The Lithuanian Vocalic System Revised - William R. Schmalstieg

The purpose of this paper is to reaffirm an earlier view that the essential contrastive features of the contemporary standard Lithuanian vocalic system are low/ mid/high on the one hand and front/back on the other hand.1 Rounding is an automatic concomitant of the back non-low vowels, except when they are preceded by (j) or a palatalized consonant. In this case the back non-low vowels are phonetically fronted and as a result the rounding must be considered distinctive in this environment.

î	û
ė	ô
â	ê

All of the so-called diphthongs can be considered as sequences of two short vowels. In addition, immediately after a (j) or palatalized consonant the front/back contrast is neutralized for the low vowels (e, a). From the morphological point of view it is easier to create underlying forms which show a contrast of */ja/ vs. */je/ (or /Ca/ vs. /Ce ), but the surface structure shows no such contrast. It would hardly be necessary to make this point except for the fact that we read in a recent article: "Bemerkt sei, dass W. R. Schmalstieg's Behauptung, dass 'primary Baltic *Ce I(< P.I.E. *Ce and *Cje) and "Cja have merged in modern standard Lithuanian as Če...' dem Tatbestand nicht entspricht".2 As her source for this information Professor Buch relies on that section of the Lithuanian Academy Grammar prepared by V. Vaitkevičiūtė.3 One can hardly accept this statement without dispute.

Reviewing the article, "Lietuvių literatūrinės kalbos fonetikos raida," by V. Vaitkevičiūtė in the collection Lietuvių kalba tarybiniais metais (Vilnius, 1967), J. Kazlauskas says (I translate): "One must indeed agree with the fact that the vowels e, e: of the standard language are pronounced with a coloring, and that after soft consonants the vowels a and a: are very much fronted, although it is not clear whether the author [V. Vaitkevičiūtė - - WRS1 believes that the e with the a coloring is the same as the fronted a and a :... That is, the vowels e, e : and a, a : are distinguished in the standard, language only in word initial position, whereas in word medial position they are in complementary distribution - after hard consonants we have a, but after soft consonants we have e (p. with a coloring)."4 Elsewhere Kazlauskas has expressed himself even more forcefully on this point (I translate) : "The fact that in the Academy Grammar. .. e and a are distinguished after soft consonants is an expression of the great conservatism (and in this case even an anachronism) in the normalization of standard Lithuanian."5

One can find similar views (although not expressed in phonemic terms) in A. Senn's Handbuch der litaui-schen Sprache, Vol. I, Heidelberg, 1966, pp. 63, 70 and J. Otrębski's Grammatyka języka litewskiego, Vol. I, Warsaw, 1958, p. 219. One should also compare the comments of Pavel Trost in his "Two remarks on Lithuanian vooalism," Acta Baltico - Slavica, Vol. III (1965) p. 184.

Professor Buch also argues for monophonematic interpretation of the Lithuanian diphthongs uo and ie on the ground that in the Dzukish dialect of the Seinai region the contrastive intonation in long monophthongs and the diphthongs uo and ie has been lost whereas fn the other types of diphthongs the contrastive intonation has been retained.6 She gives the following examples for the contrastive intonation: traukia 'drags", arc 'to plow', gėlbėc 'to save, to help', vs. saukc 'to call', vefkc 'to cry', randasi 'is found, is located'. To show the lack of contrast she quotes I'iepa 'linden' (= standard Lithuanian liepa), liepc 'to order' (= standard Lithuanian liepti), duoc 'to give' (= standard Lithuanian duoti), duos 'will give' ( = standard Lithuanian duos), J'onas (=standard Lithuanian Jonas).

Now it seems likely that this is not the only dialect in which the contrast between circumflex and acute intonation was lost in the long vowels and the diphthongs uo and ie- But as 1 understand Professor Buch's article she reasons that since the prosodic features of the long vowels and the diphthongs uo and ie are treated one way in one part of the dialects they are to be classed together elsewhere also. I can only imagine using information from one dialect to analyze another dialect, if one is trying to create some kind of underlying model from which to generate both dialects.

But in the Lithuanian case just given it would make eminently more sense to have underlying contrasts between the acute and circumflex in all long vowels and to establish a special dialect rule according to which the intonation contrasts are to be removed for the diphthongs uo, ie and long vowels.

To give an example of the results of Professor Buch's procedure: would one wish to class the Proto-Slavic sequence */au/ as one phoneme and the sequence */ar/ as two phonemes on the basis of the fact that in Russian there remains no trace of the Proto-Slavic prosodic contrasts in the etymological diphthongs and long vowels, varnas. On the other hand Russian I'ipa 'linden' = Lithuanian liepa and Russian r'uku 'hand (ace. singularis) — Lithuanian ranką, and there is no trace of the old difference between the acute and the circumflex in these Russian words.

Professor Buch's second argument for the monoquences q/ar, al/? E.g. Russian vor'ona 'crow' = Lithuanian varna, but Russian v'oron 'raven' = Lithuanian phonematic interpretation of uo and ie is as follows: "Im suedwestlichen Randgebiet des Litauischen, dem auch die litauischen Doerfer des Kreises Sejny angehoeren, ist, z. B. zuikis 'Hase", ilgas 'lang', dzirba 'er arbeitet', Stabingis 'ein Dorfname', Burbiškiui 'ein Dorfname', durpės Torf' usw. Die Dehnung had jedoch bei ie, uo nicht anders als in der Schriftsprache, nicht nur e, a, sondem auch u, i als erste Komponente der Diphthonge und diphthongischen Verbindungen bei Fallton gedehnt worden, whereas we find such a contrast for the etymological sestattgefunden: L'ietuvą acc. "Litauen", p'ienas 'Milch', p'uodas 'Topf, j'uodas 'schwarz" usw., bei denen die Angehoerigkeit an das erste bzw. dritte Akzentuierungparadigma vormaligen Fallton anzunehmen erlaubt, haben die hoerbar gleiche Aussprache von įe, uo wie die vormals schleiftonigen liepc, duos usw."7 In other words the phonemic allegiance of the /i/ in the sequence /ilC/ (in which the contrast between historical */ilC/ and /lie/ had been lost) has been shifted to /i/. This again seems to have been a fairly common phenomenon in Lithuanian.

But such an apparent reassignment of phonemes in a position of neutralization is a later development and has nothing to do with the earlier stage represented by the standard language. Phonemic re-assignments in position of neutralization are not too uncommon.8 In Proto-Indo-European the */p/ in the sequence */sp/ (cf. Latin spuo 'I spit') was presumably the same as the */p/ in the word for father (cf. Latin pater). I trust that no-one would wish to keep the /p/ of Latin pater and the /p/ in Latin spuo separate because in a different dialect of Indo-European (viz. the Germanic dialect) they are assigned to different phonemes. But this is just the kind of thing we would be expected to do if we were to keep the /i/ of Lithuanian ilgas 'long' separate from the /i/ of Lithuanian pienas 'milk') because they split apart at a later date.

But let us suppose that all dialects of Lithuanian other than the standard language had become extinct. Experience tells us that many of the world's languages and dialects have died out, so it would have been possible for the Lithuanian dialects as well. Then there would certainly be no reason not to break up /uo/ into /u/ plus /a/ and /ie/ into /i/ plus /e/. Apparently such a solution is quite acceptable even to native Lithuanians. Thus J. Kazlauskas gives us the following minimal pairs: lieti 'to pour' vs.limti 'to bend in', kietis 'hardness' vs. kirtis 'blow; stress", kietas hard' vs. kiltas 'large, fat', kuopti 'to clean' vs. kumpti 'to become crooked', kuokelė 'baton, drum stick' vs. kulkelė, diminutive of kulka 'bullet'. He then comments (I translate): "Thus it is necessary to consider the pure diphthongs as two vowels and one must regard the second element of the diphthong uo as the rather open phonetically and the short /u/ should not be netic position — after the vowel M in a diphthong (the position of neutralization of the phonemes a and o)'"9Actually, of course, the position after /u/ is indeed the position of neutralization of /a/ and /o/, but it is even more than that. It is the position of neutralization of /a/, /o/ and /e/. The same is true for position after /i/ since in that position there is no contrast between /a/, /o/ /e/. Thus we find that only ,/ie/ and /ua/ are possible, not */ia/, */io/, */ue/ nor */uo/.

In conclusion, then, I should like to affirm that I hold essentially to the system of Lithuanian vocalic phonemes which Antanas Klimas and I proposed in 1962 (see footnote 1).10

I should say that I now find questionable the following ideas presented in that article. I now doubt the status of a short counterpart of /ė/. Although such was listed in the 1962 article, it really does not exist in the system of most speakers of standard Lithuanian. Probably also the phoneme which we transcribed as /e:/ should be transcribed as /ae:/. Likewise the short /i/ is rather open phonetically and the short /u/ should not be called 'tense'.11

1. Antanas Klimas and William R. Schmalstieg, "A Note on the Vocalic Phonemes of Lithuanian, "The Slavonic and East European Review, Vol. XLI, No. 96 (1962), pp. 244 - 5. One should also see now Antanas Klimas, "Some Attempts to Inventory Lithuanian Phonemes," forthcoming in Baltic Linguistics, a book to be published by the Pennsylvania State University Press.
2. Tamara Buch, "Zur phonologischen Wertung von lit. ie, uo und lit. ė, ė," Lingua Posnaniensis, Vol. XII/XI1I (1968), p. 79.
3. Lietuvių kalbos gramatika, Vol. I, Vilnius 1965, p. 55.
4. "Matyt, reikia sutikti su tuo, kad literatūrinės kalbos balsiai e, e : tariami su a atspalviu, o po minkštųjų priebalsių a ir a : tariami kaip labai supriešakėję balsiai a ir a :, nors ir nėra aišku, ar autore tą e su a atspalviu laiko lygiu supriešakėjusiems a ir a :... Vadinasi, balsiai e, e : ir a, a: literatūrinėje lietuvių kalboje skiriami tik žodžio pradžioje, o žodžio viduryje jie sudaro papildomą distribuciją —po kietųjų priebalsių turim a, o po minkštųjų — e (e su a atspalviu)." Baltistica, Vol. IV, No. 2 (1968), p. 325.
5. "Tai, kad 'Lietuvių kalbos gramatikoje' (I, Vilnius, 1965) yra skiriama' e ir a po minkštųjų priebalsių yra vienas iš didelio konservatizmo (šiuo atveju netgi anachronizmo) lietuvių literatūrinės kalbos normalizacijoje pasireiškimų." Baltistica, Vol. Ill, No. 2 (1967), p. 238.
6. Buch, op. cit., pp. 77 - 8.
7. Buch, op. cit., p. 78.
8. Henry Hoenigswald, Language Change and Linguistic Re-construction, Chicago I960, p. 89.
9. "Taigi grynuosius dvibalsius taip pat reikia laikyti dvifone-miais, ir į antrąją dvigarsio uo dalį reikia žiūrėti kaip į fonemų o ir a archifonemą, atsiradusią aiškiai apibrėžtoje fonetinėje pozicijoje — po balsio u dvigarsyje (fonemų a ir o neutralizacijos pozicija)", in "Lietuvių literatūrinės kalbos fonemų diferencinių elementų sistema". Kalbotyra, Vol. XIV (1966), p. 75.
10. For an analysis of the prosodic features of Lithuanian see L. Dambriūnas, A. Klimas and William R. Schmalstieg, Introduction to Modern Lithuanian (New York, 1966), p. 3 and also Antanas Klimas, "Some Attempts..."
11. I should like to take this opportunity to hank Prof. Eric P. Hamp for commenting on this article. It is undoubtedly the worse for the fact that I did not accept all his suggestions. Therefore he cannot be held responsible for my mistakes.