Labeling, concord and nominal syntax in Turkish
İsa Kerem Bayırlı
November 2022

Informal Abstract: There is a line of research trying to understand the consequences of dissociating Labeling from Merge. One thing that theories of this sort dont do well is tell us how the merger of an adjective and a noun gets to be labeled, which one might take to be a price to be paid for insisting on deeper and simpler explanations. In this case, I claim, this insistence leads to an empirical discovery: Languages with idiosyncratic (i.e. non-semantic) gender systems are languages with gender concord on nominal adjuncts (here adjectives). That is because the problem of labeling associated with {aP, nP}-structures is solved by designating adjectives (more generally, nominal adjuncts) as universal probes for phi-features. This way, {aP, nP} can be labeled by the shared set of features between aP and n0. But what happens when nP lacks grammatical features? This would be your Turkish, a non-concord language. In this case, after the failure of AGREE, aP is marked as a Spell-Out domain. The paper works out some of the consequences of this claim. It seems that labeling might provide us with a predictive theory of the distribution of probes.
Format: [ pdf ]
Reference: lingbuzz/006955
(please use that when you cite this article)
Published in: Languages
keywords: labeling, nominal concord, adjectives, turkish, suspension, morphology, syntax
Downloaded:362 times


[ edit this article | back to article list ]