On the Representation of Wh-words and Foci: Evidence from Mixtec

In this paper, I explore the consequences of a cross-linguistic preference for wh-words to move across a more local focus in languages where they compete for a single position. On the basis of novel data from San Martín Peras Mixtec (Otomanguean, Mexico), I show that this general pattern does not only reflect the behavior of exhaustive foci, but in fact applies more generally to all types of foci that move syntactically. I account for this preference by positing that both foci and wh-words move when in the scope of some particle that is sensitive to the alternatives that they introduce (Alternative Particles). Wh-words are in the scope of a Question Particle (Cable, 2010), and foci are in the scope of other particles sensitive to alternatives. Q Particles bear the features [Q] and [ALT], while other Alternative Particles only bear [ALT]. A probe relativized to the Ā feature geometry [uALT-uQ] will Agree cyclically (Béjar and Rezac, 2009), first searching for the most specific segment of the probe ([uQ]), copying the entire feature bundle of a goal that bears that feature. In the event that no goal bears [Q], [uALT] probes to find the most local Alternative Particle. I argue that this analysis is superior to alternatives that posits distinct heads which probe for wh-words and foci independently. Finally, I show that the pattern is not reducible to a Focus Intervention Effect.