Sociolingvistinis vilniečių lietuvių kalbos tyrimas: konsonantizmas ir akcentuacija

LIETUVIŲ KALBOTYROS KLAUSIMAI XXVII (1988) LIETUVIŲ KALBA IR BILINGVIZMAS LAIMA GRUMADIENĖ A SOCIOLINGUISTIC STUDY OF VILNIUS LITHUANIAN: CONSONANTISM AND ACCENTUATION1 I. Consonantism The consonantism of Vilnius Lithuanian varies in essence in three respects: 1) certain requirements of standard Lithuanian consonant palatalization are violated; 2) consonant assimilation is sometimes conditioned by morphological rather than phonetic circumstances; 3) instances of elision of the consonant j occur. Consonant softness depends both on their nature and articulatory properties, and on the sound system of the dialect or idiolect as a whole (cf. Urbanavičiūtė, 1978; Zinkevičius, 1966, pp. 153–170). Acoustically, consonants are pronounced softest before long, tense front vowels, especially before the high vowel [iː]. Since, as noted earlier in the analysis of the vowel system of Vilnius speech, there is a growing tendency to pronounce long vowels with somewhat less tension, it is self-evident that consonant softness also decreases. From an articulatory perspective, consonants are palatalized especially strongly before back vowels. Since Vilnius speech is also characterized by a certain delabialization of back vowels, soft consonants before them should be especially soft—that is, acoustically very high-pitched—but this is generally not the case, especially in unstressed syllables. Apparently, the less effortful articulation of laxly pronounced vowels also entails an easier path involving less consonant palatalization. It could therefore be concluded that Vilnius speech exhibits a tendency toward less palatalization of soft consonants. One of the features of Lithuanian consonantism, its syntagmatics, is regressive palatalization of consonant clusters. In Vilnius speech, especially at morpheme boundaries, there are many instances of this rule being disregarded. Pronunciation with respect to the palatalization of peripheral vocabulary items—international words—is particularly heterogeneous. This article addresses the issue because, in terms of frequency of use, many such words are far from peripheral: they are highly conspicuous, and their number is constantly growing. In addition, it is necessary to establish which pattern is becoming prevalent and how future words will be adapted. This is not merely a problem of alto, valso, and the like. Sociolinguistic variation in consonant clusters before Vi or Vu. Regressive palatalization of consonant clusters does not have the same characteristics in all dialects (cf. Girdenis, Piročkinas, 1978, p. 25). In many cases, when k, g, and often also p, b, m occur at the beginning or in the middle of a cluster, the consonants mentioned in those clusters are not palata- 1 Continued. For a description of the methods used to study spoken Vilnius Lithuanian and an analysis of its vowel system, see: Lietuvių kalbotyros klausimai: Kalbų kontaktai Lietuvos TSR. V., 1983. Vol. 23. pp. 62–101. 132 lizuojami (Zinkevičius, 1966, p. 157). Tarmėse pasitaiko ir išimčių, o bendrinės kalbos reikalavimai gana griežti (plg. Vitkauskas, 1985). In the Lithuanian spoken by Vilnius residents, the consonants k, g, p, b, m in consonant clusters are almost never softened. In general, there is a tendency to palatalize only the last consonant in a cluster when it occurs before front vowels, or only the soft consonant before back vowels, e.g.: [l’p’t’] or [lpt’] — [ȧl’.pt’i] or [ȧl.pt’i] ‘alpti’, [ž’v’] or [žv’] — [ūž’v’ir’ę] or [ūžv’ir’ę] ‘ūžvirė’, [ks’ t’] or [kst’], [s’n’] or [sn’] — [aukš’t’es’n’is], [aukšt’esn’is], [aukš’t’es’n’is], [aukšt’esn’is] ‘aukštesnis’, [t’pl’] or [tpl’], [š’t’] or [št’] — [at’pl’ę̇š’t’i], [atpl’ę̇š’t’i], [at’pl’ę̇š’t’i], [atpl’ę̇š’t’i] ‘atplėšti’, [n’v’] or [nv’] — [r’ei.n’v’in’is] or [r’ei.nv’in’is] ‘reinvinis’, etc. Earlier studies of the speech of Vilnius schoolchildren show that two-consonant clusters are not palatalized in 17% of cases, and clusters of three to four consonants in 43% of cases (Pakerys, 1971, p. 7). In the sociolinguistic analysis of the Lithuanian spoken by Vilnius residents, the palatalization of consonant groups has been examined as sociolinguistic variables comprising the range of realizations of consonant clusters with a final soft consonant. This was abbreviated as (C’C’V’ or C’C’Vù), where C’C’ denotes clusters of two, three, or more consonants, V’ any front vowel, Vù a fronted back vowel, and v ‘or’. The variable scale has only two levels: a palatalized cluster and a non-palatalized cluster. The results should be viewed with caution, since certain compromises had to be made in order to take into account factors that seemed to us more important. Clusters were considered palatalized in all cases where at least two of their consonants were pronounced softly; that is, all palatalized two-member clusters qualified, while for three- and four-member clusters, two palatalized consonants that were not even adjacent sufficed. For example, [ȧl.’ks-n’is] ‘alksnis’ is recorded as a palatalized pronunciation of the consonant cluster. The aim was to try to investigate how variation in the sociolinguistic variable (C’C’V’ or C’C’Vù) depends not only on social factors (education, C), sociodemographic factors (age, A), and contextual style factors (B), but also to assess whether cases in which regressive palatalization does not occur are more frequent at morpheme boundaries, especially in prefixes and compound words. After the tape-recorded interviews with Vilnius residents had been transcribed, it was calculated that approximately half of the consonant clusters under investigation are palatalized, while approximately half are not. Figure 1 shows that the softening of homomorphemic and heteromorphemic consonant clusters does not differ, or differs hardly at all (cf. the confidence intervals). Variation in this sociolinguistic variable cannot be described more clearly in terms of either style or stratification, although statistically significant is the more frequent occurrence of palatalized variants of the variable in the speech of the middle generation (A₂), especially when compared with the speech of the younger generation (A₃), in which only one third of potential clusters are often palatalized. The speech of the control group of announcers and actors (C₃) is dominated by palatalized variants. Apparently, with certain reservations, the sociolinguistic variable (C’C’V’ or C’C’Vù) could be considered stratificational: sensitive to the age factor. 133 Homomorphemic clusters Heteromorphemic clusters B₁ B₂ B₁ B₂ A₁C₁ = 34 ÷ 52 43 ÷ 61 36 ÷ 58 27 ÷ 53 A₁C₂ = 30 ÷ 46 54 ÷ 70 35 ÷ 55 38 ÷ 62 A₂C₁ = 44 ÷ 60 50 ÷ 64 37 ÷ 59 48 ÷ 74 A₂C₂ = 49 ÷ 67 52 ÷ 70 36 ÷ 58 48 ÷ 72 A₃C₁ = 21 ÷ 37 35 ÷ 51 21 ÷ 41 26 ÷ 52 A₃C₂ = 22 ÷ 40 42 ÷ 60 20 ÷ 40 28 ÷ 52 C₃ = 79 ÷ 95 82 ÷ 98 69 ÷ 91 60 ÷ 86 A – age factor; A₁ – older generation, A₂ – middle generation, A₃ – younger generation; B – contextual style factor; B₁ – conversation, B₂ – reading; C – education factor; C₁ – lower-education group, C₂ – medium-education group, C₃ – highly educated group. Fig. 1. Distribution of palatalized (C’C’VᵢC’C’Vᵘ) variants (data from connected-speech recordings) The texts used to investigate (C’C’Vᵢ or C’C’Vũ) are not sufficiently comprehensive to allow a full analysis of the processes of consonant-cluster palatalization; some rarer consonant clusters may not have occurred in them. In an effort to determine the sociolinguistic norm governing the use of this variable, informants were questioned. The only conclusion that can be drawn is that the sociolinguistic norm governing the use of this variable is not conscious; that is, informants do not notice (or hardly notice) variation in (C’C’Vᵢ or C’C’Vũ), either in their own speech or in that of others. Perhaps this is why the sociolinguistic variable in question is “insensitive” to contextual styles. A tendency not to palatalize consonant clusters was observed in two cases: 1) at the boundary of compound words, e.g.: [aŋ.kštp’ip’ir’is] ‘aŋkštpipiris’, [l’ėbž’ėž’ėj] ‘lėpžiedžiai’, [s’k’ė.sv’ė’jis] or [sk’ėr.sv’ė’jis] ‘skersvėjis’, [tũ-kstantm’et’is] ‘tūkstantmetis’, etc.; 2) when the consonant cluster is not assimilated, e.g.: [išsk’ír’t’i] or [išsk’irt’į] ([išsk’irt’i]) ‘išskirti’, [uʂʂ’ũr.’pt’i] or [uʂʂ’ũr.pt’i] ‘užsiurbti’, [pũž’ž’il’is] or [pũž’il’is] ‘pušžilis’, etc. Consonant palatalization at word boundaries. In Standard Lithuanian, hard consonants are pronounced at the end of a word. In many dialects, when a word is shortened, they are retained 134 minkštieji priebalsiai (Zinkevičius, 1966, p. 169–170). Toks tarmiškas žodžio galo priebalsių minkštinimas vilniečių kalboje nėra dažnas. Mokinių kalboje fiksuota 1–25% tokių atvejų (Pakerys, 1971, p. 7). Šis klausimas 1939 m. platintoje lietuvių Tarybos anketoje dar buvo diskutuojamas (Tarties anketa, 1939, p. 64). In connected, especially rapid, speech, Vilnius speakers, while not palatalizing word-final consonants, also change the vocalism at the beginning of the next word; that is, instead of the front vowels i, e, they pronounce unusual mid-vowel allophones [ы], [э], e.g.: [kám ėr’z’in’i] ‘kám ėrzini’, [kàt bм.tum] or, without assimilation, [kàd bм.tum] ‘kad imtum’, [jæunám išman’it’i] ‘jaunám išmanýti’, etc. Such cases are particularly common when the first word becomes a proclitic, e.g.: [t’æn_эl’l’ė:s] ‘ten eilės’, [kat bм.tum] or [kàd bм.tum] ‘kad imtum’, etc. In the transcribed conversation recordings, the facts discussed occurred in the speech of many informants, but it was not possible to process them statistically. In the questionnaire on subjective evaluation of their own speech, three examples were given in which the palatalization of the final consonants of the first word varies in such a way that the vocalism at the beginning of the second word may change. Vilnius speakers were therefore asked whether, in their opinion, they pronounced kad eitum, ten eilės, kam erzini, or kad eitum, ten eilės, kam erzini. This question cannot be reliably investigated by means of a questionnaire. The fact that only three respondents (one from the actors’ and broadcasters’ group and one each representing the older and middle generations) indicated that in such cases as kam erzini, etc., they pronounce э rather than e in the second word should therefore be interpreted as indicating that respondents sense that they generally do not pronounce hard consonants at the end of a word. (Perhaps it was not without reason that there were fears that questionnaire responses would be unreliable simply because many would avoid the unusual spelling with э, bl?) Respondents indicated that they pronounced all three cases alike, without singling out any one of them—that is already a pronunciation pattern. The word combination kam erzini was included in the structured interview protocol: in contextual style B₂, that is, while reading a connected text, every informant used this combination. This made it possible to compare subjective evaluation of one’s own speech with objective speech data from the same informants. This time the results differed considerably: 43% of the older, 38% of the middle, and 52% of the younger generation pronounced [э], not [e], in the second word. Nevertheless, this unanimous view of the respondents, recorded in the questionnaire, can be explained by the fact that the absolute majority of Vilnius speakers have internalized the norm of coordinating the pronunciation of word-final consonants with the pronunciation of the following word, even though this norm is not always observed. Pronouncing hard consonants at the end of a word does not necessarily indicate the emergence of the allophones [ы], [э] in the positions mentioned; however, allegro style inevitably affects surrounding sounds more than lento style, since, when the pause between words is reduced to a minimum, front vowels change after hard consonants (cf. Кальнынь, Масленникова, 1981, с. 30). The speech data from Vilnius show that word combinations pronounced without emphasis and without a clear pause, in which the first word ends in a consonant and the second begins with a front vowel (sometimes also when it begins with a palatalized consonant), are nevertheless treated as two separate units; that is, speakers strive to mark word boundaries. This is precisely why the final consonants of the first word are pronounced hard (and, incidentally, quite often), even when the first word is a clear proclitic. How the articulation of the initial vowels of the next word changes 135 depends on whether the idiolect’s consonantism is strongly palatalized or not: the likelihood of the allophones [ь], [э] emerging is high when consonant palatalization is not strong. In Vilnius speech, a peculiar glottal stop is sometimes heard when speakers wish to pronounce word-final consonants hard while preserving the front vowels at the beginning of the next word unchanged, e.g.: [kad ‘ėi.tum] ‘kad eitum’, [kám ‘im’i] ‘kám imi’, and the like. Palatalization of consonants in non-Lithuanian words. The question of consonant palatalization in words of non-Lithuanian origin, in positions where Lithuanian rules prescribe hard consonants, remains controversial even in the codification of the standard language (cf. Pupkis, 1981; Pupkis, Pakerys, 1976; Vitkauskas, 1985; Pakerys, 1986). Few pronounce l hard in the words altas, pulsas, valsas, kultas, yet the pronunciation of very similar words from a linguistic point of view with hard l is becoming increasingly established: filmas, kultūra, paltas. This is not a new problem; the pronunciation of words of this type was also investigated in a pronunciation questionnaire in 1939. In dialects, the pronunciation of l in consonant clusters, where l occurs before hard consonants in words of non-Lithuanian origin, varies, but the pronunciation of soft l’ is very widespread (Zinkevičius, 1966, p. 157–158). The questionnaire on subjective evaluation of speech presented three such examples: did respondents pronounce filmas, kultūra, paltas, or filbmas, kultūra, palbtas. Figure 2 shows how the use of hard l in non-Lithuanian words varies according to social and sociodemographic parameters, based on data from questionnaires on subjective evaluation of speech. It should be noted that the use of each of the examples listed also differs. Only people from the older generation (42–80%) more often indicated using soft l’ in the word kultūra; representatives of other generations, especially the more highly educated, most often indicated that they used hard l in the word kultūra. The curves for the words filmas and paltas on the same graph are arranged quite similarly: here, conversely, the less educated Figure 2. Use of hard l in the non-Lithuanian words filmas (1), kultūra (2), paltas (3) (data from questionnaires on subjective evaluation of speech) Figure 3. Use of soft l in the non-Lithuanian words alchemikas (1), buhalteris (2), algebra (3) (data from questionnaires on subjective evaluation of speech) respondents indicated more cases of hard l (in the word filmas the maximum was 60%, paltas — 43%). Interestingly, even the control group of actors and broadcasters indicated that they pronounced hard l in the word paltas in only 47% of cases. 136 In the speech of Vilnius schoolchildren, around the 1970s, palatalized l’ was used in 33% of the cases under discussion (Pakerys, 1971, p. 7). Apparently, linguists’ recommendations also play a significant role in the establishment of variants. Undoubtedly, examples from contact languages and established traditions are highly significant in such cases (cf. Kasiulis, 1963), but the history of the word filmas also shows something else. The older generation used to employ the feminine form of this word, borrowed from German (now very rarely): filma. After the war, the masculine form became established quite quickly. Perhaps the fact that this word was linguistically marked for a time—that is, people consciously chose which gender form to use—also enabled the hard l’ to spread more rapidly in this word. The subjective language-evaluation questionnaire also included three examples in which the pronunciation of hard and soft l’ may vary, even though the standard language norm prescribes only the soft variant: alchemikas, buhalteris, algebra (see Fig. 3 for the results). It is interesting that Figs. 2 and 3 are somewhat similar: the middle and younger generations entirely “set apart” the word algebra, as well as the word kultūra, from other similar words. The pronunciation of the words alchemikas and buhalteris differs particularly markedly from the standard language norms: even in the group of actors and announcers, soft l’ variants are used on average in fewer than (or approximately) 50% of cases; education is not a significant factor; young people are especially inclined to depalatalize l’ in the words alchemikas and buhalteris (11–38% soft l’ variants are reported). According to A. Pakeris (1971, p. 7), Vilnius schoolchildren depalatalize l’ in 41% of the words in the cases under discussion. Subjective language data (respondents’ answers to questionnaires) were compared with objective language data from the same informants: the continuous text read by all informants during the interview also included the words filmas and algebra. The comparison showed that members of the middle and younger generations essentially indicated their actual usage in the questionnaires, whereas the older generation “embellished” it considerably: the absolute majority used the forms [f’1.l’ma] ‘filmą’ and [ā.lg’ebra] ‘algebra’. This may be a manifestation of a kind of conscious correction, since some members of the older generation were concerned with “getting rid of” the Polish [u] in place of l’ (cf. Vilniaus krašto…, 1940, p. 5). Thus, both objective and subjective language data showed that Vilnius residents vary considerably in their choice of pronunciation norms for non-Lithuanian words and are not inclined to apply the general Lithuanian sound-combination rules in these cases. This is particularly evident in the language data from the group of actors and announcers. On the other hand, the fact that the pronunciation of some non-Lithuanian words in Vilnius speech stands out markedly from that of other words with a similar pattern shows that language development is shaped not only by objective but also by subjective factors (such as codification and adoption as an accepted, prestigious form; conspicuous changes in form at one level may create favorable conditions for change in form at another level, as illustrated by the word filmas). Depalatalization phenomena. In standard Lithuanian, requirements to depalatalize consonants before e in non-Lithuanian words have occurred only rarely (where two pronunciation variants of this vowel are distinguished—in Lithuanian and non-Lithuanian words) (cf. Vaitkevičiūtė, 1964; Katiliūtė, 1965). 137 In the dialects, the patterns of depalatalization vary considerably for almost all consonants. Depalatalization phenomena in loanwords vary especially widely across dialects (see Zinkevičius, 1966, pp. 158–169; Urbanavičiūtė, 1978). Conversations with Vilnius residents yielded instances of complete and partial depalatalization of the consonants (and affricates) š, ž, dž, r in such Lithuanian words as [š́i.tas] ‘šitas’, [ž́i.notʹa] ‘žinote’, [žovʹiˈkla] ‘džiovykla’, [raikʹėj́o] ‘reikėjo’, and others. This feature was observed to occur optionally only in the speech of the older generation. Five members of this generation depalatalized consonants in their speech in practically all potential cases. All these informants were either connected with dialects in which the depalatalization rules for certain consonants operate (the eastern Vilnius highlanders and some southern highlanders), or may have been influenced by the Polish language, which they had used intensively for a long time. In the speech of two middle-generation informants and one younger-generation informant, relatives of the five older-generation speakers mentioned above, only partial depalatalization was heard. Such consonants were pronounced hard or almost hard, but the vowels following them, as if with some sort of insertion between consonant and vowel, were pronounced as though they followed a soft consonant. Tikriausiai reikėtų sutikti su Urbanavičiūtės (1978, p. 75) siūloma šio reiškinio interpretacija ir priebalsių depalatalizaciją vilniečių kalboje laikyti ilgo ir intensyvaus lenkų kalbos vartojimo pasekme, nes lenkų kalba vienais atvejais galėjo aktyvinti tarmių interferenciją, o kitais – apskritai žadinti polinkį depalatalizuoti kai kuriuos priebalsius. Consonants are depalatalized much more frequently in non-Lithuanian words than in Lithuanian ones. People of all ages pronounce them this way, but older speakers do so most often. Consonants are quite frequently depalatalized (to a greater or lesser degree) in words such as [č́e.xas] ‘čekas’, [t́e.xnʹika] ‘technika’, [ž́elʹe] ‘želė’, [ž́. mpʹerʹis] ‘džemperis’, [t́e.ntas] ‘tentas’, [profʹeˈsorʹus] ‘profesorius’, and others. The subjective language-evaluation questionnaire included 11 words in which depalatalization phenomena may occur in pronunciation (in the example Ryžikovas there are potentially two cases of depalatalization). All the words presented were non-Lithuanian. One example was selected with each of the consonants c, l and r, two with t, three with ž, and four with dž: centras, tempas, tenisas, Ženia, prožektorius, džemas, Džekas, džinsai, Džimis, šašlikas, Ryžikovas. The continuous text read by all the respondents also included the words centras, tenisas, Džekas, Džimis, Ženia, Ryžikovas. A comparison of subjective and objective language data showed that the Vilnius residents of this generation indicated very accurately in the questionnaires the pronunciation they actually used. The distribution of soft consonants in the position under study is such (see Figs. 4–7) that two more characteristic word groups can be distinguished: 1) Russian proper nouns (in the questionnaires and continuous text: Ženia, Ryžikovas) and the word šašlikas, which entered Lithuanian only through Russian; 2) other non-Lithuanian words, both proper and common nouns (džemas, džinsai, prožektorius, centras, tempas, tenisas, Džekas, Džimis, and others). The words in the first subgroup are characterized by much less frequent occurrence of soft consonants in the position under discussion: even in the actors’ and announcers’ group, in free contextual style (B₁), only 61% of all possible instances were pronounced with soft variants, whereas when reading (B₂) the figure was 86% (see Fig. 4). When subjectively evaluating their speech (in the questionnaire 138 in those), respondents also indicated no more cases of using palatalized consonants (from 0 to 44%; actors and announcers—67–78%, see Fig. 5). Incidentally, young people are somewhat more inclined to use palatalized consonants than older people; thus, in this respect, the sociolinguistic pronunciation norm of Vilnius residents is approaching the norms of Standard Lithuanian. Fig. 4. Use of palatalized c’, t’, r’, z’, dž’ before front vowels (data from recordings of connected speech) Fig. 5. Use of palatalized ž’, l’, r’ before front vowels in the words Ženia (1), šašlikas (2), Ryžikovas (3) (data from subjective language-evaluation questionnaires) Fig. 6. Use of palatalized c’ and r’ before front vowels in the words centras (1), rempas (2), temisas (3) (data from subjective language-evaluation questionnaires) Fig. 7. Use of palatalized z’, dž’ before front vowels in the words prožektorius (1), džemas (2), Džekas (3), džinsai (4), Džimis (5) (data from subjective language-evaluation questionnaires) An examination of the pronunciation of words in the second subgroup in Vilnius speech revealed three distinct features here: 1) the curves for objective speech facts (see Fig. 4) and subjective speech evaluations (see Figs. 6 and 7) are very similar, indicating that the usage reported in the questionnaires reflects actual usage; 2) subjective evaluations of the pronunciation of individual non-Lithuanian words in the questionnaires yielded a tightly clustered set of similar curves, showing that a certain pattern has taken hold of pronouncing all words of the same type alike (see Figs. 6 and 7); 3) the pronunciation of older people differs greatly (both in actual usage—8–23% palatalized con- 139 sonants, see Fig. 4, and in subjective evaluation—18–42% palatalized consonants, see Figs. 6 and 7); middle-aged and younger people actually use palatalized consonants in 75–90% of cases (see Fig. 4), while in the questionnaires they report using them in 67–100% of cases (see Figs. 6 and 7). Analysis of the questionnaire responses revealed that respondents apparently have considerable doubts about whether they pronounce a palatalized or a non-palatalized affricate dž before an e-type vowel (e.g., džemas, Džekas), since, for some reason, more highly educated members of the middle-aged generation reported using palatalized affricates in only 67% of cases, whereas less-educated people of the same generation reported using them in 90–100% of cases (see Fig. 7). Here, variation in the degree of depalatalization is the most likely explanation. The fact that nearly all older-generation and some middle-generation Vilnius residents declined to indicate how they pronounce the words Džinsai, Džimis, while not a single younger Vilnius resident—who are certainly familiar with these words—expressed uncertainty, shows how conscientiously Vilnius residents filled in the questionnaires (see Fig. 7). Consonant depalatalization before front vowels in non-Lithuanian words is not entirely a thing of the past. This is also borne out by speech facts recorded on tape from actors (among whom there were no members of the older generation): although not very distinctly, consonants were nevertheless pronounced with depalatalization in the words Džekas, Džimis, centras, tempas. Thus, it can be argued that two sociolinguistic norms for pronouncing non-Lithuanian words are characteristic of Vilnius residents. Russian proper names are pronounced with non-palatalized consonants followed by the altered vowels [ь] and [э] in 80–95% of cases, but the results of subjective speech evaluations show a discernible, though not very pronounced, tendency to use palatalized consonants here. Non-Lithuanian words of this type from other origins are clearly tending to be pronounced in accordance with Standard Lithuanian norms. The fact that older Vilnius residents pronounce them quite differently from their younger neighbours points not only to the possible influence of foreign languages, but also to the absence of more firmly codified norms. We believe that analysis of these forms once again makes abundantly clear how unnecessary, alien, and even harmful the codification of the so-called international e is for the Lithuanian language. Standardizing peripheral language phenomena (the pronunciation of non-Lithuanian words) means effectively making those norms available to the public. This is highly promising and can be implemented rapidly, since it does not affect the language system as a whole. Consonant assimilation phenomena. Assimilation of voiced and voiceless consonants (i.e., coordination of vocal-fold activity) and assimilation of fricatives (dental and alveolar) (i.e., coordination of place of articulation) are extremely strict regularities in Lithuanian dialects (Zinkevičius, 1966, pp. 170–171), and are also observed in Standard Lithuanian. The rule of regressive palatalization discussed earlier should also be mentioned here. It is unreliable to investigate voicing–devoicing assimilation phenomena on the basis of hearing alone; nevertheless, a tendency (albeit a rare one) not to assimilate voiced and voiceless consonants within a word has been observed in Vilnius speech, for example: [n’ėšdavō] ‘nėšdavo’, [l’ūpdamas] ‘lūpdamas’, [s’ėkdamas] ‘sėkdamas’, etc. Perhaps because the expected voicing of consonants does not occur, one can seemingly hear progressive devoicing of the second consonant. The examples cited show that consonant non-assimilation phenomena are possible in Vilnius speech not only at word boundaries but also at morpheme boundaries. 140 Consonant assimilation at morpheme boundaries. Cases of non-assimilation of consonants in Vilnius speech occur in various word positions, but they are most common in compound words and at the boundary between the prefixes iš-, už- and the word stem, e.g.: [l’iepž’iež’aei] ‘liepžiedžiai’, [v’īl.kdal’g’is] ‘vilkdal­gis’, [lū’pdaž’aei] ‘lūpdažiai’; [išbūd’ino·] ‘išbūdino’, [išgr’ež’e·] ‘išgręžė’, [iššūko·] ‘iššūko’, [išš’ūr.pt’i] or [iš’s’ūr.’pt’i] ‘išsiurbti’; in words with the prefix už-, the place of articulation is most often not assimilated and regressive palatalization does not occur, but voicing is generally assimilated, e.g.: [uššā:k’imu.] ‘užsakymų’, [uššukt’i] ‘užsukti’, [uš’s’ū’tas] or [uš’s’ū·tas] ‘užsiūtas’, etc. Although, in Lithuanian linguistics generally, the more widely accepted claim is that geminates are avoided in all dialects (Zinkevičius, 1966, p. 170), and that they should not occur in the standard language either (Pakerys, 1986, p. 121, Mikalajauskaitė, 1962), experimental studies of dialect materials show that degemination is not a universal rule, nor is it absolutely obligatory in the standard language (cf. Strimaitienė, Girdenis, 1978; Kruopienė, 1986). Vilnius speakers quite often, though not always, pronounce double consonants in two contexts: 1) in compound words, e.g.: [pūs’es’er’e·] or [pūs’s’es’er’e·] ‘pušseserė’, [pūssuŋk’is] ‘pūssunkis’, [švā.rrašt’is] ‘švārraštis’; 2) at the boundary between the prefixes iš-, už- and the word stem, e.g.: [iššlav’e·] ‘iššlavė’, [užž’er’ū] or [užž’er’ū] ‘užžeriu’, etc. Studies of the speech of Vilnius schoolchildren found 49% of cases of consonant non-assimilation (Pakerys, 1971, p. 8). Vilnius speakers’ geminates differ from, for example, Russian double consonants. In Russian, essentially a single long consonant is pronounced, whereas Vilnius speakers tend to mark the morpheme boundary and therefore quite often pronounce two consonants. A comparison of actual and subjective speech facts showed that Vilnius speakers in fact assimilate consonant clusters in heteromorphemic sequences much more often (in 60—82% of cases; actors and announcers — 97%, see Fig. 8) than they think they do (in 0—28% of cases; even actors and announcers — only 44—67%, see Fig. 9). This question is probably too difficult to investigate so directly by means of a questionnaire: spelling has too much influence. Nevertheless, the fact that, in actual speech, more highly educated informants assimilate consonants somewhat less often, especially when reading a text (B₂ style), would suggest that spelling has an effect. As questionnaire data indicate, a norm of not assimilating spoken consonant clusters in prefixed words is taking shape, and the fact that it is well established in the evaluative language orientations of actors and announcers indicates that, in the future, the tendency not to assimilate may become a conscious sociolinguistic norm. (At present, this is not a conscious sociolinguistic norm—they would like to pronounce unassimilated sounds, but do not: in more than 60% of cases they assimilate them.) Consonant assimilation at word boundaries. Standard-language norms require that only voiceless consonants be pronounced at the end of a word. In dialects, outside phrase-final position, the usual assimilation processes generally occur (Zinkevičius, 1966, p. 170), but deletion of word-final endings can also lead to other cases. This issue was raised in the 1939 pronunciation questionnaire. In Vilnius speech, consonants are assimilated less often at word boundaries than at morpheme boundaries, but statistical analysis of the transcribed text had to be abandoned because phenomena of information structure are also highly significant here, and their study was not an objective of the research. 141 Vilnius residents may use both [aš s’v’ieš’č’au] ‘I would brighten’, [a(s’) s’v’ieš’č’au], [báltas ž’ėž’ėl’is] ‘a white shadow’, and [bálta(s’) ž’ėž’ėl’is]. When assimilation occurs, elision of the final consonant of the first word is possible. Fig. 8. Use of assimilated consonant clusters in heteroformic sequences (data from recordings of connected speech) Fig. 9. Use of assimilated consonant clusters in the words išskaityti (1), užsukti (2), išsemti (3) (data from subjective language-evaluation questionnaires) Elision. Consonant loss is not a frequent phenomenon either in Lithuanian in general or in Vilnius speech. In Vilnius speech, however, facts of intervocalic v elision are increasingly often heard, especially in rapid speech, e.g.: [bū’ọ’m’e] ‘we were’, [tā’ọ:] ‘yours’. The consonant j is sometimes not pronounced: 1) at the absolute beginning of a word, e.g.: [is] ‘he’, [iėva] ‘willow’, [ieškọ’jæu] ‘I searched’; 2) between a prefix and a root, e.g.: [suėškọ:] ‘finds’; 3) between p, b and non-front vowels, e.g.: [rukp’ūt’is] ‘August’, [s’p’æut’i] ‘to spit’, [b’æurūs] ‘ugly’; 4) in international words, e.g.: [p’io:n’ier’ūs] or [p’ọn’ier’ūs] ‘pioneer’, [soc’el’izmas] ‘socialism’, [b’io’lọg’ijae] or [b’ọlọg’ijae] ‘biology’ (cf. Mikalauskaitė, 1975, p. 46 and Pakerys, 1986, p. 141 – the opinions differ somewhat). The consonant j at the beginning of a word, especially in the pronouns jis, ji, is usually not pronounced in the Southern Aukštaitian dialect (Zinkevičius, 1966, p. 178); cases of j elision after p, b have also been observed in the same dialect (ibid., p. 187), especially in the speech of the younger generation. According to A. Pakerys’s data, Vilnius schoolchildren do not pronounce j after p, b in 11–22% of cases (1971, p. 11). Undoubtedly, a major influence was exerted by the codified spelling norm for the roots spjau-, pjau-, bjau-, in force until 1977, which omitted j. The use of consonant j at the beginning of a word before front vowels is a difficult question. Recordings of connected speech showed that, first of all, facts of this kind are quite difficult to document because they depend greatly on the phonetic environment; therefore, positions in which j- occurs under the influence of a nonsyllabic -i at the end of the preceding word were not investigated at all. In addition, it was observed that j- elision is a fairly rare phenomenon (j- is pronounced in 77–88% of cases; see Fig. 10), although it is noticed quite readily. It was noted that this is an idiolectal feature, so the possible influence of the parents’ dialect was examined. It turned out to be a decisive factor: children from the Southern Aukštaitian dialect area failed to pronounce j-. much more often than the other informants. Three examples were included in the subjective language-evaluation questionnaire: ieškoti, ieva, iešmas. A fairly tight bundle of curves was obtained (see Fig. 11): the responses of the older and younger generations differ considerably; more precisely, only the responses of less-educated respondents from the older generation indicated that they used j- in just 20% of the possible cases, whereas the youngest generation reported 70–100%. Apparently, orthography had a particularly strong influence on respondents in group A₁C₁, since actual usage differs much less (A₁C₁ – 77–82%, A₁C₂ – 82–85%, and A₃C₁ and A₃C₂ – 84–86%; see Fig. 10)ᵃ. 142 responses from the older and younger generations, more precisely, only the responses of less-educated respondents from the older generation indicated that they used j- in just 20% of the possible cases, whereas the youngest generation reported 70–100%. Apparently, orthography had a particularly strong influence on respondents in group A₁C₁, since actual usage differs much less (A₁C₁ – 77–82%, A₁C₂ – 82–85%, and A₃C₁ and A₃C₂ – 84–86%; see Fig. 10)ᵃ. Interestingly, ratings of the pronunciation of j- in the word ieškoti differed somewhat in the questionnaires: both middle-generation respondents and actors and announcers indicated that they pronounced j- less often specifically in this word (see Fig. 11). Surprisingly, in recordings of connected speech as well, j- elision occurred most often in a word with this particular root, but this does not yet prove that respondents are truly so perceptive about their own speech. Incidentally, the words ieškoti and ieva were incorporated into a short text read by all informants (contextual B₂ style). Thus, j- elision is not frequent in Vilnius speech. In a stronger position (a stressed syllable), j- is pronounced more often than in a weak position (an unstressed syllable). Fig. 10. Use of consonant j (j- and -j-) (data from recordings of connected speech) Fig. 11. Use of consonant j (j- and -j-) in the words ieškoti (1), iešmas (2), ieva (3), spjūvis (4), rugpjūtis (5), bjaurus (6) (data from subjective language-evaluation questionnaires) The sociolinguistic norm for pronouncing j- is unconscious; Vilnius residents do not react at all when j- is not pronounced—they do not notice it. Pronouncing consonant j in words whose roots are bjaur-, pjau-, spjau- is characteristic of Vilnius speech, although recordings of connected speech nevertheless show that even elision of -j- in the relevant position is familiar to the youngest generation of Vilnius residents in particular: elision was documented in 0 to 12% of cases (see Fig. 10). This is generally an idiolectal feature, but there are few such individuals; the speech of the older generation may in this respect also have been influenced by the Polish they actively used, while the young were undoubtedly influenced by orthography. The latter explanation is supported by the fact that less-educated respondents more often reported using variants without -j-: this is undoubtedly an influence of orthography. The responses to the questionnaire of less-educated older Vilnius residents stand out in particular (see Fig. 11); in general, they are notable for their indifference to the correctness of their speech. Thus, by reporting that they use 60% of this type of word without -j-, they are following the old spelling norm. II. Accentuation Vilnius speech is characterized by the following features of accentuation: 1) approximately one fifth of words are stressed in ways that do not conform to the norms of Standard Lithuanian, 2) there is a tendency to level pitch accents, 3) a tendency toward columnar morphemic stress is becoming apparent. 143 Stress placement. Analysis of recordings of connected Vilnius speech showed that stress deviated from the norms of Standard Lithuanian in only 11–18% of possible cases (see Fig. 12). Thus, stress is not shifted away or toward a syllable all that often; something else is important: sometimes stress retraction does not follow the general requirements of the Lithuanian language (both dialectal and standard), contravening the requirements of Saussure–Fortunatov’s law. This is a violation of the fundamental laws of Lithuanian, and its cause was of greatest interest to clarify. And one more feature: non-Lithuanian vocabulary is stressed in ways that do not conform to Standard Lithuanian norms much more often than Lithuanian vocabulary is. Vilnius speech shows a tendency toward columnar morphemic stress, which may be attributable both to analogy and to the leveling of pitch accents (cf. Grinaveckis, 1977; Eidukaitienė, 1977; Robinson, 1984; Young, 1984, etc.), e.g.: ryšiai (Std. ryšiai, 4), įdomūs (Std. įdomūs, 4), spjūviu (Std. spjūviu, 2), kieme (Std. kieme, 4), sėšlius (Std. sėšlius, 4), pilką (Std. pilką, 3), liečiu (Std. liečiu), ruoši (Std. ruoši), įnirštu (Std. įnirštu), bitė (Std. bitė, 2), muses (Std. muses, 2), berželius (Std. berželius, 2), etc. The tendency toward columnar morphemic stress is particularly pronounced in words of non-Lithuanian origin, e.g.: troleibusas (Std. troleibusas, 2), gitara (Std. gitara, 2), Džimiu (Std. Džimiu, 2), architektus (Std. architektus, 2), viaduke (Std. viaduke, 2), etc. And another feature specific to Vilnius speech is hypercorrection, occurring when, in accordance with Saussure–Fortunatov’s law, stress is shifted to a short ending not only from a short or circumflex syllable but also from an acute syllable, e.g.: vienu (Std. vienu), kaulė (Std. kaulė), galvas (Std. galvas), plotu (Std. plotu), lūžtu (Std. Fig. 12. Stressing without deviation from the requirements of Standard Lithuanian (data from recordings of connected speech) Fig. 13. Stressing of the words pridengia (1), nubreiže (2), išsprendė (3) without stress retraction (data from subjective language-evaluation questionnaires) lūžtu), plyšti (Std. plyšti), etc. Hypercorrection is based on the convergence of pitch accents, and in this case its mechanism is driven by the desire to use as many prestigious forms as possible. In the subconscious of Vilnius speakers, the following often takes shape: they know that they tend to retract stress from endings in cases where Standard Lithuanian does not, and so all endings that could potentially receive stress are stressed. 144 The subjective language-evaluation questionnaire included questions about the accentuation of 33 words. Analysis of the responses showed that the greatest deviations from standard-language norms in accentuation occurred with those prefixed verb forms whose accentuation patterns in the standard language are fairly complex and vary considerably across dialects (cf. Laigonaitė, 1978, pp. 80–82). As also shown in Fig. 13, the accentuation of the words pridengia, nubrėžė and išsprendė varies greatly, with nonstandard forms fluctuating (even actors and announcers gave only 50–80% standard forms). In general, there is a tendency in such cases to accent the prefix, although the curve for the word išsprendė does not confirm this. Nevertheless, apparently there is a greater tendency to shift the accent from a circumflex root to the prefix than from an acute one. Various cases of similar accentuation occurred in recordings of connected Vilnius speech, for example: neišbrenda (standard: neišbrenda), pavelka (standard: pavelka), užlenkia (standard: užlenkia), etc. Both dialectal accentuation and other factors may have played a role (accentual colonization within the paradigm, analogy, etc.). The second group of words studied comprises words of non-Lithuanian origin. Their accentuation varied enormously (see Fig. 14). Incidentally, this was also observed in the speech of Vilnius schoolchildren (Grabauskas, 1970, pp. 19–20). The accentuation of this group of words differed somewhat by grammatical gender: non-Lithuanian feminine words (Violetà, anginà, gitarà) were more often accented nonstandardly than masculine words (troleibūsas, pedagògas, architektas). Even actors and announcers reported accenting the name Violetà on the final syllable in only 52% of cases. The accentuation of masculine nouns was also somewhat conditioned by stratificational variation: more highly educated Vilnius residents tended to use standard forms of these words more often. The words gitarà and troleibūsus occurred in a connected text read by every informant. It was calculated that gitarà was accented on the final syllable in 17% of cases, while troleibūsas was pronounced without accent retraction in 18%. Thus, once responses from one group were excluded, subjective evaluations of their own speech, reflected in questionnaire responses, essentially revealed the actual situation. This is the currently internalized sociolinguistic norm for accentuating this group of words. The third group of words examined in the questionnaires comprises the numerals septyni, aštuoni, devyni. As can be seen in Fig. 15, the bundle of curves is very compact, indicating that an accentuation pattern for these words has become established. However, the fact that responses “hover” around the 50% mark indicates that Vilnius residents are completely undecided on this issue. Fewer cases of final-syllable accentuation were recorded in connected texts than were reported in the questionnaires. Thus, no sociolinguistic accentuation norm has become established, and the prestige form is unclear; apparently, speakers therefore tend to follow the most general pattern: shifting the accent from a short final syllable to the long penultimate syllable. The fourth group of words examined in the questionnaires comprises plural and singular locative forms: namuose, akmenyse, naujose, pirtyje, darbė, Kaunè. First and foremost, the accentuation of two-syllable forms stood out: here, standard-language norms are followed almost consistently (see Fig. 16). On average, plural forms tend to be accented on the ending in only slightly more than 50% of cases, but there is a particularly marked difference between older and middle- 10. 1880 145 the references of this generation: the former are quite disinclined to stress these words on the final syllable, whereas the latter have opted for the normative stress pattern with stress on the ending. Interestingly, in connected-speech texts, most instances of non-final stress on forms of this type were indeed recorded in the speech of the oldest and youngest generations of Vilnius residents. In general, this stress pattern is realized in approximately one third of the possible cases. Thus, a clearer sociolinguistic norm has developed only for the stress of disyllabic forms: there is a tendency to stress the end of the word. The same cannot be said of multisyllabic locative forms. Fig. 14. Stressing of non-Lithuanian words Violeta (1), angina (2), gitara (3), troleibusas (4), pedagogas (5), architektas (6), without attraction (data from questionnaires on subjective language evaluation) Fig. 15. Stressing of the numerals septyni (1), aštuonių (2), devyni (3), without attraction (data from questionnaires on subjective language evaluation) The fifth group of words examined in the questionnaires consists of singular instrumental forms: su padorių, su peilių, su gėlė, su ranka (see Fig. 17). Once again, feminine forms, more often stressed on the ending, could be grouped separately from masculine forms, which less often receive stress on the ending. The responses of the more highly educated representatives of the youngest generation of Vilnius residents indicate that a certain sociolinguistic norm is emerging: they are little inclined to retain normative final stress (with the exception, admittedly, of the word ranka). The responses of actors and announcers show that the tendency to stress a non-final syllable is persistent: even they do not claim always to stress these words on the ending. The stress of the form su padorių may be related to the neutralization of the tones of monophthongs; the stress of the form su gėlė could be explained in a similar way. The stress of the form su peilių can be explained by the tendency for columnar morpheme stress to become established. The stress of the form su ranka may be explained by looking at Fig. 20, where the curve for the similarly hypercorrectly stressed form su pirmą (standard form: su pirma, 3) is shown: one may assume that disyllabic feminine forms ending in -a have acquired a generalized feature—all are now being stressed on the ending. The sixth group of words examined in the questionnaires consists of disyllabic plural accusative forms: butūs, pirštūs, penkiàs, žiniàs. One might have expected the curves for the forms butūs and žiniàs to nearly coincide, since these are similar phonological words: both have a short first syllable. But this did not happen (see Fig. 18): these two curves are the farthest apart, although they are fairly symmetrical. Perhaps the contexts in which the words are used played a role: žiniàs is usually heard correctly stressed on the radio and TV, whereas butūs 146 is more often pronounced in informal situations; it is a very common word in the speech of Vilnius residents, so its use was particularly conducive to the spread of columnar morpheme stress. Overall, the questionnaires indicated that an average of three-quarters of forms were used with final stress. Interestingly, in the text prepared for connected reading, the word butus was stressed on the ending in 72% of cases. Fig. 16. Stressing of locative forms namuose (1), akmenyse (2), naguose (3), pirtyje (4), darbe (5), Kaune (6), without attraction (data from questionnaires on subjective language evaluation) Fig. 17. Stressing of singular instrumental forms su padoria (1), su peiliu (2), su gėle (3), su ranka (4), without attraction (data from questionnaires on subjective language evaluation) Fig. 18. Stressing of plural accusative forms butus (1), pirštus (2), penkias (3), žinias (4), without attraction (data from questionnaires on subjective language evaluation) Fig. 19. Stressing of the words ateinu (1), supranti (2), nerandu (3), without attraction (data from questionnaires on subjective language evaluation) The seventh group of words was assembled according to a rather unusual principle: in connected speech, the words yra, ateinu and supranti are particularly often stressed on a non-final syllable. Although the resulting bundle of curves is fairly tight (see Fig. 19), this graph does not reflect actual usage. It only shows that words of this kind, whose stress depends on intonation and rhythm, are almost impossible to study using questionnaires. Incidentally, the words yra and supranti were also included in the reading passage. The interview results largely coincided with the indications in the questionnaires, but this may have been a coincidence. The eighth group of words examined consists of possible cases of hypercorrection. Setting aside the already discussed stress of the form su pirma, one can observe that the curves for the other forms lie in the upper third of the graph, indicating a tendency to use 10* 147 more normative accentuation paradigms. It should be noted that when speaking, Vilnius residents tend to hypercorrectly accentuate forms of the third, rather than the first, accent paradigm, e.g.: with a thick one, with the eastern one, with the expensive one, etc. Fig. 20. Accentuation and hypercorrection of the words pirma (1), pilkus (2), nendres (3), su broliu (4) (data from subjective language evaluation questionnaires) The most prominent feature of Vilnius residents’ accentuation is fluctuation and uncertainty. This indicates that major changes are taking place; in other words, the changes are emerging right now. Л. ГРУМАДЕНЕ SOCIOLINGUISTIC STUDY OF THE LITHUANIAN SPEECH OF VILNIUS RESIDENTS: CONSONANTISM AND ACCENTUATION Summary From a sociolinguistic perspective, this empirical study of the Lithuanian spoken language of native Vilnius residents showed that some shifts in the realization of consonant palatalization are observed in the consonant system; it established that consonant assimilation is determined mainly by factors of a morphological rather than phonetic nature; cases of j elision are observed, facilitated by spelling as well as by the influence of dialect speech. The accentuation of Vilnius residents is characterized, first, by deviation from the accentuation patterns of standard Lithuanian (on average, for one fifth of all words) with respect to the position of stress; second, by a tendency toward the leveling of syllabic intonations (acute and circumflex), which leads to a violation of the Saussure–Fortunatov law; and third, by tendencies toward fixed morpheme accentuation. L. GRUMADIENĖ THE SOCIOLINGUISTIC STUDY OF LITHUANIAN SPOKEN IN VILNIUS: CONSONANTISM AND ACCENTUATION Summary The sociolinguistic variances and sociolinguistic norms of Lithuanian spoken in Vilnius are analysed. The nature of speech variation has been considered in order to evaluate the role of social factors. The study was made especially of the consonantism and accentuation. A thorough 148 study of the tape-recorded speech over 70 informants was made and compared with subjective evaluation of their own speech. 20 graphs are introduced to simplify the discussion. On the basis of this data the main characteristics of consonantism are found and it is concluded that the character of palatalization differs from that found in Standard Lithuanian or in dialects; the relation between the morphological (not phonetical) factors has been established in the sphere of assimilation; the facts of elision of j are also briefly discussed. The main characteristics of accentuation are: (1) the tendency of stress attraction is recognized (about 20% in tape-recorded data); (2) the levelling of stress intonation is obtained, which increases the infringement of the Saussure-Fortunatov’s law; (3) the tendency towards columnal morpheme accentuation is observed. LITERATŪRA Edukaitienė E. V. Kupiškėnų monoftongų priegaidės (audicinis tyrinėjimas) // Kalbotyra. 1977. T. 28, sąs. 1. P. 18–23. Girdenis A., Piročkinas A. Jonas Jablonskis – dialektologas // Kalbotyra. 1977. T. 28, sąs. 1. P. 29–39. Grabauskas V. Dėl rašytinės ir šnekamosios lietuvių kalbos kultūros Vilniaus A. Vienuolio vidurinėje mokykloje // Mūsų kalba. 1970. Nr. 20. P. 16–21. Grinaveckis V. Dėl lietuvių kalbos tarmių kirčio atitraukimo kilmės // LTSR MA darbai. Ser. A. 1977. T. 2 (59). P. 129–135. Young S. The Accentuation of Colloquial Lithuanian. Unpublished dissert. University of Chicago, 1984. Kasiulis P. Dėl priebalsio l ir samplaikų ng bei nk tarimo // Kalbos kultūra. 1963. Nr. 4. P. 59. Katiliūtė I. Dar dėl kai kurių priebalsių tarties // Kalbos kultūra. 1965. Nr. 9. P. 66. Kruopienė I. Priebalsių palatalizacija ir išorės sandis šiaurės panevėžiškių tarmėje // Jaunųjų filologų darbai. 1986. T. 2. P. 139–142. Laigonaitė A. Lietuvių akcentologija. V., 1978. Mikalauskaitė E. Ar yra lietuvių kalboje dvigubų priebalsių? // Kalbos kultūra. 1962. Nr. 2. P. 58–59. Mikalauskaitė E. Lietuvių kalbos fonetikos darbai. V., 1975. Pakerys A. Kaip Vilniaus mokiniai taria priebalsius // Mūsų kalba. 1971. Nr. 6(28). P. 6–13. Pakerys A. Lietuvių bendrinės kalbos fonetika. V., 1986. Pakerys A., Pupkis A. Lietuvių kalbos bendrinė tartis: Plokštelių tekstai. V., 1976. Pupkis A. Kalbos kultūros pagrindai. V., 1981. Robinson D. F. Is Standard Lithuanian Losing Its Tone? // Folia Slavica. 1984. Vol. 7. N 1/2. P. 283–289. Tarties anketa: Nenusistoję bendrinės tarties klausimai / LKD Tarties ir kirčiavimo komisija // Gimtoji kalba, 1939. Sąs. 4. P. 64. Urbanavičiūtė Z. Priebalsių t, d, n ir b, p, m, v kietinimas lietuvių kalbos tarmėse // Kalbotyra. 1978. T. 29(1). P. 69–78. Vaitkevičiūtė V. Dėl priebalsių tarimo prieš tarptautinių žodžių balsį e // Kalbos kultūra. 1964. Nr. 7. P. 71. Vilniaus krašto...: Vieno Vilniaus korespondento pastabos: Vilniaus krašto lietuvių kalbos mokytojų dėmesiui // Gimtoji kalba. 1940. Sąs. 1. P. 5–6. Vitkauskas V. Lietuvių kalbos tarties žodynas. V., 1985. Zinkevičius Z. Lietuvių dialektologija: Lyginamoji tarmių fonetika ir morfologija (su 75 žemėl.). V., 1966. Калиньш Л. Э., Масленникова Л. И. Сопоставительная модель фонологической системы славянских диалектов. М., 1981. 149