Samplaikinių priebalsių „kietumo : minkštumo“ priešpriešos nykimas Pelesos šnektoje

LIETUVIŲ KALBOTYROS KLAUSIMAI XXXVIII (1997) Edmundas TRUMPA THE DECLINE OF THE “HARD: SOFT” OPPOSITION IN CONSONANT CLUSTERS HARDNESS: SOFTNESS IN THE PELESA DIALECT The dialect of Pelesa (Varanãvo district, Belarus), like other peripheral dialects of Lithuanian, is characterized by an interesting phonological phenomenon: consonant clusters occurring before front vowels or fronted back vowels (i.e. in the [(…C'₁)C'₂ Vi⁄Vᵘ] position) are not entirely soft. The left-hand members of consonant clusters often do not obey the law of softening assimilation, e.g.: nuzgãn.d'.i.s ‘frightened’, s'iept'.i.n'ù. ‘of seven’, sk'ĩ.l'e.s ‘holes’, g'iesm'ã.s ‘songs’, v'izg'i ‘still’, tr'ĩ.nau ‘I rubbed’, mã.rš'k'in'us ‘shirts’, šašč'ĩ ‘sixth’, g'irg'ė.t' ‘to hear’, garz'ĩ ‘tasty’. This dialectal phenomenon, which does not conform to the rules of the standard language, has been observed and described by quite a few linguists (cf. Zinkevičius 1966: 157, 163; Sudnik 1975: 65; Grumadienė 1988: 132). Such cases of regressive palatalization are usually explained by the cluster members’ lack of functional softness or by the influence of foreign phonological systems (Lithuanians and Belarusians in Pelesa pronounce the consonants r, š, ž, č, dž as hard, although they may be soft in clusters). An attempt has also been made to model the phenomenon using phonotactic rules, whereby non-palatalized clusters signal an open juncture (Strimaitienė, Girdenis 1978: 61; Girdenis 1995: 40). When discussing this phenomenon, it is particularly useful to draw on material from the Pelesa dialect, since many words in this dialect containing the clusters in question are frequent. The clusters are easy to hear and are not “obscured” by specific dialect features, so even an untrained ear can readily distinguish their soft and hard members. All the more interesting is that these clusters often have variants that attest to the process and suggest that the increase in non-palatalized clusters is not haphazard. The material was collected in 1994 in villages around Pelesa: Pavalakė, Leliušiai, Dubiniai, Piliūnai, Paditvys, and in Pelesa itself. The area, regarded as an island of Lithuanian-Belarusian bilingualism, seems to reflect several stages of linguistic change to differing degrees: from the almost complete extinction of Lithuanian (in Leliušiai), where only the very oldest residents (just a few of whom were still alive in 1994) had retained Lithuanian and thus bilingualism, through a balance between the two languages (Pelesa, Piliūnai, Dubiniai, Paditvys), to genuine, largely unimpeded dominance of Lithuanian (Pavalakė) (Sudnik 1980: 181; Vidugiris, 1988: 24). The 214 recordings on these magnetic tapes, held in the Department of Language History and Dialectology of the Institute of the Lithuanian Language (registration nos. 353–355, 360–362), capture the continuous speech (narrative or conversation) of 17 informants. According to the sociolinguistic questionnaire, all the informants, born between 1909 and 1936, were Lithuanian and had spoken Lithuanian at home from childhood and with people of their own generation (except for one, born in 1941, whose mother was Polish and who now considers herself Polish). The informants communicate in Belarusian with neighbors who speak other languages, some with their Belarusian wives, and with children and grandchildren who cannot speak Lithuanian. This may explain the differences in the richness of the informants’ speech and the variation in the number of words containing consonant clusters. Three types of consonant clusters were examined: fully palatalized ([C₁C₂ Vi/Vᵘ]), clusters entirely unaffected by assimilative softening ([C₁C₂ Vi/Vᵘ]), and entirely hard clusters ([C₁C₂ Vi/Vᵘ]) (for more details, see Trumpa 1994: 23). Each consists of two adjacent consonants, selected for analysis from a two-, three-, or four-member group (for example, the word aŋks'c'í ‘early’ yielded the clusters ŋk, ks', s'c') (cf. Sudnik 1972: 15). In proportion to the length of each informant’s recorded text, 30, 50, or 60 consonant groups were selected. The total number of examples found was 923. They are distributed as follows: a) ([C₁C₂ (…)]) – 92 clusters (437 occurrences); b) ([C₁C₂ (…)]) – 82 clusters (313 occurrences); c) ([C₁C₂ (…)]) – 46 clusters (183 occurrences). Since the informants’ sociolinguistic profiles do not differ greatly, age is a useful criterion for considering the use of variants of the juncture structures. It was observed that informants born around 1909–1920 retained the soft consonant juncture correlates at an approximate ratio of 2:1; those born in 1920–1925 retained them at a ratio of about 1.5:1; and the younger generation, born in 1925–1941, retained them at a ratio of 1:1. The causes of the decline of soft correlates might seemingly be sought in the local Belarusian dialects and explained by the effect of bilingualism; however, this view should be treated with caution because, as M. Lekomceva puts it, “the interaction of Lithuanian, Polish, and Belarusian dialects is a continuation of language contacts that began in Vilnius Region as early as the 13th–14th centuries.” (Lekomceva 1972: 127). Variants of soft and hard correlates in consonant clusters are also possible in the Belarusian dialect: s'p'avái / sp'avái ‘sing’, z'm'él'im / zm'él'im ‘we will grind’; or, explained as a Lithuanian substratum, the use of the soft r’ variant, occurring most often in a group of soft consonants: barv'óna / barv'óna ‘varnish’, p'ars't'ónik / p'ár's't'ónik ‘ring’ and so on (Lekomceva 1972: 126, 127). 215 Thus, the influence of Polish and Belarusian dialects, which have a complex consonant system, cannot be denied; however, it is necessary to look for the preconditions for change within the phonological system of the Lithuanian dialect of Pelesa itself. T. Sudnik, who studied Lithuanian, Belarusian, and Polish dialects in Lazūnai in the 1970s, established the phoneme distribution fields of these dialects and compiled detailed tables of possible phoneme clusters (Sudnik 1972: tables 10, 22, 31). The phonological systems of the Pelesa and Lazūnai dialects differ little, so a cautious comparison of the data is possible, allowing some theoretical assumptions to be made. The notations from the table prepared by T. Sudnik can be transferred to the table of junctures of the [R/R′ T′ Vᶦ/Vᵘ] type identified in this case (using the symbol “P” to mark clusters found in the Lithuanian dialect of Pelesa, “L” for clusters possible in the Lithuanian dialect of Lazūnai, and “G” for clusters possible in the Belarusian dialect of Lazūnai): p′ | b′ | t′ | d′ | k′ | g′ | c′ | č′<c′ | ʒ′ m | PLG | G | P | P | LG | LG | PLG | | PG m′ | PLG | LG | L | L | LG | LG | LG | | LG n | PG | G | P | P | PG | PG | P | | P n′ | LG | LG | PL | PL | PLG | PLG | PLG | | LG l | G | PG | | | PG | G | G | | G l′ | LG | LG | PL | PL | PLG | PLG | PLG | | PLG r | PLG | PLG | PL | PL | PLG | PLG | PLG | PLG | PLG r′₂ | P | P | P | P | P | P | P | | P 1 The presence of soft m in clusters of this type is rather problematic; its palatalization should be examined under laboratory conditions. 2 Lazūnai lies south of the boundary beyond which r has neither positional nor functional softness (cf. Urbanavičiūtė 1970: 78), whereas in Pelesa r′ retained its softness in certain rare cases (e.g., in clusters). A comparison of the data yields a dozen or so clusters, of which: 1) mʒ′ (pasamʒ′ič′i ‘to hire’), np′ (bɪn̥p′iːl′eːm ‘we poured in’), nk′ (laŋk′aːla ‘little bend’), ŋg′ (laŋg′v′aũ. ‘more easily’), lb′ (žukalb′eːc′u ‘to talk to’), lk′ (šakā.lk′e. ‘a place where firewood is chopped’) – can theoretically be explained as arising from the influence of Belarusian consonant phonotactics; 2) mt′ (krəmt′iːc′is ‘to chew’), md′ (pasəmd′aũ. ‘I hired’), nt′ (šv′eːnt′ino.), nd′ (dund′aːkl′is ‘Dundeklis’), nc′ (pašv′aːnc′iio. ‘consecrated’), nʒ′ (n′esužinōːʒ′inʲin′eːjo. ‘kept asking without finding out’), as well as r′p′ (ʒ′iːr′p′c′ ‘to work’), r′b′ (ab′d′iːr′b′t′in′eːjo. ‘was processing’), r′t′ (dabar′t′eːs ‘nowadays’), r′d′ (gar′d′in′i ‘Gardinė’), r′k′ (prɪs′p′iːr′k′i. ‘having bought’), r′g′ (paš′k′er′g′e. ‘slaughtered’), r′c′ (ar′c′aũ. ‘closer’), r′ʒ′ (g′iːr′ʒ′u ‘I hear (I heard)’) – cannot be explained by Belarusian phonological rules. Thus, clusters cannot be characterized solely as a result of language interaction; the phenomenon is much broader and probably governed by its own internal laws. 216 Therefore, account should be taken of the position occupied by clusters in the word, as well as of the structure of those clusters and the properties of the consonants that compose them. First, it should be noted that the members of the aforementioned clusters belong to different phonological syllables (and often also to different morphemes, whose boundaries coincide with syllable boundaries), cf. p’engk’i ‘five’, v’ienk’ė.m’ei ‘single homesteads’, nurinŋd’it’ ‘to gather up’, nuzgɔn.d’is ‘frightened’, turt’iŋg’i ‘rich’, skurd’ei. ‘poorly’, pirk’ă.l’e.i ‘in the shop’, v’ierp’əm ‘we spin’, á.rc’ (< á.rc’i) ‘to come’, d’ă:šimc’ (< d’ă:šimc’is) ‘ten’, šįrz’įn. ‘into the heart’. The RT consonant-juncture type (like the RS, TS, TT, RR, and SS types) has the potential to depalatalize the first member of the juncture. This type of juncture entails a certain divide between the implosive (first member) and explosive (second member) groups, and assimilation of palatalization often cannot overcome this divide: má.ršk’in’us ‘shirts’, l’á.ŋk’iška ‘Polish’, i.š’ig’á.rc’ ‘to get into’, ins’ik’le.pó.k ‘look closely’, peñ.s’įjo.s ‘pensions’, k’èpc’ (< k’èpc’i) ‘to bake’, b’arn’ų.kas ‘boy’, ká.rv’e.s ‘cows’, k’irm’ė.l’ė. ‘worm’, puss’æs’er’ės ‘female cousins’. The total number of RT, RS, TS, RR, TT, and SS type junctures makes up more than half (189 of 313) of the cases in which one consonant is not palatalized. Of these, 112 are homomorphemic clusters, which can be explained by their members belonging to different phonological syllables, and 77 are heteromorphemic clusters, which can be treated as realizations of an open juncture. Comparing partially affected by palatalization assimilation (C₁C₂) and palatalized (C’₁C’₂) RT, RS, TS, TT, RR, and SS type correlates, the palatalized forms are retained somewhat more often (usage ratio 1.89: 2.05). As noted, these are more frequently used by older speakers. But how can the remaining 124 cases of semi-palatalized clusters be explained? These are ST, SR, and TR type junctures, which are not divided into phonological syllables and only sporadically belong to different morphemes (39 instances of use): sp’ėc’æs ‘swarm(s)’, sk’ietas ‘a sliver’, drùšk’in’i.kuosa ‘in Druskininkai’, rakšč’ė.s ‘of the splinter (of a coffin)’, mō’rk’ė.l’išk’ė. ‘Morkeliškė’, áukšt’ės’n’ė.s ‘higher’, mišk’i ‘in the forest’, šašč’i ‘sixth’, bažn’į’ča ‘church’, šn’á.ka ‘speaks’, g’ėsm’ă.s ‘songs’, pàs’l’æp’ė. ‘hid’, tr’īs ‘three’, l’itv’ó.kai ‘Litvaks (Lithuanians)’, pl’įk’ėi ‘Plikiai’, (i) išpl’ė.šč’ (< (i) išpl’ė.šč’i) ‘to tear out’, bl’į.nu. ‘pancakes’, etc. Here, the reasons for the non-palatalization of the first cluster member may be sought in the articulatory basis of the Pelesa dialect. It is known that in this dialect š, ž, (s), and especially r have lost their positional and, to some extent, functional palatalization. This may cautiously explain the hardening of the first cluster members; cf. šašb’i ‘six’/ šn’á.ka ‘speaks’, žūrbi ‘looks’/ šešėzd’á:šimc’ ‘sixty’, sė.ž’i ‘sits’ (although hard s before ė, e is very rare)/ sm’ėlė ‘sand’. It may be that the depalatalization of consonants t, d, p, b (as well as m, v) is still at an initial stage (Urbanavičiūtė 1978: 76). The ongoing process is evidenced by T’®R’ type variants: tr’īs ‘three’/ tr’į:n’áu ‘I kneaded’/ trīs ‘three’. 217 Semi-palatalized ST, SR, and TR type junctures are no longer as widely used as their fully palatalized variants, apparently because they give rise to some articulatory difficulties (the natural phonological syllable boundary does not separate the first and second members of the cluster). Therefore, despite the fairly regular hardening of r, š, ž (and in certain positions, s, t, d, p, b), these consonants are more often pronounced palatalized in clusters. The usage ratio of the aforementioned partially and fully palatalized correlates of ST, SR, and TR type (i.e., (C1C2) and (C1C2)) is 1.2: 2.4. Moreover, clusters of this type tend to be pronounced wholly unpalatalized more often (108 of 183 cases). It appears that within a single syllable, cluster members with different degrees of palatalization are “undesirable.” The ratio of partially depalatalized clusters whose members belong to different phonological syllables to consonant groups belonging to the same syllable is 1.89: 1.22. Therefore, A. Girdenis’s question, “The phonological syllable boundary: construct or reality?..” can reasonably be answered in the affirmative: “(...) the phonological syllable boundary is a certain psycholinguistic reality, intuitively perceived by language users” (Girdenis 1985: 5, 9). In the receding Pelesa dialect, the phonological syllable boundary is signaled by the opposition between the members of consonant clusters with respect to palatalization. LITERATŪRA G i r d e n i s A. 1985: Fonologinio skiemens riba: konstruktas ar realybė? – Kalbotyra 36 (1), Vilnius, 5 – 11. G i r d e n i s A. 1995: Fonologija, Vilnius: Petro ofsetas. G r u m a d i e n ė L. 1988: Sociolingvistinis vilniečių lietuvių kalbos tyrimas: konsonantizmas ir akcentuacija. – Lietuvių kalba ir bilingvizmas. Lietuvių kalbotyros klausimai 27, Vilnius, 132 – 149. L e k o m c e v a M. 1972: O vzaimodeistvii fonologičeskich sistem v raione balto-slavjanskogo pograničja. – Balto-slavjanskij sbornik, Moskva: Nauka, 116 – 134. S t r i m a i t i e n ė M. G i r d e n i s A. 1978: Priebalsių junginių trukmė kaip atvirosios sandūros indikatorius bendrinėje lietuvių kalboje. – Kalbotyra 29 (1), Vilnius, 61– 69. S u d n i k T.M. 1972: Lazūnai. Litovskaja, belorusskaja i polskaja fonologičeskije sistemy. – Balto-slavjanskij sbornik, Moskva: Nauka, 15 – 115. S u d n i k T.M. 1975: Dialekty litovsko-slavjanskogo pograničja. Očerki fonologičeskich sistem (Lazduny, Deveniški, Gervjaty, Zetela), Moskva: Nauka. S u d n i k T.M. 1980: K istorii jazykovoj situacii Peljasy. – Balto-slavjanskije etnojazykovyje kontakty, Moskva: Nauka, 181 – 191. T r u m p a E. 1994: Kietumo: minkštumo priešpriešos defonologizacija priebalsių samplaikose (Pelesos šnektos duomenimis). – Lietuvių kalba: tyrėjai ir tyrimai. Konferencijos pranešimų tezės, Vilnius, 22 – 24. [P.Skardžiui paminėti] pranešimų tezės, Vilnius, 22 – 24. U r b a n a v i č i ū t ė Ž. 1970: Priebalsiai l, r, s, (ž?) prieš negalūninius ė, en lietuvių kalbos tarmėse. – Kalbotyra 21, Vilnius, 75 – 87. U r b a n a v i č i ū t ė Ž. 1978: Priebalsių t, d, n ir b, p, m, v kietinimas lietuvių kalbos tarmėse. – Kalbotyra 29, Vilnius, 69 – 78. V i d u g i r i s A. 1988: Iš etnolingvistinių santykių raidos Pelesos ir gretimose apylinkėse. – Lietuvių kalba ir bilingvizmas. Lietuvių kalbotyros klausimai 27, Vilnius, 21 – 36. 218 Zinkevičius Z. 1966: Lietuvių dialektologija. Lyginamoji tarmių fonetika ir morfologija (su 75 žemėl.), Vilnius: Mintis. DEPHONOLOGISIERUNG DER OPPOSITION VON NICHTPALATALEN: PALATALEN KONSONANTENGRUPPEN DER MUNDART VON PELESA Zusammenfassung Ziel dieses Artikels ist es, Ursachen für die nur partielle Palatalisationsassimilation bei den Konsonantengruppen vom Typ [C₁° C₂° Vi / Vu] aufzuzeigen. Einige Konsonantengruppen bleiben hart. Bisher wurden die Fälle der Depalatalisierung durch die einflüsse fremder Sprachen auf das phonologische System der litauischen Mundarten sowie durch das Fehlen einer funktionellen und positionbedingten Palatalisierung in den litauischen Mundarten erklärt. In der Mundart von Peleša stimmen die Grenzen der Palatalisierung gewöhnlich mit den phonologischen Silbengrenzen überein. 219