This dissertation aims to push forward our understanding of syntactic displacement, a phenomenon—analyzed here as movement, or Internal Merge (Chomsky 2004)—in which an element occurs in one position in surface syntax but occupies some other position covertly. One open question about movement is what its “driving force” is. Three prominent hypotheses are Greed (a constituent moves to satisfy a feature of its own—Bošković 2007, a.o.), Enlightened Self-Interest (a constituent moves to satisfy a feature of its own or of the head to whose specifier it moves—Lasnik 1995, a.o.), and Labeling (a constituent moves so that every relevant constituent can be labeled—Chomsky 2013, a.o.).
This dissertation argues for option two, Enlightened Self-Interest, on the basis of P’urhepecha data elicited from native speakers. P’urhepecha is an isolate of Michoacán State, Mexico; the variety investigated here is Janitzio P’urhepecha (JP). Following an investigation (Chapter 2) of JP finite-clause syntax, Chapter 3 argues that JP allows the subject to move to any of a wide array of specifier positions in the inflectional layer (“profligate subject movement”), and that this movement is driven by features of clausal functional heads, not by a feature of the subject itself or by the Labeling Algorithm (LA). Chapter 4 investigates JP quantifier float, which further supports the results of Chapter 3 and provides evidence that (JP) floated quantifiers are stranded adnominal elements, not adverbials. Chapter 5 argues that JP allows hyperraising to object (subject-to-object raising from finite clauses), and that this involves two steps of purely altruistic (higher-head-driven) movement, rather than being driven by properties of the moving element or by the LA.
Both profligate subject movement and hyperraising to object in JP, then, are driven by features of c-commanding heads, supporting Enlightened Self-Interest over Greed and Labeling. Furthermore, JP hyperraising to object provides evidence that the A/Ā-distinction does not emerge from two classes of syntactic positions, but is instead a consequence of features on particular functional heads. Finally, if indeed Internal Merge is feature-driven in the way argued for here, then we expect External Merge to be also—contra the Free Merge hypothesis.