LIETUVIŲ KALBOS SANDAROS TYRINĖJIMAI

LIETUVIŲ KALBOTYROS KLAUSIMAI, XXVI (1987)

KAZIMERAS GARŠVA

PRINCIPLES FOR TRANSCRIBING DIALECT FOLKLORE

J. Balčikonis did considerable work, recording folklore in dialect from 1907 onward, translating it from other languages into Lithuanian, and publishing it. His observation that “collecting folklore […] requires great dedication and skill. If someone has collected it but recorded it poorly or failed to arrange it, its value is slight” remains relevant to this day¹.

While organizing the folklore collection of O. Bluzmienė (from the village of Talačkoniai, Pasvalys district), recorded in the Northern Panevėžys dialect, I became convinced that there are no universally accepted principles for transcribing Lithuanian folklore in dialect, and that the existing ones could also be improved. Folk creativity is collected by people of varied educational backgrounds and interests, and sometimes even a linguist may find it difficult to choose the best way to record it.

As early as 1969, K. Aleksynas, language editor for folklore publications, wrote that “many problems arise when folklore texts are published for the public. […] There is a growing need to consider these questions more broadly and seriously”². Three phonetic problems arise when recording and publishing folklore: 1) how to record a work—in dialect or in the standard language; 2) how many dialect features to represent—all of them or only some; and 3) which signs to use to mark dialect sounds.

The statement in the instructions “Guidelines for Folklore Collectors” that “nothing in folklore material may be changed or corrected arbitrarily”³ was not entirely consistent with another, tentative statement in the same publication: “if one knows the local dialect well, it is desirable to record folklore texts in dialect”⁴. This would mean that someone who does not know the dialect of the area visited (and that is true of most collectors), and who does not wish to struggle with special signs, may translate everything into the standard language at their own discretion. Because of a proofreading error, a similarly worded statement even slipped into the official instructions⁵, although it is well known that changing a dialect often destroys even the natural sound of a song⁶. Writing, for example, vištos instead of vištos “vištõs” is the same as writing pamiródyk — man paródyk (“show me” instead of “show me”). Only the first distortion belongs to the phonetic level, while the second belongs to morphology. In a later

¹ Balčikonis J. Selected Writings. — V., 1978, vol. 1, p. 83.

² Aleksynas K. On a Few Questions of Folk-Tale Textology. — In: Folk Creativity. V., 1969, p. 250.

³ Nurodymai tautosakos rinkėjams. — V., 1958, p. 6.

⁴ Ibid., p. 11.

⁵ Aleksynas K. Recording Folklore Works. — In: How to Collect Data on Folk Speech (Methodological Advice). V., 1973, p. 10.

⁶ Lietuvių liaudies dainynas. Vaikų dainos. — V., 1980, t. 1, p. 7.

108

In the folklore-collection instructions, linguistic requirements were formulated more precisely, but, in accommodating inexperienced collectors, still not strictly enough (“The value of the text will be greater if it is recorded while preserving at least those dialect features that can be expressed using the letters of the standard language”⁷).

It is appropriate to publish literature for children in the standard language. In other popular publications, in our view, at least songs should be presented in dialect (when a dialect differs greatly from the standard language, narrative and sometimes shorter folklore can also be published in the standard language, changing only regular sound shifts). Scholarly publications should be prepared in dialect. That dialect phonetics is one of the most important signs of folklore’s distinctiveness and does not greatly hinder comprehension of a work is demonstrated by our ethnographic ensembles. Folklore recorded accurately is a valuable source for linguists as well.

There are two ways of departing from the authentic sound inventory of folklore: compared with the standard language, by not marking an additional sound, and, much more often, by representing the same sound differently. Deviations of both kinds arise consciously and unconsciously. For example, for the sake of simplicity, folklorists do not mark vowel lengthening, although it can easily be inferred.

Authors who omit certain phonetic features (for example, the inserted j) explain that these are nonessential and that this is the standard-language norm. But what is being read is not a text in the standard language but in some dialect; moreover, the consonant j is not inserted in all subdialects. It is difficult to judge which subdialect features are essential and which are not; consequently, the same inserted j is marked in some books (published in the same year)⁸ and omitted in others⁹. For in traditional Lithuanian songs, j is often not only added before front vowels, but also inserted between two front or back vowels, e.g., pajauga ‘grew up’. For those who want to learn to sing that song accurately or study the remnants of jotation, this phenomenon will be important.

From a phonetic standpoint, it is less objectionable not to mark consonant voicing or devoicing, nor their omission in clusters, or the widening of the vowel e before a hard consonant. The sound features in these positions can usually be inferred. Voiceless consonants (in the prefix iš- and so on) are very rarely voiced before voiced consonants. Somewhat more overstated (without adequate research) is the claim that at the juncture between the elements of a prefixed or compound word, and likewise when a proclitic or enclitic meets an independent word, only one of two identical or assimilated consonants is pronounced. In a 1985 survey of 36 fifth-year Lithuanian philology students, only 17 (47%) said isigando, usnigo. In the speech of older people, less influenced by writing, this pronunciation is more common, but it is unlikely ever to have reached 100%. A. Salys had already written that adjacent similar consonants can be retained; this is also mentioned in the academic “Grammar of the Lithuanian Language”¹⁰. In subdialects that completely shorten the prefix i- and have not adopted the prefix ant- in its place (for example, in the Panevėžys—

⁷ Patarimai tautosakos rinkėjams (Metodiniai nurodymai). — V., 1974, p. 5; 1940 m. J. Balio išleistame „Tautosakos rinkėjo vadove“ (K., p. 61—76) plačiai aptarta paprastoji ir net mokslinė fonetinė rašyba.

⁸ Daukantas S. Žemaičių tautosaka. — V., 1983, p. 9.

⁹ Lietuvių liaudies dainynas. Vestuvinės dainos. — V., 1983, t. 2, p. 9; Čiulba ulba sakalas. — V., 1983, p. 8.

¹⁰ Lietuvių kalbos gramatika. — V., 1965, t. 1, p. 116.

109

—žiškiai), the pairs išoka ‘jumped out’ — iš.oka ‘jumped out’ are distinguished (the consonant š in the second word is slightly longer), and especially išiak ‘get out’ — iššiak ‘get out’. In the latter case (when consonants of different qualities occur next to each other), speakers who articulate clearly and wish to distinguish the meaning—and some East Aukštaitians (Širvintiškiai) and Southern Dzūkians¹¹—may retain the consonant š fully (or assimilate it to s). If only two or one consonant is marked everywhere, actual facts are leveled, and linguistics loses necessary data. Moreover, folklorists usually mark the omission of consonants at the boundary between two words (cf. an tėvelio dvaro, etc.)¹², but do not do so consistently within a word.

It would be good to mark the widening of the vowel e before hard consonants as well. Otherwise, to reconstruct the authentic pronunciation, the reader must consider which dialect the word belongs to and in what position the vowel e occurs. Moreover, in pretonic syllables this vowel may occasionally remain unwidened. In the North Panevėžys dialect, in the same stressed position (before a hard consonant), the vowel i also widens to o (cf. lėp’ ‘you climb’ — l’ôp ‘I climb’); thus, if the widening of i is marked, the widening of e should consistently be marked too.

So far, folklore of the North Panevėžys dialect prepared with phonetic precision has not been published in a mass edition, either in the standard language or in dialect (presumably the same is true of other more complex dialects). Printed as: “O jūs šuns šuns, kam suėdėt visas bandels!?” (in dialect O jūs šun’s šun’s, kam suėdėt vioss bundials, and not everything has been restored to the standard language); mūsų lauktai šalime (= šal imais), tokio tai dieną (= tokią tai dieną), pas man (= mane), ji (= jin), per (= par) and so on.¹³

That a more authentic, fuller version of folklore can also be recorded on magnetic tape and transcribed is evident from a comparison of two versions of a song by O. Bluzmienė¹⁴.

1

Old man Išgarglino went out, killed a little sparrow, and gobbled it up. He plucked two little barrels of feathers, salted two little tubs of meat. A farmhand came to drive him to the manor; he wanted to swallow the whole sparrow.

2

The old man went out, Ižgarglina; he killed a little sparrow, } 2 gobbled it up. He plucked feathers from two little barrels, salted meat in two little tubs } 2. The farmhand came to drive him to the manor; he wanted to swallow the sparrow } 2 whole.

¹¹ K. Morkūno, J. Karaciejaus informacija.

¹² Lietuvių liaudies dainynas, t. 2, p. 74, 170, 174, 198—202 ir t.t.; Čiulba ulba sakalas, p. 173, 180, 194, 205 ir t.t.

¹³ Bluzmienė O. Apie turgus ir žmones. — Kn.: Liaudies kūryba, t. 1, p. 201—203.

¹⁴ The first was recorded in 1965 (Lietuvių liaudies dainynas, vol. 1, pp. 270—271), the second in 1984 (K. Kalibatas) and 1985 (K. Garšva).

110

He baked a little thigh, roasted it crisp, gorged the manor folk, and stuffed them to bursting.

He baked a little thigh, roasted it crisp, gorged the manor folk, and stuffed them to bursting. } 2

Some qualified folklorists perhaps think that, to show the specificity of a dialect in a work, it is enough to retain one or another word pronounced dialectally. From the point of view of language, such a text is somewhat more useful than one recorded in the standard language, but it is misleading about the speech variety¹⁵.

To justify errors in recording and printing dialectal folklore, folklorists often exaggerate the influence of the standard language on dialects. In reality, the standard language has had no significant effect on the speech of many old people. But when a recorder finds standard pronunciation easier to understand and more convenient, more than one informant, to some extent, begins to accommodate themself to the “educated” recorder. It should be the other way around. One should immediately tell the person to speak as they normally do and not to embellish anything. But when standard-language forms are naturally interspersed in songs, and occasionally in narratives, they should not be changed, as folklorists emphasize (J. Balčikonis used to render song texts in Dzūkian¹⁶).

Worst of all is when sounds from other speech varieties are deliberately inserted into a dialect. In “Lietuvių liaudies dainynas,” the Panevėžys dialect’s word-final vowel e, derived from ė, is written ia, although this distorts the dialect (in an open ending the vowel e is always closed) and two letters are written instead of the one required. If one does not wish to write the dialectal ending -e, it would be somewhat better to mark the standard-language ending -o—this would avoid creating nonexistent forms and misleading the reader. In editing the language of the publication, the endings of many Eastern Aukštaitian dialects (Širvintiškiai, Anykščiai, Utena, and Vilnius varieties) have been generalized. But in Panevėžys-dialect sentences Oi tu moteriške, Ar nematei kiškio?, Dave šūria¹⁷ the ending vowels e are entirely identical, yet are written differently (e and ia).

Shortened unstressed vowels should not be marked with nasal letters or long vowels either. Sometimes the vowel o is also written u, e as i, ū, and so on, for example: botagų, kuokų, tų liepeli¹⁸, linbi (= liõne), and so on.

In Panevėžys-dialect word endings, the words vaike “children” and vaikė “to the child” cannot both be marked identically (with the vowel ė). It is impossible to make do with the symbol ɫ, because that vowel also occurs after other consonants. It would be better to mark the sound e not with ė, but with the familiar and readily understood e. Folklorists could use a simplified transcription devised by linguists, in which the only new letters would be ė and ʋ, along with signs for length (e.g., ã, ã.) and palatalization (ʹ).

The clearest approach is to write the text without exception as it is heard—as it is spoken in the clear style (without omitting identical consonants in adjacent words,

¹⁵ Back in 1976, we proposed in “Literatūra ir menas” that folklore from northern Lithuania not be recorded in a half-Dzūkian manner, but that appeal went largely unheeded. When folklore books are published, linguists are invited to consult, but not all their comments are taken into account.

¹⁶ Ambrasas-Sasnava K. Paminklas Druskininkams ir profesoriui. — Kalbos kultūra, 1985, sąs. 48, p. 64.

¹⁷ Lietuvių liaudies dainynas, vol. 1, pp. 251, 508, etc.

¹⁸ Ibid., pp. 109, 234, 235, 607.

111

e.g., the child said). In scholarly publications, at least those that also have a linguistic profile, it is appropriate not to restore omitted consonants or those that have become voiced or voiceless. This spelling had already been used in publishing folklore in the editions by A. Baranauskas (1922), “Tautos ir žodis” (1923, 1924), “Lietuvių kalbos tarmių (chrestomatija)” (1970), and “Joniškėlio apylinkių šnektos tekstai” (1982). In “Druskininkų dainos” (1972), J. Balčikonis later probably also abandoned the phonetic principle of spelling in many places without good reason. Short endings were not replaced with nasal letters, nor narrow e with open ia, in the folklore editions of S. Daukantas (1846), non-Lithuanians A. Schleicher (1857), A. Leskynas and K. Brugmann (1882), A. Becenberger (1882), A. Niemi and A. Sabaliauskas (1912), and others¹⁹.

This article examined the final printed form of folklore works, without considering the extent to which it was shaped by the transcriber, the publisher, and other complex circumstances involved in preparing the text. Poorly transcribed texts recorded only by hand, of course, are generally impossible to edit properly, but tape recordings should be checked more carefully. Just as no one can replace the folklorist when classifying folklore works and preparing authoritative explanations, or the musicologist when transcribing a song’s melody, so a dialectologist can and must have the final word on the language of the text²⁰.

PRINCIPLES OF TRANSCRIBING DIALECT FOLKLORE

Summary

Since 1907, J. Balčikonis recorded folklore in various dialects, translated it into Lithuanian from other languages, and published it (“Druskininkai Songs” was published in 1972). Three kinds of phonetic problems arise when recording folklore works and preparing them for publication: 1) how to record a work— in the dialect represented or in the literary language; 2) how to reflect the features of a given dialect—fully or partially; and 3) which symbols to use for dialect sounds. In scholarly publications, it is appropriate to reflect all the main features of a dialect: not to restore unpronounced consonants or those that have become voiced or voiceless, and to indicate widened vowels i, u, e, (= ö, o, ä), etc. Inconsistent representation and distortion of phonetic features are unacceptable: writing ia instead of e, etc. Only those elements of other dialects and the literary language that the informant themselves pronounces for particular reasons are justified; reduced unstressed vowels should not be represented by nasal letters or long vowels, nor should u be written instead of o, i instead of e, or ы, ё, etc.

¹⁹ Quoted from: K. Aleksynas, “Linguistic and stylistic features of Lithuanian folk songs.” In: Literature and Language. Vilnius, 1971, vol. 11, pp. 32–281.

²⁰ I am grateful to K. Aleksynas and B. Stundžienė for their valuable comments.

112