|Title:||A Markedness Subhierarchy in Syntax: Optimality & Inversion in Spanish|
|Comment:||Shorter version published in P. Barbosa et al. (eds.), Is the Best Good Enough? Optimality and Competition in Syntax, MITWPL and MIT Press, 1998.|
|Abstract:||In this paper, inversion (verb-subject word order, seen as head movement) in Spanish wh-questions is shown to be conditioned by two factors that vary from speaker to speaker (based on introspective judgment survey data collected by the author): (i) the "argument-hood" of the wh-phrase (a subcase of which was first noted by Torrego 1984), and (ii) whether the wh-question is direct (matrix) or indirect (subordinate). I propose an OT analysis of this variation, expanding on the work of Grimshaw (1997).
I argue that Grimshaw's OP-SPEC constraint ("syntactic operators must be in specifier position") must be decomposed into a universally-ranked subhierarchy of constraints defining a relational scale of different types of wh-phrases: arguments (quiÃ©n 'who', quÃ© 'what'), spatio-temporal locations (dÃ³nde 'where', cuÃ¡ndo 'when'), manners (cÃ³mo 'how') and reasons (porquÃ© 'why'). The possible interactions that the members of this subhierarchy can have with constraints on movement are the source of the first observed conditioning factor on inversion: if less argument-like wh-phrases must be in specifier position and thus require inversion (= more movement), then more argument-like wh-phrases will also.
Grimshaw distinguishes a general constraint on movement (STAY) and another specifically prohibiting movement into the head of a subordinate clause, building on Rizzi & Roberts (1989) and McCloskey (1992). This distinction is the source of the second observed conditioning factor on inversion: if inversion is require in a subordinate clause with a particular type of wh-phrase, then it will also be required with the same type of wh-phrase in matrix clauses.