Lietuvių kalbos trumpųjų ir ilgųjų balsių kiekybės ir kokybės etalonai

JURGITA JAROSLAVIENĖ Institute of Lithuanian Language  QUANTITY AND QUALITY OF SHORT AND LONG VOICES IN THE LITHUANIAN LANGUAGE STANDARDS  1. INTRODUCTION There are not many consistent instrumental studies of the quantity and quality of short and long vowels in the common Lithuanian language. However, both the old (carried out in the second half of the 20th century) and modern methods[1] are still relevant not only from the point of view of pure linguistics. All of them differ in their objectives and tasks, research material, sample, sound recording and examination methodology, number of informants, gender, etc. Therefore, as is known, when carrying out new experimental studies and comparing the data with the corresponding data from other studies, it is important to remain methodical, because the quantitative and qualitative parameters of vowels depend on various linguistic and non-verbal factors, which without taking into account can be obtained very contradictory data and accordingly draw contradictory conclusions. Mostly, various researchers have studied and described in one aspect or another the pronounced and (or) unpronounced vowels of the common Lithuanian language, pronounced by the informant men in single-syllable, two-syllable, three-syllable and four-syllable words (see Pakerys 1982: 31–34 and others. ; Kliukienė 2002: 73–78, 2003: 81–91, 2005: 361–367; Girdenis 2003: 222; Kaukėnienė 2002: 41–59, 2003: 35–47, 2004a, 2005: 65–71; Балшайтите 2004: 169–175; Kliukienė, Raudžiūtė 2013: 111–119 et al. (Some of the results and conclusions of spoken research will be discussed later in this article.) However, we do not have comprehensive synchronous (and diachronic) instrumental and typological studies of the quality and quantity of vowels in spoken language. More consistent studies of certain sounds, their characteristics, aspects in the context of the change of the official (public) Lithuanian language, etc. (cf. Kruopienė, Kačiuškienė 2008: 38–42; Vaicekauskienė 2011: 191–207; Nevinskaitė 2013: 118–136; Čičirkaitė 2014: 45– 58 and others.) In other words, there is still a lack of a more detailed instrumental analysis of the influence of various factors (public and non-public spoken language) on the quantity and quality of vowels. After all, speech acts may contain less acoustic information than it would be necessary to accurately recognize certain phonemes, and everything that is actually missing is usually reproduced according to other implied information – context, content of statements, etc. (Girdenis 2003: 224 and min. lit.). Furthermore, researchers have pointed out that in certain cases it is not enough to study only the length of vowels, because it is the quality that can be more stable than the length (cf. Vaitkevičiūtė 2001: 79–80; Kazlauskienė 2002: 33–35, 2009: 27; Kaukėnienė 2004a: 62–63; also see Pakerys 1982: 94, 2003: 26). For example, it has been established that the contrast of the quantity of specific vowels varies greatly, and statistically significant can be even a difference of 15-20 ms in the length of vowels (long vowels in this case are approximately 1.1-1.2 times longer than short ones), if the experiment was carried out properly and methodically, i.e., selected speakers of the same tempo, they read the comparative words (preferably of the same or very similar phonological structure) with the same intonation, under similar conditions, etc. (Kazlauskienė 2002: 34; cf. Kaukėnienė 2003a: 46, 8 p.). The sounds that could be used as a basis for the analysis of the above-mentioned variantity of vowel allophones, especially when explaining the sensitive issues of the quantity and quality of vowels in the norm of pronunciation[3], and which could be considered as prototypes of common language or as a goal when speaking about exemplary pronunciation as a value (consonant culture[4]) and language policy in the second decade of the 21st century in general[5], have not been studied in isolation[2], i.e. separately pronounced. Recent scientific research shows that, from the perspective of geolinguistics, one of the most important features of modern society is literacy: on the one hand, the common language functions in communities as a linguistic factor (i.e. a condition) that determines the change and transformation of regional variantity, and on the other hand, the image of common language as a prestigious code has no less influence on the development and change processes of regional language variants (for more information see Aliūkaitė, Mikulėnienė 2014: 257–262; also cf. Zabarskaitė 2015; Mikulėnienė 2015 and others). In view of the need for research to help determine what vowel length ratios should or could be considered as the norm of common language and what exemplary (articulatory) standards of vowel quality are recommended, the aim of this article is to examine isolated (without any context, as naturally as possible) pronounced long and short vowels of constant articulation of common Lithuanian as possible standards of common Lithuanian pronunciation, investigating both the quantitative opposition of vowels (vowel length ratios) and the most important qualitative characteristics, based on instrumental research data[6]. The standard of general Lithuanian pronunciation is also necessary to improve the defective Lithuanian language teaching methodology: it is desirable that the audio teaching tools meet the minimum requirements of general Lithuanian pronunciation (see Miliūnaitė 2013: 208)[7]. In this article, the common Lithuanian language is considered to be the language variant that is used as a standardized language variant for the purposes of public life and culture (see LKE 2008: 87–91). As is known, it has been formed on the basis of the western Kaunas dialect of the Alytaus region[8].  2. RESEARCH MATERIAL AND METHODOLOGY There are 14 vowel phonemes in the common Lithuanian language (DLKG 2005: 23):   These separately (isolately) pronounced long and short vowels of the present common Lithuanian language (the short [ɔ] is also  assigned to the periphery of the phonological system) are the object of this study (see also table 1). However, during the study it became clear that the informants do not imagine the short [eerd] (the “two-faced” phoneme /erd/ variant, or otherwise – allophone) as a separate, i.e. isolatedly pronounced, and tend to identify it with the simple [e] (phoneme /e/ variant), therefore the isolated [eerd] was not pronounced for the study without any context[9]. It is believed that the non-superior (medium) ascending vowel [eerd] is prevented from establishing itself in the common language system by its specific use (moreover, its purpose is not distinctive), its relatively low frequency (facultativity) and the fact that it is marked in writing with the same letter as the short [e] of the lower ascending vowel (for more details see Pakerys 2003: 32–34 and min. lit.; also cf. Girdenis, Pupkis 1978: 61–62; Girdenis 2003: 191) and for other reasons. No other letter is used to mark two short vowels of different quality, cf. table 1 (cf. Pakerys 2003: 29, 2 scheme). Outside the boundaries of this study, ie, uo remain isolated in the phonetic system of the Lithuanian language, and according to recent experimental studies (see Girdenis 2009: 213–242), ie, uo function as independent long uppercase phonemes of variable elevation (Girdenis 2009: 240). They are not examined in this article because of a different methodology of research and description. (2005: 23–26) and in Aleksas Girdenis’ Theoretical Foundations of Lithuanian Phonetics (2003: 209, table 23) the variable articulations /ie/, /uo/ are assigned to non-ascendent vowels (also cf. Bacevičiūtė 2002; Pakerys 2003; Kaukėnienė 2003b, 2004a). However, it should be noted that the assignment of the variable articulated long vowels /ie/, /uo/ to vowel phonemes of one or another elevation is conditional. As can be seen in Table 1, the common Lithuanian vowel phoneme in red is also written in the International Phonetic Alphabet (IPA) characters in square brackets (see IPA 2005). Thus, the traditional Lithuanian vowels of constant articulation /i, e, a, ɔ, u, i·, ų·, e·, a·, o·, u·/ are represented by the following TFA characters: /ɪ, ɛ, ɐ, ɔ, ʊ, iː, eː, æː, ɑː, oː, uː/. These vowels (i.e., except for /an/, /ie/, /uo/ – according to TFA /e/, /iɛ/, /uɔ/ respectively) will be described in the following using TFA characters.  TABLE 1: Phonetic system of Lithuanian vowels (traditional characters in black, TFA in red) and letters representing vowels (green)[11]  The International Phonetic Alphabet (IPA) is a standardized universal alphabet used to describe and study the pronunciation of sounds in languages around the world. The International Phonetic Alphabet (IPA) has been adapted for Lithuanian sounds, for example, in the 1997 edition of the Lithuanian Grammar (LG 1997).[1] This article uses a slightly modified TFA transcription, i.e. it is based on the newer (2005) version of the TFA. In table 1, vowels are divided according to differential characteristics (see Girdenis 2003: 191, etc.)[12], which will be discussed in more detail when discussing the results of the spectral analysis. The study included 12 informants (6 women and 6 men, aged from 20 to 50 years), representatives of the common language without speech defects (most of them have a higher philological education, all of them use the common language daily or almost daily not only in the public space, but also speak one or several dialects, foreign languages) from all the regiolects of the Lithuanian language area (Figure 2, for more details see Aliūkaitė, Mikulėnienė 2014: 257, etc.), so that the research material would be as objective as possible and the results obtained would reflect as realistic as possible the situation of the pronunciation of vowels in the common Lithuanian language. (and graphs) are aggregated data (average) of all respondents.   2 PAV. Regions of the Lithuanian language area[13] The recordings were made in enclosed rooms with a professional  dictaphone (Tascam HD‑P2) and a directional, headband-mounted vocal microphone (AKG C 520). The informants were asked to pronounce the material at normal speed and as naturally as possible. The recordings were then transferred to computer memory and saved in.wav format (a total of over 5,000 files). The segmentation of the examined elements was carried out with free computer sound processing and analysis programs: Praat 5.3.63 and later versions (developed by Paul Boersma and David Weeninkas) and WaveSurfer 1.8.8.p4 (developed by Kåre Sjölander and Jonas Beskow). For further analysis of the instrumental data, MS Excel was used, for example: according to the formula of Anthony Bladon and Gunnar Fant (Bladon, Fant 1978: 3; see also Fant 1983: 7) statistical averages (in hertz and barts) were calculated, standard deviation (in hertz), effective value of the second form (F2′) (in barts), which requires not only the first two forming values, which usually determine the basic acoustic characteristics of vowels, but also the third, fourth forming values and the main tone (frequency).The results were converted into barts according to Hartmut Traunmüller's formula (Traunmüller 1988: 97).Graphically, vowels are represented not only in acoustic (in hertz), but also in psychoacoustic F2'/F1 space (barts scale) – it is believed that such a picture of the vocal system more accurately conveys the peculiarities of human hearing, i.e. In the psychoacoustic space (see Figures 3, 4; also cf. Figure 5) the vowels under study are marked with circles with a diameter of 1 bar, the distances between which indicate the degree of difference in the perceived quality (Iivonen 1987). In the description of the vowel system of the common Lithuanian language, the tonotopic distances (bars) between the first form and the main frequency (tone) (F1f0) and between the first and second forms (F2F1) are also used: such differences are closely related to the openness of the vowels and to the characteristic of the anterior and posterior order of the vowels (Miller 1989: 2119). The values of the measured formants (generalized data) and acoustic parameters (compactness, bemolarity, tonality, tension) of female vowels are presented both original, i.e. unnormalized (Table 4; for male pronunciation data see Table 7), and normalized according to coefficients of 17% (Table 3) and 14% (Table 4) in order to compare the results obtained as objectively as possible with the corresponding male pronunciation data (for universal coefficients see Fant 1975: 8).The normalization coefficients are calculated based on the data of [iː] second formants and on the basis of specific formants data that depend on the entire vocal tract or on a particular cavity (see Fant 1975: 8; Grigorjevs 2008: 45‑61). In order to better describe the qualitative relationships between long and short vowels in Lithuanian, the FORMANT2 program was used. PAS[14], which calculates the following acoustic parameters (numeric values) on the basis of the average values (Hz) of the first, second and third formants: bemolarity, compactness, tension and tonality (see tables 4 to 7). As mentioned in the introduction of the article, the results of the quantity and quality of voices are compared with some previous studies (Table 3, Figure 7).  3. ANALYSIS OF THE RESULTS OF THE INVESTIGATION The results of the measurement of the length of vowels (see table 2) show that the vowels [ɐ], [ɛ] are the longest of the short vowels and [æː], [ɑː] are the longest; the short vowels [ɪ], [ʊ] are the shortest of the short vowels and [iː], [uː] are the longest. The other vowels ([ɔ] of the short vowels and [eː], [oː] of the long vowels respectively) occupy an intermediate position in terms of length: they are shorter than the corresponding short and long low vowels, but longer than the corresponding high vowels. Pakerys 1982: 43–48 and min. lit. ).  TABLE 2: Total number of vowels in Lithuanian language: aggregated male and female data (female vowels N = 60, male vowels N = 72)   As can be seen from the summarized results presented in Table 2, long vowels pronounced by women and men are 2.1 times longer than short vowels. The confidence intervals of the corresponding short and long vowels do not overlap (this and some other statistical information described below are not presented in the table), and the difference in their quantity is statistically significant (the highest confidence level). When comparing the data of men and women, a tendency was highlighted that men pronounce isolated vowels longer, but this difference is small and statistically insignificant, and the confidence intervals coincide in many places. However, it was noted that the number of female vowels varies more (from 7 to 20%) than men (from 6 to 14%).  TABLE 3: The ratio of the total number of short and long vowels in Lithuanian according to some previous studies  POSITION A. PAKERYS (1982: 93, 95, 112) L. KAUKĖNIENĖ (2004a: 62) D. BALŠAITYTĖ (2004: 174) Pronounced short: long (cross-short: acute: circumflex) 1:1,72:2,09  1:2,2  1:1,94 Short: long, unsharpened 1: 1,76 1: 1,4 –  As mentioned at the beginning of this article, other researchers have focused on the number of syllables in a single word (from one to four) rather than on the number of sounds in a single word. However, as we can see (see table 3), the ratios of the quantity of pronounced short and long vowels (except for acute vowels) are quite similar to those described in this article. After examining the results of different studies in more detail, it became clear that the ratios of the length of short and long vowels in words range from 1.9 to 2.4 times, only acute vowels are longer than pronounced short vowels from 1.5 to 1.9 times. In the unpronounced position, the differences between short and long vowels may be small at all, but they turn out to be statistically significant: ranging from 1.1 to 1.9 times. Nevertheless, it is important to base the phonological opposition of phonemes on differences in quality, taking into account various linguistic and non-linguistic factors. As is known, phonological elements of language are absolute, not relative: sounds of similar or identical spectral structure can be perceived differently if their auditory environment differs, because, as mentioned, in certain cases, the acoustic information may be less than it would be necessary to accurately recognize phonemes (Girdenis 2003: 224).  3.2. MAIN QUALITATIVE CHARACTERISTICS OF GENERAL VOCAL SPECTRUMS OF LITHUANIAN LANGUAGE The analysis of the results showed that the vowel spectrum depends on the specific speaker, and the differences in the vowel spectrums pronounced by male and female are also influenced by the smaller resonator size of females. However, it should be noted that some relatively constant relationships always remain (cf. Girdenis 2003: 222). In this article, only generalized data on female and male pronunciation are described. As can be seen, after the speakers were normalized, the differences in female and male voices decreased to a minimum (cf. table 4 and tables 5–7). Table 5 was selected to describe the results of female pronunciation, the other tables (i.e. tables 4 and 6) are presented for comparison, and they could also be useful for carrying out new instrumental studies. The results of male pronunciation are presented in table 7 and figure 4. The first form of [iː], [eː], [ɪ] pronounced by all informants (both male and female[15]) is higher than 2000 Hz, and the first form is very low: from 285 Hz to ∼450 Hz (see tables 4–7). After examining the distance between the first and second forms and between the second and third forms, it was observed that of all the vowels (both long and short) in Lithuanian the greatest distance is between the vowels [iː], [eː], [ɪ] first and second forms, and the smallest distance is between the second and third forms of these vowels. Of all the long vowels pronounced by all informants (both male and female[16]): [uː], [oː], [ʊ], [ɔ] both forms are relatively low (see tables 4–7): from about 300 Hz to 600 Hz F1 and from 630 Hz to 1100 Hz F2. Of all the long vowels, the smallest distance is between the vowels [oː] and [uː] first and second forms, and the greatest distance is between the second and third forms of these vowels. The same pattern is seen in the case of the short vowels [ɔ] and [ʊ]. [æː] and [ɛ] forms are higher at about 1700 Hz and above, i.e. higher than the second [ɑː], [ɐ] forms and lower than the second [iː], [eː], [ɪ] forms (see tables 4-7). Of all long and short vowels, the third form of [ɑː] and [ɐ] is the closest to the first, and the third form of [iː], [eː] and [ɪ] is the furthest from the first.  TABLE 4: Common values (in hertz) of short and long vowel forms in Lithuanian and their assessments: unnormalized female pronunciation data   TABLE 5: General values of the short and long vowel forms of Lithuanian language (in hertz) and their assessments: normalized female pronunciation data (k = 17 per cent)   TABLE 6: General values of the short and long vowel forms of Lithuanian language (in hertz) and their assessments: normalized female pronunciation data (k = 14 per cent)   TABLE 7: Common meanings of short and long vowel forms in Lithuanian language (in hertz) and their assessments: summary of male pronunciation data   The acoustic parameters of the vowels calculated from the average values of the first three forms (see tables 4–7) are associated with the following characteristics: ‘high’ – ‘low’, ‘compact’ – ‘diffuse’, ‘bemolar’ – ‘ordinary’, ‘tense’ – ‘non-tense’. According to the pitch (timbre), isolatedly pronounced [iː], [eː], [æː] and [ɪ], [ɛ] are considered to be high, high-timbre sounds (their second form is higher than 1500 Hz, and the degree of tonality is positive[16]), compared with the corresponding low-timbre[17], or, more precisely, non-high vowels [ɑː], [oː], [uː] and [ɐ], [ɔ], [ʊ] (the second form of these vowels is less than 1500 Hz, and the degree of tonality of all low-timbre sounds is negative, except for female [ɐ]) (see also the phonological classification of vowels in table 1). According to male and female pronunciation data (see second form and degree of tonality in the tables), of the long vowels of the common Lithuanian language, the highest timbre is [iː], and the lowest timbre vowels are [uː] and [oː]. Of the short vowels, the highest timbre is [ɪ], and the lowest timbre vowels are [ʊ] and [ɔ]. It should be emphasized that the lowest timbre vowels also have the highest bemolarity index (see tables 4–7). According to acoustic classification, vowels [oː], [uː], [ɔ], [ʊ] are considered bemolar: [u:] has the highest bemolarity index, so it can be assumed that the lips are most active when pronouncing this sound. The most compact sound is [ɑː]: as mentioned, the first two forms of this vowel are not far from each other and generally from the central part of the spectrum (the numerical value of compactness is the highest). At the very edges of the spectrum are [iː] and [uː] – non-compact, or the most diffuse, vowels (their spectral energy is scattered at the edges of the spectrum, and the numerical values of compactness are the lowest). Of the short vowels, the most compact is the low [ɐ] (the highest compactness index of all short vowels), and at the edges of the spectrum, but slightly closer to the center than the diffuse long [iː] and [uː], are the first and second forms of the short [ɪ] and [ʊ]. The numerical values of long and short vowel stress in Lithuanian (see stress index in tables 4–7) show that the long vowels [iː], [uː], [oː] have a much higher average value of this parameter than the corresponding short vowels [ɪ], [ʊ] and [ɔ]. However, as could be expected (see also Girdenis 2003: 222, 226), the differences in stress between long [æː], [ɑː] and their short counterparts [ɛ], [ɐ] are considerably smaller.  Pakerys A., Pupkis A. 2004: Common pronunciation of Lithuanian language: texts of a CD, Vilnius: Gimtasis žodis. RocaI.,JohnsonW.1999:Acourseinphonology,BlackwellPublishingLtd. TraunmüllerH.1988:Analyticalexpressionsforthetonotopicsensoryscale.–Perilus8,Stockholm: University of Stockholm, 93–102. Vaicekauskienė L. 2011: The official Lithuanian language is a dead language: Journalists on the Lithuanian language standard on television and radio. – Darbai ir dienos 55, 191–207. Vaitkevičiūtė V. 2001: Basics of Lithuanian Pronunciation and Dictionary, Vilnius: Pradai. "Decision of the State Commission for the Lithuanian Language on the Minimum Requirements for the Lithuanian Common Consonant". Access on the Internet: http://www.vlkk.lt/aktualiausios‑temos/tartis‑ir‑kirciavimas/apie‑skyriu‑tartis‑ir‑ kirciavimas (accessed 2015‑11‑10). Decision of the State Lithuanian Language Commission “On the Approval of the Programme for the Use, Standardization and Dissemination of the State Language”, 2015‑04‑30, No. N‑2 (157). Available online: http://www.vlkk.lt/vlkk‑ nutarimai/nutarimai (accessed 2015 11 10). Zabarskaitė J. 2015: World Literacy Day: The Relationship of Lithuanian Language and Digital Society Today, 2015‑09‑08, article prepared by Dovilė Balčiūnaitė. Accessed online: http://www.fm99.lt/pasauline‑ rastingumo‑diena‑lietuviu‑kalbos‑ir‑skaitmenines‑visuomenes‑santykis‑siandien/ (accessed 2015 09 11). БалшайтитеД.2004:Спектральныехарактеристикигласных(монофтонгов)современного Lithuanian Literary Language. – Respectus philologicus 5(10), 169–175. Fant G. 1964: Acoustic theory of speech formation. Moscow: Science. ПиотровскийР.Т.1960:Ещеразодифференциальныхпризнакахфонемы.–Вопросыязыкознания6, 24–38.  Received 17 November 2015 Adopted on 16 December 2015   JURGITA JAROSLAVIENĖ Institute of Lithuanian Language Petro Vileišio g. 5, LT–10308 Vilnius, Lithuania [email protected] [1] Modern experimental sound studies (usually investigating the acoustic characteristics of sounds) are usually carried out with special devices or computer programs. The most important methods are the analysis of oscillograms, spectrograms, measurements of sound duration, basic tone and intensity. If auditory experiments are not discussed, the term instrumental phonetics is used to describe instrumental tests only (Bacevičiūtė 2009: 202). [2] It should be noted that the vowels pronounced isolatedly at the natural pace of speech are not to be confused with Daniel Jones’ cardinal (or otherwise – cardinal) vowels (Ladefoged 1975: 194–199; Jassem 1973: 190, table 13.1) and the isolated vowels of Lithuanian (and other languages) and dialects pronounced according to special requirements: the cardinal vowels are used as theoretical constructs, a kind of reference system for describing and comparing the vowels of specific languages or dialects, but do not exactly match the corresponding sounds of any particular language (e.g., cf. Roca, Johnson 1999: 126; Jaroslavienė, Kaukėnienė 2004: 25–45; Kaukėnienė 2004b: 199– 211; Murinienė 2007: 116–118; Girdenis 2008: 29–32 and others). [3] Among the most basic pronunciation requirements of the Lithuanian language, we can mention at least a few: to distinguish short vowels from long ones in both accentuated and unaccentuated syllables; to distinguish vowels e, ė, o from ie, uo in accentuated and unaccentuated syllables; to distinguish the firm-started prefix of diphthongs and mixed diphthongs from the firm-started prefix; to observe the rules of pronunciation of hard and soft consonants, etc. ). For more information about pronunciation rules see. Decision of the State Lithuanian Language Commission (hereinafter – VLKK) “On the Minimum Requirements for Lithuanian Common Consonants”; see also Girdenis, Pupkis 1978; Girdenis 2003: 24–25; Pakerys, Pupkis 2004 and others. [4] Or in other words – the language of culture (see Mikulėnienė 2015), as a sign of added value, represented not only by language cultures knowledge, but also correct, clear pronunciation. [5] As is well known, the main objective of the state language policy is to strengthen the status and prestige of the state language, to promote its development, to ensure the functionality of the Lithuanian language in all areas of public life, to maintain it on the language market (see the VLKK resolution “On the approval of the programme for the use, standardization and dissemination of the state language”, 2015‑04‑30, No. N‑2 (157)). [6] The results discussed in this article were calculated at the Institute of Lithuanian Language in 2013–2015 during the implementation of the project “Contemporary Baltic languages: spectral sound characteristics (instrumental study)”, funded by the Lithuanian Scientific Council. Juris Grigorjevas, Dr. Inesė Indričānė, Dr. Jurgita Jaroslavienė (project leader) and Dr. Jolita Urbanavičienė. [7] The issue of correct, clear pronunciation is relevant not only for Lithuanian, which is classified as one of the least spoken languages in the world, but also for major languages such as English. For example, see online English Phonetics and Phonology <[https://www.facebook.com/English‑Phonetics‑and‑Phonology‑268299689848577/?fref=ts]> (accessed 2015‑10‑28). [8] There is still a lack of extensive discussion and established opinion on the current concept of common language, standard language, plain language, usage and areas of use, and the relationship with the state language. For a brief overview of the usage of such terms as written language, common language, literary language, standard language, see Cidzikaitė 2002: 280–282 and there min. lit. [9] In order to investigate whether the pronunciation (especially the quality) of the vowel [ad] (i.e. phoneme /ad/ allophone) of foreign-origin words in similar positions differs from the simple short [e] (i.e. phoneme /e/ allophone), an instrumental and auditory examination is planned with the selection of appropriate research material. According to Antanas Pakerius (2003: 33), the functional independence of the vowels [ad] and [e] can be verified by comparing the word forms gèst (object-verb gèst dgs. origin.) and gèst (verb gèsti conjugation): 75–80 per cent of the vowels [ad] and [e] are independent. listeners correctly recognized those isolated words. [10] Examples of isolated long vowels with variable pitch can be heard by pressing “[ie]” and “[uo]” respectively. [11] The marking of the softness sign before the letters of the a, ą type is symbolic (to maintain consistency, we should also write after the letters of the ú, ů, u, o type), in order to show that in this case the sound is used after the soft consonant, and therefore its quality changes. Two short vowels of different quality marked equally in the writing are outlined in red (for more on the letters see Pakerys 2003: 29, 3 exn. and min. lit. ; Girdenis 2003: 194; DLKG 2005: 24–25). [12] As is known, only such a classification of phonemes and such a system of differential features and hierarchy can be effective, which allows to describe as consistently and simply as possible the relationships of all phonological elements and the functioning of the entire phonological system (Girdenis 2003: 201pp.; also cf. Pakerys 2003: 32–37; DLKG 2005: 21–26). [13] Picture taken from Lithuanian dialects of the beginning of the 21st century: geolinguistic and sociolinguistic research: maps and their commentaries, compiled by Danguolė Mikulėnienė, Violeta Meiliūnaitė, authors D. Mikulėnienė et al., Vilnius: Briedis, 2014. [14] FORMANT2 software. PAS was developed by Alex Girdenis, based on Raymond Piotrovsky (Пиотровский 1960: 24-38) proposed method. [15] The results of unnormalised and normalised data are not discussed in detail. [16] The higher the timbre, the greater the numerical value of the second formant and the tonal coefficient (degree) indicating the pitch of the sound. [17] Lithuanian vowels [ɔː], [uː], [ɔ], [ʊ] (as well as [uɔ]) are realized by two kinds of allophones: both hard and soft consonants (and /j/) can precede these vowels. After soft consonants (and /j/) are used contraction variants of vowels: at the beginning of their pronunciation, the tongue is significantly moved forward, but then usually retreats to the position of the main (‘ordinary’) allophones (see Girdenis 2003: 192 and min. lit., 229; Jaroslavienė 2014: 76–77). [18] Figure drawn from https://www.internationalphoneticassociation.org/content/ipa‑chart (accessed 02/09/2015). Labialized (labial) vowels are located on the right side of each margin, unlabialized (non-labial) on the left. [19] In the figure, the circles represent long vowels, the triangles represent short vowels. 