Assoc. Prof. Dr. Vladimir Zhobov, Assistant Prof. Georgi Kolev, Assistant Petyr Shishkov
Sofia University

 
The aim of this paper is to test a frequently stated but rarely backed claim about Turkish influence on Bulgarian language: that the eastern dialects were more heavily influenced than the western (see for example Mirchev 1978: 86). This hypothesis is almost invariably mentioned in discussions about the origin of the renarrative mood in Bulgarian and the related question of the development of the imperfective l-participle. It is a well known fact that such participle is found only in eastern dialects. We will not take sides in the argument about native vs. foreign origin, but we can note that such argument can easily get circular: the new participle and more complex rennarative forms appeared under Turkish influence and their presence in the east testifies to a stronger Turkish influence on eastern dialects. And: the new participle and more complex rennarative system appear only in dialects that were more strongly influenced by Turkish, therefore they must have appeared under Turkish influence.
The material on which this paper is based was collected during the accomplishment of a project on Bulgarian dialectometry sponsored by the Volkswagen Stiftung (more information can be found at http://www.sfs.uni-tuebingen.de/dialectometry and also in Prokić et al. 2009). The phonetic data consist of 157 words collected from 197 sites from Bulgarian dialects. The primary source for the data was the archive of the Ideographic Dialect Dictionary of Bulgarian, compiled in the Bulgarian Department of Sofia University.
We will test the claim by looking at an allophonic variation in Bulgarian, which in our opinion was borrowed from Turkish: the softening of final velar stop after front vowels. In Bulgarian velars are softened before but not after front vowels: [`k’еsten] ‘chestnut’ but [e`zik] ‘tongue’. In Turkish velars (and also l) are softened both before and after front vowels: [kal`mak] ‘stay’ but [kjelje`bekj] ‘butterfly’ (Gylybov 57: 15–17). Such final softness of the voiceless velar stop is found in Bulgarian dialects, but in order to prove that it is a result of a borrowed allophonic variation is has to be demonstrated that the final velar is soft only after front and not after back vowels. The data from the Bulgarian Dialect Atlas do not allow such conclusion, because the maps present the final softness in the word език (BDA, I: map 50; II: map 61; III: map 59), but there is no consistent information about the pronunciation of final /k/ after back vowels.
Four of the words in our phonetic list contain final velar stop preceded by a front vowel: /e`zik/ ‘tongue’, /eche`mik/ ‘barley’, /pone`delnik/ ‘Monday’, /cho`vek/ ‘man’. Five of the word contain final velar stop preceded by a back vowel or a consonant: /dyl`bok/ ‘deep’, /`kamək/ ‘stone’, /`petək/ ‘Friday’, /`pjasək/ ‘sand’, /vəlk/ ‘wolf’. This is enough ground to formulate a rule, however inconclusive. Is has to be noted that in many eastern dialects the word for ‘man’ appears as /chi`ljak/, with a back vowel, and the word for ‘barley’ appears with another suffix – /en/. On the other hand, the word for ‘stone’ appears with a front vowel before the final k in many Western dialects – /`kamik/ – but the softening of final /k/ after front vowel proved unknown in the west.
The results clearly show that the allophonic alternation of the final velar stops exists only in Eastern dialects. Soft final velar in both /e`zik/ and /eche`mik/ is found in districts of Asenovgrad, Chirpan, Kazanlyk, Kotel, Nova Zagora, Pavlikeni, Pomorie, Popovo, Preslav, Razgrad, Shumen, Silistra, Svishtov, Tutrakan, Tyrgovishte, Varna. Another result of the analysis is that in most cases the rule applies only to stressed syllables. The word /pone`delnik/ is recorded in only 6 sites (out of 29). It has to be noted that vowel in unstressed vowel are subject to reduction, quantitative and in many cases qualitative. In additional 27 sitеs the word /e`zik/ is pronounced with soft consonant but the word for ‘barley’ has another suffix. They are found in the districts of Ardino, Asenovgrad, Chirpan, Devin, Goce Delchev, Harmanli, Haskovo, Ivajlovgrad, Krumovgrad, Malko Tyrnovo, Pyrvomaj, Smoljan, Stara Zagora. The number of the sites in which both words have stressed front vowel before the final velar, but only one of them ends in soft consonant is negligible (see the map).
As Benjo Conev pointed out, such final softness appears only in dialects which preserved the old final soft consonants (Conev 1984: 436). However the inventory of the old final softness did not include velar stops. Soft final velar stops appeared later in a number of eastern dialects (Devin, Kotel, Malko Tyrnovo) as a result of internal development, namely as reflexes of palatalized alveolar stops. In such cases however they are found both after front and back vowels: [zekj] ‘son in law’ and [pəkj] ‘road’. The original velar stop is softened only after front vowels, and therefore we assume that the most likely explanation for this variation is Turkish influence.
It is also well known that western dialects have by far fewer soft consonants, but they have soft counterparts of the velar stops, so the reason for the lack of final softening is not shortage of soft consonants in the inventory.
Final softness of other consonants is common in all dialects and the difference between eastern and western dialects concerns the inventory rather than the positional constrains on soft consonants. In this sense the Turkish rule did not lead to a new, previously unknown distribution of a distinctive feature.

In their influential book on language contact Sarah Thomason and Terrence Kaufman offer a borrowing scale in which they correlate the degree of structural change in the borrowing language with the intensity of the contact. We reproduce here only the features concerning phonetics:
1.    Casual contact: lexical borrowing only;
2.    Slightly more intense contact: Appearance of new phonemes with new phones, but only in loanwords;
3.    More intense contact: Phonemicization, even in native vocabulary of previously allophonic alternations (especially true of those that exploit distinctive features already present in the borrowing language). Prosodic and syllable structure features.
4.    Strong cultural pressure: Introduction of new distinctive features represented in native vocabulary, and perhaps loss of some contrasts; new syllable structure constrains, also in native vocabulary; a few natural allophonic and automatic morphophonemic rules, such as palatalization and final obstruent devoicing.
5.    Very strong cultural pressure: added morphophonemic rules; phonetic changes (i.e., subphonemic changes in habits of articulation, including allophonic alternation); loss of phonemic contrast and of morphophonemic rules. (Thomason, Kaufman 1988: 74-75).

Lexical borrowings from Turkish are found all over Bulgarian language territory. In terms of phonetic influence, all Bulgarian dialects contain features that correspond to the second and the third degree of intensity. The voiced post-alveolar affricate was introduced in the system as a result of internal development only in the western-most dialects (as reflex of *dj, e.g. /`predzha/ ‘yarn’). A softer variety of the same sound appears in Razlog area as a result of the so-called new first palatalization: /`dzhizdav/ ‘pretty’. In other dialect it was introduced through loanwords, primarily Turkish and now is almost ubiquitous in Bulgarian, with the exception of some southeastern dialects.
The introduction of phonemic soft correlates of velar stops in Bulgarian was due almost entirely to Turkish loanwords and never spread to native vocabulary. The relatively highly marked front rounded vowels were not introduced in the Bulgarian vocalic system. After velar consonants they were adapted by phonemicization of previously allophonic variation (the velars had, and still have, soft allophones before front vowels) and the front rounded vowels were rendered as combination of soft velars and back rounded vowels. That happened throughout all the language territory. Only in the eastern dialects, in which the soft correlation encompasses all but the post-alveolar consonants, an already existing distinctive feature was used and the front rounded vowels were adapted in the same way. In the western dialects none-velar consonants had neither soft correlates nor soft allophones and the front rounded vowels were rendered only by back vowels and the features frontness was lost. In this way the Turkish kömür ‘coal’ and kütük ‘log’ were adapted as /kju`mjur/, /kju`tjuk/ in the eastern, and /kju`mur/, /kju`tuk/ in the western dialects.
The Turkish phonetic influence in the cases discussed so far remain in the second and arguably third degree of intensity of contact. The only feature that can be interpreted as a testimony to “strong cultural pressure” in the above scale is restricted to dialects spoken by Bulgarian Muslims: new syllable structure constrains in native vocabulary. In the dialect of Tihomir, in the Eastern Phodopes, a vowel is inserted in initial clusters of obstruent and sonorant: /pu`luk/ ‘plough’, /de`reban/ ‘tiny’, /pala`til/ ‘paid’ (in fact all the words begin with a stop). In many cases even the vowel harmony is observed. Kabasanov is explicit that this is Turkish influence (Kabasanov 1963: 39). Similar forms exist in the dialects south of Smoljan and Stojkov points out that they are found in Muslim villages (Stojkov 1993: 131). It must be noted however that these are not strict phonetic rules and they allow many exceptions. According to Kabasanov there is no inserted vowel in the plural forms: /pu`luk/ but /`pluguve/ (Kabasanov 1963: 38). There is no vowel harmony in the dialects south of Smoljan. The prothetic vowel [i] before initial cluster of obstruents ([i`strah] ‘fear’) is very rare in eastern dialects and never follows the vowel harmony.
Other phonetic features that can be attributed to Turkish influence include the pronunciation of the vowel /ə/ in northeastern dialects and the use of front rounded vowel in many eastern dialects. The raised and retracted pronunciation of /ə/ was initially reported by Ljubomir Miletic for the Moesian dialects (he used the term “strongly dark” Miletich 1989: 31) and is cited as a relevant feature of this group in later classifications, for example by Stojko Stojkov, who uses the term “velar ъ” (Sojkov 1993: 102). However, such pronunciation of /ə/ seems to have disappeared: according to the second volume of The Bulgarian Dialects Atlas it is found in only two villages (BDA, II, Part 2: 25). The most likely explanation is that the field-workers were influenced by Miletich’s authority.
The front closed rounded vowel appears in native vocabulary as an allophone of the vowel /i/ adjacent to labial or post-alveolar consonants and its pronunciation is optional. It invariably alternates with back rounded vowel preceded by palatalized consonant ([po`lüvam] and [po`ljuvam] ‘to water’). More open front rounded vowels are unknown in native vocabulary.
Bulgarian and Turkish have coexisted for centuries on the Balkans and inevitably there was strong linguistic influence. Due to the dominant position of Turkish, the influence was mostly in one direction – from Turkish to Bulgarian. However, on the phonetic level it remained relatively limited, compared to the lexical. Marked phonetic categories such as front rounded vowel were not introduced (the development of the soft correlation is beyond the scope of this paper), the syllabic structure was marginally affected in small areas, and prosodic features like stress remained uninfluenced. Besides the intensity of the contacts, the typological dissimilarities of the two languages also played a role in the process.

BDA: Български диалектен атлас. Т. І 1964; Т. ІІ 1966; Т. ІІІ 1975; Т. ІV 1984. Издателство на БАН.
Conev 1984: Беньо Цонев. История на българския език. Т. ІІ. Наука и изкуство.
Gylybov 1957: Гълъб Гълъбов. Граматика на турския език. Фонетика морфология и синтаксис. Наука и изкуство.
Kabasanov 1963: Стайко Кабасанов. Един старинен български говор. Тихомирския говор. Издателство на БАН.
Miletich 1989: Любомир Милетич. Източнобългарските говори. Издателство на БАН.
Mirchev 1978: Кирил Мирчев. Историческа граматика на българския език. София.
Prokić et al. 2009: Jelena Prokić, John Nerbonne, Vladimir Zhobov, Petya Osenova, Kiril Simov, Thomas Zastrow, Erhard Hinrichs. The Computational Analysis of Bulgarian Dialect Pronunciation. – Serdica Journal of Computing, Vol. 3, Number 3, 269-299.
Stojkov 1993: Стойко Стойков. Българска диалектология. Издателство на БАН.
Thomason. Kaufman 1998: Sarah Grey Thomason, Terrence Kaufman. Language Contact, Creolization, and Genetic Linguistics. University of California Press.