Another look at wh-questions in Romance: the case of medrisiotto and its consequences for the analysis of French wh-in-situ and embedded interrogatives

This article has three main descriptive goals: A. We shall show that there are (at least) two types of wh-doubling in Romance, one involving a {clitic wh-, wh-phrase} pair, the other a {‘weak’ wh- element, wh-phrase} pair. Correspondingly, we shall show that in addition to the well known syntactic split between wh-words and wh-phrases in many Romance and Germanic languages, Mendrisiotto displays a hitherto undocumented tripartite distinc-tion among wh-items: in addition to clitic and strong (‘tonic’) wh- items (see Poletto & Pollock (2004b) it also has weak wh- words, in the sense of Cardinaletti & Starke (1999). B. In order to begin to account for this tripartite distinction we shall argue that the internal structure of wh-items contains several distinct functional features that can be spelled out as the various elements in¬volved in the various wh-doubling configurations of Mendrisiotto. C. As in our previous work on Illasi and Monno (cf. Poletto & Pollock (2004b) ) we shall show that the struc¬tures and derivations at work in Mendrisiotto shed light on better-known Romance languages, French among them. Just as Illasi suggested very strongly that clitic que (‘what’) in French was paired with a null (strong) associate, Mendrisiotto suggests equally strongly that quoi (‘what’) is the lexical counterpart of the null associate of the weak form cusa (‘what’) in Mendrisiotto. Various advantageous consequences will be seen to follow from this analysis which will derive all the ‘odd’ properties of French quoi and ce+que questions and suggest a new revealing analysis of the syntax of French wh-in situ construc¬tions . Beyond its attempt to describe and explain previously undocumented facts –a worthy enterprise in its own right without which the second goal could never be reached–, this article can thus be read as an exercise in micro-comparative syntax attempting to show how the uniquely preserved dialects of Northern Italy can shed light on longstanding problems and puzzles in the syntax of better known Romance languages. The article is constructed as follows: section 1 sums up previous results and analyses on overt and covert wh-doubling. Section 2 introduces the Mendrisiotto data we shall be concerned with and section 3 offers our analysis of these data. Section 4 and 5 extends our findings to French interrogative syntax: French in situ questions in section 4, embedded questions, cleft and ‘diable’ (‘the hell’) interrogatives in section 5.