Prienų šnektos izoliuotieji ir rišliosios kalbos balsiai: akustiniai ir artikuliaciniai požymiai

JURGITA JAROSLAVIENĖ Institute of the Lithuanian Language Research areas: dialectology, phonology, experimental phonetics. VOWELS IN ISOLATION AND CONNECTED SPEECH IN THE PRIENAI DIALECT: ACOUSTIC AND ARTICULATORY FEATURES Acoustic and articulatory features of vowels of the subdialect of Prienai pronounced both in isolation and in sentences ABSTRACT The article describes the main acoustic and articulatory features of vowels in isolation and connected speech in the Prienai dialect. The research results show that vowels in the Prienai dialect pronounced in isolation and in natural speech differ, but the relationships between the vowels remain the same. The acoustic and articulatory properties of most vowels pronounced in isolation are more prominent than those of the corresponding sounds pronounced in sentences. A comparison of the spectral characteristics of vowels in isolation in the Prienai dialect and Daniel Jones’s cardinal vowels suggests that some Prienai dialect vowels differ little in their acoustic and articulatory properties from the corresponding primary cardinal sounds. A comparison of the qualitative features of sounds pronounced in isolation in the Southern Aukštaitian dialect of Kuciūnai, the Western Aukštaitian Šiauliškiai dialect of Pašūsvys, the Northern Kauniškiai dialect of Lukšiai, the Eastern Aukštaitian Vilniškiai dialect of Svirkai, the Southern Samogitian Raseiniai dialect of Eržvilkas, the Northern Samogitian Telšiai dialect of Akmenė, and the Eastern Kauniškiai dialect of Prienai showed that the vowels in isolation in the Prienai dialect deviate least from the corresponding primary sounds of D. Jones. ANNOTATION The article analyses the main acoustic and articulatory features of vowels of the subdialect of Prienai pronounced both in isolation and in sentences. Isolated vowels differ from those KEYWORDS: vowel, sound, formant, spectrum, acoustic and articulatory properties. KEYWORDS: vowel, sound, formant, spectra, acoustic and articulatory features. Straipsniai / Articles 65 JURGITA JAROSLAVIENĖ articulated in a natural flow of speech in the subdialect of Prienai, yet the vowel interrelations remain unchanged. The acoustic and articulatory features of the majority of sounds pronounced in isolation are more prominent in comparison with the respective sounds pronounced in sentences. The performed comparison between spectral characteristics of isolated vowels of the subdialect of Prienai and those of cardinal vowels by Daniel Jones allows stating that some of isolated vowels of the subdialect of Prienai should be contrasted with the primary vowels. Having compared qualitative features of isolated sounds from the Southern Aukštaitian Kučiūnai subdialect, the subdialect of Pašušvys of the Western Aukštaitian Šiauliai dialect, the Northern Kaunas Lukšiai subdialect, the Švirkai subdialect of the Eastern Aukštaitian Vilnius dialect, the Eržvilkas subdialect of the Southern Samogitian Raseiniai dialect, the Akmenė subdialect of the Northern Samogitian Telšiai dialect and the Eastern Kaunas Prienai subdialect, it was noticed that isolated vowels in the subdialect of Prienai are least deviated from their counterpart sounds by D. Jones. 0. INTRODUCTION 0.1. Research objective The aim of this study is to investigate and describe the acoustic and articulatory characteristics of the vowels of the Prienai dialect, spoken in isolation1, using the PRAAT computer sound-analysis program developed by University of Amsterdam researchers P. Boersma and D. Weenink; to compare the data obtained with universally recognized vowel-quality standards pronounced in the same way—Daniel Jones’s cardinal vowels2—and with the vowels of connected speech in the Prienai dialect and other dialects: Lukšiai in the Western Aukštaitian Kaunas subdialect (Bacevičiūtė 2000: 13–17; 2004: 158–174), Pašušvys in the Western Aukštaitian Šiauliai subdialect (Kazlauskaitė 2002a: 64–67; 2002b: 69–71), Kučiūnai in the Southern Aukštaitian dialect (Leskauskaitė 2000: 89–93; 2004: 85–96), Švirkai in the Eastern Aukštaitian Vilnius subdialect (Urbanavičienė 2004: 65–80; 2005: 76–93), and Akmenė in the Northern Samogitian Telšiai subdialect (Murinienė 1998: 91–105; 2007: 109–127), 1The acoustic and articulatory characteristics of vowels pronounced in isolation are more general and more clearly defined, and better reveal the nature of vowels, because when pronouncing them it is easier to separate them from the effects of other sounds, as well as intonation, speech rate, and other factors. 2The formant values for Daniel Jones’s cardinal vowels were taken from a work on acoustic phonetics by the Polish phonetician Wiktor Jassem (1973: 190, Table 13.1). As is known, Jassem’s formant values were calculated using an older methodology. Seven years ago, when the cardinal-vowel formants were recalculated using Praat 4.1.14, it was observed that formant values calculated by different methods do not coincide; although the relationships among the cardinal vowels remain the same, Jassem’s data should be considered more reliable (for more detail, see Jaroslavienė, Kaukėnienė 2004: 25–38). 66 Acta Linguistica Lithuanica LXIV–LXV Prienų šnektos izoliuotieji ir rišliosios kalbos balsiai: akustiniai ir artikuliaciniai požymiai pietų žemaičių raseiniškių Ežžvilko (Atkočaitytė 2002: 169–173)3 – izoliuotai ir frazėse ištartų balsių duomenimis. 0.2. Research material and methodology The following words used in the dialect under study were selected for the study of vowels pronounced in isolation and in connected speech: tiːsė. ~ tišė, tẽtė. ~ tėtė, tẽse. ~ tęse, tãːsto. ~ tęsto, tãːso. ~ tašo, stõːtu. ~ stōtu “stōtú”, pũːstu. ~ pūstu “pūstū”, kẽːpe. ~ kėpė, kãːpto. ~ kėpto, kãːso. ~ kąso, tiksi ~ tiksi, kèpsi ~ kèpsi, kèps ~ kèps, kàst ~ kàst, kòstas ~ Kòstas, skùst ~ skùst. To study vowels in connected speech, coherent sentences, mostly three or four words long, were constructed using these words; they were written in random order on sheets using semi-dialectal spelling, cf.: Vaikai tąso tą šunį, Greitai kepsi tą žuvį, Vaikai tęse tą darbą, etc. Three older-generation male informants from Prienai, without speech defects, read the research material directly into a computer at a normal speaking rate, three to five times each, for the experiment on vowels pronounced in isolation and in connected speech. The speakers prolonged the so-called isolated vowels to approximately 1 s. The sentences and words were transferred to the computer’s memory. The most successful recordings were selected for analysis. The material was saved as * wav files. The PRAAT 5.0.34 computer program was used to segment the elements under study, selecting the options Formant & LPC, followed by To Formant, Track and Down to TableOfReal; formant values were determined at the midpoint of the steady-state portion of the segment (rounded to the nearest 10 Hz). Before measurement, the spectra of the high back vowels were filtered, i.e., frequencies above 3500 Hz4 were removed. For the spectral analysis, first, a natural sound- or word-type spectrogram of varying degrees of greyVisible Speech was generated from the excised segments under study, with the upper frequency limit set to 3500 Hz. The formant trajectories were then measured automatically and, if necessary, plotted, again with the upper frequency limit set to 3500 Hz. Often the natural spectrogram and the plots of formant dynamics were combined into a single image, precisely aligned with respect to frequency and duration. The formant values of the isolated vowels were assessed, and the qualitative characteristics of the segments 3Studies devoted to the dialects listed above investigated the spectra of dialectal vowels pronounced in isolation using the PRAAT computer sound-analysis program. It is important to note that in the studies of these dialects, with the exception of the description of the Švirkai dialect of Eastern Aukštaitian Vilnius, the formant value was taken to be the mean of all formant values across the isolated segment. As in the description of the Švirkai dialect, this study measured the middle—the purest—part of the isolated sound (cf. Urbanavičienė 2005: 76–77 and the literature cited there). The steady-state portion of the segment was obtained by cutting a 500 ms section from a sound artificially lengthened to approximately 1 s. 4The high frequencies were removed using the same PRAAT sound-analysis program by selecting the options Spectrum, To Spectrum and Edit, followed by Publish band-filtered sound and then Edit. Straipsniai / Articles 67 JURGITA JAROSLAVIENĖ The parameters (compactness (C), flatness (b), tonality (T), and tension (įt)) were calculated using Aleksas Girdenis’s FORMANT2.PAS program, developed on the basis of Raimondas Piotrovskij’s formulas (Piotrovskij 1960). The mean formant values for the same sound across all informants are reported in the tables in this article; graphs plotted using EXCEL2007 are also presented. 1. RESULTS OF THE STUDY OF ISOLATED VOWELS The results of the instrumental study of isolated vowels and vowels in connected speech in the Prienai dialect are discussed and analyzed below. 1.1. Acoustic and articulatory characteristics of isolated vowels in the Prienai dialect Examples of spectrograms of isolated vowels in the Prienai dialect are presented in Figures 1–13. Based on the results obtained (see Table 1) and taking the spectrograms into account, the vowels of the Prienai dialect can be classified on both acoustic and articulatory grounds. First, the acoustic features of isolated vowels are reviewed. In terms of timbre height, which depends on the oral resonator, the vowels [i̇], [i], [ė], [e], [e̞], [e] pronounced in isolation are considered high-timbre sounds: their second formant is higher than 1500 Hz, and their tonality indices are considerably higher than those of the low-timbre vowels [ɑ], [ȧ], [a], [ȯ], [ɔ], [u̇], [u]—the second formant of these vowels does not reach 1500 Hz (cf. Figs. 1–13). The degree of tonality of all low-timbre sounds, except the short [a], FIG. 1. Spectrogram of the vowel [i̇] pronounced in isolation FIG. 2. Spectrogram of the vowel [u̇] pronounced in isolation 68 Acta Linguistica Lithuanica LXIV–LXV Prienų šnektos izoliuotieji ir rišliosios kalbos balsiai: akustiniai ir artikuliaciniai požymiai FIG. 3. Spectrogram of the vowel [e̜] pronounced in isolation FIG. 4. Spectrogram of the vowel [o̜] pronounced in isolation FIG. 5. Spectrogram of the vowel [e̜] pronounced in isolation FIG. 6. Spectrogram of the vowel [a̜] pronounced in isolation FIG. 7. Spectrogram of the vowel [e̜] pronounced in isolation FIG. 8. Spectrogram of the vowel [a̜] pronounced in isolation negative. The isolated [e̜], [e̜] and [e̜], compared with the high- and low-timbre vowels, occupy an intermediate position in acoustic space: their second formants are 1770 Hz, 1710 Hz and 1820 Hz, respectively, while their tonality indices are much lower than those of [i̜], [i̜], [e̜], Straipsniai / Articles 69 JURGITA JAROSLAVIENĖ FIG. 9. Spectrogram of the voice [j] pronounced in isolation FIG. 10. Spectrogram of the voice [u] pronounced in isolation FIG. 11. Spectrogram of the voice [e] pronounced in isolation FIG. 12. Spectrogram of the voice [ɑ] pronounced in isolation FIG. 13. Spectrogram of the voice [ɔ] pronounced in isolation FIG. 14. Spectrogram of the voice [ɔ] in connected speech but considerably higher than [ɑ], [a.], [ɑ], [o], [ɔ], [u], [u]. These [e]-type allophones are often also regarded as intermediate sounds in other dialects (cf. Bacevičiūtė 2004: 166; Kazlauskaitė 2002b: 56; see also Girdenis 2001: 341; Urbanavičienė 2005: 78 et al.). 70 Acta Linguistica Lithuanica LXIV–LXV Prienų šnektos izoliuotieji ir rišliosios kalbos balsiai: akustiniai ir artikuliaciniai požymiai TABLE 1. Formants and evaluations of the isolated vowels of the Prienai dialect5 V F₁ (Hz) F₂ (Hz) F₃ (Hz) C b T tension index [iˑ] 280 2460 3210 712 107 828 1890 [i] 390 2020 2720 772 107 585 850 [ɛˑ] 360 2390 2780 750 107 756 1310 [eˑ] 640 1770 2540 850 106 285 450 [eː] 670 1710 2610 859 106 223 490 [e] 560 1820 2650 829 106 349 530 [aˑ] 780 1270 2530 915 106 −88 540 [aː] 790 1300 2590 914 106 −83 580 [a] 710 1330 2430 895 107 11 450 [oˑ] 410 730 2210 890 113 −231 1150 [ɔ] 480 910 2300 885 111 −125 810 [uˑ] 330 660 2090 868 115 −200 1420 [u] 400 870 2320 862 112 −89 910 The isolated [iˑ] has the highest pitch in the dialect under study (F₂ = 2460 Hz, T = 828), while the lowest are the long [uˑ] (F₂ = 660 Hz, T = −200) and [oˑ] (F₂ = 730 Hz, T = −231). It should be noted that the lowest-pitched vowel has the highest index of lip-rounding (115), meaning that the lips should be most active when pronouncing it. According to the acoustic classification, this vowel belongs to the class of rounded vowels. The isolated [oˑ], [ɔ], and [u] should also be considered rounded; compare, respectively: b = 113, 111, 112. It is noteworthy that the isolated high-pitched vowels of the Prienai dialect are pronounced with a higher pitch, and the low-pitched ones with a lower pitch, than, for example, those in the northern Kaunas subdialect of Šakiai (cf. Bacevičiūtė 2004: 160). Acoustic sound properties, as indicated by the dispersion of the spectrum, are related to the vertical movement of the tongue upward or downward (cf. Girdenis 2003: 193–195; Pakerys, 2003: 23): the lower the tongue is lowered when pronouncing an isolated vowel, the higher its first formant. As can be seen, in the lowest frequency range are the first formants of [iˑ] and [uˑ] (fig. F₁ = 280 Hz, 330 Hz); for the other isolated vowels—[i], [ɛˑ], [oˑ], [ɔ], and [u]—F₁ ranges from 360 to 480 Hz. Interestingly, the first formant of the long [eˑ] is lower than those of the short [u] and especially [i] (fig. F₁ = 360 Hz, 400 Hz, 390 Hz). The low-pitched [aˑ], [aː], and [a] have the highest first formants (fig. F₁ = 780 Hz, 790 Hz, and 710 Hz), meaning that the tongue is lowered the most when producing these sounds (cf. Figs. 1–13). 5 The symbols mentioned in this and the following tables mean: F₁, F₂, F₃ – values of the first three formants in the middle portion of the vowel, rounded to the nearest 10 Hz; C – compactness index; b – flatness index; T – tonality index; tension index – tension index. Straipsniai / Articles 71 JURGITA JAROSLAVIENĖ Taking the position of both formants into account, isolated vowels can be classified according to their degree of diffuseness, or compactness⁶. As can be seen in Table 1, the most compact are the low-pitched vowel of inherent length [ɑ], as well as [ɑː] and [ɑ] (fig. C = 915, 914, and 895)⁷: their first two formants (F₁ and F₂) lie close to one another in acoustic space and, in general, close to the central part of the spectrum (see Figs. 6, 8, 12). At the very edges of the spectrum, [i] and [u] have their first two formants F₁ and F₂—these are non-compact, or the most diffuse, vowels. The tension index makes it easy to establish that long vowels are pronounced with considerably greater tension than the corresponding short vowels; compare: [i] tension = 1890, [i] tension = 850; [uː] tension = 1420, [u] tension = 910; [oː] tension = 1150, [ɔ] tension = 810. The [eː], [e.], and [e] and [ɑː], [ɑ.], and [ɑ] of the Prienai dialect differ least in tension; compare: [eː] tension = 450, [e.] tension = 490; [e] tension = 530; [ɑː] tension = 540, [ɑ.] tension = 580, [ɑ] tension = 450); indeed, as can be seen, the tension index of short [e] is higher than that of [e.] and [eː]. The following presents an articulatory interpretation of the isolated vowels of the dialect. The high-pitched sounds [iː], [i], [i̞ː], [eː], [e.], [e] should be considered front vowels; when they are pronounced, the tongue is shifted toward the front part of the mouth. The [ɑː], [ɑ.], [ɑ], [oː], [ɔ], [uː], [u] should be considered back vowels—they are pronounced with the tongue withdrawn toward the back of the mouth. However, even when vowels of the same series are pronounced, the tongue’s position differs: of all the vowels, [iː] is the most front, while [uː] and [oː] are the most back. By height, the isolated sounds of the dialect divide into high ([iː], [i], [uː], [u]) and low ([ɑː], [ɑ.], [ɑ], [eː], [e.], [e]) vowels. The isolated [ɛː] also lies close to the high vowels; [oː] and [ɔ] should be considered more intermediate, mid-height sounds, although the F₁ of [oː] differs by only 10–20 Hz from the first formants of [u] and [i]. In terms of mouth openness, the long [iː] is the closest vowel in the dialect, while [ɑː] and [ɑ.] are the most open. The potential articulatory features and relationships of the isolated vowels are directly represented by the acoustic space diagram (see Fig. 15), whose abscissa plots the differences between the second and first formants (F₁–F₂), while the ordinate plots the values of the first formants (F₁) (cf. Ladefoged 1975: 173 ff.; Clark, Yallop 1999: 268; Ladefoged, Maddieson 2002: 284–286; Girdenis 2008: 31, n. 18, etc.). The diagram shows that the distances between isolated front sounds are considerably greater than those between back sounds: the left side of the triangle, representing the relationships among front vowels from [i] to [e.], is noticeably longer than the right side, connecting back vowels from [ɑ.], [ɑ] to [u] (cf. Fant 1964: 114; Murinienė 1998: 103, etc.). The greater distances between front vowels indicate that their horizontal tongue movement toward the teeth tends to be more pronounced than that of back vowels. The long 6 Dėl kompaktiškumo indekso dar žr. Urbanavičienė 2005: 79, 102 išn. 7 Taip pat žr. Pakerys 2003: 26. 72 Acta Linguistica Lithuanica LXIV–LXV Prienų šnektos izoliuotieji ir rišliosios kalbos balsiai: akustiniai ir artikuliaciniai požymiai FIG. 15. The isolated vowels of the Prienai dialect in acoustic space [i˙] is a sound with the most fronted articulation; when it is pronounced, the tongue is shifted furthest toward the front of the mouth. The isolated [o˙] and [u˙], [ɔ] and [u] differ very little in degree of backness: [o˙] and [ɔ] are only slightly further back than [u˙] and short [u], respectively. Of all the sounds, the long [o˙] and [u˙] are shifted furthest toward the back row. In some other dialects whose isolated vowel systems have been analysed according to similar principles, [o˙] (in the Pašūsvys dialect) and [u˙] (in the Lukšiai and Kučiūnai dialects) also occupy the extreme position (see Leskauskaitė 2004: 93; Kazlauskaitė 2002b: 62; but cf. Crothers 1978: 97; Urbanavičienė 2005: 83). Compared with the [o˙], [ɔ], [u˙] and [u] discussed above, the low [a˙] and [a.] are least back; the inherently long [a˙] is slightly “darker”. 1.2. Comparison of the spectra of isolated vowels in the Prienai dialect and Daniel Jones’s cardinal vowels Next, the dialect’s vowels pronounced in isolation are compared with the cardinal vowels of the English phonetician Jones, which consist of primary (₁[iː], ₂[eː], ₃[ɛː], ₄[aː], ₅[ɑː], ₆[ɔː], ₇[oː], ₈[uː]) and secondary (₉[yː], ₁₀[øː], ₁₁[œː], ₁₂[ɶː], ₁₃[ɒː], ₁₄[ʌː], ₁₅[ɤː], ₁₆[uː], ₁₇[iː]) vowels Straipsniai / Articles 73 JURGITA JAROSLAVIENĖ and ₁₈[u̟])⁸. The primary vowels ₁[i̞], ₂[e̞], ₃[ɛ̞], ₄[a̞], ₅[ɑ̞] are unrounded, while ₆[ɔ̞], ₇[o̞], ₈[u̞] are rounded. The secondary vowels correspond to them in terms of lip position in the opposite way: ₉[y̞], ₁₀[ø̞], ₁₁[œ̞], ₁₂[ɛ̞], ₁₃[ɒ] are rounded, while ₁₄[ʌ̞], ₁₅[ɣ̞], ₁₆[ɯ̞] are unrounded. ₁₇[ɨ̞] and ₁₈[ɯ̞] are intermediate close vowels: the former is unrounded, the latter rounded. The qualitative parameters of these vowels are presented in Table 2 (Jassem 1973: 190, Table 13.1). Jones’s cardinal vowels are theoretical constructs and therefore do not exactly correspond to the respective sounds of any particular language or dialect. Girdenis (2008, 29–30; see also Girdenis 2001: 340–343 and Murinienė 2007: 116–117) has provided the most detailed description of their auditory impression and approximate similarity to sounds in Lithuanian and other languages. The author of this article has also listened to these cardinal vowels (cf. Jaroslavienė, Kaukėnienė 2004: 33; Jaroslavienė 2010: 146–147). The listener noted that ₂[e̞] and ₇[ɔ̞] resemble [ɛ̞] and [o̞] in the southern Western Aukštaitian dialects (especially Zanavykian), while ₁[i̞] and ₈[u̞] resemble [i̞] and [u̞] in the same dialect. Cardinal ₅[ɑ̞] is somewhat similar to Zanavykian [ɑ]; ₃[ɛ̞] resembles Standard Lithuanian [e̞] (only it is not as open). Comparing the qualitative sound parameters (see Tables 1 and 2) and taking into account the comparative diagram of the cardinal and isolated-vowel systems of the dialect (see Fig. 16), it can be argued that the isolated unrounded unmarked vowels [i̞] and [ɪ] in the dialect do not exhibit the extreme front and close articulation characteristic of the corresponding cardinal vowels. The long vowel is distant both from primary unrounded ₁[i̞] and from secondary rounded ₉[y̞]. It coincides with the former [i̞] only in its degree of unmarkedness (cf. b = 107 and 107). The short [i̞] also does not differ from this cardinal vowel in unmarkedness, but it ([i̞]) is shifted even further toward the centre of acoustic vowel space than the long vowel. Thus, short [i̞] is a lower-pitched, more back, and more open sound than [i̞]. This is also confirmed by the acoustic properties of [i̞] and [ɪ] (see Table 1). It should be noted that the isolated [ɛ̞] in the dialect is further front and higher than [i]. This long vowel is pronounced in the dialect with highly tense articulators. Similar tendencies have also been observed in the Lukšiai dialect. For these reasons, [ɛ̞] could be considered a close vowel (cf. also Bacevičiūtė 2004: 160, 166; Girdenis 2003: 222). In terms of tenseness and compactness, isolated [ɛ̞] occupies an intermediate position between its corresponding primary cardinal vowel, unrounded ₂[e̞], and secondary rounded ₁₀[ø̞]. In acoustic vowel space, long [ɛ̞] is nevertheless shifted closer to secondary ₁₀[ø̞] (cf. [ɛ̞] and ₁₀[ø̞] F₁ = 360 and 350 Hz, F₂ = 2390 and 2320 Hz); these vowels do not differ at all in unmarkedness (cf. [ɛ̞] and ₁₀[ø̞] b = 107 and 107). It should be noted that isolated [ɛ̞] and ₁₀[ø̞] differ most in tenseness—the dialect vowel is more tense (cf. [ɛ̞] and ₁₀[ø̞] tens. = 1310 and 1070; ₂[e̞] tens. = 1800). 8 For a description of the cardinal vowel system and planimetric models, see Jaroslavienė 2010: 146, 153 n.; Jaroslavienė, Kaukėnienė 2004: 25, n. 1. 74 Acta Linguistica Lithuanica LXIV–LXV Prienų šnektos izoliuotieji ir rišliosios kalbos balsiai: akustiniai ir artikuliaciniai požymiai T A B L E 2. Formants and indices of Daniel Jones’s cardinal vowels V F₁ (Hz) F₂ (Hz) F₃ (Hz) C b T tens. 1. [iː] 210 2750 3500 667 107 1012 2540 2. [eː] 380 2630 3050 748 105 776 1800 3. [ɛː] 590 2280 2700 818 105 514 1070 4. [aː] 870 1750 2700 892 104 115 820 5. [ɑː] 800 1050 2720 947 107 −295 970 6. [ɔː] 550 850 2500 917 110 −279 700 7. [oː] 400 730 2300 886 113 −237 1070 8. [uː] 270 615 2200 845 117 −196 1415 9. [yː] 220 2550 3100 680 108 979 1930 10. [øː] 350 2320 2600 751 107 772 1070 11. [œː] 520 1950 2500 815 106 466 470 12. [ɶː] 790 1650 2600 885 105 122 540 13. [ɒː] 710 900 2850 952 108 −398 1160 14. [ʌː] 570 940 2700 907 109 −241 830 15. [ɤː] 450 850 2500 883 111 −192 700 16. [ɯː] 280 850 2250 810 115 59 1120 17. [ɨː] 240 1550 2400 730 112 613 370 18. [ɤː] 270 1370 2500 756 112 444 360 It is more difficult to determine which cardinal vowels the compact front isolated [eː], [eː] correspond to—they are considerably distant not only from the primary cardinal ₃[ɛː], but also from the secondary cardinal ₁₁[œː]. These vowels coincide with the eleventh [œː] only in flatness (b = 106); their other qualitative parameters differ from those of the corresponding Jones vowels. It is noteworthy that [eː] is somewhat higher, closer, and fronter than the corresponding positionally long vowel (cf. Bacevičiūtė 2004: 166). The short [e], compared with [eː] and [eː], is fronter and higher; in terms of the position of its first formant, [e] (F₁ = 560 Hz) differs little from the primary third [ɛː] (F₁ = 590 Hz) and the back secondary vowel ₁₄[ʌː] (F₁ = 570 Hz). In terms of its second formant, [e] (F₂ = 1820 Hz) is the least retracted from the primary ₄[aː] (F₂ = 1750 Hz) and secondary ₁₁[œː] (F₂ = 1950 Hz). On the basis of acoustic and articulatory properties, the dialect’s [eː], [eː], and [e] (cf. Figs. 5, 7, 11) should be considered the most retracted front vowels; according to vertical tongue displacement, they can be classified as mid vowels. The low-timbre compact non-flat [ɑː], [ɑː], located at the bottom vertex of the triangle of the dialect’s vowels pronounced in isolation, occupy an intermediate position in the spectrum Straipsniai / Articles 75 JURGITA JAROSLAVIENĖ FIG. 16. Isolated vowels of the Prienai dialect (●) and D. Jones’s cardinal vowels (○) position between the primary unrounded [ɑː] and the secondary rounded [ɔː]. These dialect vowels and Jones’s vowels are closest in their first formant, or vertical tongue displacement; in terms of the second formant, [a.], [aː] are shifted closer to the primary back unrounded [ɑː], cf.: [aː] F₁ = 780 Hz, F₂ = 1270 Hz, [a.] F₁ = 790 Hz, F₂ = 1300 Hz and [ɑː] F₁ = 800 Hz, F₂ = 1050 Hz, [ɔː] F₁ = 790 Hz, F₂ = 1650 Hz. It should be noted that of all the dialect’s [a]-type vowels, the long nasal-origin [aː] has the lowest second formant (cf. [aː], [a.], [a], respectively, F₂ = 1270 Hz, 1300 Hz, 1330 Hz). These vowels also differ in their first formants—the positionally long [a.] has the highest F₁, so the tongue should be lowest when it is pronounced; the short [a] has the lowest F₁ value (cf. Figs. 6, 8, 12). It is noteworthy that in terms of vertical tongue displacement, the short vowel coincides completely with the thirteenth cardinal [ɒ] (F₁ for both = 710 Hz). In compactness, the dialect’s [a.], [aː] (see C = 914, 915) are closer to Jones’s primary back vowels, cf.: [ɑː] and [ɒ], C = 947 and 917. The short [a] is closer in this respect (C = 895); it can be assumed that the mouth opens less when articulating it. The isolated [aː], [a.], [a] have very low tonality, but it is expressed by values that are not as low as those of the cardinal [ɒ] and [ɑː]. The short [a] coincides with the primary fifth [ɑː] in degree of flatness (107); [aː], [a.] are less flat (106). 76 Acta Linguistica Lithuanica LXIV–LXV Prienų šnektos izoliuotieji ir rišliosios kalbos balsiai: akustiniai ir artikuliaciniai požymiai In summary, the dialect’s isolated [ɑˑ], [ɑ.], [ɑ] are back low unrounded vowels. In terms of their spectral properties, they can be compared not only with secondary but also with primary cardinal vowels, although they coincide with neither. The low-timbre flat [ɔˑ] stands out most among all the sounds pronounced in isolation in the dialect (cf. Fig. 4)—as can be seen, it is articulated in the same part of the mouth as its primary cardinal counterpart, the rounded ₇[ɔˑ]: cf. [ɔˑ] F₁ = 410 Hz, F₂ = 730 Hz; ₇[ɔˑ] F₁ = 400 Hz, F₂ = 730 Hz. The qualitative parameters of these vowels are also similar; cf. [ɔˑ] and ₇[ɔˑ], respectively: T = −231 and −237, b = 113 and 113, C = 890 and 886, tens. = 1150 and 1070). Thus, the dialect’s [ɔˑ] is characterized by an extreme back articulation; in terms of vertical tongue displacement, this vowel should be considered a mid vowel. The short [ɔ], belonging to the dialect’s periphery, is also a low-toned intermediate flat vowel, but has a higher timbre than [ɔˑ] (cf. Figs. 4, 13); moreover, [ɔ] is less tense and less flat than the corresponding long vowel mentioned above. In flatness, compactness, and tonality, the short [ɔ] differs least from the secondary fifteenth [ɣˑ] (for both, b = 111; respectively, C = 885 and 883, T = −125 and −192). In tenseness, however, the vowel in question should be compared with the secondary ₁₄[ʌˑ]. [ɔ] is also associated with this vowel and with Jones’s thirteenth vowel in terms of the degree of downward horizontal tongue displacement (cf. ₁₃[¤], ₁₄[ʌˑ], [ɔ], respectively, F₂ = 900 Hz, 940 Hz, 910 Hz). In terms of vertical tongue displacement, [ɔ] occupies an intermediate position between the secondary ₁₅[ɣˑ] and the primary cardinal ₆[ɔˑ] (F₁ values, respectively, = 480 Hz, 450 Hz, 550 Hz). The low-toned, noncompact, flat [uˑ] also stands out among the dialect’s isolated vowels (cf. Fig. 2). First, it should be noted that, as with [ɔˑ] and [ɔ], both of its first formants are lower than those of the corresponding short [u] (cf. Fig. 10). Thus, the long [uˑ] has a considerably lower timbre and is more labialized. In tonality and tenseness, this vowel is very close to the corresponding primary cardinal ₈[uˑ], e.g.: T – −200 and −196, respectively; tens. – 1420 and 1415, respectively (cf. [u] T = −89, tens. = 910). Only in flatness is [uˑ] comparable to Jones’s secondary vowel ₁₆[uˑ] (b = 115). The short [u] is much less flat (b = 112); in this respect [u] coincides with the secondary unrounded ₁₇[ɨˑ] and rounded ₁₈[ɯˑ]. In terms of its position in acoustic vowel space, the second formant of long [uˑ] is closer to the primary eighth [uˑ]; in terms of the first formant, that is, vertical tongue displacement, [uˑ] occupies an intermediate position between the eighth [uˑ] and the front secondary ₁₀[øˑ], cf. [uˑ] F₁ = 330 Hz, F₂ = 660 Hz, ₈[uˑ] F₁ = 270 Hz, F₂ = 615 Hz, ₁₀[øˑ] F₁ = 350 Hz, F₂ = 2320 Hz. The short isolated [u] (F₁ = 400 Hz, F₂ = 870 Hz) is least distant in acoustic vowel space from the primary cardinal ₇[oˑ] (F₁ = 400 Hz) and the secondary ₁₅[ɣˑ] (F₁ = 450 Hz, F₂ = 850 Hz). Having reviewed data from studies of corresponding isolated vowels in other Lithuanian dialects, it can be stated generally that vowels pronounced in isolation in the Kučiūnai dialect of Southern Aukštaitian are likewise comparable in their acoustic properties with both primary and secondary Straipsniai / Articles 77 JURGITA JAROSLAVIENĖ cardinal vowels. The isolated vowels of the Pašušvys dialect of the Western Highland Šiauliškiai are, in this respect, closest to Jones’s secondary vowels. Extreme articulation is likewise not characteristic of the Lukšiai dialect of the Northern Kauniškiai. The same can be said of the Svirkai dialect of the Eastern Highland Vilniškiai, the Akmenė dialect of the Northern Samogitian Telšiškiai, and the Eržvilkas dialect of the Southern Samogitian Raseiniškiai. Comparative analysis of the data shows that the isolated high-pitched vowels of the Prienai dialect are pronounced with a higher timbre, and the low-pitched ones with a lower timbre, than, for example, in the Šakiai dialect of the Northern Kauniškiai. 2. RESULTS OF THE STUDY OF VOWELS IN CONNECTED SPEECH Below, the isolated vowels of the Prienai dialect are compared with the corresponding sounds produced in natural speech. The same allophones were selected for the study as for the experiment on isolated sounds. The rather diphthongal [æː], [æ.] and [e] produced in natural speech were also studied. Table 3 gives the mean values of the medians for all the realizations studied. 2.1. Comparison of the acoustic properties of isolated vowels and vowels in connected speech Both isolated vowels and vowels in connected speech have, in addition to an initial and a final (i.e., transitional) portion, a stationary sound portion, which may disappear entirely in natural speech for various subjective reasons (particularly in vowels produced at a faster tempo and also in unstressed position). Adjacent sounds exert their greatest influence on the transitional portion of vowels—the beginning and end of the sound (in this regard, cf. Jaroslavienė 2010: 238 ff.; see also Ladefoged, Maddieson 2002, and others). This is clearly illustrated by the spectrogram examples (see Figs. 14, 17–20; for spectrograms of vowels produced in isolation, see Figs. 1–13).9 The experimental results show that isolated vowels and vowels produced in natural speech differ to a greater or lesser extent (see Tables 1 and 3; cf. also Figs. 1–14, 17–20). In terms of timbre (cf. F2), vowels in connected speech first divide into high-timbre vowels ([iː], [i], [eː], [eˑ], [e.], [e], [æː], [æ.], [e]) and low-timbre vowels ([ɑː], [ɑ.], [ɑ], [oː], [ɔ], [uː], [u]), just like isolated vowels. It is noteworthy that the high-pitched [iː], [i], [eː] 9 Daugiau rišliosios kalbos balsių spektrogramų pavyzdžių žr. Jaroslavienė 2010: 74 tt. Diftongoidų [ie], [uo] ir žemutinių [a], [e] tipo balsių spektrogramų pavyzdžių ir kokybės bei kitų požymių tyrimą žr. Jaroslavienė 2010: 158 tt. 78 Acta Linguistica Lithuanica LXIV–LXV Prienų šnektos izoliuotieji ir rišliosios kalbos balsiai: akustiniai ir artikuliaciniai požymiai FIG. 17. Spectrogram of the word tĩsę. ~ tĩsę FIG. 18. Spectrogram of the word tìksi ~ tìksi FIG. 19. Spectrogram of the word tãso. ~ tãso FIG. 20. Spectrogram of the word tẽsę. ~ tẽsę the second formant is, respectively, 140 Hz, 40 Hz and 110 Hz lower, whereas [eˑ], [eˑ], [e] are, respectively, 60 Hz, 40 Hz and 40 Hz higher than those of the corresponding isolated sounds. In this respect, F₂ is also higher for all low-timbre vowels. Taking spectral spread into account, the vowels are distributed as follows: the highest first formants (from 510 Hz to 780 Hz) occur in connected-speech [a], [aˑ], [a] and [eˑ], [eˑ], [e], [æˑ], [æˑ], [e]; the lowest occur in [iˑ], [i], [uˑ] and [u] (from 330 Hz to 450 Hz). As with the isolated vowels, the connected-speech vowel [eˑ] also has a fairly low F₁ (cf. isolated and connected-speech [eˑ] F₁ = 360 Hz and 380 Hz). The isolated [aˑ], [aˑ], [a], [eˑ], [eˑ], [e] and [i] have higher F₁ values. The first formant is higher for the other vowels when produced in connected speech. Indeed, the non-rounded [oˑ] and [ɔ] of connected speech (as also when produced in isolation) occupy an intermediate position for the first formant; cf. [oˑ], [ɔ] respective F₁ values = 440 Hz, 500 Hz. As with the isolated vowels, the vowels with the highest tonality and least compactness, or greatest diffuseness, are [iˑ], [eˑ], [i] (cf. respective T values = 740, 679, 612; C values = 739, 759, 764); [o] and [uˑ] have the lowest degree of tonality (cf. respective T values = -202, -166), while the most compact vowels in the dialect are the low-pitched [aˑ] and [aˑ] (cf. respective C values = 909, 908). Straipsniai / Articles 79 JURGITA JAROSLAVIENĖ TABLE 3. Formants and evaluations of vowels in connected speech in the Prienai subdialect V | F₁ (Hz) | F₂ (Hz) | F₃ (Hz) | C | b | T | įt. [iˑ] | 330 | 2320 | 2970 | 739 | 107 | 740 | 1460 [i] | 360 | 1980 | 2660 | 764 | 108 | 612 | 780 [ɛˑ] | 380 | 2280 | 2870 | 759 | 106 | 679 | 1270 [ɛ] | 570 | 1830 | 2670 | 830 | 106 | 343 | 570 [eˑ] | 630 | 1750 | 2690 | 848 | 106 | 257 | 570 [e] | 510 | 1860 | 2640 | 815 | 106 | 410 | 510 [æˑ] | 610 | 1620 | 2660 | 851 | 106 | 209 | 390 [æ.] | 670 | 1640 | 2680 | 862 | 106 | 175 | 490 [e] | 570 | 1730 | 2700 | 835 | 106 | 289 | 500 [aˑ] | 770 | 1310 | 2580 | 909 | 106 | −64 | 540 [a.] | 780 | 1340 | 2530 | 908 | 106 | −41 | 470 [a] | 700 | 1370 | 2600 | 889 | 106 | 13 | 430 [oˑ] | 440 | 790 | 2260 | 890 | 112 | −202 | 1010 [ɔ] | 500 | 1040 | 2190 | 874 | 110 | −5 | 770 [uˑ] | 350 | 720 | 2170 | 866 | 114 | −166 | 1260 [u] | 450 | 920 | 2240 | 873 | 112 | −76 | 890 Vowels in connected speech, like isolated vowels, differ in their degree of flatness: [u] and [uˑ], as well as [ɔ] and [oˑ], are considered flat (cf. flatness indices 112 and 114, and 110 and 112, respectively), = whereas the corresponding isolated vowels—except [u], apparently because of more active lip movement—have higher flatness indices (cf. [uˑ], [ɔ], [oˑ]: flatness indices 115, 111, 113), = which means that they are more strongly labialized. The degree of flatness is the same for [u] pronounced in isolation and in natural speech. Compared with isolated [u], [uˑ], [ɔ], [oˑ], the corresponding vowels pronounced in natural speech also have a lower degree of tenseness, and the ratio of tenseness between long and short vowels is somewhat reduced in this respect; cf.: tenseness indices for [uˑ] and [u], 1260 and 890; = for [oˑ] and [ɔ], 1090 and 770. = They are not as highly tense as isolated and high-toned [i], [ɛˑ], and especially [iˑ] (see Tables 1 and 3 for vowel tenseness indices). The most tense of the isolated and connected-speech vowels is long [iˑ] (cf. tenseness indices for isolated and connected-speech [iˑ], 1890 and 1460). = For the other long, short, and half-long vowels, i.e. those of the [a], [e] type, tenseness is not as important a feature; however, other qualitative parameters differ—a fact that seriously encourages doubt as to whether vowel duration can indeed be regarded as a distinctive feature of phonemes. Other dialect researchers hold the same view (see Murinienė 2007: 127 et seq.). In summary, it should be noted that the acoustic features of vowels pronounced in isolation are generally more pronounced. The only exceptions are [e], [eˑ], [e]—when these vowels are pronounced in isolation 80 Acta Linguistica Lithuanica LXIV–LXV Prienų šnektos izoliuotieji ir rišliosios kalbos balsiai: akustiniai ir artikuliaciniai požymiai the place of more open articulation is emphasized; this is apparently also related to the lower value of the second formant for the corresponding vowels pronounced in natural speech. It is precisely for this reason that the isolated pronunciation of [æˑ], [æː], [ę̄] was not investigated. 2.2. Articulatory interpretation of vowels in connected speech The articulatory interpretation of vowels in connected speech is presented below. The sounds of connected speech in relation to isolated vowels are depicted in acoustic space in Figure 21. A comparison of isolated vowels and vowels in connected speech (cf. Fig. 21) shows that the same back vowels are distinguished by horizontal tongue displacement in natural speech conditions—the so-called low-timbre sounds [ɑˑ], [ɑː], [ɑ], [oˑ], [ɔ], [uˑ], [u], which are pronounced with the tongue retracted into the back part of the mouth. Fig. 21. Isolated vowels (●) and vowels in connected speech (▲) in the Prienai subdialect Straipsniai / Articles 81 JURGITA JAROSLAVIENĖ Their tonality is, as mentioned, very low and, with the exception of [a], expressed by negative values. The tongue is shifted toward the front of the mouth when pronouncing the so-called high-toned sounds [i], [i], [ę̇], [e], [e.], [e], [æ], [æ.], [e]; thus, they are to be considered front vowels—the tonality of all of them is expressed by positive values, and [i] is to be considered the most front. It should be noted that in connected speech [i], [i], and [ę̇] are sounds with a more retracted articulation, whereas [e], [e.], and [e] are fronter than the corresponding isolated vowels. In terms of vowel frontness, [æ], [æ.], [e] are articulated further back than [e], [e.], [e] and further forward than [a], [a.], [a]. In terms of vertical tongue movement, [u], [u] are to be considered high back vowels, and [i], [i] high front vowels. [ę̇] should also be assigned to this vowel height, but it is not as close and closed as the long [i], and therefore is formally considered a mid vowel. [o], [ɔ] are also assigned to the intermediate category. The other vowels—[e], [e.], [e], [æ], [æ.], [ę], [a], [a.], [a]—are considered low in terms of tongue height; however, for example, [a], [a.], [a] are not as distinct as the corresponding isolated sounds in their articulatory properties of height and frontness. It is interesting to note that the quality of not only isolated vowels but also low vowels of nasal origin and vowels lengthened under stress, as pronounced in sentences, differs to a greater or lesser degree. For example, the connected-speech vowel [a] has a somewhat more retracted articulation and is darker than [a.]. This positional-length vowel may be more strongly affected by adjacent sounds, since when pronounced naturally it is somewhat shorter than the long [a] (for more detail, see Jaroslavienė 2006: 245–264; 2010: 173–188). The same applies to front low vowels: the low vowel of inherent length has a lower timbre than [æ.]; in the case of [e] and [e.], the sound with the lower timbre is the positional-length [e.] (cf. also the corresponding results for the Lukšiai dialect (see Bacevičiūtė 2004: 176)10). 3. CONCLUSIONS Taking into account the acoustic and articulatory properties of isolated vowels and vowels in connected speech in the Prienai dialect, as well as the differences and similarities between Jones's cardinal vowel system and the vowel system of the dialect under study (see Figs. 16, 21 and Tables 1–3), and comparing the data from corresponding studies of the Prienai dialect and other Lithuanian dialects, several general conclusions can be drawn: • In terms of acoustic and articulatory properties, some isolated vowels of the Prienai dialect ([o], [u]) can be compared with primary ([₇øː], [₈uː]) vowels, while others ([i], [i], [ę̇], [a], [a.], [a], [ɔ], [u]) can be compared with both primary and secondary, or only secondary, cardinal vowels in Jones's 10 See also Girdenis 2000: 349; 2003: 194; 2005: 45–55. 82 Acta Linguistica Lithuanica LXIV–LXV Prienų šnektos izoliuotieji ir rišliosios kalbos balsiai: akustiniai ir artikuliaciniai požymiai cardinal vowels. Relative to the primary cardinal vowels, most vowels are shifted to a greater or lesser extent toward the center of the acoustic space (cf. especially the isolated [e˙], [e˙.], [e]). • The most extreme articulation in terms of frontness and height is characteristic of the mid back vowel [o˙] of the Prienai dialect—its spectral properties almost completely coincide with those of the primary vowel ₇[o˙]. The high back vowel [u˙] also tends, though to a lesser extent, toward extreme articulation; its primary cardinal counterpart ₈[u˙] is nevertheless more pronounced. • A comparison of the qualitative features of isolated sounds in the Kuciūnai Southern Aukštaitian, Pašušvys Western Aukštaitian Šiauliškiai, Lukšiai Northern Kauniškiai, Svirkai Eastern Aukštaitian Vilniškiai, Eržvilkas Southern Samogitian Raseiniškiai, Akmenė Northern Samogitian Telšiškiai, and Prienai Eastern Kauniškiai dialects reveals one common tendency: a preference for avoiding extreme articulations of cardinal vowels. Nevertheless, the isolated vowels of the Prienai dialect are the least distant from the corresponding primary sounds of Jones. • The vowels of the Prienai dialect differ when isolated and when pronounced in natural speech, but the relationships among the vowels remain the same: [i˙], [i], [e˙], [e˙], [e.], [e], [æ˙], [æ.], [e] are considered front vowels, while [a˙], [a.], [a], [o˙], [ɔ], [u˙], [u] are back vowels. According to the first-formant values, [i˙], [i], [u˙], [u] stand out as high vowels, joined by [e˙]; the vowels [o˙], [ɔ] are mid, while [e.], [e˙], [e], [æ˙], [æ.], [e], [a˙], [a.], [a] are considered low. • The acoustic and articulatory properties of most vowels pronounced in isolation are more pronounced than those of the corresponding sounds pronounced in sentences. A general tendency was observed for high vowels in connected speech to be pronounced more open, and low vowels more closed, than the corresponding isolated vowels. In addition, front vowels in connected speech (except [e˙], [e.], [e]) are slightly more retracted, while low vowels are fronter than the corresponding isolated vowels. Only the connected-speech vowels [e˙], [e.], [e] have a fronter articulation and a somewhat brighter timbre than the corresponding vowels pronounced in isolation. In the dialect under study, as in other Western Aukštaitian Kauniškiai dialects, [æ˙], [æ.], [e] are articulated further back than [e˙], [e.], [e] and further forward than [a˙], [a.], [a]. L I T E R A T U R E A t k o č a i t y t ė D a i v a 2002: Pietų žemaičių prozodija ir vokalizmas. Vilnius: Lietuvių kalbos instituto leidykla. B a c e v i č i ū t ė R i m a 2000: A study of isolated vowels in the Lukšiai subdialect. – Kalbotyra 48–49(1), 5–18. B a c e v i č i ū t ė R i m a 2004: Šakių šnektos prozodija ir vokalizmas. Vilnius: Lietuvių kalbos instituto leidykla. Straipsniai / Articles 83 JURGITA JAROSLAVIENĖ Clark John, Colin Yallop 1999: An Introduction to Phonetics and Phonology. Oxford, Cambridge: Blackwell. Crothers John 1978: Typology and Universals of Vowels Systems. – Universals of Human Language 2, ed. by J. H. Greenberg. Stanford (Calif.): Stanford University Press, 95–152. Fant 1964 – Фант Гуннар. Акустическая теория речеобразования. Москва: Наука. Girdenis Aleksas 2000: Kalbotyros darbai 1. Vilnius: Mokslo ir enciklopedijų leidybos institutas. Girdenis Aleksas 2001: Kalbotyros darbai 3. Vilnius: Mokslo ir enciklopedijų leidybos institutas. Girdenis Aleksas 2003: Teoriniai lietuvių fonologijos pagrindai. Vilnius: Mokslo ir enciklopedijų leidybos institutas. Girdenis Aleksas 2005: Adutiškio šnektos balsių [aː]/[aː] trukmė ir spektrai. – Baltistica 40(1), 45–55. Girdenis Aleksas 2008: Frydricho Kuršaičio priegaidės: (Akustinė rekonstrukcija). – Baltistica 43(3), 381–404. Jaroslavienė Jurgita 2006: Prienų šnektos žemutinių balsių kiekybė ir kiti skiriamieji bruožai. – Baltistica 41(2), 245–264. Jaroslavienė Jurgita 2010: Rytinių kauniškių Prienų šnektos fonologija (Daktaro disertacija). Vilnius: Vilniaus universitetas. Jaroslavienė Jurgita, Kaukėnienė Lidija 2004: Danielo Joneso kardinalinių balsių spektrinės charakteristikos. – Kalbotyra 53(1), 25–38. Jassem Wiktor 1973: Podstawy fonetyki akustycznej. Warszawa: Państwowe wydawnictwo naukowe. Kazlauskaitė Rūta 2002a: Pašūšvio šnektos fonologinė sistema: vokalizmas ir prozodija (Humanit. m. dr. disert.). Vilnius: Vilniaus universitetas. Kazlauskaitė Rūta 2002b: Pašūšvio šnektos izoliuotų balsių spektrų tyrimas. – Kalbotyra 51(1), 61–78. Ladefoged Peter 1975: A Course in Phonetics. New Yourk etc.: Harcourt & Jovanovich. Ladefoged Peter, Ian Maddieson 2002: The Sounds of the World’s Languages. Malden: Blackwell Publishers. Leskauskaitė Asta 2000: Kučiūnų šnektos izoliuotų balsių spektrinės charakteristikos. – Kalbotyra 48–49(1), 83–94. 84 Acta Linguistica Lithuanica LXIV–LXV Prienų šnektos izoliuotieji ir rišliosios kalbos balsiai: akustiniai ir artikuliaciniai požymiai Leskauskaitė Asta 2004: Pietų aukštaičių vokalizmo ir prozodijos bruožai. Vilnius: Lietuvių kalbos instituto leidykla. Murinienė Lina 1998: Akmenės šnektos izoliuotų balsių spektrinė analizė. – Kalbotyra 47(1), 91–105. Murinienė Lina 2007: Rytinių šiaurės žemaičių fonologija: vokalizmas ir prozodija. Vilnius: Lietuvių kalbos institutas, Vilniaus pedagoginis universitetas. Pakerys Antanas 2003: Lietuvių bendrinės kalbos fonetika. Vilnius: Enciklopedija. Piotrovskij 1960 – Piotrovsky Raimond G. Once again on the differential features of the phoneme. – Вопросы языкознания 6, 24–38. Urbanavičienė Jolita 2004: Svirkų šnektos izoliuotų balsių akustinė analizė. – Kalbotyra 53(1), 65–80. Urbanavičienė Jolita 2005: Svirkų šnektos (Rytų aukštaičių vilniškių) fonologinė sistema: Vokalizmas ir prozodija (Humanit. m. dr. disert.). Vilnius: Vytauto Didžiojo universitetas, Lietuvių kalbos institutas. Acoustic and articulatory features of vowels of the subdialect of Prienai pronounced both in isolation and in sentences SUMMARY The article analyses the main acoustic and articulatory features of vowels pronounced both in isolation and sentences of the Eastern Kaunas Prienai subdialect. The research materials were recorded in closed premises using a computer directly (using directional microphone). Every sentence or word was recorded 3 – 5 times under natural tempo of speaking. Speakers would pronounce isolated vowels for approximately one second. Sentences and words were further transferred to computer memory and saved using the .wav file format. Segmentation of the analysed elements was performed using the sound processing and analysis software PRAAT (5.0.34 version) of the program developed by University of Amsterdam researchers, P. Boersma and D. Weenink. In addition to the sound analysis software PRAAT, other programs were applied for evaluation of the experimental data. The research results show that isolated vowels differ from those articulated in a natural flow of speech in the subdialect of Prienai, yet the vowel interrelations remain unchanged: [i̇], [i], [ė], [e], [e̩], [ɛ], [æˑ], [æ.] and [e̩] should be regarded as front, and [ɑˑ], [a.], [a], [ȯ], [ɔ], [u̇] and [u] as back vocalic sounds. High-rise vowels are [i̇], [i], [u̇], [u]; the sound [ė] is adjoined to them; middle-rise vowels are [ȯ] and [ɔ], and the vowels [e̩], [ė], [e], [æˑ], [æ.], [e̩], [ɑˑ], [a.] and [a] should be treated as being low. Straipsniai / Articles 85 JURGITA JAROSLAVIENĖ The acoustic and articulatory features of the majority of sounds pronounced in isolation are more prominent in comparison with the respective sounds pronounced in sentences. A general tendency was noticed that high vowels are pronounced more openly in connected speech, and low vowels are respectively closer than respective isolated vowels. Besides, front vowels (except [ɛ], [e.] and [e]) are slightly more back and low ones are more back in connected speech as compared with their isolated counterparts. Only the vowels [ɛ], [e.] and [e] have a more front articulation and a slightly lighter pitch than respective vowels pronounced in isolation. The performed comparison between spectral characteristics of isolated vowels of the subdialect of Prienai and those of cardinal vowels by Daniel Jones allows stating that some isolated vowels of the subdialect of Prienai ([o̞], [u̞]) should be contrasted with the primary vowels (₇[oː], ₈[uː]), while others ([iː], [i], [ɛ], [a], [ɑ], [ɐ], [ɔ], [ʊ]) with both primary and secondary or secondary cardinal vowels by Jones both according to their acoustic and articulatory features. It is complicated to distinguish which cardinal vowels could be counterparted with the isolated vowels [ɛ], [e.] and [e], as they are located in the middle section of the acoustic space of sounds according to their rise. Compared with primary cardinal vowels, the majority of vowels are shifted in the middle direction of acoustic space. Extreme articulation according to the horizontal and vertical position of the tongue is the most typical of the back middle-rise vowel [oː] in the Eastern Kaunas Prienai subdialect, as its spectral features almost totally coincide with those of the primary vowel ₇[oː]. The back high-rise vowel [uː] is only slightly less inclined to extreme articulation – its primary counterpart ₈[uː] is still more prominent compared with it. Having compared the qualitative features of isolated sounds from the Kučiūnai subdialect of Southern Aukštaitian, the Pašūšvys subdialect of the Šiauliai dialect of Western Aukštaitian, the Lukšiai subdialect of Northern Kaunas, the Svirkai subdialect of the Vilnius dialect of Eastern Aukštaitian, the Eržvilkas subdialect of the Raseiniai dialect of Southern Samogitian, the Akmenė subdialect of the Telšiai dialect of Northern Samogitian, and the Prienai subdialect of Eastern Kaunas, a general tendency was observed: an inclination to avoid extreme articulation of the cardinal vowels. However, isolated vowels in the Prienai subdialect deviate least from their counterparts in Jones's system. Received on October 24, 2011. JURGITA JAROSLAVIENĖ Institute of the Lithuanian Language P. Vileišio g. 5, LT-10308 Vilnius, Lietuva [email protected] 86 Acta Linguistica Lithuanica LXIV–LXV