Abstract
The virtue of imperfection. Slavic-Turkic contact in the Balkan linguistic area
Barbara Sonnenhauser (Zürich)
The languages of the Balkan linguistic area share particular features that are not found in the other members of their respective families. These features are usually regarded as displaying a ‘tendency towards analytism’, instantiated in particular by the coding of grammatical information in terms of ‘particle + inflected verb’ (cf. Fielder 2004 on ‘particle analytism’), (1a). However, these languages also exhibit synthetic structures, (1b).
| (1) | a. | da | napiša (Bulg) | b. | na-pis-va-x (Bulg) |
| sbjv | pref.write.pfv.prs.3sg | pref.down-write-ipfv-aor.1sg |
Both analytism and synthetism are ascribed to two quite distinct features that equally characterize the Balkan linguistic area: long term intensive language contact on the one hand, and geographical isolation on the other. Accordingly, Hinrichs (2004) cites the Balkans as an area of creolisation and increasing morphological simplification, Nichols (1992) as an area exemplifying morphologically complex residual zones. These seemingly contradictory positions can be related to the application of the problematic notion of ‘word’, and the scarceness of qualitative data illuminating the mechanisms of language contact and multilingualism. The former involves the danger of equating description and observation by judging the data against the background of contemporary linguistic abstraction. The latter impedes insight into the diverse contact scenarios and the processes underlying morphosyntactic changes.
Pulevski’s (1875) trilingual dictionary of Macedonian, Albanian and Turkish is among the rare sources providing access to multilingual speakers’ perceptions and conceptions of the languages they are exposed to. Focusing on Macedonian (Mc) and Turkish (Tr), the present paper illustrates the empirical and theoretical insight that can be drawn from this dictionary.
Examples such as (2)–(4) show the multi-directionality of copying strategies (see Johanson 1992) and question the assumption of ‘analytic coding’. In (2), the Mc subjunctive complementizer da and the aux.1sg appear as orthographically integrated with the lexical verb reče ‘say’, very much like the Tr equivalent (with demek ‘say’ as lexical verb). The Tr word-form comprises the complementizer ći and renders the subjunctive meaning by a volitional marker, obviously copying the Slavic pattern.
| (2) | da-su-rekol (Mac) | ći-de-sai-dim (Tr) | (Rečnik 131) |
| comp.sbjv-aux.1sg-say.ptzcp | comp-say-vol-pst.1sg |
(3) illustrates structural parallels in the verbal domain. The perception of the Mc l-participle and the Tr miş-participle, (3a), as structurally equivalent can be seen by their employment for the formation of verbal tenses in (3b), where Mc ‘stem+imperfective+l’ corresponds to Tr ‘stem-aorist-miş’. This parallel triggered the innovation of the l-form based on the imperfect stem in Mc (Gołab 1960), (3c).
| (3) | a. | dal, dalje (Mc)
give.pf.l-ptcp.m.sg, –pl |
vermiš, vermišler (Tr)
give.prs.miş-ptcp.sg, –sg, –pl |
(Rečnik 131, 146) |
| b. | daval, davalje (Mc)
give.ipf.l-ptcp.m.sg, –pl |
verermiš, verermiš, vererljermiš (Tr)
give.aor.miş.sg, –pl |
||
| c. | pišel, pišele (Mc)
write.ipf.imp.l-ptcp.m.sg, –pl |
jazarmiš, jazarmiš, jazarlarmiš (Tr)
write.aor.miş-ptcp.(m.)sg, –pl |
The structural parallels observed in (3) are the basis for the correspondence of analytic vs. synthetic past tense forms in Mc and the opposition of miş– vs. di-pst in Tr, (4). Moreover, both underlie the functional difference of confirmative vs. non-confirmative interpretations, (5):
| (4) | a. | dal su, daljeste (Mc)
give.pf.aor.l-ptcp aux.1sg, –aux.2pl (‘analytic’) |
vermišum, vermišsiniz (Tr)
give.miş-ptcp.(aux.)1sg, –2pl |
(Rečnik 131,138) |
| b. | miluvav, miluvaje (Mc)
love.ipf.pst.1sg, –3pl (‘synthetic’) |
severdim, serverlerdiler (Tr)
love.aor.pst.1sg, –3sg, 3pl |
| (5) | Od kade možeme dase naučime […] za carstvana, kako se, osnovalje, ilji propadnalje […] (Mc) | Nereden urele biljiriz ći […] padišaxljiklar, nasl kurulmiš jaot fet olmišlar […]. (Tr) |
| ‘Where can we learn from those kingdoms, how they were (supposedly) founded, how they (supposedly) perished [tr. were conquered].’ (Rečnik 52) | ||
The bidirectional structural assimilation suggests that instead of being complexified by adding material or simplified by abolishing overt distinctions, morphology is re-used and re-structured. This makes the assumption of ‘analytic coding’ highly questionable – the more so, as Pulevski lists structures like (2) under ‘D’ (da) instead of ‘R’ (reče).
These observations suggest that simplicity and complexity are not contradictory, but emerge as effects of an increase in morphological transparency (in the sense of Trudgill 2011: 21) – the most salient and probably most effective outcome of ‘imperfect’ multilingualism.
References
Fiedler, W. (2004): Der südosteuropäische Typ des grammatischen Analytismus – die ‘Balkanische Partikelkonstruktion’ im Verbalsystem. In: Hinrichs, U. (ed.): Die europäischen Sprachen auf dem Wege zum analytischen Sprachtyp: Wiesbaden, 363-398.
Gołąb, Z. (1960): The influence of Turkish upon the Macedonian Slavonic dialects. Folia Orientalia 1, 26–45.
Hinrichs, U. (2004): Südosteuropa-Linguistik und Kreolisierung. Zeitschrift für Balkanologie 40/1, 17-32.
Johanson, L. (1992): Strukturelle Faktoren in türkischen Sprachkontakten. Stuttgart.
Nichols, J. (1992): Linguistic diversity in space and time. Chicago and London.
Pulevski, Đ. (1875): Rečnik od tri jezika. S. makedonski, arbanski i turski. Kniga II. Beograd
Trudgill, P. (2011) Sociolinguistic typology. Oxford.