Žemutinių prigimtinio ir padėtinio ilgumo balsių skirtumai Lukšių šnektoje

LITHUANIAN KALBOTYROS QUESTIONS XL (1998) Rima Institute of Lithuanian Language, Vilnius, Lithuania IR DIFFERENCES BETWEEN AUXILIARY LENGTH BALLS IN PIKES 1. In the linguistic literature we usually find the common opinion that the vowels [a', e] < *-an-, *-en and [a”, e] < *-a-, *-ė are not different. This applies both to the system of the common language, where “in accentuated syllables the vowels of historical and positional length [a-], [e] are pronounced equally” (Pakerys 1995: 33; also cf.: Vaitkevičiūtė 1960: 207, 217; Vaitkevičiūtė 1961: 31, 37; LKG 1: §17, §23; Mikalauskaitė 1976: 17), and to the closely related South-West Highland dialect, in which the “monophthongs a”, e* < 4, ę do not differ in length from the accentuated elongated a, ė” (Zinkevičius 1966: 80). However, in many dialects, not excluding the South-West Highlands, the opposition of natural (historical) and positional vowel lengths has been observed for quite some time. A. Baranauskas (Baranovskij 1898: 15) and especially K. Jaunius, who has clearly stated that the vowels elongated due to accentuation “in Kaunas district are of medium length (i.e. slightly longer than the short ones, but not equal to the long ones)” (Jaunius 1970: 83), have already spoken about the double number of low vowels a century ago. Notes defending the existence of the opposition of long and half-long vowels in Kaunas dialect are also found in later works (Girdenis 1971: 205; Zinkevičius 1980: 98; Pakerys 1986: 33), and some of them also mention the differences in quality: vowels of natural length may be more open and perhaps even slightly nasalized (Girdenis 1995: 172; also cf.: 1971: 205; 1981: 128). This type of vowel has been studied experimentally. The conclusions of B. Simanavičienė's (1993: 46-53) study of the Veliuoniečių schnekt concerning the dependence of [a-, e] < *-a-, *-ėprepositions on the tinted or muffled context" allow us to think about the quantitative differences of positional and natural vowels. The existence of quantitative opposition in Igliauka’s speech is proved by instrumental and auditory” investigation by A. Kazlauskienė (1996: 128-130). The results of this study indicate that positional long vowels have a true strong-end prefix only next to harsh consonants. In the postposition or preposition of a muffled consonant, the preposition is more similar to a strong preposition. * It is especially important that this difference is perceived by the dialect speakers themselves. Therefore, the words must differ in certain phonologically significant features. 102 In this thesis we tried to experimentally investigate the quantity and quality of low vowels ir In the talk. Ši spoken according to the current dialect ją classificationbelongs to the northern part of the western highland Kaunas (Zinkevičius 1994: 26), o according to the old K. Jauniaus — A. The division of Sali dialects this would be Western Highlands Veliuoniškiai, because they systematically shorten open unpronounced vowels [0", e:], e.g.: sa.ku, da.ri, vaiku, sd.uli and so on. Snekt Centre — Lukšiai Town (Šakių raj.), located 8 km į east of Šakiai ir 52 km į west of Kaunas. 2. The word pairs chosen for the research drill —kraštas, šėla —kala, patrėšęs —patrēšs, patrēšus —patrēšus, and with them to form short sentences, well illustrating the meanings of the investigated pairs. Most of the time, the investigated words were in the penultimate position: aštrūz gra-Staz birdavu || visas krd. 2000. pp. 10. "S.A.S.E.- R.S.E.R.S.- E.R.S.E.R.- S.E.R.S.E.- R.S.E.R.S.- E.R.S.E.R.- S.E.R.S.E.- R.S.E.R.S.- E.R.S.E.R.- S.E.R.S.E.- R.S.E.R.S.- ". It's a bit of a mess. 12|| smdrkei patré-Ses ta. pieva|| būvu gerai patrdėšus taz_braskdes || ta__o-belis patrdė.šus įeu būvu. The sentences were mixed in random order, between which were inserted sentences irrelevant to the subject of the study. The material was recorded on a tape recorder from the speaker A. Bacevičius, born in 1930, and the author of the text was put directly into the computer memory. Both speakers are representatives of the investigated schneck. During recordings, the informant, after listening to a pronounced sentence (including those irrelevant to the investigation), would repeat it freely, in as neutral an intonation as possible. It is not possible to avoid the unnatural speech intonation that occurs during reading." The sentences were repeated five times. Later, the recordings from the tape recorder were also transferred to the computer memory and processed with P. Kasparaitis’s special speech sound analysis program “Kalbam44”. The program displays an oscillogram of the input signal, according to which the signal can be split into separate sounds; the length of the selected audio segment is automatically measured by the program itself. The same program allows to perform spectral analysis of sounds: on the screen is visible dynamic signal spectrogram, change of forming trajectories and segment spectral section. 3. The data have been evaluated statistically. After length measurements, the arithmetic mean x, standard deviation s, showing how far the values of the test characteristic are from the mean, and the confidence interval of 9596 were calculated. In addition, the significance of the results was checked by the Student criterion. These calculations were performed by computer according to A. Girdenis program. 3.1. The results are presented in Tables 1-2. * Cf. I. Mažulienės (1996: 46-47) work methodology. 103 Table 1 Length of vowels in individual words Examples n x (ms) s (ms) 95% int drilled t edge 10 203 29 182 + 224 5,42 > 3,92 10 151 8 145 + 157 freeze 10 209 13 199 + 218 6,40 > 3,92 fish 10 173 12 165 + 182 patrésus 10 203 23 185 + 219 6,37 > 3,92 patrésus patrésus 10 223 17 210 + 235 9,56 > 3,92 patrésus 10 152 9 145 + 159 10 161 11 153 + 169 Table 2 Generalized Vowel Length Examples n x (ms) s (ms) 95% int. t [a] < *-an-20 206 22 195 + 216 7,29 > 3,57 [a] < *-a-20 162 15 155 + 169 [er] < *-and20 213 22 202 + 223 10,18 > 3,57 [er] < *-ė-20 156 11 150 + 160 longer 40 209 9 194 + 224 7,18 > 5,96 shorter 40 159 10 143 + 176 We see that there is a clearly ir significant difference in length between the vowels of historical padetic length: the dependency intervals do not intersect anywhere, do not overlap. The values of the Student criterion also indicate that this difference is significant because they are greater than the critical value of 99.996 in all cases. The quantitative difference is even clearer when we summarize the results (see table 2): the confidence intervals again do not cross anywhere, and the values of the Student criterion are much higher than the critical value (two times for posterior vowels and almost three times for anterior vowels). Total ratio of average duration of shorter and longer vowels — 1 : 1,3. 3.2. The quantitative difference between vowels of natural and positional length is also illustrated by a dot diagram. By placing the length dimensions (ms) of the positional length vowels on the ordinate axis of the first quarter of the coordinate system, and the natural length vowels in the corresponding pair on the abscis axis, we will obtain a graphic image that perfectly illustrates the differences in the quantity of these vowels (see Figure 1). * For clarity, the coordinate system reference point is not 0 ms but 100 ms. 104 250 225 200 175 150 125 100 10 12 15 17 20 22 25 0 505050 longer vowels (ms) shorter vowels (ms) Figure 1. All the points are on one side of the diagonal, and some are even very far away from it, so the two variants are very different. 4. The qualitative parameters of these vowels were also analyzed. The program “Kalbam44” analyzed the spectral sections of the sound segments, and then the averages of the formants were evaluated by a special program for evaluating formants created by A. Girdenis, which, according to certain formulas (see Piotrovskij 1960: 29), calculates the indices of compactness (C.10*), bemolarity (b.10*) and tonality (T) (a larger T indicates a higher vowel timbre, a smaller and especially negative one — a lower one). 4.1 Since the differential characteristics of vowel sounds are determined by the first two forms, and the other forms often convey only the individual characteristics or expressive content of the speaker, the first forms are mainly used to analyze the acoustic characteristics of sounds. Male and female voice spectra were studied separately, as the acoustic characteristics of male and female voices differ significantly, although some constant relationships remain. The results of the spectral analysis of male and female voices are also similar, only the female voice forms are significantly higher due to the higher voice timbre. Therefore, we provide measurements of the male voice that give a clearer picture. The quality of the vowels of the anterior and posterior articulation is worth discussing separately. Table 3: Vowels [21 < *-an and [a] < *-a4male voice-forming values” For example: F, (Hz) F, (Hz) F; (Hz) C.10° T drill 675 1326 2410 888 34 edge 685 1353 2320 889 61 Sala 675 1320 2564 888 3 kala 703 1340 2525 893 5 3 Plg. female voice drill F,=728 Hz, F,=1410 Hz; edge ¥,=785 Hz, F,=1498 Hz; freezes F,=728 Hz, F,=1369 Hz; fish F,=785 Hz, F,=1398 Hz. 105 As we know, all the vowels of the posterior row have a low timbre, which is due to the wide and large mouth resonator, formed when the tongue retreats back. This is also shown by the second form in the low frequency area. However, the natural length [a] type vowel F, is still slightly lower than the positional length vowel F,. This means that vowels of historical length have a lower timbre, which is due to a greater backward movement of the tongue. Vowels of natural length also have a slightly lower first form, so they have a slightly higher pitch and are more open. Lower, from F; The more distant position of the first two forms is also a sign of [a:] <*-affinity, so we can say that the spectral analysis confirms the hypothesis about the remaining nasalization of these vowels. There are also differences in the significance of the third formants: for vowels of historical length they are considerably higher. In terms of spectrum scattering, the low-pitched vowels are the most compact, because both forms are not too far from the middle of the spectrum (x 1000 Hz). Historical long vowels are slightly less compact, but lower-tone sounds (see Table 3). These vowels are neutral in terms of bemolarity (b.10* =168), because the lips do not participate in the articulation. We can also see the forming relationships in the graph (see Figure 2). 1700 1200 + 700 200 1 2 3 Figure 2. Spectra of vowels [2] < *-a@ (----) and [a] < *-4-(—) in the words gręžtas and kraštas’ The vowels [e] and [ee] are relatively unstable in articulation, and therefore their acoustic characteristics vary. Particularly unstable vowel [22]: at the beginning of pronunciation F, and F> are within the [e]-forming boundaries, and at the end this sound becomes very similar to [a'] — open, more compact, lower timbre. Therefore, the segmental spectrum of these vowels was measured at three locations: at the beginning, middle and end of the sound. We see the calculations in the table. * Similar would be the graphic representation of the words šėla —kala. 106 Table 4 | Vowels [e] < *-en-/[e] < *-ė ir [2 <*-a2r-/[a2] <*-2-male voice-forming values” Examples F, (Hz) F, (Hz) F; (Hz) C.10° T 576 1930 2580 828 399 patreses 605 1787 2529 842 320 patreses 626 1621 2379 857 247 587 1775 2471 839 337 615 1714 2414 849 269 610 1605 2326 855 259 613 1627 2451 853 246 patrésus 690 1412 2349 885 90 patréSus 627 1448 2297 869 677 1351 2353 887 59 623 1555 2363 860 216 163 678 1377 2249 886 95 We see that the formantinés characteristics of all [e]-type vowels change during pronunciation. The first form almost consistently rises, the second — falls. This is particularly evident before the hard consonant (in position /—C). 5.2.1. Comparing the low vowels in position /—C’, it is evident that the F of the vowels of historical length remains higher than the positional length throughout the whole time of pronunciation — they are of higher timbre, more forward articulation, more tense (the formants are further away from the “neutral” vowel spectrum). The positional length voice is less tense, backward, lower timbre.” At the beginning of the pronunciation the positional vowel has the higher first form, but at the end the natural vowel F jumps higher; positional vowel it drops slightly down. The long-term long-term long-term long-term long-term long-term long-term long-term long-term long-term long-term long-term long-term long-term long-term long-term long-term long-term long-term long-term long-term long-term long-term long-term long-term long-term long-term long-term long-term long-term long-term long-term long-term long-term long-term The compactness of these vowels is also variable. At the beginning and in the middle of the pronunciation, the positional vowel is more compact, and at the end, the natural vowel. These vowels are very unstable in terms of tonality, but no matter how they bequiet, the historical and positional differences between long vowels remain. At the beginning and in the middle of the pronunciation, the historical vowel has a much higher tone, and at the end of the pronunciation, the positional vowel has a slightly higher tonal index. 5.2.2. When comparing these low vowels in position /—C, their instability is even more pronounced. At the beginning of the pronunciation, the difference between the first forms is small, slightly higher positional length vowel F; in, but further it rises very slightly, O natural length vowel F; jumps much higher. In the middle of the segment, the forms differ by more than 60 Hz. However, at the end of the pronunciation of the difference vi- " Three forming digits indicate the measurements at the beginning, middle and end. * The female voice indicates the same forming relationships. ’ In terms of quantity, these vowels differ the most. 107 sounds disappear. Historical length of vowel only F, at the beginning is higher than in position ir /—C’; later the forms pass the low į frequency range, sounds become low timbre, the ratio of forms is o the same as [a] type vowels: the higher is the second form of the positional vowel. So at the end of the segment of lower timbre is the voice of historical length [2]. The upper third form has always been used for vowels of historical length. At the beginning, the positional length vowel is more compact, in the middle the natural length vowel, at the end — the difference is vague. Tonality index clearly shows both o dynamics and differences of sounds jų (see section 4.4). 4 Table 1). Start again as in position ir /—C’, a much higher tone is a vowel of historical length, in the middle of the o segment at the end of the ir higher tone is a vowel of positional length, similar to ir [a] type vowels. The bemolarity index in this case also does not show any differences (b.10*=167), because the lips do not participate in the articulation of these vowels either. The corresponding relationships are also shown in the graphs of the forming trajectories (See section 4.4.) 3 and Figure 4. ). 2200 1700 fizz. | 1200 + 700 + 200 1 2 3 Figure 3. Change in the trajectories of the vowels [e] < *-er+ (----) and [e] < *-é& (—) formandiy in position /-C' 1700 1— 1200 + 00 . .., caiventis = 200 ; 1 2 3 Figure 4. Change in the trajectories of the vowels [2] < *-22r7-(--- --) and [2] < *-2£-(—) at position /—C. 108 6. Differences ir in natural positional length of low vowels are shown in the graphs of ir the basic tone and intensity (see section 4.4). 5-10 Figure 1. ). The graphs were drawn using the computer speech sound analysis program WINCECIL22. "T (mA LEB Ei Gi SIR m | SM SRL (ES Si BBI IN SEO EO Vnt BIT MBT VB VE a | 140Hz/Lg | shades Figure —r— 5. Balsio [2] < *-or+ intensity (in conventional dB) main tone ir (Hz logarithms) in words gra-stas 15 P————————————T “1280 b 140HziLg A km Ak A Ls Ai a BE SY SSS I SS ny Sy "| I RT fig. 6. Balsio [a7] < *-4intensity main ir tone in word krd.stas 65Hz/Lg fig. 7. Balsio [e] < *-efdintensity main tone ir in patré-sces . 109 US 4 Tam (O am ESS Ark Figure 8. Vowel [e] < *-ē intensity and main tone in patre.ses 30 | | 290 m. (iii)’ i 0 | | | 145Hz]See Figure 9. Vowel [#7] < *-z4intensity and main tone in patra-Sus Ng —————— 0 i i B ES RV Na (Eat Co EE Gia ms am TR fe To VB Figure 10 Intensity and basic tone of vowel [#1] <*-a in patrdė.šus We can see that the basic tone curves of both the front and back vowels of historical and positional length differ: the curve of all historical vowels is ascending, the positional — descending or ascending-- descending. Slightly higher up at the end of the sound, the pitch curve of the posterior vowels rises, so the pitch range of these vowels is wider (see Figure 5). Of the positional length vowels, the [e] curve at the end of the sound is the lowest — the range of the basic tone is the widest. 110