Panevėžiškių fonetikos raida abipus Latvijos sienos

ACTA LINGUISTICA LITHUANICA XLVII (2002), 9–18 The Development of Panevėžiškiai Phonetics on Both Sides of the Latvian Border* KAZIMIERAS GARŠVA Institute of the Lithuanian Language, Vilnius The article attempts to reconstruct the main features of the phonetic development of the Panevėžiškiai dialect on the basis of place names and personal names recorded between the 16th and 18th centuries. Attention is paid to the narrowing of long stressed /a:/; the treatment of nasal vowels and diphthongs; the broadening of unstressed /i/, /u/; stress retraction; depalatalisation of /l/ before e-type vowels; lengthening of short stressed vowels; shortening of long unstressed vowels; monophthongisation of /ie/, /uo/; the change of fronted unstressed /a/, /u/ to /i/; reduction of /i/, /u/, /e/, /a/ in final syllables; and vowel insertion in clusters ending in /l/, /r/, /m/, /n/, /v/. 0. INTRODUCTION In Latvia and northern Lithuania, between Klaipėda and Biržai, stress is generally retracted and short word-final vowels are reduced. Along the Latvian border between Pāvilis and Breslauja, stress is generally not retracted, so this phenomenon may have been caused not by the influence of Latvian, but by Finnic influence transmitted through the Curonian and Semigallian languages. The Curonians and Semigallians inhabited this area until the 16th century. For approximately 200 years, Lithuanians who had migrated from the south also lived alongside them, adopting stress retraction, ending reduction, and certain other dialect features from the local inhabitants. The Lithuanian Semigallia region includes the North Panevėžiškiai dialect and the northern parts of the Šiauliai and Telšiai dialects. Dialect isoglosses show that there was an intensive expansion of the Central Aukštaitian dialect between Šiauliai and Žagarė, which formed a wedge separating the Panevėžiškiai and Telšiai dialects. The old Samogitian–Aukštaitian boundary may have run near Lygumai. Parish archives indicate that slight stress retraction and very rare ending reduction in the Panevėžiškiai dialect may have begun in the 16th century, when Semigallian influence was already limited, and intensified particularly in the 17th century (stress retraction) and the 19th century (ending reduction), when the Semigallians had been fully assimilated. Thus, Semigallian influence developed after their language had disappeared. The aim of the article is to attempt, on the basis of written sources, to determine approximately when the most important phonetic features of the author’s native North-Western Panevėžiškiai subdialect arose— * This article was prepared on the basis of papers delivered at the M. Daukša and J. Kazlauskas conferences at Vilnius University. The author thanks his Teachers, Prof. A. Girdenis and Academician Z. Zinkevičius, for their comments. 10 | K A Z I M I E R A S G A R Š V A …ma. This is a very complex task, because sources written in Polish and Latin do not accurately reflect the Lithuanian language. There are no scholarly articles on the development of the phonetics of the northwestern Panevėžys subdialects. The dialect lies in the western part of the Eastern Aukštaitian area and is transitional between the Samogitian and Aukštaitian dialects. There is no consensus on the development of the phonetics of these dialects. For example, in 1971 Prof. A. Girdenis (2000: 209) wrote that the Samogitian dialect had separated from the Aukštaitian dialect before the mid-15th century, whereas in 1998 he stated that this had happened no later than the 7th century (Girdenis, Pabrėža 2000: 7). A. Salys (1992: 364) placed this development 1,000 years later, in the late 16th–early 17th century. In preparing this article, about 3,000 place-name records and 1,000 surname records from the 16th–18th centuries were studied in the baptismal and marriage registers of the churches of Linkuva, Pakruojis, Klovainiai, Joniškis, Joniškėlis, Vaškai, and Pušalotas. The dialect texts recorded over the past 150 years provide a considerable basis for assumptions (see Garšva 2000: 173–174). Most of the data come from the 17th-century archive of Linkuva parish. The married couples recorded there came from a large part of the northwestern Panevėžys subdialect area (the Mūša river basin), from Pašvitinys to Pasvalys and from Lauksodis to Klovainiai; consequently, the archives attest to the speech of roughly 10 neighboring parishes. Their speech differs very little from one another. Across the Lithuanian border lay the parishes of Brūnava and Būdberga, as well as parts of the parishes of Žeimelis, Joniškis, and Nemunėlio Radviliškis. Historical church registers are the only written monument of the subdialect. Their entries are imprecise, since words were adapted to the articulation of Slavic languages. Priests usually did not know Lithuanian; thus, for example, the nominative singular was recorded as de Giefe ‘Degesiai’ (three times, Lnkv KĮ No. 1, 2ᵇ, 87ᵃᵇ, 1677, 1711; insignificant variants are not distinguished¹), de Wonogalie ‘from Žvanagala’ (83ᵃ), de Isznory ‘from Noriai’ (the Lithuanian prefix was written together, 1671, 92ᵇ). Most often, the Slavic prefix s was added to Lithuanian place-names (perhaps the scribe was dictated to by a “translator” who spoke Polish): de Skizerij 85ᵃ ‘from Kyzeriai’, de Smiciski 80ᵃ ‘from Mičiškiai’, de SNakiski 23ᵃ ‘from Nakiškiai’ (three times, 23ᵃ, 17ᵃ, 82ᵃ), de Stitanij 74ᵇ ‘from Titoniai’, d. Sturczyn 4ᵇ ‘from Turčinai’, z Żudek 98ᵇ ‘from Udeku’, de Szudaki 97ᵃ ‘from Udeku’, de Zwaldeyki 60ᵇ ‘from Valdeikiai’, and so on. 1. STRESSED VOWEL o < ā Many of the present-day villages in the Mūša region came into being in the 16th–17th centuries, and the northwestern Panevėžys subdialect took shape, along with its local varieties, including the central Linkuva variety. As early as the 16th century, the stressed Proto-Lithuanian long vowel *ā had narrowed to o in the dialect: Rimeikonis 1555, Puknioniu, Titoniu 1599 JNš (figures indicate the year of recording; the abbreviation indicates the parish) KĮ No. 1, 5ᵃ. In the 17th century, the vowel o was most often written in the place-name suffixes -oniai, -onys, -orai: Blekonie (5 times), Jawidony, Junelony, Juszkony, Gaylony, Kurniony, Mieguciony (6 times), Meilony, Labory (5 times), Mosczony, Puknioni (5 times), Rymszony (2 times), Titonie (30 times), Tryszkony (3 times), Wisgonie, Woronie (4 times), and so on. 1 Historical Place-Name Index of the Institute of the Lithuanian Language. Numbers indicate leaves (a – recto, b – verso). Place-names from the sources were collected by Rimas Garškauskas and the article’s author. A source leaf is cited when the source is first mentioned and is not repeated with other examples from the same source. The leaf and page numbers are omitted for some examples; in such cases, only the locality and year of recording are given. PANEVĖŽIŠKIŲ FONETIKOS RAIDA 11 Instead of the suffix -on, the suffix -un was often used, e.g.: Blakuny ‘Blekoniai’, Gailunie ‘Gailioniai’ (and Kapciony ‘Kapčiūnai’), Levszkuny ‘Laiškoniai’, Mieguny ‘Miegoniai’, Miegociuny ‘Megučioniai’, Narykuny ‘Nereikonai’ (and Planiuny ‘Plonėnai’), Ropuny ‘Roponiai’, Sterkuny ‘Starkoniai’, Tituny ‘Titoniai’, Titkuny ‘Titkoniai’, Tryszkuny ‘Triškoniai’, Wayniekunie ‘Vainekoniai’ (about 15 place names). Instead of the vowel o, e was sometimes written: a: Daunari ‘Daunoriai’, Jowidany ‘Javidoniai’, Labary ‘Laborai’, Mieguciany ‘Megučioniai’, Saban ‘Sabonių’, Pukniany ‘Puknioniai’, Ropanis ‘Ruponiai’, Tryszkany ‘Triškoniai’, Tytany ‘Titoniai’ (the vowel ė was sometimes written as i; cf. Butynie (3 times) 1676 ‘Butėnai’, Bininy ‘Binėnai’). In the oikonymic suffix -(i)ūnai, the vowel u was retained in 13 place-names or represented by o in the same place-names (10 forms): Buciuny / Buciony (3 times) ‘Bučiūnai’, Butniuni / Butniony ‘Butniūnai’, Gayzuny / Gaiziony (17 times) / (6 times) ‘Gaižūnai’, Gedziony ‘Gedžiūnai’, Kapciuny / Kapciony ‘Kapčiūnai’, Katkuny ‘Katkūnai’, Pikciuny / Pikcionie ‘Pikčiūnai’, Podziuny / Podziony ‘Puodžiūnai’, Tautkuny ‘Tautkūnai’, Ziniuny ‘Ziniūnai’. The two forms here may represent two different variants of the same place-name, with the suffixes -(i)ūnai and -(i)oniai. In some cases, the suffixes of one or another place-name may also have been confused. The parallel suffix form is unattested for some place-names only because they are mentioned just once or twice. The two variants of certain suffixes (and their confusion) in place-names may be supported, in part and indirectly, by mid-20th-century surnames in Linkuva, which, for example, were pronounced Gasiliūnas, Pečeliūnas but officially written Gasiulionis, Pečelionis. Authentic older forms of place-names are recorded in the inventories: Linkawa 1618 (Jnš KĮ 218ᵇ) (the form is still used in the local variety, officially Linkuva), Janiszkiey (= Janiškiai) 1669 ‘Joniškėlis’. Some place-names were used both with and without a newer suffix; compare: Balsiai (Balsie 1677) – Balsiškiai (Bolszyszkiu), Barysai (Barys) – Barysoniai (Borysany 1676), Kalpokai (Kolpaki 1677, 1712) – Kalpokiškiai (Kolpakiszki 1672, 1676), Lumbiai (Lumbie 1673) – Lumbeliai (Lumbe le 1679), Mašniai – Mašnelės (Mesznele 1676), Plikiai (gen. Plikiu 1675, 1678) – Plikėliai (Plikiele), Sidaugai (Sidaugie, Sydaukis 1677) – Sidaugiškis (Sydaugiszki s), Šikšniai (Szyksznie) – Šikšnioniai (Sixniany 1677), Tričiai (Trycze) – Tričiūnai (Tryciuny 1674), Turčiai (Turcze 1676) – present-day Turčinai, Žeim is – Žeimėlis. 2. DIPHTHONGS In this region, the diphthongs am, an, em, en began to narrow, apparently in the 16th century; the narrowing was most intense in the 17th century, and the process may have ended in the 18th century. Compare Montotovič, Minkunovič 1555 (LNKV), Jankunow 1599 (Jnš KĮ No. 1) and Jonk unow 1604 (53ᵃ), Krampiu 1612 / Krompiu 1600, 1617 / Krumpiu 1599 (Jnš KĮ, Pociūnai). Prof. Z. Zinkevičius also holds a similar view (1994: 51, 65; the isophone of narrowing in an-type sequences reached the Panevėžys dialect area from the east only six centuries later). In the Šiauliai dialect, the diphthongs am, an, em, en are preserved; therefore, in some sources these vowels are narrowed for other reasons (Polonization, influence from neighboring Lithuanian dialects, etc.). In the second half of the 17th century, more diphthongs had already undergone narrowing. For example, in the extinct oikonym *Sangailiškiai, an was preserved in 7 records, 12 | KAZIMIERAS GARŠVA while the narrowed form occurred in 8. In this case, on was written 5 times and un 3 times. The diphthong in the name Jomantonys was preserved twice and narrowed 7 times (on 4 times and un 3 times) in 1668–1771. In Zigimantiškiai, an was recorded as on and un three times each. Compare also Bancze 1666 and Boncze 1678 (Klov), Domontaytys ‘Domantaitis’ 1671 (LNKV KĮ No. 1, 29ᵇ), Łonguenaytys ‘Lengvinaitis’ (24ᵇ), Ponkaycia ‘Pankaitė’ (27ᵃ), Wapraÿtis ‘Vanpraitis’ (1677, 87ᵇ), Punkraytys ‘Pankraitis’ 1671 (24ᵃ), Szwumbarÿs ‘Švambarys’ (28ᵃ), Walunczunas ‘Valančiūnas’ (1ᵇ). In the Šiauliai and Panevėžys dialect areas, the nasalization of the vowel ą disappeared and the vowel narrowed during the 17th–18th centuries; compare Dranseikia 1612 / Draseikiow ‘Drąseikių’ 1616 (JnŠ), Szoszgale ‘Žasgalai’ 1673 (LNKV KĮ No. 1, 96ᵃ), Žaysynaytis 1672 (36ᵃ). By the end of the 18th century, the diphthongs an, am, and probably en and em as well had also narrowed in the Lygumai parish. This is indicated by the forms Jokontyczia, Monszaycie, Montjurgie, Narmontas, Pronckunas, Vismonty, and others, recorded in 1763–1781. That these diphthongs had narrowed by at least the mid-18th century is shown by forms in which the diphthongs had not yet developed: Jokans 1776, Preymans 1784. By that time, an had already narrowed further in some instances; compare Tinkunas 1765. However, the origin of the forms Andreszunas 1773, Antanaytis, Pranskunas 1782, Stankuns 1784, Walancius 1764, Walantyns 1780, Tamkayte 1782, Stambbras 1782 (and Stombrayte 1781 / Stumbraite 1780) is not entirely clear. Some of these forms may have been recorded in a non-dialectal or inaccurate way (Antanas, Walantynas / Walontinas were also written in 1860); Slavic names such as Anton may have influenced them, etc. Compare also Endre(a,y)szunas, Endreszuniene 1778, Wenslajunas 1772, and so on. As early as the 17th century, the second component i of the diphthong ei may have been omitted; compare Nareykany 1700, Narekany 1700, 1701 (JnŠK) ‘Nereikoniai’, Benoraicių 1609, 1611 / Beinoraicziu 1614, 1616 (JnŠ) ‘Beinoraičiai’. The diphthongs en, in were used in the 17th–18th centuries: Endresiunow 1605 (63ᵇ), Endruszkayciu 1601 (22ᵃ), Senkunu 1601 (JnŠ KĮ No. 1), Gendrotaycia 1672 (LNKV KĮ No. 1, 30ᵃ), Wenckunas 1671 (25ᵃ), Indrunas 1671 (28ᵃ), Swiadrozynitis 1671 (27ᵃ) ‘Šventażentis?’, Endrelunas, Endrysziunas, Endresiunas 1772. In the place name Viendžindžiai, en was recorded 9 times at the end of the 17th century (< ien), ien 7 times (the vowel i was probably added solely because of Polish spelling), and in 6 times. In the Joniškis parish (Šiauliai dialect), the sounds ie, uo may already have been rendered as e, o/u in the 16th century: Lepaičiu 1599, Leparow ‘Lieporų’ 1613, Szlestykow 1599 / Szlostiku 1603 / Słusti(k)ow 1613, 1616 / Szluostikiu 1604 ‘Šluostikių’. Proper names from the Panevėžys dialect with equivalents of ie, uo were recorded in the 17th century: Iudmisewicz, Dewaytys (but Awizienutis). In the Lygumai dialect, unstressed diphthongs ie and, apparently, uo may have been preserved or monophthongized: Motiejuns / Motejuns 1777–1783, Urniezicia / Urnezicia, Usienis / Usenis 1767–1781, Jodpetrayte ‘Juodpetraitė’, Jozenie ‘Juozienė’, Judicie 1776 ‘Juodyčia’ (the diphthong uo, represented by the letter o, may also have been preserved—stressed syllables were represented in the same way). 3. UNSTRESSED VOWELS i, u, o From the earliest sources of the 16th century onward, the unstressed vowels i, u were occasionally broadened not only in the Panevėžys subdialect area but also in the Šiauliai area: Dwilaičiu 1601 / Dwelaičiou 1600, 1606, 1611, Dwilia PANEVĖŽIŠKIŲ FONETIKOS RAIDA 13 / Dweliu 1616 Jnš, Szperakų 1615, Melwidžių 1613, 1616 (Satkūnai), Juszkayciu 1599, 1612 / Joszkaiciu 1610, cf. also: Noraiciu 1600 / Nurajczu 1613, 1615 (Jnš KĮ). In the 17th century these vowels were broadened perhaps somewhat less frequently than today. In the Linkuva parish records, the place name Girbutkai (< Gürbūčiai) is recorded 25 times as Girbutki and 16 times as G(i)erbutki; Zimbiškiai is recorded 30 times as Zymbiszki and only 8 times as Žembiszki; Diržiai, 20 times as Dyrze and only 5 times as Derzie. Other rare examples of broadening of vowel i include: Katenu 1675 (102ᵃ) ‘Katinų’, Lencow 1676 (73ᵇ) ‘Linkava’, Skierеле 1675, 1678 (22ᵇ, 60ᵃ) ‘Skireliai’, Wiesiejекow 1675 (101ᵃ) ‘Višeikių’, Typiony 1672 (32ᵇ) ‘Stipinai’ (LNKV), Wėldunij 1671 (KLOV KĮ No. 3, 48ᵇ) ‘Vildūnų’, Andrelunas 1677 (LNKV KĮ No. 1, 87ᵃ) ‘Andreliūnas’, Melenaytys 1671 (27ᵃ) ‘Mėlinaitis’, Žiegaytys (25ᵇ) ‘Žigaitis’, Boteles 1671 (92ᵇ) ‘Butelis’, Garnes 1671 (28ᵃ) ‘Garnys’, (but Pintolis (26ᵇ) ‘Pintulis’, etc.). The unstressed vowel u was broadened in a similar way: Gostanie – 1672 (95ᵇ) ‘Gustoniai’, Boywieniu 1676 (103ᵇ) ‘Buivėnų’, Trompanie 1672, 1678 (16ᵇ) ‘Trumpėnai’, Androtaycia ‘Andriulaičia’, Androtaytys (twice), Indrotaycia ‘Indriulaičia’, Bodreonis ‘Budrionis’, Dargozianas ‘Dargužūnas’, Janotaycia ‘Janulaičia’, Jesolunas ‘Jasiuliūnas’, Matosaytys ‘Matusaitis’, Plokaycia ‘Plūkaĩčia’, Plokaytys ‘Plūkaitis’²; cf. also: Puntokany 1701 / Pontakany 1699, 1700 (Jnšk 10ᵇ, 18ᵇ, 25ᵇ) ‘Pantakonai’. The books of the Lygumai parish show that by the end of the 18th century many features of the northwestern Panevėžys subdialect (and of the Lygumai and neighboring local dialects) had become established. The data were collected from books dating from 1763–1786. Short vowels i, u were broadened to ę, ǫ, for example: Miszkuns / Meszkuns 1777, 1780, Sireykie / Sereikie 1776–1785, Syreikiele / Sereikiele 1765–1783, Szikszcie / Szekscie 1764–1780 (the sound ę was represented by the letter e), Kusztayte / Kosztycia 1764–1780, Jasionaycie 1764 ‘Jasiūnaičia’, Jordaycie ‘Jurdaičiai’, Lukosziuns / Lukoszaytis / Lokoszaytis 1774, Mantjurgie / Montjorgie 1764, Jorgaycie 1764 ‘Jurgaičiai’, Gutjurgaite / Godpetriene 1764, 1777, Puczayte / Poczayte 1780–1783, Stumbrayte / Stombrayte 1780, etc. Not only pre-stress but also post-stress vowels were broadened, although they occur less often in speech; cf. Melvelene 1776 ‘Milvilienė’. The 1860 register of the Lauksodis parish displays the same features of the northwestern Panevėžys subdialect as the 18th-century books of the Lygumai parish. There are given names and surnames with broadened pre-stress (Tamolonis, Kunegunda) and post-stress (kunegs, Narbotas, Motiejės, Marcionis, Pintolis) vowels u, i. Sometimes the vowels in these positions were not broadened either (as today—Ludviks, Kleryks, Nikodims, Wilips). In the 16th–17th centuries, the pre-stress vowel o had perhaps already been somewhat shortened and occasionally narrowed; cf. Beinoraiciu / Benuraicia 1616 / 1599, 1613 (Jnš KĮ), Nurgielu 1600, Petrošunu 1600, 1673 / Petrušunu 1673. The name of the village Oniūnai is recorded 12 times with the old vowel (10 times with o and 2 times with a) and three times with u (twice more with i); Tolminiai is recorded 14 times with o and 3 times with u. In the name of the village Ropūnai, only o (9 times) and a (5 times) are written; cf. also: Dowidu / Duwidu 1679, Truszkuni 1711, etc., Jonaytys, Jonetaytys, Moteiunas, Powilaytys, Simonaytys 1671, Atkuciūnas ‘Atkočiūnas’, Grutaicia ‘Grotaičia’, Junaicia ‘Jonaičia’, Junaitis ‘Jonaitis’, Junitaytys, Tutoraycia ‘Totoračia’ 1671. 2 According to Prof. A. Girdenis, the more open vowels i, u may also have been written with the symbols e, o. 14 KAZIMIERAS GARŠVA In the 18th century, in Lygumai parish, unstressed o (and probably ė as well) was also retained or reduced to a closer vowel u (ė > i); cf. Juvanytė 1778 ‘Jovaišaitė’, Tamoszatyene / Tamaszayte 1782, Tomkenie 1779, Tumkickie 1765. By 1860, more examples of a narrowed pre-stress vowel o were already found in Lauksodis parish: Bajurionas, Jokubas / Jukubas, Lukusziuns, Lukuszewicze, Petruszu-nene. In the records of this period, there are particularly many fronted ū vowels that became i: Bėje-wiczis, Gelažis, Izidoris, Karalis, Lukoszis, Motejis / Motejes, Monkowiczis, Narkewiczes, Tamoszis, and so on. Many such forms are also found in A. Baranauskas’s texts recorded at the end of the 19th century. From the 18th century to the present, the development continues whereby fronted vowels ă, ū after a palatalized consonant become i; cf. Wiejas / Wiejis ‘Vėjas’ 1763–1782, Urnezis ‘Urniežius’. As today, in the 17th century fronted pre-stress u occasionally became i (Ciciuyka ‘ciuciuko’ 1674), but was usually retained (Pažiuraitytis, Stasiulautys, Stasiulonis, Žiukaytys 1671). In Žiemgalian examples, the vowels e and i were also occasionally inserted between consonants: Ianiskekiel ‘Joniškėlis’, Senureloy ‘Sniūrelį’ 1700 (Jnšk 17ᵇ, 19ᵇ, 27ᵇ), Kivedery 1671 ‘Kvedarai’, Posiwityn 1671 ‘Pašvitinys’ (for more on Žiemgalian anaptyxis, see Šlavas 1996: 11–12; Zinkevičius 1999: 87; Girdenis 2000 b: 268–271). 4. CONSONANT CORRESPONDENCES. STRESS In the 17th century, consonant l could be hardened before e-type vowels: Gialažiunas 1675 (62ᵇ) ‘Geležiūnas’, Sietaykanij 1674, 1677 (53ᵃ, 86ᵇ) ‘Šileikoniai’ (in the vicinity of Vaškai and Alksniupiai), Szylaykonis 1678 (8ᵇ) ‘Šileikonis’. In the second half of the 17th century, l began to harden in the clusters lbe, lde: in the place name Valdeikiai (written as Waldeykany eight times), l was marked as soft 20 times (Waldeyki) and as hard only twice (Waldaykie) (21ᵇ, 42ᵃ). Also cf. Greycinis 1677 (85ᵃ) (Graitinis today). The forms Dewiniargie (7 instances) and Dewinelgi (5 instances), Daunory (6 instances) and Dalnoris (2 instances), and Dalgaylicia 1677 have been recorded (Pautarakas 1671 would suggest rare confusion of sounds r–l and u–l, which persists to the present day). One or another form may also have been written this way by chance. In 18th-century oikonyms, the consonants k and č were also somewhat confused; cf. Mikiškių 1678 ‘Mičiškių’, Valdajcziu 1673 / Valdajkiu 1677 ‘Valdeikių’ (Lnkv KĮ, p. 20ᵃ, 42ᵃ). By the second half of the 17th century, stress may already have been retracted. This is indicated by the forms Typiony < Stipione ‘Stipinai’ 1672 (32ᵇ), Garnes < Gārnes ‘Garnys’ (28ᵃ), Siuxnaytis < Šiokšnaitis (cf. š’ọkšn ‘šiokšna’) ‘Šikšnaitis’ 1672 (35ᵇ), with long endings reduced (written as -iai with y or e, etc.). Short vowels i, u, e, a were then lengthened and stressed, while unstressed long vowels i·, u·, e·, o·, ė·, a· were shortened. 5. REDUKUOTI TRUMPIEJI GALŪNĖS BALSIAI On the basis of the reduction of short vowels in word endings, six Eastern Baltic areas can be distinguished: 1) the northern Panevėžys dialect speakers, who drop almost all short ending vowels; 2) Latvians, who retain the stem vowels of i- and u-stems, cf.: brālis, medus (northern Panevėžys dialect bró·l's, infixed form brá·l' ‘brother’, mèds ‘honey’); 3) the Biržai dialect speakers and the peripheral dialects of the Panevėžys area; 4) the northern Samogitians, who retain remnants of word endings: a) having stress on endings of... PANEVĖŽIŠKIŲ FONETIKOS RAIDA 15 … (cf. ak'ės ‘eye’), b) vowels shortened from long vowels (vārnᵃ ‘crow’), and c) vowels between 2–3 consonants if this would produce more difficult-to-pronounce sound clusters (bās'əs ‘barefoot’, kākl'əs ‘neck’); 5) Western (Central) Highlanders, who do not pronounce the remnants of short vowels, and 6) the Utena and Vilnius dialect speakers and others, who drop the short vowel in the infinitive, etc. (cf. dž'u·t' ‘to dry’). The recording in the 1600 Joniškis baptism register, Petrošiuns ‘Petrošiūnas’ (p. 10ᵇ), would indicate an old tendency to reduce short word-final vowels. The aforementioned Petrošiūnas is associated with the vanished village of Užtupai and Petrošiūnai, situated between Linkuva and Pakruojis. In the 17th-century archive of Linkuva parish, surnames have word endings, with the sole exception of the surname Gulbinis (1678, p. 17ᵃ). Entries of names and surnames in the 1860 records of Lauksodis church (Jonas / Jons, Motiejis / Motejis, etc.) suggest that the reduction of short vowels in closed endings intensified in the first half of the 19th century, but at that time it affected only about a third of endings (Garšva 2000: 166). In Pakruojis parish, about 30 km to the south, only 3 out of 200 proper names had a reduced vowel in 1858. “Murmured vowels” have been marked only for the northern Panevėžys dialect, although reduction of short word-final vowels is strongest in this dialect. The theory of “murmured vowels” rests on three arguments involving consonant clusters before final -s: 1) the consonant t before -s does not turn into the affricate c, cf.: bā·ts ‘shoe’ (but kā·t'es / kā·t's / kā·c's ‘cats’); 2) the voiced consonants b, d, g before -s do not become voiceless, cf.: acc. pl. bó·bs ‘women’; bā·ds ‘famine’, blō·gs ‘bad’ (but the whole paradigm helps preserve these consonants; cf. also: bó·b ‘woman’ – bó·bos, bā·d ‘famine’ – bā·da, blō·g ‘bad’ – blō·ga ‘bad’, dau·g ‘much’); 3) the consonants š, ž, s do not merge with -s into a single consonant of the same duration, cf.: lā·šs ‘drop’, lā·žs ‘corvée’, lā·ss ‘lesus’ (lā·s ‘pecks’, but here too these penultimate consonants are supported by other forms of the same word, cf.: gen. sg. lā·ša, lā·ža; lā·s'es ‘having pecked, having pecked’ etc.). One argument for this theory might also be that circumflex vowels do not form acute diphthongs with sonorants, cf.: kā·l ‘he/she/it cuts’ (and kā·lam ‘we cut’, kā·lns ‘mountain’, kāl / kal ‘of cuts’). When short word-final vowels are lost, Žemaitian voiced consonants before -s become voiceless (cf.: rāks ‘horn’, bāups ‘toad’), the clusters t + s, d + s become c (bāc ‘famine’), and secondary diphthongs arise (vēj / vēi ‘wind’). One cannot require the same to happen in other dialects (for example, the Panevėžys dialect), since the distribution of vowels i, u and other features also differ in these dialects (cf.: in the Telšiai dialect, aš lòpò ‘I peel’, tò lùpi ‘you peel’, òns lòp ‘he peels’; in the northern Panevėžys dialect, aš lòp, tò lòp', jòs lù.p). The two word-final features mentioned above are also characteristic of the Šiauliai dialect (cf.: drau̇gs ‘friend’, lāds ‘ice’; daržs ‘garden’, sau̇ss ‘dry’, kāu̇ss ‘ladle’), but “murmured vowels” are not proposed for it. This may also be explained by the preservation of unusual consonant clusters and consonant voicing, and by the opposition between short and long s. A scientific hypothesis must not omit data that contradict it. First, it is mainly in the Panevėžys dialect that, after palatalized consonants (often also after front vowels; cf. kā·r'vės ‘cows’, sidė·s ‘will sit’, naktī.m ‘for the nights’), the word-final -s' is soft. The softness of the final consonant -s' in the pairs ar·kels ‘plough’ – ar'kels ‘horse’, kā·lens ‘you will imprison’ – kā·lens ‘prisoner’ can also be explained by its position after a palatalized consonant. But, if the “murmured vowel” theory is accepted, one would have to explain the palatalization of -s as occurring after a front “murmured vowel” (or before it, since pronunciation does not allow us to determine— 16 | KAZIMIERAS GARŠVA where that “murmured vowel” should be). On the basis of pronunciation, one would then have to transcribe synchronically kā·len's'b ‘you will imprison’ – kā·len's'b ‘prisoner’ (soft final consonant), jaū-č'os'b ‘I feel’ – jaūč'os ‘I felt’, whereas diachronically (as is usually done) – kā·len's'b – kā·len'b's, jaūč'os'b – jaūč'os, and so on. After a palatalized consonant, in the words gāl' ‘end’, nòr's' ‘although’, we would likewise have to write a front “murmured vowel”, even though it is unclear how it could have “arisen” here. The hardening of the dialect’s final consonant -s' continues to this day. In several localities (Linksmūčiai, Šiaudiniai, Žeimelis), this consonant is still pronounced softer than throughout the rest of the dialect area; cf. the singular first-person forms of the subjunctive barč'es' / barč'os ‘I would scold’, third person bartòs' / bartus ‘he/she would scold’; first person of the past frequentative tense bār'os' / bār'os ‘I scolded’, and so on. Contrary to what is claimed, at the end of a word the consonants j, v become i, u, showing that no vowel follows them; cf. kó·j ‘leg’, kó·is' ‘legs’, vé·is' ‘wind’, ló·u ‘bed’, klā·us ‘maple’, tė́us ‘father’, l'ōn.kau ‘Linkava’ (the analogy of other forms requires that the consonants mentioned be retained here as well). At the end of a word, the vowel e (ə) (after a hard consonant) or e (after a soft consonant) is inserted into clusters of the consonants s, š, b, p, d, g, k and the sonorants l, r, m, n, v that are difficult to pronounce; cf.: gai·sers ‘fire’, vāps'ev ‘wasp’, kābel's' ‘hook’, bàsel's' ‘moth’, tr'ù·k-šems ‘noise’, dù·gens ‘bottom’, dògen' ‘at the bottom’, etc. If a syllabic “murmured vowel” had been present here, there would have been no need to insert a new vowel. In 1850, only about 30% of the short vowels had been reduced (completely eliminated) in the Lauksodis area; in 1930, about 50%; and in 2000, 99%. Minimal pairs sometimes demonstrate not phonological differences, but merely slight phonetic differences between them. Murmured vowels in Northern Panevėžys dialect disappeared across most of the dialect area (excluding the margins) over the last 300 years and should be regarded as a diachronic, rather than a synchronic, category. 6. CONCLUSIONS The article discusses the recorded forms of 3,000 place names and 1,000 surnames in the sixteenth–eighteenth-century registers of the churches of Linkuva, Klovainiai, Joniškis, Joniškėlis, and others. Texts written in Polish and Latin did not reflect the dialect with complete accuracy. Conclusions can therefore be drawn only approximately and cautiously. The stressed long vowel *ā may have narrowed to o in the Panevėžys dialect as early as the sixteenth century. The diphthongs am, an, em, en and the nasal vowels ą, ę apparently narrowed during the sixteenth–eighteenth centuries. The unstressed vowels i, u were already being widened in old seventeenth-century texts. The vowel o (and, apparently, ė) only began to be narrowed at that time; cf. Oniuni / Uniuni ‘Oniūnai’. At that time, the unstressed diphthongs ie, uo were monophthongized. The unstressed fronted ū was sometimes rendered as i, and the vowels e, i were occasionally inserted between consonants (cf.: Ianizsekiel ‘Joniškėlis’, Posiwityn ‘Pašvitinys’); the second component of the diphthong ei was omitted (Nareykany / Narekany ‘Nereikoniai’). The consonant l was probably hardened in the dialect in the seventeenth century. Stress in the Panevėžys dialect was retracted in the second half of the seventeenth century; cf.: Typiony < Stipione ‘Stipinai’, Garnes < Gañnes ‘Garnys’. The more detailed phonetic development of the narrower (central) dialect area—that of Linkuva and the neighboring subdialects—may have proceeded as follows: PANEVĖŽIŠKIŲ FONETIKOS RAIDA | 17 1) The narrowing of sounds of the a, e, an, en type around the sixteenth century; cf.: Montovovič, Minkunovič 1555 (based on church archives); 2) The narrowing of stressed long a· to o· by the sixteenth–seventeenth century; cf.: Rimeikonis 1555, Labaris / Laboris 1679; 3) Hardening of the consonant l before vowels of the e type and in the clusters lbe, lde around the seventeenth century: *lantà < lentà, Waldeyki 1677 > Valdaikė ‘Valdeikiai’; 4) Retraction of stress around the seventeenth century; cf. *lantà > *loǹtà > lońt(a); 5) Lengthening of the stressed short vowels i, u, e, a; 6) Shortening of the unstressed long vowels iː, uː, ēː, oː, eː, aː; 7) Widening of the vowel i to ę: Zembiški 1674, Gierbutki 1676, Pezenis 1765; 8) Monophthongization of unstressed ie, uo and narrowing of e, o (more frequent in the southern part of the subdialect area): Wiendzindzie / Wendzindzie 1677, Woytekunas 1738, Binkiawiczene 1793; Tolminij 1677, Tulminy 1711, Truszkany 1711; 9) Change of fronted unstressed ū, ā to i (cf. sęrb'ala ‘siurbėlė’, vėlnis ‘velnias’). The second stage in the development of the subdialects continued after the plague of 1708–1711 (also based on abundant subdialect texts): 1) qualitative reduction of short i, u, e, a in the final syllable in the 18th–20th centuries (in the personal names and surnames recorded in the Lauksodis parish in 1860, a third of the short vowels had already disappeared); 2) insertion of vowels ę, ę/ę into a consonant cluster before l, r, m, n, v in the 19th–20th centuries; 3) further lowering of the vowel ę (< i) to ọ before a hard consonant in the 19th–20th centuries. The dialect formed in the 16th–18th centuries and changed slightly thereafter. This is a hypothesis that will encourage further research3. There is insufficient data to fully substantiate the earlier changes. Some of the oldest features of the dialect may be explained by a Selonian language substrate. The Latvian (Courland) border had little influence on the traditional dialect. It merely encouraged language interaction. ABBREVIATIONS Jnš – Joniškis Jnšk – Joniškėlis KĮ – baptismal record book Klov – Klovainiai Lnkv – Linkuva LITERATŪRA GARŠVA, K. 2000: Šiaurės vakarų panevėžiškių „murmamieji balsiai“. Acta Linguistica Lithuanica 43, 165–175. GIRDENIS, A. 2000 a: Kalbotyros darbai 1, Vilnius: Mokslo ir enciklopedijų leidybos institutas. GIRDENIS, A. 2000 b: Kalbotyros darbai 2, Vilnius: Mokslo ir enciklopedijų leidybos institutas. 3 Profesoriaus A. Girdenio nuomone, šiame straipsnyje minimi kai kurie garsų pakitimai turėjo įvykti daug anksčiau. 18 KAZIMIERAS GARŠVA GIRDENIS, A., PABRĖŽA, J., 2000: Tarmių reikšmė. Žemaičių žemė 1, 6–7. SALYS, A. 1992: Raštai 4, Roma: Lietuvių katalikų mokslo akademija. ŠLIAVAS, J. 1996: Žiemgalių pėdsakais, Vilnius: „Žiemgalos“ draugija. ZINKEVIČIUS, Z. 1994: Lietuvių kalbos dialektologija, Vilnius: Mokslo ir enciklopedijų leidykla. ZINKEVIČIUS, Z. 1999: Žiemgaliai ir jų kalba. Žiemgala. Žiemgalos krašto praeitis 2, Kaunas, 83–88. Kazimieras Garšva Lietuvių kalbos institutas P. Vileišio g. 5, 2055 Vilnius, Lietuva kgar [email protected] Gauta 2002 10 01