Keywords: antilocality, internally headed relative clause, syntactic ergativity, Northern Tujia, Sino-Tibetan
TL;DR: Northern Tujia (and many other languages!) exhibits a ban on relativizing transitive subjects that cannot be explained along the lines of syntactic ergativity.
Abstract: This paper is concerned with a curious relativization restriction in Northern Tujia (Sino-Tibetan, China). In externally headed relative clauses (EHRCs), all arguments can be relativized; in internally headed relative clauses (IHRCs), all but transitive subjects can be relativized. This ban on transitive subject IHRCs is reminiscent of the Ergative Extraction Constraint (EEC), a signature of syntactic ergativity. I argue that ergativity-based approaches to this IHRC restriction are not tenable given that Northern Tujia does not show ergative alignment and lack object inversion. Crucially, both EHRCs and IHRCs are island-sensitive, but only IHRCs are subject to the restriction. As an alternative, I propose that the relativization restriction in question is correlated with clause size: IHRCs are structurally smaller than EHRCs, and only the former exhibits EEC-like patterns due to an antilocality constraint on A’-extracting transitive subjects. Beyond Northern Tujia, this analysis has several conceptual and empirical advantages: it provides straightforward and principled explanations for some cross-linguistic generalizations about syntactic ergativity. It also captures the typology where all languages that disallow transitive subject RCs—regardless of their case alignment—exhibit structural reduction.
Submission Number: 197
Loading