. Instances have been recorded in which l is not hardened before e, ė, e.g.: lēpšai ‘lepšės’, tokie grybai, klēvas¹⁵, gēle.s ‘gėlės’, klepšēt, plepējimas ‘puskėjimas’ (: kōše.s plepējimas), Plečkēnas (surname). N has hardened l also before the intact root diphthongs en, em in the words lēnda, lenkām, lenkās, lenktas, iš lēnga., suglēmba, plēntas and before in < en in the words lintā ‘lenta’, lintēla.s, Linkijōj, nesilūnka, lińk ‘lenk, gink, saugok’, lingviā¹⁶. Before e and another consonant, l is sometimes hard, e.g., kulnāliai ‘kulneliai’,

¹⁴ Plg. Lietuvių kalbos tarmės (Chrestomatija). V., 1970, p. 27; Ancītis K. Aknīstes izloksne. Rīgā, 1977, p. 123.

¹⁵ Compare also the form klēvas, found in all the other East Aukštaitian dialects that harden l before e, and probably derived from kliavas (see Būga K. Rinktiniai raštai, vol. 3, p. 90; Zinkevičius Z. Lietuvių dialektologija, p. 160), as well as Latvian klavs ‘maple’.

¹⁶ In the Skrudaliena Belarusian dialect, as mentioned above, the Lithuanian-origin proper names we recorded—the surname Dunķūn and the place name Lyntōpiški—are undoubtedly Slavicized adaptations of the Lithuanian dialect forms Linkūnas (< Lenkūnas), Lintōpiškiai (< Lentōpiškiai). That Lithuanian words and names tend to harden root l as they become Slavicized is also evident from other Lithuanian loanwords used in Belarusian dialects. For instance, a stream in the Švenčionys area and the town near it, already in Pastovys district of the Byelorussian SSR, are called by the local Švenčionys Lithuanians Lintūpis (Ik Leītupis), i.e., with a soft l. But the Belarusians of the Lentupis area itself and official documents (maps, gazetteers, and

227

kalnaliūkas ‘hillock’, but balsvėne.s || The hardened l also occurs before word-final -i < *-ēn, *-ēn. After this l, a mid-front vowel is always pronounced here, bl, e.g., kni.gebli, kiaūbli; galbi ‘at the end’; žōbl. ‘grass’, vinēbl.; prišābl.s ‘having caught a chill’, nejišbāl.s, užpūolbl. ‘attacked’, prižēbl. ||

The softness of l in the word stem before dialectal in < en and intact en (for instances of the latter, see p. 233), and its hardness in that dialect before word-final -i (< *-ēn, *-ēn), show that the hardening of l and the development of en > in differed in the two cases. In the word stem, the first component of the diphthong en here, as in many other neighboring Eastern Aukštaitian dialects, narrowed quite rapidly under the influence of the assimilating second component, consonant n¹⁷. Consequently, narrowing in this way, this en had changed into in even before the beginning of the hardening of l. By the time l began to harden, the conditions for hardening l before this in (< en) no longer existed, and it remained soft in this dialect. The development of *-len at the end of a word is fundamentally different. By then, reduction of endings (especially acute-accented ones) and the disappearance of old tautosyllabic diphthongs must already have begun. Therefore, the first component of the word-final diphthong en (vowel e), much less affected by the assimilation of the second component (consonant n), weakened as a result of ending reduction, narrowed much more slowly than in the stem. With this narrowed, but not yet changed into i, first component (vowel ę), the diphthong en would have encountered the hardening of l¹⁸. Thus, this diphthong (en) still retained the conditions that allowed the l preceding it to harden. The en itself, affected by this hardening, changed into en, whose ę, increasingly less affected by the weakening n, gradually narrowed further, becoming not the high-front i but a mid-front vowel bl.

On the other hand, the ending -lbl could perhaps also be the result of analogy with other forms of the same words: on the model of ža.lė, šaūla. and žōbl., šālbl. ||

The stem-final ending -la. (< *lē, lk lē) of the 3rd person singular past tense of verbs has most likely analogically influenced the stem endings of the 1st and 2nd persons singular in the dialect, and accordingly changed them to -lau, -lai, e.g., acikėlau ‘I got up’, papūolau ‘I ended up’, papūolai ‘you ended up’. On the other hand, this could also be the influence of the morphology of Latvian verbs with past-tense ē-stems, cf. Latvian veda ‘led’, vedu ‘I led’, vedi ‘you led’; cf. also the form pērvedau (< *pervedjau) in the dialect under study, ‘I translated’ (for the semantics, cf. Russian переводить ‘to translate’), which probably arose under the influence of Latvian morphology.

In loanwords, l is not hardened before e, ē in the dialect, e.g., davlėnija ‘pressure’, kalēda.s ‘Christmas’, kazlēnas ‘a kind of mushroom’, klebōnas, klēvas (cf. Belarusian хлеў) ‘barn’, pakalēnija ‘once’, palēkas ‘left-handed’, plēmai ‘spots’, skle.nyčia.s ‘glasses’, sklēpas

rašai etc.) call the stream Лынтупка, Лынтуп, and the town Лынтупы, i.e., with a hard l. Thus, the hardness of l in this case, as also in Лынкӯн and Лынтонішки, is undoubtedly the result of Slavicization of a Lithuanian word. That the l of a Lithuanian word can be pronounced hard before any in solely because of Slavicization is well demonstrated by cases we recorded in the Lithuanian dialect of Lazūnai, e.g., išblīs’c’ ‘to crawl out’, išblēzdinē ‘used to crawl out’, išblindo ‘crawled out’. Such hardening of the l in these words cannot be explained in any way by facts of the Lithuanian language itself. This is undoubtedly only a result of the Belarusian adstrate.

See Kazlauskas J. Possibilities and limits of establishing the relative chronology of changes in the Baltic languages. — Baltistica. Supplement I, 1972, p. 84; Zinkevičius Z. On the narrowing of the first component of Lithuanian tautosyllabic -an combinations. — Ibid., p. 227—228; Grinaveckis V. On the origin of *-ān equivalents of Lithuanian *-ā in the stem, noun and pronoun endings. — LKK, 1978, vol. 18, p. 127, 135.¹⁷

¹⁸ Plg. Morkūnas K. Rytų aukštaičių pietinės tarmės fonetika. — LKK, 1960, t. 3, p. 43.

228

‘cellar’, tūpele.s (cf. Latvian tupeles) ‘shoes’, vaspalėnija ‘inflammation’. Thus, these loanwords became widespread in the dialect relatively recently.

A hard l was also heard in the words: Suvalkija.j ‘in Suvalkija’, alksnis, žvilgsnis¹⁹. In certain cases, unlike in lk l, an unhardened form occurs in the dialect:

a) In the stem-final position of 2nd person singular imperative forms (with and without the formant k), e.g., jiškū.l’k | jiškū.l’ | sumal’k | sumal’, pi.l’k | pi.l’;

b) At the end of the word in certain function words, e.g., atgal’, dėl’ (cf. Latvian dēl), gāl’ (and gál), kāde.l’ (cf. Latvian kādē) ‘why’, tegul’, vēl’;

c) In certain words before another hard consonant or consonant cluster, e.g., žal’sva.s ‘greenish’, Kal’kūnai ‘Kalkūnai’ / Kālku.ne. ‘Kalkūnė’;

d) In loanwords, e.g., bāl’ka ‘beam’, bū.l’ba.s ‘potatoes’, butel’kōs ‘in bottles’, kastrū.l’ka.n ‘in a little pot’, kul’kōs ‘bullets’, va.si.l’ka.s ‘cornflowers’. In the latter case, the soft l must have entered the dialect along with the loanword itself.

4. Hardening and palatalization of other consonants

In the speech of younger speakers of the dialect, there is a tendency in certain cases to harden r, s, š, v, ž and the affricates c, č before i, ie, e, ia, although in many other words the same consonants are pronounced softly in an analogous position. We observed the following words with hard consonants: čėrui ‘for the car’, dukrė ‘daughter’, duńčka. ‘of the roof’, grabėstai (grebėstai, sn.), inžbniėrius ‘engineer’, kōšbl. ‘porridge’, lēpšai ‘lepšės’, mašb.ln.ka.s ‘little machines’, Papūšė ‘place-name near Kalkūnai’, pasėta, pasisės, pre (alongside pri) ‘toward’, rōžbl ‘rose’, rōže., sėkla, sāsėrbs (sėseris, sn.) ‘sisters’, sėja, vaisla.s ‘varieties’, varbīt’ ‘to drive’, žbina., šaurē ‘north’, šblfero. ‘of slate’, but atāveže., āžeru., jiše.jau ‘I went out’, maišelis, na.rėcia, nūmuše., reikia, rūdame.se.s, senātve., šeimyna, šeši, šešias, šve.sėsne., umsispūre. ‘dug in’, užėjo. ‘came in’, veršiai, viešiakelia. ‘of the highway’, žėme., and so on. In many cases, this hardening of consonants is probably a recent phonetic phenomenon, only beginning to take hold in the dialect, undoubtedly caused by the influence of the phonetics of neighboring dialects in contact—Latvian, Belarusian, and Russian²⁰. However, in the words krakenai ‘colostrum’, skraģžde.s ‘swallows’, mėnasi. ‘in the month’, priragėjau ‘I looked after’, rāgis ‘it seems’, sānas, sanēliai ‘grandparents’, sānis ‘old man’, sanyšte. ‘old age’, sūsani ‘grows old’, the consonants, as in neighboring Lithuanian dialects in Lithuania, are evidently hardened by all speakers of the dialect from of old.

In the speech of older people, the pronunciation of some word-final consonants as soft is also observed here, though not entirely consistently:

1) The consonant of the secondary ending is usually soft, e.g., aĩn’ ‘go’, jīm’ (alongside pajīm) ‘take’, nėš’ ‘carry’, pame.gén’ ‘try’; mān’, kasdiēn’, pa.piēt’;

⁹ According to Prof. Dr. M. Rudzytė of the Latvian State P. Stučka University, in Latvian too—and especially in its High Latvian dialect—the tautosyllabic l in this case is pronounced much harder than in other positions.

²⁰ In an analogous position, these consonants are also pronounced hard in all the dialects in contact—Slavic and High Latvian.

229

2) In the speech of older people, t is soft in the ending of present-tense converbs, e.g.: šakunt’, žinunt’ (alongside gi.vėnumt, dabojunt, nj., dž.);

3) In words new to the dialect, the root ž is sometimes also pronounced softly, e.g., žeizdu. ‘a wound’, žeidžia ‘plays’.

5. Addition (insertion), loss (dropping), and disappearance of consonants

At the beginning of a word, before j, ie, j is added almost regularly, e.g., jįrgi ‘also’, iškuta, jįlsta, jįmtų, jį.lgas, jįšimi, jiena, jįeška. ja. ‘looked for’, but no such addition occurs in place names (cf. Ilgis ‘name of a lake’). Probably by analogy, j is also added to the ending of the old locative namõ, which has become differentiated in opposition to other forms: namój, cf. ailiój ‘in a row’.

It has been observed that j is also added at the beginning of some words in the neighboring Belarusian dialect, before i, e.g., įindbikám (< индбикам) ‘to turkeys’.

Usually j is also inserted between two vowels at a syllable boundary, e.g., pasijį.lga. ‘grew tired’, pasijįnš, pajįlsis ‘rest’, ažuje.me. ‘occupied’, mujįnt’, neprisijima, nejiš-gýda.ma, nejįsėsi ‘you won’t manage’.

An intrusive v sometimes occurs between u and e, e.g., nuvėja.m ‘we went’, suvėinam ‘we meet’. At the beginning of a word, s may be added, e.g., skragždė.s ‘swallows’, stvėre.s ‘held on’.

Some loanwords beginning with the vowel a are pronounced in three ways: with an added g, h, or without it, e.g., arbótu. (sn., dž.) /garbótu. / harbótu. (nj., rt.), armóšku. / garmóšku. / harmóšku. ‘accordion’.

In the first-person singular form of the past simple, the verb root and the suffix -ėj- tend to be lost in the dialect, although full forms also occur less often, e.g., nediau ‘I did not put’, pradiau ‘I began’, padiau ‘I helped’, negaláu ‘I could not’, nena.ráu ‘I did not want’, neragiáu ‘I did not see’, raviáu / ravėjau, ma.kiáu / ma.kėjau, mi.láu / mi.lėjau, praguláu / pragulėjau || Forms mi.láu, guláu, etc., show that the loss of -ėj- is relatively recent, occurring after the sequence lė had hardened in this dialect, as in some other Lithuanian dialects on the periphery²¹. Forms of the other persons in this tense are the same as in 1k, e.g.: ma.kėjai, ma.kėja., ma.kėja.m, ma.kėja.t ||

In the singular locative forms of masculine nouns, j usually disappears together with the entire syllable, e.g., Daugavpili. ‘in Daugpils’, undený ‘in water’, pavakary ||

A voiceless consonant at a morpheme boundary before a voiced consonant, as elsewhere in Lithuanian dialects, is always voiced through regressive assimilation (e.g., trizdešimt); a voiced consonant before a voiceless one becomes voiceless (e.g., dį.rpk). The consonants š, ž disappear in the dialect when they occur before s, š (e.g., asįkrečia < ažįkrečia ‘gets infected’, išivede. < išsivede.., išáuke. ‘called out’).

6. Various consonant changes

The consonant n before b, p has invariably become m, e.g., um burókais, um púse.s, um piėtu., um pečiū ||

²¹ Cf. E. Grinaveckienė. Some More Characteristic Features of the Apso Dialect. – Proceedings of the Academy of Sciences of the Lithuanian SSR. Series A, 2(9), 1960, p. 177.

230

Most likely under the influence of Latvian, the consonant š in some words has sometimes become s, e.g., svarù ‘clean’, svariai ‘cleanly’, iš kur ‘from where’, latviškai ‘in Latvian’, le.tuviškai | le.tùviškai, išeina | išeina, as | aš ||

The consonant h is rare in the dialect and usually occurs in the speech of young people, cf. the example harmošku. Otherwise, g is more often found in its place, e.g., gichtāru. ‘hectare’, while sometimes h is dropped altogether, e.g., arauš (< German heraus) ‘out’.

It happens that in foreign words in the dialect, the consonant l is sometimes pronounced instead of word-initial j or r, e.g., lubilėiny (< юбилейны, regressive assimilation of consonants) ‘anniversary’, lādu.ga (< радуга) ‘rainbow’.

III. VOWELS AND DIPHTHONGS

1. Short vowels

a) The old short stressed stem vowels a, e are generally long in the dialect, e.g., nāmas, nėṡa. Unlike in the standard Lithuanian dialect, a, e are half-long (to our ears, they differ little from long vowels) in verb prefixes, in the infinitive of two-syllable verbs and in tense forms derived from it, in the pronouns mano, tavo, savo, and in comparative forms of adjectives, e.g., kãste., kãzdamas, nẽšte., ištept’, nẽneša, atãneše., mánā., tãva., sãva., gardẽsnis, vi.rẽsni.sai || It happens that stressed a, e in verbs are sometimes pronounced short, e.g., nẽšaukia, nẽmi.li, suprãšk, nupẽšk. In Latvian place names and some individual words, owing to the influence of Latvian, and in borrowings from Slavic languages, stressed a, e may be pronounced short, e.g., Kãplava, Krẽveņi, Mẽdinas (place names), bãlunde.s ‘daisy’; zapalẽnija. It has also been observed that stressed e in an open syllable is pronounced somewhat more open, with a slight a coloring, by younger speakers of the dialect, e.g., kãpe.m, lintẽla.s, whereas older speakers slightly raise stressed e before a syllable containing u or a fronted a, pronouncing it with a slight ie coloring, e.g., miẽtu. ‘years’, apẽja. < apẽja. ‘went around’. I have heard e pronounced as the diphthong ei instead of stressed e, e.g., ske.deila.s (< skiedelės) ‘splinters’.

Word-initial e and ei invariably become a, ai, e.g., ãsu. ‘supposedly’, ãst ‘to be’, ãina ‘goes’, ailiój ‘in a row’.

b) Old stressed stem vowels i, u are generally lengthened to half-long, e.g., nù-mire. (rt. and nùomire., cf. Latvian nomira), būvus, ùšlipe.m ‘we climbed up’, iškepe., apinẽšta, išlupe., ažiùterše., ažiùgine. ‘forbade’. In a few rare cases, stressed stem vowels i, u are not lengthened (most likely under Latvian influence), e.g., nesigirdi, tùpele.s ‘little shoes’, nù-guru. (cf. Latvian muguru) ‘back’, etc. The u in foreign words is sometimes pronounced long, e.g., sùtka.s (< суткú) ‘days’.

c) The stressed stem vowel o in foreign words occurring in the dialect is generally short, e.g., kìpiatòkas ‘boiled water’, dòmikas ‘frame beehive’, burdòvas ‘dark red’, živatnavòdstva ‘livestock’ (e.g.: karvytė., avẽla. — and all māna. živatnavòdstva), tvòragas ‘curd’, baciònaì ‘storks’, although in older loans it is pronounced long (šampijōnai). When unstressed, this o has become a short a, e.g., nerekomendùjas ||

231

2. Long vowels

In the dialect, the old unstressed long vowels *ā, *ē, *ī, *ū in the word stem and ending have most often shortened to half-long vowels, e.g., ža.lė, raša.t, me.lyniu., ža.li.nai, māte.m, ma.kēja., se.jimu., pla.kštīte.s ‘plates’, dabartēlia.s ‘now’, bliu.dēlin, vālgi.mas, kriūkšti.te., gāudi.te., su.nai, mi.lėjau || Under primary (and secondary retracted) stress, the old ī, ū have remained intact, while ā, ē have become long o, ē, e.g., žōbl., švēre.s, praplōve., kokis ‘tree’, žōbli, rēkū; perkūnija, nejīsgyda.ma, gīdī.te., šaknīte.s, dūmai, knyga, sūnus. However, it has been observed that older speakers of the dialect and those who speak Lithuanian better often retain broad pronunciation even in stressed ā, ē, e.g., ēda ‘eats’, žiemās ‘winters’, sēja ‘sows’, na.rējau ‘I wanted’, tēvas ‘father’, duktē ‘daughter’, gēda ‘shame’. These are most probably living relics of the old Žadininkai dialect inherited from antiquity, linking it with the Lithuanian dialect of the Ciskodo area²² and maintained under the influence of Latvian, cf. Latvian āda ‘skin’, ēst ‘to eat’. Among younger speakers of the dialect, the old unstressed ā, ē, ī, ū tend to be shortened to short vowels, e.g., kurēna, nerūkiau, rūkyte., apiniai ‘hops’, rūsių ‘kinds’, kruvōn, vālgimas, gydītajis, virēšnīsi, tilēk, silōs ‘strength’, rūkiijimas ‘smoking’. In the nominative plural endings of nouns and the third-person future-tense endings of verbs, the old ā, ē, ī, ū are always short: žūšis ‘geese’, ausis ‘ears’, akis ‘eyes’, tūūkstantis ‘thousands’; gē-rus ‘good’, teisi.ngus ‘righteous’, bāltus ‘white’; nestavēs, nusibās, susanēs, pаведzīs ‘will show’, pridigs, pridažīs, surūgs || At the end of new dialect words, under the influence of standard Lithuanian, a half-long o is occasionally pronounced instead of the usual half-long a. (< *ā), e.g., šībfero. ‘of slate’.

3. Diphthongs

a) In the dialect, unstressed ie, uo have most often become half-long e, a, e.g., ve.na.likis ‘eleventh’, Va.ke.tija.i, le.tūviškai ‘in Lithuanian’, ašta.niū, ge.dōdava., ūža.las, but among older speakers they are often pronounced intact as ie, uo, e.g.: išjuoka.s ‘mockery’, puodūkas, diedūtis ‘mullein’, pievōs ‘into the meadows’, pietais ‘with lunch’, Juozūkas, kviečiūos, skiedēla.s ‘splinters’, žiemās, vartūos, juodējī, mētuos, pērvāžuoji, n.žūolū, vabuolai ‘beetles’, lietāus || Alongside these, at the end of stressed and unstressed words, a weakly labialized, lax o (written ǝ) is also quite often heard instead of uo, e.g., akōp.s ‘in the trenches’, Kalkūn.s | Kâlkūn.s ‘in Kalkūnai’, šītūs ‘in these’, pabēge.-lio.s ‘among the refugees’, plentōs ‘on the highways’, undō ‘water’, rudō ‘autumn’, akmō ‘stone’, šū ‘dog’. Less often, such ǝ occurs instead of uo in the word root, e.g., ǝga ‘berry’, ǝ.gauti ‘to pick berries’, ǝ.dai ‘mosquitoes’²³.

Probably under the influence of standard Lithuanian or neighboring Lithuanian dialects, unstressed ie, uo are pronounced intact in words that have recently entered the dialect, e.g., rankšluostis, stāltiese. ‘tablecloth’; tiesiōg, sliekai. Instead of stressed uo, long ū is pronounced in the new dialect word grūdīs ‘December’ (probably on the model of Polish grudzień ‘idem’).

In the dialect, ie, uo under secondary retracted stress are also always intact, e.g., netiēsi, viēdri ‘small bucket’, skiedā ‘splinter’; Juōžū, juōkū ||

²² Cf. Ciskodo Lithuanian dialect lobas ‘lobas’, krōsnis ‘stove’ (Būga K. Rinktiniai raštai, vol. 3, p. 571).

²³ Sometimes such ǝ is also pronounced in place of the stressed ending -au, e.g.: nežinǝ ‘I don’t know’.

232

At the end of a word, the old acute-dipthongs ie, uo have generally become short i, u (e.g., aiši; aini; guliū). In the singular second-person imperative without the ending k, this short i has disappeared entirely, e.g., ai’n ’go’, pame.gén’ ’try’. The full ie is pronounced in the adverbialized old locative artiė ’near’.

Alongside the infinitive with the soft suffix -t’ (: ait’), among older speakers a form with the suffix -te. (: aite., nėšte.), which phonetically most likely developed from -tie²⁴. is widespread.

b) In the dialect, the stressed and unstressed stem diphthongs am, an, em, en, as in neighboring East Aukštaitian dialects in Lithuania, have generally developed into um, un, im, in, e.g., bumbytė., kuṃpas ‘corner’, tumšumas, užšūmpi.s ‘swollen’, aplunkytė.‘ balundė ‘pigweed’, bālunde.s ‘pigweeds’, brū.ngi, dabójunt, duṇgtis, duṇtis ‘teeth’, gi.vėnunt, kū.nda, kuntrybė., lungai, luṅkas, nusguṇda. ‘was frightened’, pū.nciu. ‘of the fetters’, pusūntra., ruṅkā, skanduōnis (cf. skandinti) ‘drowned person’, šiūnde. ‘today’, tūks-tuntis ‘thousands’, tū.nkiai, tū.nkus ‘dense’, ū.ndeni. ‘water’ (acc.), ūṅksti, untienà, ū.ntis, untrū, vėrdunčiu. ‘boiling’; diṅgta ‘covered’, gi.vī.nte., liṅk ‘bend, keep watch’, Linkijoj ‘in Poland’, lintėlą.s, lintū, nesilinkja, nuskīnda., nuskīndusiu., pī.mpe., pink-tādieni., and also umsispyrė. < antsispyrė. ‘dug in’, pīñktas, švīntas, tī.msta, ži.n-tas || However, it happens that in words new to the dialect, secondary diphthongs am, an, em, en are also pronounced intact, e.g., sánde.lis, raṅkdarbiu., raṅkšluostis, balañdis; mañ’, támsta, némne.s (< nemunės) ‘such mushrooms’, sénmerge., sénpane., vėžam, ve.-nam, ka.žnai (sl.) ‘in every’, jaunámsiam ‘to the young one’, latviám, ažerañ, lenkám, pame.gén’ ‘try’; āmžius, paskañbink, iš leṅgva.

Cases have been recorded in the dialect in which the development of these diphthongs is optional in particular words, e.g., brū.ngus, but brángiam, liṅk ‘bend, keep watch’, but sūlenktas, Linkijoj, but lénkiškai, léṅkas, unds, ū.ndeni, but vandeńis, pīñktas, but peṅki, dzia.vī.nt’ and dzia.vėnt’, diṅgta, but padeṅgt’.

It has also been observed that, in attempts to speak correctly, en is sometimes restored even in in that did not originate from in, e.g., mán’ ‘I happen upon such a tooth’; susitenki ‘you meet’.

In the dialect, the infix an in verb roots generally also becomes un, e.g., nesu-prūta, nesuprūntū, prūntū ‘I understand’ (but neprantū is also said), atsirūnda. The en, em infix in verb roots here usually remains intact, e.g., užsteṅka, pateṅkitas ‘satisfied’, išteṅki, nesugleñmba, atāskreṅda (although alongside these, skruṅ-da < skrañda < skreñda are also said). Verb-root en arising through vowel alternation does not become in either, e.g., reṅkam, susirenka, leṅda, neišleṅda.

Sometimes the diphthong -án- in the prefix sán- has developed into uo, e.g., sūoma.ne. ‘consciousness, understanding, awareness’ (: saneliūtė. already without sūoma.ne.s: here pasākė., here nežī-na.; būva, pri visai sūoma.nei and mire.). Although contacts between Latvian speakers and the dialect under study have been good both in the past and today, as the phonetics shows, Latvian samana ‘consciousness’, samaz-gas ‘pamagzos’, etc., could not have influenced the formation of the dialectal form sūoma.ne.s. The Latvian dialects (with which, as mentioned earlier, the dialect under study has not had and does not havesamana ’sąmonė’, samaz-gas ’pamagzos’ and kt.

²⁴ It should also be noted that, in rare cases, the dialect also has an infinitive form with the suffix -tei, e.g., lygintei ‘to compare’, važiuotei ‘to drive’, kalbėtei ‘to speak’, rašytei ‘to write’. It is hard to believe that this suffix -tei is some relic of the old *-tei; cf. also the infinitives nėstei ‘to carry’ and pl.hei ‘to pour’ in other East Aukštaitian dialects. This is most likely a newly added postpositional particle (cf. also bėgtienai ‘to run’ in East Aukštaitian dialects); see also Lietuvių kalbos gramatika. V., 1971, t. 2, p. 398—399.

233

(direct connections) rare (and of not entirely clear origin) forms suõmazgas ‘Spülicht’, suõvara ‘das Querholz auf der Egge’, suõvardis ‘der Namensvetter’, suõvista ‘Kindsbinde’, etc.²⁵ are also unlikely to have influenced the emergence of suõma.ne.uo in the dialect under investigation from án. The case of suõma.ne. in the dialect under investigation could perhaps be explained as an independent outcome of the development of the diphthong an in that dialect, arising, possibly, by the same route as uo developed from *-ān in other Eastern Aukštaitian dialects in the forms gerõju ~ gerąja, gerõsius ~ gerąsias, or Lazūnai vuõžuolas from *ānžuolas²⁶. On the other hand, perhaps this suo- is a relic of the old prefix suo- < *sō?²⁷

Foreign words with am, an, em, en generally remain unchanged in the dialect, e.g., ánglas, apendi-citas, beñkis ‘bench’, centimetrų., cimeñtas, fundameñtas, grántas (< Latvian grants) ‘gravel’, kamándava.ja., kamandujù, nerekamendijas, penicija, reñdu. ‘rent’, strumeñtas ‘instrument’, studeñte., šampijōnai, valniánka.s ‘mushrooms, woolly’, vantrò-bu. ‘intestines; entrails’, venzlai ‘hand luggage, belongings carried by hand’. We have observed that an has become um in the loanword punčiáka.s ‘stockings’. This would indicate that the latter borrowing entered the dialect relatively long ago.

c) The old circumflex diphthongs an, en, in, un, corresponding to Standard Lithuanian ą, ę, į, ų, have here become u, i, which are long when stressed and usually half-long when unstressed, e.g., úža.las ‘oak’, žùsis ‘geese’, ku.snelíukas ‘a small piece; a little bit’, nuzgąsta ‘gets frightened’, tru.sù ‘I crackle’, neaužšùla ‘does not freeze’, apsigrįžtam ‘we turn around’, skįsta ‘sinks’, tavįs ‘you’ (gen. sg.), tį ‘there’, ãsu. ‘are’, išėjì. ‘having gone out’, bùvis, dãris ‘having done’, žiẽmu. ‘winter’ (acc. sg.), mộčiu. ‘mother’ (acc. sg.), nepažįstù ‘I do not recognize’, bẽganti. ‘running’, vỹru. ‘men’ (gen. pl.), nesumintù. It has been noted that younger speakers of the dialect usually shorten unstressed word-final u, i of this origin, e.g., smáugis ‘grown-up’, dãris ‘having done’, láidis ‘having let’, numirìs ‘having died’. In compound words, the old circumflex an reflex occasionally occurs in two forms—half-long u and half-long a (under the influence of Standard Lithuanian!), e.g., viẽnu.kart, kita.kart, ùntra.kart. In certain cases the old diphthong en here also becomes in, e.g., nuskìnsta ‘sinks’ (alongside skįsta).

The old in in prefixes has generally become i, e.g., įkišė. ‘put in’, i.kùs ‘will bite’, although it too is preserved in rare cases: insidaužiau ‘I bruised myself’ (perhaps owing to the analogy of unsispyrė. ‘dug in’?).

In third-person future-tense forms, the dialect’s u, i (< an, en, in, un) are always short, e.g., i.kùs ‘will bite’, nuzgùs ‘will get frightened’, sugrìs ‘will return’, nuskìs ‘will sink’, pasìus ‘will become mad’.

d) The old word-final acute *-ān, *-ēn in the singular instrumental and locative forms of nouns and pronouns have become u, i, e.g., kadù ‘when’, tadù ‘then’, mộčiu ‘mother’ (instr.), griñčiu ‘in the cottage’, kãrvi ‘cow’ (acc.), žẽmi ‘earth’ (acc.), ãžeri ‘in the lake’, miški ‘in the forest’, galì ‘at the end’, vežìmi ‘in the wagon’, karštùmi ‘in the heat’, sciuri (cf. Latvian stūris) ‘in the corner’. The locative endings *-ēn have been entirely lost before j and s, e.g., žẽme.j, paežerėj, karioñmine.j, bažnyčia.j, kataroj ‘in which’, Va.ke.ti.ja.j, Ri.goj

²⁵ Žr. Mūlenbachs K. Latviešu valodas vārdnīca. Rediģējis, papildinājis, turpinājis J. Endzelīns. Rīgā, 1927—1929, t. 3, p. 1138—1139; Endzelīns J. Latviešu valodas gramatika. Rīgā, 1951, p. 685—687.

²⁶ Cf. Grinaveckis V. On the Origin of the Correspondences of *-ān in the Endings of Lithuanian *-ā-stem Nouns and Pronouns, pp. 132—135.

²⁷ Plg. Fraenkel E. Litauisches etymologisches Wörterbuch. Heidelberg—Göttingen, 1962, S. 935—936 ir ten nurodytą literatūrą.

234

‘in Riga’, ma.kýkla.j, ža.lėj, paciōj, pirmōj, ailōj, tōkia.j, forma.j, balōj, liga.nija.j ‘in the hospital’; bul’bōs ‘among the potatoes’, dideliuōs, rugiuōs, dvejuōs, kve.ciuōs ‘among the wheat’, miškuōs, pabaliuōs, paduškōs (Slavic) ‘among the pillows’, plaukuōs || For cases in which j also disappears after *-ėn, see p. 230.

In the ending of the I-stem inflected forms, *-ān has developed into uo before j and into o before s, as in the neighboring Eastern Aukštaitian dialects²⁸, e.g., su pirmuoju dukteriu; su senóju kumelbi; unksi.vōsias bū.l’bas pasad.dina.m; pirmōsias ōgas vālga.m ||

In the dialect, the stressed *-ān of the archaic acute accent in certain singular instrumental and plural accusative forms of pronouns has yielded as many as three different reflexes—ūo, -ū, and -ù, e.g., mōma (the queen bee) tuos bitis kamandava.ja; lepšės vērda su skūru, juos nėlupa; žarnas jįskuta, pripucia ir juos atvercia (= išverčia); nėmumes (kelmučius, tokius grybus) ne rāuma, a juos nupjauna; višta pėri kókius tris savaites, kai pa. jū pãdeda pinčiolika šešiolika kiaušinių. ||

In our view, the presented pronoun-form reflex -ū in place of *-ān is entirely compatible in origin and form with forms found in neighboring Eastern Aukštaitian dialects (Zarasai and its environs), such as the inflected adjectives baltūju ~ baltūja, gerūsias ~ gerūsias²⁹. In our view, the reflex -ū could have arisen only from -ūo, most likely through word-final reduction associated with the retraction of some of the stress onto the preceding long syllable (cf. jūos, tūos and kókius). Where there were no conditions for word-final -ūo to shorten, it remained intact (jūos, tūos).

The presence of different *-ān reflexes in pronoun endings is undoubtedly explicable by the different position of an in those pronoun forms. Where the pronoun form was monosyllabic and no other sound followed an, the an sequence (consonant n) would have disappeared most slowly, lengthening and narrowing the first sequence—which had developed into u—for the longest time.

e) In the dialect’s prefix per-, the first element of the acute-stressed diphthong er has narrowed in all cases and is pronounced as ē, e.g., pēršalb.s ‘having caught a cold’, pērnėst, pēršiukta, pērtrauki.s, pērvardina ‘gives another name’, pērinčiu. mētu ‘of last year’, pērvede.m, pērkalbi ‘talks over’, pērdave., pēreite., pērverte. ‘translated’, pērpjauna, pērvažiuoja, pērgi.venimu ‘experience’. The same ē is pronounced in the loanword fērma.s ‘farms’ (probably because of the positional similarity of p and f in the dialect). The first element of the circumflex-stressed diphthong er is likewise narrow in the words cēntras, nevėtas, vēršim. Cases of narrowing in acute-stressed er are also characteristic of neighboring Lithuanian dialects in Lithuania³⁰.

It has been observed that the first element of the diphthong er (especially when unstressed) has often developed into ie here as a result of considerable narrowing, and the diphthong itself has become a triphthong, e.g., Kiergalvỹs ‘Kergalvys’ (surname), Kiērgėlis ‘Kergelis’ (surname), miergaičių ‘of the girls’, Vierdāniene ‘Verdanienė’ (surname). In isolated cases, the narrow long ē developing into ie is also known in the neighboring Latvian dialect of Aknīste³¹.

²⁸ Cf. Grinaveckis V. On the origin of the correspondences of *-ān in the endings of Lithuanian *ā-stem nouns and pronouns, pp. 132–135.

²⁹ Cf. Grinaveckis V. On the origin of the correspondences of *-ān in the endings of Lithuanian *ā-stem nouns and pronouns, pp. 126–127, and the literature cited there.

³⁰ Cf. Lithuanian Dialects (Chrestomathy), p. 320.

³¹ See Ancītis K. Aknystės izlokšne, pp. 66–67.

235

4. Cases of vowel omission

a) When vowels or diphthongs belonging to different morphemes come together, one of them may be omitted, e.g., (the second) nepsimoka < neapsimoka; net išs aukė. < neatsišaukė., netvaikysi < neatvaikysi ‘you will not drive away’; (the first) neįeisi < neišs ieisi ‘you will not manage’, priseina < prisieina ‘one has to’, pasilks < pasilgs, but suáugi.s ‘grown-up’. When two identical vowels occur at a morpheme boundary, or when the first morpheme contains a or u and the second i, both vowels remain, and j is inserted between them (see p. 230).

b) At a morpheme boundary, the vowel -i of the unstressed reflexive particle may also usually be omitted, e.g., paskalbėte. ‘to talk’, pasklausytum ‘you would listen’, pazdúodat ‘you give in’, but unsispyre.: insidaužiau ||

IV. STRESS AND INTONATION

a) The dialect’s stress system is not entirely stable and is fairly innovative in many respects. In the speech of the oldest dialect speakers, the original position of stress is generally preserved with near-regularity, e.g., tvirtà, rankà, varlàs, dangčių, neprunti, primirštu, umkšti, kartùs, bjaurì, pinkčiùs, pinkčiàs, kalbù, minkšti, girdžiu, žiūriu, jauni. However, words with a short stressed ending preceded by a long syllable containing ē, o, ō, tu never remain unchanged, e.g., dievi, skiedà ‘splinter’, momà ‘mother’, miestùs, tvorù³², nėra, žmogùs, teliokùs, kietùs, ramunėlius ‘chamomiles’, kruopàs, viedrì ‘bucket’. Moreover, in such long pre-stressed syllables, a second, secondary retracted stress is usually heard quite clearly, or the impression is created that stress is somewhat divided between the two syllables, e.g., netiėsi, blýnūs, žiūriu, tõli, žmõna, gūšta, liēkū (cf. Latvian lieku) ‘I put’, tyli, nebedà³³. These examples would suggest that stress began to be retracted here before unstressed long vowels shortened and merged. In general, however, secondary retracted stress is not always clear or strong. Greater vocal force sometimes falls on the final syllable (especially in the endings of the 1st and 2nd person singular future tense—mirsiù, aisi). Despite this, partial stress retraction is sometimes observed in the dialect, particularly in the speech of younger speakers, with no stress at all left on the final short syllable, e.g., viščiõkus, agurkùs, grabėstùs, grinčia, mirtìs, širdis, pirštùs, vaikùs. The fact that stress in the dialect is sometimes truly unstable is also indicated, among other things, by cases in which words with a fixed stress pattern and an acute root sporadically shift stress to a short ending, e.g., žiedù ‘with a flower’, žiedùs, ‘flowers’, melžiù ‘I milk’³⁴. This is probably because partial stress retraction is beginning to occur in the dialect, and its stress system is becoming increasingly unstable under the influence of both neighboring Lithuanian dialects and Latvian Ik, as well as other languages. Among other things, sporadic stress retraction from a short ending to…

³² This phenomenon was not recorded in the Lithuanian dialect of Ciskoda described by K. Būga. The example he recorded there is tvārà (see Būga K. Rinktiniai raštai, vol. 3, p. 571).

³³ Similar stress patterns were previously recorded in neighboring Lithuanian dialects in Lithuania and the Byelorussian SSR; see Grinaveckienė E. Some Characteristic Features of the Apso Dialect, pp. 175—176 and other sources cited there.

³⁴ Cf. Grinaveckis V. New Data on the Stress Patterns of the Western Dzūkian Dialect. — Lietuvos TSR Mokslų Akademijos darbai. Series A, vol. 4(41), (1972), pp. 182—185.

236

a short preceding syllable, e.g., mažiūčius ‘very small’, galines ‘rear’, muses ‘flies’, neti.rūmus (cf. Latvian netīrumi ‘excrement’) ‘dirt, manure’, vidutinius ‘average; middle, from the innards’, kurziūkų ‘smoked sausages’. However, such cases are relatively rare.

Due to the influence of Latvian, a new tendency is observable in the dialect’s stress system: to stress the first syllable of a word (more precisely, the word stem), e.g., mėzgima knyga ‘knitting book’, gailėjau, giminė, įdoma, mėšlais, balandis, lāpus, rūgitas, nusiramintė, mūkina, rožytes, sliekai || However, this retraction of stress is neither systematic nor regular. As most examples show, words exhibiting it also retain the length of the stressed syllable characteristic of the original (Lithuanian) stress pattern. Although Latvian stress is still fairly rare here, parallel forms (mostly newer words), stressed in two ways according to the rules of the Lithuanian and Latvian stress systems, are nevertheless becoming more frequent in the dialect, e.g., baltai | baltai (adverb), ma.kýkliu. | mȯkykliu. ‘schools’, Va.ke.tija.j | Vōke.tija.j, pa.rūojas | pārūojas (the latter example shows a very recent Latvian stress pattern, which arose in the dialect only after unstressed ā had shortened; cf. Latvian pāroties) ‘they mate’, isima.kinau | isīmōkinau ‘I learned’, nereikia | nė-reikia, visā akis | visā akis, te.vaī | tēvai ||

Words of foreign origin entered the dialect with their own stress, and they are generally stressed that way, e.g., akōpus ‘trenches’, du klasus ‘two classes’, rasōrna celiega ‘a spring wagon’, vantrōbus ‘intestines’, sōtkas ‘acres’, živatnavōdstva ‘livestock’. This also shows that foreign words flooded into the dialect relatively recently. However, the stress of the dialect’s borrowing punčiākas ‘socks’ indicates that the word was borrowed long ago and has adapted to Lithuanian stress.

b) In the dialect, pitch accents are distinguished relatively clearly. The second element of falling-accent diphthongs is very prominent here, while the first is usually darkened and somewhat reduced. This darkening is particularly evident in the diphthong au, whose first element is pronounced with even a slight o quality, e.g., šāukiam, dāūg, pašāūks. Sometimes one even gets the impression that falling-accent au at the end of an open word is pronounced with a certain falling contour, e.g., rukiāū ‘I smoked’, sakāū || The same applies to the falling-accent diphthong ai. Because of reduction, its first element is very similar to a (especially clearly so at the end of a word), e.g., sakāi | sakai, laukāi | laukāi || In open and closed word-final position, stressed ɔ < uo and au, it seems to our ears, are pronounced in the dialect with a fairly clear abrupt accent, e.g., šš ‘dog’, šītš ‘these’, plentōs ‘on the highways’; nežinō ‘I don’t know’.

The first element of acute-accent diphthongs i, u + l, m, n, r is generally lengthened to half-long, e.g., šū.lni. ‘well’, ki.lpa, ši.rdi., čū.lba, kū.rke.s ‘turkey hens’, suši.lda., pati.nsta || Diphthongs derived from acute-accent a, e + m, n seem to have merged in length with the dialect’s old acute-accent i, u + m, n, e.g., brū.ngi, kū.nda, pī.mpe. ||

Sometimes falling-accent -il- has changed to -iul- in the dialect, as in some other Eastern Aukštaitian dialects, e.g.: šiulta ‘warm’.

The dialect’s stressed half-long i and u, as well as a and e in prefixes, disyllabic infinitives and the tenses derived from them, comparative forms, and the pronouns mano, tavo, have a mid accent,

237

savo a, e, e.g., stipra. ‘strong’, i.pi.nimas ‘stinging nettle’, númeta, le.túviskai; pãmeta, kás-te., gerẽsnis, mãna.

V. CONCLUSIONS

1. The Lithuanian dialect of Daugavpils District in the Latvian SSR, in its phonetic (and other) features, is closely connected with the entire Lithuanian dialect area and is its continuation.

2. The dialect belongs to the eastern Dzūkian žalininkai, but also has features characteristic of the žadininkai.

3. While sharing features with the dialect of Lithuania’s eastern Dzūkija, it also has quite a few phonetic features peculiar to it alone (inconsistent Dzūkian pronunciation according to both features, inconsistent hardening of certain consonants, the diphthong uo becoming o in certain cases, manifestations of the cut accent, retraction of part of the stress force from a short ending to the preceding long syllable, manifestations of Latvian stress patterns, etc.). It also has features that distinguish it from the Lithuanian dialect of Ciskodas studied by K. Būga in 1911 (Rēzekne District, Latvian SSR), and the facts about the dialect under study presented here somewhat refine K. Būga’s comparative data on it.

4. The dialect shows a comparatively strong influence from Lithuanian and Latvian standard languages and from the dialects of the surrounding Slavic languages (Belarusian, Polish, and Russian).

5. The dialect is clearly declining.

6. In the future, the entire dialect system should be examined in greater detail, since even the data on it published here are, in our view, topical and important for historical and contemporary Lithuanian dialectology, phonetics, lexicography, and the study of contacts between Lithuanian and Latvian, as well as between the Baltic and Slavic languages.

Э. ПРИНАБЕЦКЕНЕ

PHONETICS OF THE LITHUANIAN DIALECT OF THE ENVIRONS OF DAUGAVPILS (Latvian SSR)

Abstract

We studied the isolated Lithuanian dialect of the left bank of the Daugava River (Daugavpils District, Latvian SSR) during a dialectological expedition in 1977, one might say, for the first time. However, some comparison of its dialect data had also been made by the Lithuanian linguist K. Būga, who in 1911 surveyed the isolated Lithuanian dialect of the environs of Ciskoda (now Rēzekne District) in Latvia.

On the basis of the latest dialect data, the article establishes that, in its phonetic system, the dialect under study is closely connected with the area of Lithuanian dialects in the Lithuanian SSR and is a continuation of it. The dialect belongs to the dialect of the so-called Eastern Dzūkai–Žalininkai, but alongside this it also exhibits certain features characteristic of the so-called Žadininkai dialect.

³⁵ The author sincerely thanks Prof. Dr. M. Rudzīte of the Latvian State P. Stu<č>ka University, who kindly agreed to read this article and made valuable comments.

238

The dialect has many features peculiar to it alone. These include: a) somewhat irregular “dzūkavimas”—the pronunciation of the affricates с, dz before historically front vowels and before vowels advanced toward the front; b) the somewhat irregular hardening of certain consonants; c) the change of the diphthong uo into the monophthong ə in certain cases; d) traces of a shortened intonation; e) partial retraction of stress from a short ending to the long penultimate syllable; f) elements of Latvian stress; g) certain differences between its phonetics and the phonetics of the Lithuanian dialect of Ciskoda.

Using specific examples, the author of the article also establishes that the dialect exhibits strong interaction between the linguistic systems of both the Lithuanian and Latvian standard languages and those of the surrounding Slavic-language dialects—Belarusian, Russian, and Polish.

It is also established that the Lithuanian dialect under study is endangered.