The Diachronic Typology of Relative Clauses

Abstract

This thesis investigates the diachronic behaviour of relative clauses across a broad sample of constructions from genetically and geographically diverse languages. Previous studies of change in relative clause constructions have most frequently been restricted to individual languages or language families. By comparing such studies with each other and with the historical records of languages that have less commonly been the focus of diachronic syntactic works, I examine the strength of evidence for developments that are predicted by earlier literature to be "e;natural"e; or even "e;universal"e; pathways of change (for example, various sources of relative clause markers, the development of hypotaxis out of parataxis, shift from prenominal to postnominal relative clause position). I also look for evidence of changes that synchronic typological studies of relative clause constructions might lead us to expect to find (i.e., diachronic variation in the same parameters by which relative clause types distinguish themselves synchronically). I conclude that the sources of relative clause markers and the results of the extensions of these markers into other constructions are more varied then has generally been thought to be the case, including, for example, such sources as classifiers and discourse markers. Changes in other features of relative clauses, however, such as verb forms, embeddedness, and the relative position of the relative clause and its head tend to be remarkably stable over long periods of time. The factor that appears to have the greatest influence on whether changes in these otherwise stable features do occur is language contact. Features of relative clauses, markers, and even entire constructions can be copied from other languages, competing with pre-existing constructions until in some cases one replaces the other, and in others the two are redistributed according to considerations such as restrictiveness, animacy, case role or similar. These results point to the importance of incorporating the effects of language contact into models of language change rather than viewing contact situations as exceptional. There are also implications for the definition of relative clauses, their syntactic structures, and the relationships between the different"e;subtypes"e; of this construction.