Dar kartą apie dvigarsių dėmenų i, u kiekybę: ar tikrai neilgėja bendrinei kalbai artimiausiame šnektų areale?

Bendrinė kalba / Standard Language 2025, Nr. 98, p. 1–16 Turinio nuoroda / Contents link ISSN 2351-7204 (elektroninis / online) ONCE AGAIN ON THE QUANTITY OF THE ELEMENTS OF MIXED DIPHTHONGS i, u: DO THEY REALLY NOT LENGTHEN IN THE DIALECT AREA CLOSEST TO THE STANDARD LANGUAGE? Once Again on the Quantity of Elements i, u in Mixed Diphthongs: Do They Really Not Lengthen in the Dialect Area Closest to the Standard Language? RIMA BAKŠIENĖ Institute of the Lithuanian Language Email: [email protected] ORCID ID: https://orcid.org/0000-0002-6716-0776 Research interests: experimental phonetics, phonology, geolinguistics. https://doi.org/10.35321/bkalba.2025.98.05 ABSTRACT The article analyses the quantity of the elements i, u in the northwestern part of the Western Aukštaitian subdialect of Kaunas, encompassing the local dialects of Jurbarkas, Šakiai, and Vilkaviškis, whose phonological and prosodic structure is most often relied upon in codifying the norms of standard pronunciation. The standard pronunciation of acute diphthongs of this type was discussed more actively several years ago, when a set of International Phonetic Alphabet (IPA) symbols was proposed for Standard Lithuanian. At that time, it was envisaged that the first elements could be either short or semi-long in less formal speech. However, the question remains unresolved, making it important to continue investigating the prosodic features of these diphthongs in local dialects traditionally considered an area where the elements i, u do not lengthen, setting it apart from the Lithuanian language system as a whole. Based on the results of studies conducted using experimental phonetic methods and evaluated with Student’s t-test, the article concludes that in the northern part of the Kaunas subdialect, in the Jurbarkas area and partly in the Šakiai surroundings, the elements i, u in both acute and circumflex diphthongs should be considered semi-long, while the lengthening of acute elements in the area described becomes more pronounced northward. The established prosodic patterns of the Kaunas subdialect should undoubtedly also be taken into account when proposing standard pronunciation variants. KEYWORDS: Lithuanian dialects, Western Aukštaitian subdialect of Kaunas, pitch accent, mixed diphthong, vowel quantity. ABSTRACT The article analyses the quantity of the elements of mixed diphthongs i and u in the northwestern part of the Western Aukštaitian subdialect of Kaunas, encompassing the local dialects of Jurbarkas, Šakiai, and Vilkaviškis. The phonological and prosodic structure of these regional dialects is most commonly used as a basis when codifying the norms of standard pronunciation. There is still ongoing debate about the established pronunciation of the acute syllables of this type in the standard language (whether they are short or semi-long), so it is essential to identify their prosodic features in those local dialects that have traditionally been considered an area where the elements i and u do not lengthen and which stands out from the rest of the Lithuanian language system. Based on the findings of studies conducted using experimental phonetic methods and evaluated with the statistical Student’s t-test, it can be concluded that in the northern part of the Kaunas subdialect, specifically the Jurbarkas area and partly the surroundings of Šakiai, the elements i and u in both acute and circumflex mixed diphthongs should be considered semi-long. Moreover, the lengthening of acute elements Copyright © 2025 Rima Bakšienė. Published by the Institute of the Lithuanian Language. This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. // Published by the Institute of the Lithuanian Language. This article is open access, distributed under the terms of the “Creative Commons” Attribution Licence, which permits unrestricted use, distribution and reproduction of the content in any medium, provided that the author and source are indicated. Received: / Received: 2025-04-22. Accepted: / Accepted: 2025-05-25. 1 becomes more pronounced as we move northward within the described area. The identified prosodic patterns of the Kaunas subdialect should undoubtedly be taken into account when establishing standard pronunciation variants. KEYWORDS: Lithuanian dialects, Western Aukštaitian of Kaunas, pitch accent, semi-diphthong, vowel quantity. INTRODUCTION In the work of dialectologists, the northern subdialects of Western Aukštaitian Kaunas are universally identified as the closest to Standard Lithuanian (Zinkevičius 1994: 27–29; 2006: 165–167; LKTCh 2004: 38; VAKKKA 2005: 36–38, etc.), and their phonological structure is commonly used as a basis for codifying standard pronunciation norms (for more detail, see Pupkis 2020: 14–18; Bielinskienė et al. 2014: 24–34). Although opinions emerged in the early 21st century that, due to various factors, the pronunciation basis for the formation of the standard language is increasingly becoming that of large cities, especially the capital (see Vaicekauskienė 2017; Čičirkaitė 2019, etc.; for criticism, see Pupkis 2020: 117–146), traditionally the correctness of the standard language is still assessed by comparing it with the phonological system of the northern part of the Kaunas subdialect. One element of the sound system that remains relevant to standard Lithuanian pronunciation is acute diphthongs with the elements i, u—their pronunciation has been the subject of considerable discussion among language standardizers (for an overview, see Pupkis 2020: 170–173), but the main normative variant still prescribes pronouncing such diphthongs without lengthening these elements, keeping them short, although a slightly lengthened pronunciation is now also offered as a less formal style variant (see VLKK 2021). For the vast majority of Lithuanian subdialect speakers, it is customary to pronounce diphthongs in the position under discussion by lengthening (and for Samogitians, also broadening) their elements i, u. Traditional dialectological studies defined the subdialects around Vilkaviškis, Šakiai, and Jurbarkas as the only area that does not lengthen these elements, alongside small parts of the Highlanders of the Klaipėda region and Western Aukštaitian Šiauliai (see LKA 1982: map no. 20). This position began to be questioned more consistently after dialect research based on experimental phonetic methods developed in Lithuania, particularly in the work of experimental phoneticians of the early 21st century, most of them students of Professor Aleksas Girdenis (for an overview, see Bakšienė 2014: 197). However, the question of the duration of acute elements i, u in the mentioned Kaunas subdialects has still not been definitively answered; therefore, this article returns to a more detailed experimental study of these sound units. The aim of the study is to analyse, using experimental phonetic research methods, the duration of the diphthong elements i, u in the northwestern part of Western Aukštaitian Kaunas and, on the basis of statistical parameters, to assess its relationship with accent types. The study objectives: 1) to identify the specific quantitative features of diphthong elements i, u in the subdialects studied; 2) on the basis of the duration data obtained, to survey the quantitative relationships of the elements under discussion throughout the northwestern (traditionally non-lengthening) VAK1 area; 3) to discuss general patterns in the quantity of this type of diphthong element and address whether they are lengthened or not in the area analysed. The research material consisted of targeted sound recordings from the peripheral subdialects of the VAK area under discussion—Jurbarkas and Vilkaviškis. Six speakers (3 men and 3 women, born and raised in the locations studied, fluent in the local subdialect, and middle-aged—mean age approximately 53) were recorded in each subdialect. Acute and circumflex2 diphthongs with the elements i, u were recorded for the study in the following artificial sentences: the bridge over the river broke: that pot is still warm; it was meant for her: those hands are splitting in autumn; that child is very stuffed: that skilandis is tightly stuffed; yesterday we bought a cupboard: in summer those flies are swarming; that monster was directly (ambushed): he came completely rotten; 1 See the end of the article for all abbreviations. 2 The duration of the first elements of acute diphthongs can be assessed only by comparing them with the corresponding elements of circumflex diphthongs; this is particularly important in subdialects with a continuous accent. 2 already (un)built that stove: all the wealth was lost; that table had been (pushed) aside: it was completely crooked; Kunca came to us: in summer that child (thaws). The recordings were made in 2014,3 with the article’s author and one representative of the local dialects taking part; the latter would say the required sentence, and the speaker would repeat it freely 3–5 times. Unsuccessful attempts were discarded during analysis of the material; three repetitions by the speaker were taken for the study of each diphthong, giving a total of 288 realizations (144 acute and 144 circumflex diphthongs). The results obtained are additionally compared with studies of the Šakiai dialect conducted in previous years, in order to cover comprehensively the entire area in which the acute elements i, u potentially do not lengthen. The study was carried out using experimental phonetic methods: the required sounds were segmented using the PRAAT computer program, segment durations were measured, and the data were then processed using the EXCEL computer program, which was used to calculate mean duration values and plot graphs. The absolute duration of the entire diphthong and of its first element was calculated in milliseconds (ms), as was the relative duration of the first element, expressed as a percentage of the duration of the entire diphthong4. The duration samples being compared were evaluated using a statistical parameter frequently used in experimental phonetics, the so-called Student’s criterion (English t Test), which indicates whether the difference between two groups being compared is statistically significant. Statistical evaluation was also performed using EXCEL, by selecting the commands Data > Data Analysis > t-Test: Two-Sample Assuming Equal Variances. Duration data were chosen to be presented in informative box plots, which also display certain statistical parameters of the samples being compared: the first and third quartiles, median, mean, and minimum, maximum, and outlier values (for more detail, see Kruopis 1993; Čekanavičius, Murauskas 2000; Lapinskas 2010). In assessing the novelty and relevance of the study, it should first be noted that the entire Kaunas subdialect area in which the elements under discussion do not lengthen is analysed comprehensively. This area comprises three more distinct regions—Vilkaviškis, Šakiai, and Jurbarkas—which, in addition to sharing the absence of lengthening, also exhibit quite a number of distinguishing features (for more detail, see Bakšienė 2022); therefore, answering the question of the duration of the diphthong elements in question requires analysing data from all the regions. A new, thorough study of the duration of diphthongs in the Kaunas subdialect would also be relevant to the codification of standard pronunciation norms, since, as noted, there is still no final consensus on this aspect of standardization. 1. EARLIER STUDIES OF DIPHTHONGS IN THE KAUNAS SUBDIALECT The duration of the elements i, u in diphthongs of the Western Aukštaitian Kaunas subdialect is not a new topic; dialectologists have addressed it in several earlier studies. As mentioned in the introduction, in major and landmark works on dialectology of the twentieth century, where dialect features were most often described on the basis of auditory perception alone, the northern and northwestern dialects of the Kaunas subdialect (in A. Salys’s classification, part of Northern Western Aukštaitian, the Veliuona dialect, the Zanavykai dialect, and part of Western Kapsai) were distinguished from the rest of the Lithuanian language area, with the claim that the elements i, u in acute diphthongs do not lengthen there, but remain short (LKA 1982: map no. 20; see Fig. 1). 3 Nedidelė dalis šios medžiagos jau analizuota, žr. Bakšienė 2014. 4 When discussing relative duration, one may also calculate what proportion of the entire word the segment in question constitutes (cf. Švageris 2015: 12 et al.), but this was not possible in the present case because the pairs under study were not absolutely minimal: they differed not only in accent, but also in other sounds, the number of sounds, and so on. Duration data were expressed as percentages rather than in other relative units also because the approach followed earlier studies, with the aim of comparing the results for several dialects. 3 Fig. 1. Kaunas subdialect area in which the elements i, u in acute diphthongs do not lengthen (excerpt from LKA 1982: map no. 20) With the development of experimental phonetic research in Lithuania, dialect features began to be re-verified using a more objective methodology. In the early twenty-first century, a considerable number of experimental studies of the sound systems of the Kaunas subdialect were carried out, including studies of the duration of diphthong elements i, u. The conclusions of these studies almost universally stated that the elements of acute diphthongs lengthen in the area under discussion and become semi-long, while retaining the quality of short vowels—as they lengthen, they tend not to become closer, as they do throughout the southeastern area of Lithuanian dialects, but to become more open, much as in the western Samogitian dialects, although the degree of opening is said to be much weaker. It should also be noted that in the northern Kaunas subdialects, the elements i, u in circumflex diphthongs, lengthened because of the sustained accent, were also almost universally considered semi-long and usually more open, e.g.: [²ˈʊˌɪˑlkˢ] ~ vìlkas, [²ˈpʲɪˑrkt] ~ pìrkti. Thus, in syllables with either of the two phonological accents, these elements were treated as equivalent in duration; lengthening in syllables with the sustained accent was even regarded as a more pronounced and systematic prosodic feature (for more detail, see Bacevičiūtė 2004: 89; Bacevičiūtė, Leskauskaitė 2004: 12–14; Vaišnienė, Bacevičiūtė 2005: 292–302; for an overview, see Bakšienė 2014). However, the most frequently cited studies generally analysed data from one particular dialect, or compared them with data from the southeastern part of the Kaunas subdialect, where the elements in question are clearly lengthened; no comprehensive study of the area where they do not lengthen had been conducted. 2. RESULTS ON THE OVERALL DURATION OF DIPHTHONGS First, the results of measurements of the overall duration of diphthongs with the elements i, u were analysed6. A look at the box plots showing the distribution of overall-duration data for acute and circumflex diphthongs in the Vilkaviškis and Jurbarkas dialects (see Fig. 2) reveals that both dialects exhibit shared features as well as different duration patterns. 5 Dėl transkripcijos TFA ženklais žr. Bakšienė, Čepaitienė 2024. 6 Only absolute-duration data are discussed in this section. 4 What the two dialects have in common is that circumflex diphthongs tend to be longer than acute ones, as indicated first of all by their higher median values in both cases. However, the difference between diphthongs with different tones is noticeably greater in the Vilkaviškis dialect: here, the central value for the duration of acute diphthongs is 178 ms and that for circumflex diphthongs is 194 ms, whereas in Jurbarkas the values are 164 and 168 ms, respectively. Mean values also differ quite clearly in the Vilkaviškis dialect: 182 ms for acute diphthongs and 192 ms for circumflex ones. In the Jurbarkas dialect, there is no difference in the means at all—the mean duration of both acute and circumflex diphthongs is 168 ms. The differences between the dialects are also clearly revealed by the main duration ranges, which lie between the first and third quartiles (the diagram “boxes,” excluding the “whiskers”): in Vilkaviškis, the first- and third-quartile values for circumflex diphthongs are noticeably higher than those for acute diphthongs, whereas in Jurbarkas the main ranges for diphthongs of both tones coincide, with the circumflex-diphthong data falling entirely within the acute-diphthong range. Thus, it is likely that in the Jurbarkas dialect there is no significant difference in the overall duration of acute and circumflex diphthongs in the group under discussion. The diagrams also show that, overall, the duration of all the diphthongs measured is greater in the Vilkaviškis dialect than in the Jurbarkas dialect. However, when assessing absolute-duration data, this is not a very important difference; it may also be related to individual speaker characteristics. Fig. 2. Distribution of overall-duration values of diphthongs7 Comparing the mean overall durations of pairs of diphthongs with different tones (il – il, ir – ir...), (see Fig. 3) shows the trends already discussed (a clearly greater difference between acute and circumflex diphthongs in the Vilkaviškis dialect; all diphthongs in the Vilkaviškis dialect tend to be longer than those in Jurbarkas), but also brings out new features that substantiate the general patterns noted above. In the Vilkaviškis dialect, circumflex diphthongs are longer in 7 pairs (out of the 8 compared); only in one case (ùn – un) is the duration greater for 7 In the box plots, the “boxes” in the middle represent the range of sample values between the first and third quartiles. The median (the middle sample value; in this case, since the number of values is even (144), the median is calculated as the mean of the two middle values) is indicated by a horizontal line, and the arithmetic mean by the symbol × inside the “box.” The beginning and end of the “box whiskers” coincide with the minimum and maximum sample values (for more details, see Lapinskas 2010). 5 acute diphthong. In the Vilkaviškis examples, the greatest difference between acute and circumflex diphthongs was recorded for the pair ùm – ũm (24 ms), and the smallest between the members of the pair ìr – ĩr (3 ms). In the Jurbarkas dialect, the pattern is quite different: in as many as 5 of the 8 pairs, the acute diphthongs are longer, but their duration differences from the circumflex diphthongs are much smaller; only the acute ùn stands out, being 20 ms longer than the corresponding circumflex diphthong. Overall, there is more variation in the Jurbarkas dialect, suggesting that total duration is not a significant feature for differentiating diphthong intonations here. It is also noteworthy that the total duration of diphthongs in both dialects (regardless of intonation) is generally strongly related to their component sounds: diphthongs with the second elements n, m (mean total duration 189 ms) are considerably longer than diphthongs with l, r (166 ms). Fig. 3. Mean total durations of individual diphthongs To verify the trends identified in the figures, a statistical assessment of the duration samples was conducted using the Student’s t-test (t Test); total-duration data were compared separately by intonation and by study location (see Table 1). Two statistical parameters are of primary importance in the assessment: the p value (the difference between samples is considered statistically significant if it is less than 0.05), and the Student’s t-test statistic (T) and the calculated critical value at the 95 percent probability level (t) (a statistically significant difference is confirmed when T > t)8. Variance and standard deviation (obtained by taking the square root of the variance) may also be relevant as additional parameters, indicating how far the values in the sample under study deviate from the mean (see Lapinskas 2010 for more detail). The tables show that the statistical assessment confirmed the patterns discussed. Only the difference between acute and circumflex diphthongs in the Vilkaviškis dialect is considered statistically significant (see Table 1), since in this case p < 0.05 (0.00733341), and the T values are greater than t (2.7006357 > 1.96829326). The difference in total duration between acute and circumflex diphthongs in the Jurbarkas dialect is considered statistically insignificant, since the p value is greater than 0.05 (0.9469643), and T < t (0.0665778 < 1.9682933). 8 Since a two-tailed alternative is being assessed, the minus sign before the T value need not be taken into account. 6 Table 1. Statistical assessment of the total duration of diphthongs Location | Vilkaviškis | Jurbarkas | Intonation | Statistical parameters | Acute | Circumflex | Acute | Circumflex | Mean (ms) | 182,146 | 192,188 | 168,715 | 168,493 | Variance | 1030,545 | 960,321 | 903,702 | 700,573 | Standard deviation (ms) | 32,102 | 30,989 | 30,062 | 26,468 | Number of measurements | 144 | 144 | 144 | 144 | T value | –2,7006357 | 0,0665778 | p value (T < = t) (two-tailed alternative) | 0,00733341 | 0,9469643 | t critical value (two-tailed alternative) | 1,96829326 | 1,9682933 To summarize the data on the total duration of diphthongs with elements i, u in the Vilkaviškis and Jurbarkas dialects, two points are important: the duration of circumflex diphthongs was found to be statistically significantly greater in the Vilkaviškis dialect, and diphthongs in general tend to be longer there than in the Jurbarkas area. No reliable differences between acute and circumflex diphthongs were identified in the Jurbarkas dialect. Comparing these results with earlier data from studies of the Šakiai dialect, it can be said that the pattern of total-duration relationships is more similar in the Jurbarkas and Šakiai areas: acute diphthongs with elements i, u tend to be longer there, but the difference is small and statistically insignificant; diphthong duration relationships vary depending on the sounds forming the syllable nucleus (see Bacevičiūtė 2004: 86–88 for more detail). The differences between the Vilkaviškis and Jurbarkas–Šakiai areas are likely due to the continuous intonation that exists in the latter9 (LKA 1982: map no. 106). 3. RESULTS OF THE ABSOLUTE DURATION OF THE FIRST ELEMENT A more precise answer to the question of whether the elements i, u lengthen in the dialects under study was expected from an analysis of the duration of the first diphthong element. As mentioned in the introduction, the absolute and percentage duration of the first elements was measured. Box plots of absolute duration values (see Fig. 4) clearly reveal differences among the dialects. The following patterns revealed by the plots are the most important: in the Vilkaviškis dialect, the duration of all diphthong elements studied is somewhat greater, but these are absolute-duration data that may be affected by the individual characteristics of speakers and other factors; in the Vilkaviškis dialect, no clear difference is apparent between acute and circumflex first elements: the first-quartile, median, and mean values might suggest that circumflex diphthong elements tend to be longer, but their value range falls within the range of values for acute elements, and the ranges overlap completely; the relationship between duration values in the Jurbarkas dialect is clearer and more consistent: all the visible parameters—the first quartile, median, and mean—are in a somewhat higher range for acute diphthongs, so their first elements tend to be slightly longer; however, the difference is also small, as most of the duration samples for acute and circumflex elements overlap. 9 As can be seen, the duration of diphthong elements tends to depend on several factors, most notably the prosodic system of the particular dialect, the pitch accent, and perhaps the sounds that make up the syllable. Further research should examine the differentiating power of these factors using a wider range of statistical methods, for example, two-way or multifactorial analysis, which researchers have recently applied to identify models of sound interaction in the standard language (see Švageris, Urbanavičienė 2024). However, this is not the subject of this article, but of a broader study. 7 Fig. 4. Distribution of absolute duration values for the first diphthong element A graph showing the relationships between the mean values of elements in diphthong pairs (see Fig. 5) once again confirms the patterns discussed above. The figure shows that in the Vilkaviškis dialect, the means for circumflex diphthong elements were higher in 6 of the 8 pairs studied; in the Jurbarkas dialect, the means for acute diphthong elements were higher in 5 of the 8 pairs. This confirms the tendency for circumflex diphthong elements to be longer in the area around Vilkaviškis i, u, and for acute elements to be longer in the Jurbarkas area. On the other hand, the graph shows that the relationships between the means are not consistent: there are also cases contrary to the tendency discussed, and in some instances the duration difference between acute and circumflex elements is very small indeed. Fig. 5. Mean absolute duration values for the first element of individual diphthongs The data for this part of the study were also evaluated using the Student’s t-test (see Table 2). Based on the parameters mentioned above, it can be stated that the difference in absolute duration between the elements i, u of diphthongs bearing different pitch accents could be considered statistically significant only in the Jurbarkas dialect: here T (2,012787915) > t (1,968293255), while p < 8 0.05 (0.045073953), but it can be seen that the calculated values of the statistical criteria differ little from the critical values, and thus the difference between the samples being compared is small. The difference in duration between the diphthong components i, u in the Vilkaviškis dialect according to pitch accent proved statistically insignificant: T (1.299146024) < t (1.968293255), and p (0.194940153) > 0.05. Table 2. Statistical evaluation of the absolute duration of the first diphthong component Locality | Vilkaviškis | Jurbarkas | Pitch accent | Acute | Circumflex | Acute | Circumflex | Statistical parameters | Mean (ms) | 92.666 | 96.395 | 84.840 | 81.048 | Variance | 641.860 | 544.646 | 287.953 | 223.054 | Standard deviation (ms) | 25.335 | 23.338 | 16.970 | 14.934 | Number of measurements | 144 | 144 | 144 | 144 | T value | −1.299146024 | 2.012787915 | p value (T <= t) (two-tailed alternative) | 0.194940153 | 0.045073953 | Critical t value (two-tailed alternative) | 1.968293255 | 1.968293255 To summarize the findings of this part of the study, the first-component i, u duration patterns of diphthongs with different pitch accents are not uniform across the dialects compared. In the Vilkaviškis dialect, the components i, u of circumflex diphthongs tend to be longer, but the difference in their duration is not statistically significant; thus, in these areas, the acute and circumflex diphthong components do not differ in duration in any substantial way. In the Jurbarkas dialect, the components of acute diphthongs tend to be longer, and the difference in their duration is also statistically confirmed. Compared with data from earlier analogous studies of the Šakiai dialect, it is clear that the Šakiai dialect occupies an intermediate position: here, the components i, u of acute diphthongs are more often longer (a similarity to the Jurbarkas area), but no statistically significant duration differences were found between diphthong components bearing different pitch accents (a similarity to the Vilkaviškis area) (for more details, see Bacevičiūtė 2002: 88–91). Thus, the preliminary data indicate that lengthening of the first components i, u of acute diphthongs in the northwestern Kaunas subdialect becomes more pronounced moving northward. 4. RESULTS FOR THE PERCENTAGE DURATION OF THE FIRST COMPONENT In analysing the duration of sound units, researchers most often rely not so much on absolute as on some form of relative data, since these are considerably more objective and nuanced, and less affected by a speaker’s individual characteristics (cf. Pakerys 1982: 44). As already noted, this study analyses the percentage duration of the first diphthong component, i.e. the proportion of the total diphthong duration that it accounts for. The results for percentage duration partly confirmed the trends discussed earlier (for example, that the first components of all diphthongs in the Vilkaviškis dialect are longer than those in Jurbarkas), but also brought to light new aspects of the relationship between diphthongs and pitch accents. The box plots (see Fig. 6) show that, in terms of percentage duration, the components i, u of acute diphthongs tend to be longer in both dialects compared: in every case, the values for both the first quartile and the medians and means are higher. However, in the Vilkaviškis dialect, the duration difference between acute and circumflex components is barely perceptible; the ranges of duration values overlap almost entirely, whereas the difference is much more pronounced in Jurbarkas. 9 Fig. 6. Distribution of percentage-duration values for the first diphthong component The mean percentage durations of the first component in the individual diphthong pairs compared also confirm the identified trends (see Fig. 7). First, it should be noted that in this group of diphthongs with components i, u, there are no very large differences between acute and circumflex components; their duration ranges from 40–60% of the total diphthong duration. The shortest is the circumflex component m in Jurbarkas, at 40.6%, while the longest are the acute components r in Vilkaviškis and r in Jurbarkas, both at 59.1%. The graph shows clear intrinsic-duration differences (depending on the sounds that make up the syllable nucleus rather than so much on pitch accent): the diphthong components i, u tend to be longest next to the sonorant r and shortest next to m, n. Differences associated with pitch accents are also apparent: in the Vilkaviškis dialect, acute diphthong components are longer in 5 of 8 cases, and in Jurbarkas in as many as 7 of 8; thus, this most objective measurement confirms stronger lengthening of acute components in the Jurbarkas area. Fig. 7. Mean percentage durations of the first component in individual diphthongs 10 Statistical assessment of the first elements’ percentage duration also shows that in the Jurbarkas dialect, the elements i, u of acute diphthongs are statistically significantly longer than those of the corresponding circumflex diphthongs (see Table 3): in this case, T (2,549156057) > t (1,968293255), while p (0,01132068) < 0,05. The statistical difference between the compared duration values in the Vilkaviškis dialect is insufficient: T (0,594895476) < t (1,968293255), while p (0,552383706) > 0,05. Table 3. Statistical assessment of the percentage duration of the first element of diphthongs Locality | Vilkaviškis | Jurbarkas | Accent | Acute | Circumflex | Acute | Circumflex | Statistical parameters | Mean (%) | 50,761 | 50,090 | 50,593 | 48,414 | Variance | 104,447 | 79,023 | 55,545 | 49,641 | Standard deviation (%) | 10,220 | 8,890 | 7,453 | 7,046 | Number of measurements | 144 | 144 | 144 | 144 | T value | 0,594895476 | 2,549156057 | p value (T <= t) (two-tailed alternative) | 0,552383706 | 0,01132068 | Critical t value (two-tailed alternative) | 1,968293255 | 1,968293255 Summarizing the percentage-duration results, this part of the study can be said to have reliably confirmed the greater lengthening of the elements i, u of acute diphthongs in the Jurbarkas dialect. According to data from an earlier study of the Šakiai dialect, no reliable difference in the percentage duration of the corresponding first elements was found in the Šakiai area: in some cases the elements of acute diphthongs were longer, in others those of circumflex diphthongs; the statistical assessment also did not confirm a difference (for further details, see Bakšienė 2014: 208–211). The difference between duration-data samples supported by statistical criteria is an important argument for treating the acute elements i, u in the Jurbarkas dialect as semi-long, particularly given that, as in the Šakiai area, the Jurbarkas area has a continuous accent, which, in circumflex diphthongs, emphasizes not only the second but also the first element. However, the acute elements i, u are even longer than the latter, which are emphasized by the continuous accent10. 5. SUMMARY AND INTERPRETATION OF RESULTS In summarizing the results of this and the earlier study of the Šakiai dialect, attention should first be drawn once again to the differing status of the dialects studied with respect to the realization of circumflex syllables: in the Jurbarkas and Šakiai areas, the circumflex has a distinct allotone—the continuous accent—in which both elements of diphthongs are emphasized, whereas the Jurbarkas dialect has no continuous accent; there, circumflex prominence is concentrated only on the second element of the diphthong (see LKA 1982: map no. 106). These differences in allotones undoubtedly determine the quantity relations of the diphthongs analysed. 1. The results of the study of the i, uoverall duration of diphthongs with elements revealed that in the Vilkaviškis dialect circumflex diphthongs are somewhat longer; their duration difference from acute diphthongs is also statistically significant. This is most likely due to the more pronounced lengthening of only the second element of circumflex diphthongs, since the first elements of both acute and circumflex diphthongs 10 In earlier standard dialectological works, the first elements in diphthongs bearing a continuous accent were customarily marked as semi-long i, u, whereas in acute diphthongs in the area under discussion they were considered short. This is a rather paradoxical assessment, since the first elements of circumflex diphthongs were considered longer than those of acute diphthongs (for a broader overview, see Bakšienė 2014: 195–197). 11 The duration of the elements in the Vilkaviškis area is very similar (see below). In the Jurbarkas and Šakiai areas, the overall duration of the diphthongs is almost independent of the accent, although acute diphthongs tend to be slightly longer; however, this difference is not statistically reliable. Thus, the overall duration data for the diphthongs analysed distinguish the Vilkaviškis area from the Šakiai–Jurbarkas area. 2. The summary data on the duration of the first element of the diphthongs under discussion (the data on both absolute and percentage duration essentially showed the same trends) (see Tables 4 and 5) made it possible to differentiate the dialects under study even more clearly: in the Vilkaviškis dialect, there is no substantial difference between the elements of acute and circumflex diphthongs; in the Jurbarkas area, the elements of acute diphthongs were found to be longer, and the difference is statistically significant; the data from the Šakiai dialect are inconsistent: in terms of absolute duration, the elements of acute diphthongs appear longer, whereas in terms of percentage duration, those of circumflex diphthongs are somewhat longer, but in neither case are the differences statistically significant. On the basis of these findings, it can be argued that the lengthening of the elements i, u in acute diphthongs increases westward in the western part of the Kaunas subdialect as one moves north. Table 4. Ratios of absolute duration of vowels and first elements of diphthongs Duration | Vowels i, u (ms) | Tvp i, u (ms) | Tvg i, u (ms) | Duration ratios | Vowels i, u: tvp i, u: tvg i, u | Dialect | Jrb | 66 | 85 | 81 | 1 | 1,29 | 1,23 | Šk | 63 | 102 | 98 | 1 | 1,62 | 1,56 | VIkv | 75 | 93 | 96 | 1 | 1,24 | 1,28 Table 5. Ratios of percentage duration of diphthong elements Duration | Tvp i, u (percent) | Tvg i, u (percent) | Duration ratios tvp i, u: tvg i, u | Dialect | Jrb | 50,6 | 48,4 | 1,05 | 1 | Šk | 53,6 | 55 | 1 | 1,03 | VIkv | 50,8 | 50,1 | 1,01 | 1 3. To answer more clearly the question posed at the beginning of the study—whether the elements of acute diphthongs in the area described should be considered short or semi-long—an attempt was made to compare them, at least minimally, with the stressed short vowels i, u in the word stem of the corresponding dialects (see Table 4). Admittedly, this comparison cannot be considered highly reliable, since the data were collected in different studies11 and only absolute duration values are used; nevertheless, certain preliminary duration trends can be discerned. It can be seen that in all the dialects examined, the elements of both acute and circumflex diphthongs are considerably longer than vowels. The greatest difference in the duration of vowels and diphthong elements is found in the Šakiai dialect—approximately one and a half times—while in the other two dialects the difference is somewhat smaller, with diphthong elements about 1.3 times longer than vowels. This comparison confirmed the tendency for the elements of acute diphthongs to be lengthened more in dialects with a continuous accent, that is, in the Šakiai and Jurbarkas areas; here they tend to be longer than circumflex elements, which are universally treated as semi-long because of the continuous accent. Thus, at least in the northernmost part of the Kaunas subdialect, the elements i, u of acute diphthongs should also be considered semi-long, whereas the Vilkaviškis area requires more extensive research on this issue in the future. 11 Šakių šnektos balsių i, u trukmės duomenys imti iš R. Bacevičiūtės (2004) straipsnio; Jurbarko ir Vilkaviškio atitinkamų balsių trukmė apskaičiuota iš kitam tyrimui (žr. Baksienė 2022) įrašytos medžiagos. Suprantama, kad skirtingu metu ir iš skirtingų kalbėtojų įrašyti duomenys gali būti tinkami tik labai preliminarioms pastaboms. 12 4. When the results of these studies are assessed in the context of the entire Lithuanian language area, it can be argued that, in their pronunciation of acute diphthongs with the elements i, u, the northern Kaunas dialects are no different from other Lithuanian dialects or from the system of acute diphthongs as a whole: just as diphthongs with the elements a, e (see Bakšienė 2014 for more detail), the diphthongs analysed here are also pronounced lengthened in this area, although, as already noted, they most likely retain the qualitative features of short vowels12. This prosodic feature of the Kaunas subdialect should also be taken into account when codifying the norms of standard pronunciation (if the phonological structure of this subdialect is still being used as a basis), since, in principle, there are no dialects in the Lithuanian language area that do not lengthen the elements under discussion; pronouncing them as short elements here would therefore contradict the general structure of the prosodic features of acute diphthongs. LITERATŪRA Bacevičiūtė Rima 2004: Šakių šnektos prozodija ir vokalizmas [Šakiai Subdialect: Its Prosody and Vowel System], Vilnius: Lietuvių kalbos institutas. Bacevičiūtė Rima, Leskauskaitė Asta 2004: Akūtinių dvigarsių pirmųjų dėmenų i, u trukmė aukštaičių šnektose [Quantity of First Components [i], [u] of Acute Diphthongs in Aukštaitian Subdialects]. – Baltų ir kitų kalbų fonetikos ir akcentologijos problemos, ats. red. G. Akelaitienė, Vilnius: Vilniaus pedagoginis universitetas, 9–20. Bakšienė Rima 2014: Vakarų aukštaičių kauniškių dvigarsiai: kiekybės santykiai [Diphthongs of Western Aukštaitian of Kaunas: Quantitative Ratio]. – Acta Linguistica Lithuanica 71, 193–214. Bakšienė Rima 2022: Vakarų aukštaičiai kauniškiai: arealai ir diferenci­niai požymiai [Western Aukštaitian of Kaunas: Areas and Differential Features], Vilnius: Lietuvių kalbos institutas. Bakšienė Rima, Čepaitienė Agnė 2024: Tarptautinės fonetinės abėcėlės rašmenys lietuvių tarmėtyroje [International Phonetic Alphabet Characters in Lithuanian Dialectology], Vilnius: Lietuvių kalbos institutas. DOI: https://doi.org/10.35321/e-pub.135.tfa-lietuviu-tarmetyroje. Bielinskienė Agnė, Kazlauskienė Asta, Rimkutė Erika, Tamošiūnaitė Aurėlija 2014: Lietuvių bendrinė kalba: normos ir vartosena [Standard Lithuanian Language: Norms and Usage], Kaunas: Vytauto Didžiojo universitetas. Čekanavičius Vydas, Murauskas Gediminas 2000: Statistika ir jos taikymai I [Statistics and its Applications I], Vilnius: Vilspa. Čičirkaitė Ramunė 2019: Vilniaus kalbos fonetinių ypatybių socialinės reikšmės: stigmatizuojamas ilginimas ir santykis su žiniasklaidos kalba [Social Meanings of Phonetic Variation of Vilnius Speech: Stigmatised Lengthening and Relation to Media Language]: daktaro disertacija, Vilnius: Lietuvių kalbos institutas. Kruopis Julius 1993: Matematinė statistika [Mathematical Statistics], Vilnius: Mokslo ir enciklopedijų leidykla. Lapinskas Remigijus 2010: Pats trumpiausias taikomosios statistikos kursas su R [The Shortest Course in Applied Statistics using R], Vilnius: Vilniaus universitetas. Prieiga internete: https://web.vu.lt/mif/a.reklaite/files/2013/02/2010.01-Pats-trumpiausias-su-R.pdf. 12 Rengiamame platesniame tyrime, kurio dalis yra šio straipsnio analizė, numatoma išsamiai ištirti būtent Vilkaviškio ir Jurbarko apylinkių dvigarsių dėmenų i, u kokybės požymius, nes anksčiau šiuo atžvilgiu daugiausia dėmesio buvo skirta tik Šakių šnektai. Paminėtina, kad minėtoje VLKK rekomendacijoje, nurodančioje aptariamųjų dvigarsių tarimą, tuo atveju, kai laisvesniame bendrinės kalbos stiliuje neprieštaraujama pusilgiams jų dėmenims, teigiami tik kokybiškai trumpieji pirmieji dėmenys i, u (TFA žymimi [ɪ] ir [ʊ], pvz.: [ˈpʲɪlkes] ~ pilkas, [ˈkʊrʲtʲi] ~ kūrti) (plačiau žr. VLKK 2021). Taigi aptariamųjų dėmenų kokybės klausimas taip pat išlieka aktualus. 13 LKA 1982: Lietuvių kalbos atlasas 2: Fonetika [Lithuanian Language Atlas 2: Phonetics], red. K. Morkūnas, Vilnius: Mokslas. LKTCh 2004: Lietuvių kalbos tarmių chrestomatija [Lithuanian Dialects Reader], sud. R. Ba-cevičiūtė, A. Ivanauskienė, A. Leskauskaitė, E. Trumpa, Vilnius: Lietuvių kalbos instituto leidykla. Pakerys Antanas 1982: Lietuvių bendrinės kalbos prozodija [Prosody of the Standard Lithuanian Language], Vilnius: Mokslas. Pupkis Aldonas 2020: Lietuvių bendrinė tartis. Norminimo istorija ir dabartis [Standard Lithuanian Pronunciation: the History and Presence of Standardisation], Vilnius: Vilniaus universiteto leidykla. Švageris Evaldas 2015: Lietuvių ir latvių tarmių monoftongų priegaidžių akustiniai požymiai: lyginamoji analizė [Acoustic Characteristics of Monophthong Pitch Accents in Lithuanian and Latvian Dialects: a Comparative Analysis]: daktaro disertacija, Vilnius: Vilniaus universitetas. Švageris Evaldas, Urbanavičienė Jolita 2024: Lietuvių kalbos balsių formantinę struktūrą ir trukmę lemiančių veiksnių sąveika bei hierarchija [Phonetic Distribution of Vowel Quality and Quantity in Standard Lithuanian: a Factorial Analysis], Vilnius: Lietuvių kalbos institutas. Vaicekauskienė Loreta 2017: Du sąmonės lygmenys, dvi geriausios kalbos, vienas standartas [Two Levels of Consciousness, Two Best Languages, One Standard]. – Lietuvių kalbos idealai, sud. L. Vaicekauskienė, Vilnius: Naujasis židinys-Aidai, 111–142. Vaišnienė Daiva, Bacevičiūtė Rima 2005: Dėl pietų žemaičių raseiniškių ir šiaurinių vakarų aukštaičių kauniškių dvigarsių [i, u + l, m, n, r] [Concerning the Semi-diphthongs [i, u + l, m, n, r] in Southern Žemaitian of Raseiniai and North-Western Aukštaitian of Kaunas]. – Kalbos istorijos ir dialektologijos problemos 1, vyr. red. D. Mikulėnienė, Vilnius: Lietuvių kalbos institutas, 292–303. VAKKKA 2005: Vakarų aukštaičiai kauniškiai ir Klaipėdos krašto aukštaičiai [Western Aukštaitian of Kaunas and Aukštaitian of Klaipėda region]: mokomoji knyga, sud. R. Bacevičiūtė, D. Mikulėnienė, V. Salienė, Vilnius: Lietuvių kalbos institutas. VLKK 2021: Rekomendacija „Dėl pagrindinių bendrinės lietuvių kalbos tarties normų“ [Recommendation on Basic Norms for Pronunciation of Standard Lithuanian Language] (2021 m. gruodžio 16 d. Nr. PN-18 (k-26)). Prieiga internete: https://www.vlkk.lt/aktualiausios-temos/tartis-ir-kirciavimas/26-pagrindiniai-tarties-reikalavimai. Zinkevičius Zigmas 1994: Lietuvių kalbos dialektologija [Lithuanian Dialectology], Vilnius: Mokslo ir enciklopedijų leidykla. Zinkevičius Zigmas 2006: Lietuvių tarmių kilmė [Origin of Lithuanian Dialects], Vilnius: Lietuvių kalbos institutas. SUTRUMPINIMAI Jrb – Jurbarkas ms – milisekundė proc. – procentas T – Studento kriterijaus reikšmė t – kritinė 95 proc. tikimybės reikšmė p – 0,05 reikšmingumo lygmens reikšmė 14 Šk – Šakiai tvp – akūtas, tvirtapradė priegaidė tvg – cirkumfleksas, tvirtagalė priegaidė VAK – vakarų aukštaičiai kauniškiai Vlkv – Vilkaviškis RIMA BAKŠIENĖ ONCE AGAIN ON THE QUANTITY OF ELEMENTS I, U IN MIXED DIPHTHONGS: DO THEY REALLY NOT LENGTHEN IN THE DIALECT AREA CLOSEST TO THE STANDARD LANGUAGE? Summary The article analyses the quantity of the elements i and u in mixed diphthongs in the northwestern part of the Western Aukštaitian subdialect of Kaunas, which encompasses the local dialects of Jurbarkas, Šakiai, and Vilkaviškis. The phonological and prosodic structure of these regional dialects is most commonly used as a basis for codifying the norms of standard pronunciation. As the established pronunciation of the acute syllables of this type in the standard language is still a subject of debate (whether they are short or semi-long), it is essential to identify their prosodic features in those local dialects that have traditionally been considered an area where the elements i, u do not lengthen and which stands out from the rest of the Lithuanian language system. The study employed experimental phonetic methods, and the statistical significance of differences in duration values across samples was assessed using the Student’s t-test. The total duration of both acute and circumflex mixed diphthongs, as well as the absolute duration of their initial elements i and u in milliseconds (ms) were analyzed. Additionally, the percentage duration of the initial elements relative to the total duration of the mixed diphthong was calculated. It was found that in terms of total duration, only the circumflex mixed diphthongs in the Vilkaviškis local dialect are slightly longer, and the difference in their duration is statistically significant. In the areas of Jurbarkas and Šakiai, the total duration of mixed diphthongs is almost independent of pitch accent, although acute mixed diphthongs tend to be slightly longer. However, this difference is not statistically reliable. The results of the study of the initial element duration revealed no significant difference between the acute and circumflex mixed diphthong elements in the Vilkaviškis local dialect. In the Jurbarkas area, the elements of acute mixed diphthongs were found to be longer, with the difference being statistically significant. The data from the Šakiai local dialect are inconsistent. Thus, the lengthening of i and u in acute mixed diphthongs in the western part of the Kaunas subdialect becomes more pronounced toward the north. A comparison of the findings in the present study with data from previous research on the duration of short vowels i and u shows that, in all the local dialects examined, the elements of both acute and circumflex mixed diphthongs are noticeably longer than the corresponding vowels. The lengthening of the elements of acute mixed diphthongs tends to be stronger in local dialects with a continued pitch accent, namely in the area of Šakiai and Jurbarkas. Therefore, in the northernmost part of the Kaunas subdialect, the elements i and u in acute mixed diphthongs can also be considered semi-long. Meanwhile, more detailed studies on this question are still needed in the Vilkaviškis area. The summarized results of the study confirmed that in terms of the pronunciation of the elements i, u in acute mixed diphthongs, the northern local dialects within the Kaunas subdialect do not differ either from other Lithuanian local dialects or from the overall system of acute mixed diphthongs; here, too, the elements are pronounced with lengthening, while retaining the qualitative features of short vowels. Thus, it is likely that 15 there are no local dialects within the entire Lithuanian language area in which these elements remain unlengthened, as the pronunciation of short elements would contradict the overall prosodic structure of acute mixed diphthongs. 16