Colin Brown:

Multiple wh-clefts and their answers

Content questions with multiple wh-expressions, such as ‘whosaw what?’ admit two different types of answer: single-pair answers andpair-list answers. This does not hold for cleft questions with multiple wh-expressions,such as ‘who is it that saw what?’ which allow only a single-pair answer. Inthis talk I pose the following questions: (i) why do interrogative clefts withmultiple wh-expressions not permit pair-list answers? and (ii) how do weaccount for the availability of single-pair answers for multiple wh-clefts? Isuggest that (i) is explained by adopting Dayal’s (2002) account of multiplewh-questions with pair-list answers, which requires the structurally lowerwh-expression to undergo covert movement to a scope-taking position. I suggestthis covert movement is not available in the cleft-question as the presupposedclause is analyzed as a null-headed relative clause (Percus 1997), andtherefore is an island to movement. Turning to (ii), Reinhart’s (1998) choicefunction analysis of wh-in-situ can account for the availability of single-pairanswers without appealing to island-violating covert movement. I extend thisanalysis to a language (Sm’algyax; Tsimshianic) which does not allow multiplewh-questions at all, and suggest that this ban is explained if, in thislanguage, (i) wh-questions are structurally cleft-like constructions, and (ii)there is no wh-in-situ.