Regresyvinė balsių asimiliacija kretingiškių šnektose

LIETUVIŲ LEKSIKOS IR TERMINOLOGIJOS PROBLEMOS LIETUVIŲ KALBOTYROS KLAUSIMAI. XXIX (1991) ZOFIJA BABICKIENĖ REGRESSIVE VOWEL ASSIMILATION IN THE KRETINGA SUBDIALECTS 1. Introductory remarks*. The law of vowel assimilation is a very important phenomenon in the Samogitian dialect: even the classification of the northern Samogitian subdialects—Kretinga and Telšiai—is based on it (Girdenis, Zinkevičius, 1966, 142; Zinkevičius, 1978, 112–114; Grinaveckis, 1973, 13, 142–145, Girdenis, 1981, 42, 50). Vowel assimilation has been studied more extensively only in the Telšiai and Varniai subdialects (Girdenis, 1962; Rokaitė, 1961; Vitkauskas, 1963, 1974; Skirmantas, Girdenis, 1972). Closer observations and experiments have shown that assimilation is also found in the speech varieties of Kuliai, Budriai, and Darbėnai, which belong to the northern Samogitian dounininkai subdialect of Kretinga. In the Kretinga subdialect, vowel assimilation is not as pronounced as in the Telšiai subdialect. In almost all positions where Telšiai speakers have i, u (e.g., singular nominative vė̃nės “nail,” mõsės→musis “fly”; plural accusative vinis “nails,” mùsis → musis “flies,” Mžk¹), speakers of the Kretinga subdialect pronounce closer ę and o variants; compare the following minimal pairs: singular nominative bẽtės → bitis “bee”; plural accusative bẽtis → bitis “bees”; present tense, singular, first person knę̃šǫ “I dig”: second person knęši “you dig”; plural nominative rõgė “rye”: plural genitive rùg’ūm “of rye”; present tense, singular, first person pl’ǫ.rpǫ→pliurpù “I babble”: second person pl’ǫ.rpi “you babble.” In the second examples of the pairs, vowels ę and o are closer and higher (for the transcription, see p. 24). The tendency toward assimilation is also readily noticeable when ę is lengthened in Kretinga speech varieties. Compare aš rė̃.šǫ → aš rišu “I will sew”: to rė̃.ši “you sew”; aš patę̃.ksǫ → aš patiksu “I will please”: to patę̃.ksi “you will please”; drė̃.skes “rags”: drį̃.skiṣ “rag”; žlė̃.be “cross-eyed woman”: žl’į̃.bi “cross-eyed”; spę̃.tres “squints”: sp’į̃.tris “squinting.” In this case, before a high vowel, a very close, tense vowel is pronounced, reminiscent of the standard language’s ė. In the speech varieties under study, this vowel is strongly diphthongized and quite similar to the diphthong ie. Comparing vowel assimilation in the Kretinga and Telšiai subdialects, we see that its essential difference is that in the Telšiai subdialect assimilation changes the phoneme, whereas in Kretinga speech varieties new variants of the same phoneme arise—closer e and o allophones. It seems that Telšiai speakers may have passed through such a stage as well (cf. Kazlauskas, 1969, 32), but later, for certain reasons, they replaced the intermediate sounds with clear-cut ones. If this assumption is correct, the phenomenon of vowel assimilation in Kretinga speech varieties must be older. * The author sincerely thanks Prof. A. Girdenis, who encouraged her to write this article and also provided many valuable critical comments, which she has endeavored to take into account. 1 Place-name abbreviations are the same as those used in the academic “Dictionary of the Lithuanian Language.” 15 No. | Vowel | Syllable volume | Position | Parameters | Pra F₁ 1 | ẹ | 6 | —C′ | x ± s | 345 ± 14 | | | | v | 4,14 | | | | ÷ | 330 ÷ 360 2 | ẹ | 11 | —C | x ± s | 400 ± 35 | | | | v | 8,45 | | | | ÷ | 380 ÷ 420 2. Research methodology. The recordings of the Kretinga subdialects were analyzed not only auditorily; the instrumental-spectrographic method was also used quite extensively2. This method was used to investigate the acoustic properties of individual vowels and diphthongs, especially the positional variants of the phonemes /ę/, /e/, /ọ/. For the first time, the following text was used: ātē·t drẽ.s̃kis õntā, ātē·t drẽ.s̃kēs õntā, tūoki drẽški atvès, tūokę drẽškę atvès, tūokę drẽškę laūko, aš tum drẽšk’u laūko…* The following text was selected for the second time: aš pẽlūs jeu alāus, aš patẽ·ks’o pūonams, ātē·t drẽ.s̃kis õntā, jeu nudrẽ.s̃ka Stāpuons, to palẽ·ksi vẽšūs, lāuk tuo plẽ.ke šētā, to pẽlis tuo alāus, ātē·t drẽškes õntā, to jeu pẽli šētā, jeu nuplẽ.ka Stāpuons, lāuk tuo žlẽ.bę šētā … The speakers3 read the texts into a tape recorder with a declarative intonation. In investigating texts from the Kuliai subdialect, one recording was made at Vilnius University’s Audiovisual Laboratory on a “Tembr” tape recorder (tape speed 190.5 mm/s). Another was made in an apartment in Kaunas, using an “Orbita” tape recorder (tape speed — 97.5 mm/s) and an MD-52-A directional microphone. Dynamic “Visible speech”-type spectrograms were filmed at Vilnius University’s Laboratory of Experimental Phonetics. A spectrograph designed at Kaunas Polytechnic Institute, KP1-69, was used. The spectrogram frequency range was 50—10000 Hz, tape speed 63.3 mm/s, analysis time 4 ms, and time marks 20 ms. The formant centers of vowels and diphthongs on the spectrograms were measured at the beginning, middle, and end of the microfilms using a “Mikrofilm” reader with a lens, di- 2 The author of this article investigated texts from her native Kuliai subdialect, which encompasses the villages of Blidakių, Reiskių, Pāalančio, Unkurių, and others, using this method; the Darbėnai subdialect was investigated by A. Kalninaitė. 3 Two men who knew their dialect well were selected as speakers. * For diphthong (broken) intonation (notation), see p. 24. 16 Table 1 džia | Middle | End F₂ | F₁ | F₂ | F₁ | F₂ 1830 ± 28 | 380 ± 28 | 1855 ± 29 | 400 ± 20 | 1855 ± 19 1,56 | 7,55 | -1,58 | 5,17 | 1,04 1800 ÷ 1860 | 350 ÷ 410 | 1820 ÷ 1825 | 380 ÷ 420 | 1835 ÷ 1875 1725 ± 60 | 445 ± 35 | 1780 ± 60 | 465 ± 35 | 1725 ± 50 3,54 | 7,95 | -3,2 | 7,16 | 3 1685 ÷ 1765 | 420 ÷ 465 | 1745 ÷ 1815 | 445 ÷ 485 | 1690 ÷ 1760 with a tenfold magnification. In the spectrograms, duration was measured using time markers. During measurement, the serial number of the formant’s middle filter or of two adjacent filters was recorded, and was later converted into hertz (Hz) using a special table. In this way, more than 200 spectrograms were measured and examined. After all measurements had been made, the following statistical parameters were calculated: arithmetic mean (x̄), variance (S²), standard deviation (s), 95% confidence interval (÷), and coefficient of variation (v). The significance of the results was assessed using Student’s criterion (Урбах, 1975). All parameters and criteria were calculated according to the formulas given in the book (Урбах, 1964, 131—133, 166—171). 3. Results of the spectrographic study 3.1. The vowel e “i” Table 1 shows fairly clear assimilation of the vowel ę — singular nominative žlė.bis “žlibis”: singular genitive žlė.bę “žlibio”. Data for forms with k (drė.skis: drė.skę) were excluded from the calculations because it was observed that the palatalized consonants k’, š’ modify the quality of the vowel ę. The confidence intervals for the second formants do not overlap at the beginning, middle, or end. The difference between the intervals is greatest at the end. Table 2 also shows assimilation of the half-long stressed vowel ę “i”. The following examples are compared here: the vowel ę — singular nominative drė.skis “driskis” and ę plural nominative drė.skes “driskės”. We see that the greatest difference between the confidence intervals of the second formants occurs at the beginning—105 Hz; it is considerably smaller in the middle—65 Hz—and decreases further at the end. These statistically significant results also show fairly clear diphthongization of the vowel ę, which arose in the Kretinga subdialect because of consonant palatalization. The beginning of the vowel, which resembles i, is more readily compatible with that palatalization. 2. Lietuvių kalbotyros klausimai 29 t. 17 No. Vowel Syllable volume Position Parameters Pre F1 1 ę 6 — C’i x ± s 430 ± 19 v 4,5 ÷ 410 ÷ 450 2 ę 6 — C’e x ± s 435 ± 39 v 7,1 ÷ 400 ÷ 470 No. Vowel Syllable volume Position Parameters Beginning F1 1 ę 6 — C’i x ± s 360 ± 56 v 15,7 ÷ 300 ÷ 420 2 ę 9 — C’e x ± s 355 ± 53 v 14,75 ÷ 310 ÷ 400 3 ę 8 — Ca x ± s 400 ± 60 v 15 ÷ 350 ÷ 450 18 2 lentelė beginning Middle End F₂ F₁ F₂ F₁ F₂ 2250 ± 67 425 ± 18 2275 ± 38 445 ± 14 1900 ± 65 2,98 4,27 1,7 3,2 3,42 2180 ÷ 2320 405 ÷ 445 2235 ÷ 2315 430 ÷ 460 1920 ÷ 2060 2030 ± 42 430 ± 19 2095 ± 69 445 ÷ 27 1925 ± 44 2,0 4,5 3,29 6,2 2,3 1985 ÷ 2075 410 ÷ 450 2020 ÷ 2170 415 ÷ 475 1880 ÷ 1970 Table 3 beginning Middle End F₂ F₁ F₂ F₁ F₂ 1805 ± 45 355 ± 31 1875 ± 62 375 ± 33 1840 ± 88 2,49 8,7 3,3 8,96 8,4 1755 ÷ 1850 320 ÷ 390 1810 ÷ 1940 340 ÷ 410 1750 ÷ 1930 1790 ± 49 380 ± 55 1845 ± 52 395 ± 60 1815 ± 95 2,75 14,5 2,8 15,3 5,24 1750 ÷ 1830 330 ÷ 425 1800 ÷ 1890 345 ÷ 445 1735 ÷ 1895 1590 ± 65 470 ± 39 1675 ± 35 470 ± 63 1590 ± 48 4,14 8,4 2,12 13,5 3 1535 ÷ 1645 435 ÷ 500 1645 ÷ 1705 415 ÷ 525 1550 ÷ 1630 2* 19 In fact, the softness and hardness of consonants considerably modify the vowel spectrum. Before a soft consonant, a vowel acquires a higher timbre (Girdenis, Kubiliūtė-Kliukienė, 1982, 36). The difference between an assimilated and an unassimilated vowel is especially noticeable when the latter is followed by a hard consonant, cf. žļė.bis “žlibis”: apžļė.bęs “apžlibęs”: apžļė.ba “apžlibo”, drė.s kis “driskis”: drė.skęs “driskęs”: nudrė.ska “nudrisko”. In the third examples of the pairs, the vowel ę strongly resembles Russian э or the eastern dialectal ė used after a hardened l. Even consonant pronunciation differs. When the ending contains a high vowel, the consonant before ę is clearly soft. The consonant softens only very slightly before ę when it precedes vowels of the ę type. And the consonant is completely hard before ę when it precedes a hard consonant, cf. spē.tris “spitris”, spē.tres “spitrės”: spę.tra “spitro”. Table 3 presents the results of calculations obtained by measuring the spectrum of the half-long vowel ę “i” in words such as drė.skis “driskis” and trė.dis “tridis”; in the second set, drė.skę “driskio” and trė.dę “tridžio”; in the third, nudrė.ska “nudrisko” and sotre.da “sutrido”. Statistically significant results were obtained by comparing the confidence intervals of the second formants in the first (1755 Hz, 1810 Hz, 1750 Hz) and third (1535 Hz, 1645 Hz, 1550 Hz) rows, or in the second (1750 Hz, 1800 Hz, 1735 Hz) and third rows. The assimilation of the vowel ę “i” in the Kūliai dialect is clearly shown in the graph, where the vertical axis marks the frequency of the spectrograms (Hz), and the horizontal axis marks the formant-center points—at the beginning, middle, and end. Data from several trials were used for the graph, with their mean calculated. Confidence intervals for curve A: beginning — 1962±47; middle — 2002±43; end — 1865±57; for curve B — beginning — 1848±50; middle — 1906±60; end — 1822±64; for curve C — beginning — 1590±65; middle — 1675±35; end — 1590±48. Assimilation of the vowel ę “i”. A — ę — C’i; B — ę — C’e; C — ę — Ca Assimilation of the vowel ę “i” is also found in the Darbėnai dialect. Particularly significant results of the spectrographic study were obtained by comparing the vowel ę “i” in positions before front and back vowels. The confidence intervals of the second formant for the vowel ę “i” before a front vowel are as follows: at the beginning — 2540 Hz, in the middle — 2300 Hz, at the end — 2250 Hz; before a back vowel, considerably lower results were obtained: at the beginning — 2170 Hz, in the middle — 1870 Hz, and at the end — 1970 Hz (Kalniūtė, 69). 20 The patterns of assimilation of the vowel ę “i” described here can be seen in the inflection of the words: rė.štę “rišti”, bętęs ~ bitis “bitė”, trę.dis “tridis”. 1. Present tense Future tense 1a. Kūliai Darbėnai Kūliai Darbėnai 1a. rėšǫ rešo re.šo rė.šu 2a. rėši riši rė.ši rė.ši 3a. rėš reš reš reš 2.⁴ Singular Plural V. bętęs trę.dis bįę.tis trędē K. bętęis trę.de bęt’ū’m trįę.d’u N. bętęi tręđ’ōu bętįms trędēms G. b’įę.ti trįę.di bętis tręd’ūs Įn. bętę tręd’ō bętęmis trędēs Vt. bętīē trędīe bętīēs tręd’ūs Assimilation of the diphthong ęi “ie” is especially clear in syllables that are unstressed in the standard language but sometimes receive retracted stress in the dialects under study, e.g.: kvęitis “kvietys”: sǫ kvęit’ǫ; rėibībę “riebybė”: rėibōs “riebius”; stęipis ~ stie pies “stiebiesi”: stęibōus ~ stiebos “stiebuosi”. Assimilation of the diphthong ęi “ie” is facilitated by the influence of the second component: after all, even in words such as kęimę “kieme” and kęim’ā “kiemai”, the vowel ę is already narrower than, for example, in kę.rmes ~ kirmis “kirmėlė” and pę.rmā “pirma”. 3. Singular Plural Kūliai Darbėnai Kūliai Darbėnai V. kvęitis kvītis kvęitē kvītē K. kvęitę kvęitę (kvīti) kvęit’ū’m kvīt’ū N. kvęit’ōu kvęit’ōu kvęitēms kvęitēms G. kvęiti kvęiti kvęit’us kvīt’us Įn. sǫ kvęit’ō sǫ kvęit’ō sǫ kvęitēs sǫ kvęitēs Vt. kvęitīē kvęitīē kvęit’ūs kvīt’ūs Variation in the allophones of the vowel ę “i” in the Kretinga dialects has both shared and some differing features. The Darbėnai dialect uses a short ī before i-type vocalism, e.g.: bītims “bitėms”, kiši “kiši”, riši “riši”. In the Kūliai dialect, a slightly more open ę-type allophone is pronounced. 3.2. The vowel ę “e” Vowel e assimilation has also been observed in the Kretinga subdialect. Let us compare the following pairs in the Kuliai dialect: sg. acc. s’ēni “old man,” ēžī “hedgehog”: sēnę “old woman,” ēžę “female hedgehog”; sg. nom. gerē.snis “better”: sg. gen. gerē.snę “of a better one”; sg. acc. vaik’ēli ~ 4These are paradigms of the Kuliai dialect. 21 No. | Vowel | Syllable volume | Position | Parameters | Start | F₁ 1 | i̇e | 10 | – C′i | x ± s | 400 ± 28 | | | | v | 7.1 | | | | – | 380 ÷ 420 2 | e | 10 | – Cọ | x ± s | 430 ± 40 | | | | v | 9 | | | | ÷ | 400 ÷ 460 little boy“boy,” sg. gen. little girl “of the little child.” Before high vowels, a diphthong-like vowel is pronounced, strongly resembling the diphthong ie in Standard Lithuanian. This is also confirmed by the results of the spectrographic study of vowel e, presented in Table 4. Vowel e was examined in the following pairs: mi̇e.di “tree” (acc.): mėde “tree” (gen.); pi̇ė.ti “shoulder” (acc.): pėte “shoulder” (gen.); vaikė.li “little boy” (acc.): vaikelė “little child” (gen.); lapiė.li “leaf” (acc.): la-pėlė “leaf” (gen.). The largest difference between the confidence intervals for the second formants of assimilated and unassimilated e occurs at the beginning of the vowel’s articulation: 1760 Hz: 1615 Hz; the difference is negligible at the end of the vowel’s articulation. A. Kalniūtė’s data on vowel e assimilation in the Darbėnai dialect are quite significant. The confidence intervals for the second formants of assimilated vowel e are: at the beginning—2330 Hz; in the middle—2270 Hz; at the end—2110 Hz; for the unassimilated vowel: at the beginning—2140 Hz; in the middle—1990 Hz; at the end—1910 Hz (Kalniūtė, 78). Significant statistical results show that in the Darbėnai dialect the formants of vowel e before vowels of the i and e types lie at different levels. In the Kretinga dialects, vowel e, like ę, has several distinct variants: before a hard consonant it is very open ɛ, e.g., fut. 1st sg. 1a. ne.šo ~ nėšu “I will carry,” nume.so ~ numėsu “I will throw down.” Before a consonant preceding a vowel of the ę type, e is pronounced, resembling Standard Lithuanian e, e.g., numė.tės “having thrown down,” nutė.pės “having smeared.” A close ę, resembling a short international ę (nėši “you carry,” gėri “good,” kėpi “you bake”) and a clearly diphthongized i̇e (numė.si “you will throw down,” nuti̇e.psi “you will smear”) are used when the ending contains a high vowel. Long ę (grėžo ~ grėžų “I will drill,” grė.šo ~ grėšu “I will drill”) differs from the close, tense ę occurring before i-type vocalism (grėži “you drill,” grėši “you will drill”). The dialects under study exhibit a hierarchy of allophones of assimilated vowel e: i̇e, ę, ę, etc. 22 Table 4 tion | Middle | End F₂ | F₁ | F₂ | F₁ | F₂ 1810 ± 65 | 495 ± 69 | 1750 ± 71 | 580 ± 69 | 1695 ± 80 3.6 | 14 | 4 | 12 | 4.7 1760 ÷ 1860 | 440 ÷ 545 | 1695 ÷ 1805 | 530 ÷ 630 | 1635 ÷ 1755 1700 ± 116 | 545 ± 58 | 1690 ± 72 | 605 ± 71 | 1665 ± 53 6.86 | 10.75 | 4.26 | 11.8 | 3.2 1615 ÷ 1785 | 500 ÷ 585 | 1640 ÷ 1740 | 555 ÷ 655 | 1625 ÷ 1705 Consonant hardness and softness are distributed similarly: before vowels of the ε type, the consonant is completely hard; before e, it is slightly palatalized; and before ę or ‘e-type vowels, it is completely soft. 3.3. The vowels o and u Observations confirm that u-type vowels are also assimilated in the Kretinga dialects. The diphthongs oi < ui, ou < uo, used in an unstressed syllable that has received stress retracted from a stressed ending, are most affected by assimilation. This is confirmed by paradigms 4 and 5: kộudis ~ Kuodys, trộikis ~ Truikys (surnames). 4. Singular Plural Kulių Darbenų Kulių Darbenų Nom. kộudãs kộudãs kộudė kộudė Gen. kộudę kộudę kộud’ù’m kộud’ū Dat. kộud’ộu kộud’ộu kộudėms kộudėms Acc. kộudi kộudi kộud’ūs kộud’ūs Instr. sộ_kộud’ộ sộ_kộud’ū sộ_kộudė·s sộ_kộudės Voc. kộudi kộudi 5. Singular Plural Kulių Darbenų Kulių Darbenų Nom. trộikis trộikīs trộikė trộikė Gen. trộike trộike trộik’ū’m trộik’ū Dat. trộik’ộu trộik’ộu trộikėms trộikėms 23 Gen. trọikì trūikì trọik’ùs trūik’ùs Instr. sò_trọik’ò sò_trọik’ù sò_trọikēs sò_trọikēs Voc. trọikì trūikì trọikē trūikē The results of the spectrographic study in the Darbėnai dialect showed an unmistakable tendency toward assimilation of the u-type vowel. For example, when examining an o-type vowel occurring before high vowels, the following confidence intervals for the second formants were obtained: at the beginning — 1400 Hz, in the middle — 1310 Hz, at the end — 1370 Hz; in the Ce position, the results were: at the beginning — 1390 Hz, in the middle — 1250 Hz, at the end — 1240 Hz (Kalniūtė, 81). Assimilation of the vowel u is also observed in the following word pairs: jōukà “jokes”: jōukū·m “of jokes”; pōrvà “mud”: pō·rvū·m “of mud”; kōtà “tassels”: kōtū·m “of tassels”; jōudà “black”: jōudībė “blackness”; trōpō “little”: trōpiēs “little ones,” and so on. Assimilation is slight when the vowel is lengthened under the ancient accent, e.g., skọ·si “you will peel”: skọ·sęs “you will peel yourself”; ọ·tis ~ utys “fleas”: ọ·tęs; krọ·pis → krupis “does not grow”: krọ·pęs ~ krupės “toads.” The difference between an assimilated and an unassimilated vowel is very slight when the ending contains a short stressed high vowel, e.g., singular nominative ọ·tęs ~ utis: plural genitive ọ·tis ~ utis “louse.” 4. Conclusions 1. In the Kretinga subdialects, nondiffuse e and e-type vowels have the following distinct variants: a) Before high vowels and palatalized consonants, these vowels are very close and diphthongal in nature — ẹ, īe. b) Before non-palatalized consonants, much more open e-type vowels are pronounced, resembling Russian э. c) Before softened consonants and in other positions (e.g., before e and e), ẹ- and e-type vowels are pronounced. 2. Assimilation is also characteristic of u-type vowels and the diphthongs that contain them. 3. The conclusions of auditory observations are confirmed by a spectrographic experiment. In the first (a) case, vowels ẹ and e have a much higher second formant than in the third (c) case, and especially than in the second (b) case. Notes. Because of a shortage of characters, in some places assimilated e- and o-type vowels are marked only as e, o, instead of the narrower ones (with two dots below). *The printing house marked the falling accent of the diphthongs au, ou, ie, uo not on both components, but only on the first one LITERATŪRA Girdenis A. Balsių asimiliacijos reiškiniai Tirkšlių tarmėje || Kalbotyra, 1962. T. 4. P. 141—150. Girdenis A., Zinkevičius Z. Dėl lietuvių kalbos tarmių klasifikacijos || Kalbotyra, 1966. T. 14. P. 139—147. 24 Girdenis A. Hierarchinė šiaurės žemaičių tarmės klasifikacija || Baltistika, 1981. T. 17(1). P. 42—51. Girdenis A., Kubiliūtė-Kliukienė R. Regresyvinis priebal­sių palatalizacijos poveikis balsių spektrui šiaurės žemaičių tarmėje || Kalbotyra, 1982. T. 33(1). P. 30—37. Grinaveckis V. Žemaičių tarmių istorija (fonetika). V., 1973. Kalniūtė A. Darbėnų šnektos prozodija ir vokalizmas. Diplominis darbas. V., 1978. Kazlauskas J. Dėl baltų *ei ir *o lietuvių žemaičių tarmėse || Baltistika, 1969. T. 5(1). P. 29—35. Rokaitė B. Kai kurie nauji žemaičių tarmių fonetikos dalykai || Lietuvių kalbotyros klausimai, 1961. T. 4. P. 149—153. Skirmantas P., Girdenis A. Progresyvinė balsių asimiliacija pietinėse „dounininkų“ šnektose || Kalbotyra, 1972. T. 2(4). P. 91—96. Vitkauskas V. Dėl regresyvinės balsių asimiliacijos kai kuriose šiaurės rytų žemaičių dūnininkų tarmėse || Lietuvių kalbotyros klausimai, 1963. T. 6. P. 269—273. Vitkauskas V. Nauji regresyvinės balsių asimiliacijos veikimo faktai Kuršėnų šnektoje || Kalbotyra, 1974. T. 26(1). P. 81—85. Zinkevičius Z. Lietuvių kalbos dialektologija. V., 1978. Урбах Б. Ю. Биометрические методы. М., 1964. Урбах Б. Ю. Статистический анализ в биологических и медицинских исследованиях. М., 1975. 3. БАБИЦКЕНЕ РЕГРЕССИВНАЯ АССИМИЛЯЦИЯ ГЛАСНЫХ В КРЕТИНГСКИХ ГОВОРАХ Резюме В статье исследуется регрессивная ассимиляция (гармония) гласных в кретингских говорах. В результате субъективных и объективных методов фонетического исследования было установлено, что перед гласным верхнего подъема произносится очень узкий напряженный гласный, напоминающий i литературного языка. Это различие особенно ярко выражено в позиции перед твердым согласным, ср. ž̇i̇.bis ~ ži̇bis букв. „слепец“: apž̇lē.bes ~ apž̇lēbęs букв. „ослепнувший“; apž̇lē.ba ~ apž̇libo букв. „ослеп“; drė̇.skis ~ driskis букв. „босая“; drė̇.skes ~ driskes букв. „босики“; nudrė̇.ska ~ nudrisko букв. „раздиралась“. В словах apž̇lē.ba, nudrė̇.ska гласный ẹ в кретингских говорах напоминает русский э, или гласный, употребляемый в восточных говорах литовского языка после твердого l. Выявлено, что в кретингских говорах недиффузные гласные типа ẹ и e имеют три четко выраженные варианта: а) перед гласными верхнего подъема и мягкими согласными эти гласные произносятся закрыто и несколько дифтонгизируются; б) перед твердыми согласными встречаются более задние варианты, несколько напоминающие русское э; в) перед смягченными согласными, за которыми следуют гласные типа ẹ, e, произносятся ẹ, e. Ассимиляции подвергаются и гласные типа ø, а также дифтонги, первым компонентом которых являются исследованные гласные. Результаты аудитивного наблюдения подтверждаются данными спектрографического анализа. В подавляющем большинстве случаев в позиции (а) вторая форманта F₂ гласных ẹ и e находится в более высокой зоне частот, чем в позиции (в) и в особенности — (б). Если сравнить ассимиляцию гласных в кретингских и тельшяйских говорах, то основное их различие заключается в том, что в тельшяйских говорах в результате ассимиляции изменяется фонема, а в кретингских появляются новые варианты той же самой фонемы. 25