The interaction of syntax, prosody, and discourse in licensing French wh-in-situ questions

The current experiment addresses the proposal by Cheng and Rooryk (2000) that wh-in-situ questions in French are marked by an obligatory rising contour, which is the result of an intonation morpheme [Q:] in C. Twelve native French speakers participated in a production study in which they produced the target interrogatives, along with a range of similar sentences. While most participants were perceived to assign wh-in-situ questions a sentence-final rise, a minority was not. Moreover, the rise associated with wh-in-situ was smaller than the rise exhibited in yes- no questions, which C&R claim to be licensed by the same morpheme. Given that these two results are unexpected under C&R's account, we conducted a further acoustic analysis of the productions, which revealed that for sentences lacking a sentence-final rise, the in situ wh-word had an elevated high pitch accent. A statistical analysis shows a negative correlation between the height of the pitch accent assigned to the wh-word and the presence and height of the sentence-final rise, indicating that instead of the sentence-final rise for wh-in-situ questions being optional, it may instead be variable and predictable by focus placed on the wh-word, for discourse reasons. We discuss three possibilities for the status of the intonation morpheme concerning yes- no and wh-questions and the role of information structure in French wh-in-situ questions.