Turkish as spoken in the Netherlands (NL-Turkish) sounds “different” (unconventional) to Turkish speakers in Turkey (TR-Turkish). We claim that this is due to structural contact-induced change that is, however, located within specific lexically complex units copied from Dutch. This article investigates structural change in NL-Turkish through analyses of spoken corpora collected in the bilingual Turkish community in the Netherlands and in a monolingual community in Turkey. The analyses reveal that at the current stage of contact, NL-Turkish is not copying Dutch syntax as such, but rather translates lexically complex individual units into Turkish. Perceived semantic equivalence between Dutch units and their Turkish equivalents plays a crucial role in this translation process. Counter to expectations, the TR-Turkish data also contained unconventional units, though they differed in type, and were much less frequent than those in NL-Turkish. We conclude that synchronic variation in individual NL-Turkish units can contain the seeds of future syntactic change, which will only be visible after an increase in the type and token frequency of the changing units.
(Received January 05 2007)
(Revised June 18 2007)
(Accepted August 10 2007)
* We would like to thank the audience at the 2007 UWM Linguistics Symposium on Formulaic Language and the 6th International Symposium on Bilingualism, as well as Elma Nap-Kolhoff, Maria Mos and anonymous reviewers for their valuable comments. The research reported was made possible by a grant from NWO (Netherlands Organization for Scientific Research, Grant 016-024-012).