English allows conjoined wh-questions to have two different readings. The English sentence Which person does John like and Mary admireulcorner involves ATB movement and is understood either as single questions requiring one set of individuals that are liked by John and admired by Mary (which person x, likes (j,x) & admires (m,x)) or as coordinated wh-questions, allowing distinct individuals that John likes and Mary admires (which person x, likes (j,x) & which person y, admires (m,y)). I argue this ambiguity is explained by the two key assumptions about wh-movement in Chomsky (1995): (1) Movement is copying. (2) wh-phrases consist of a wh-element and a nominal restrictor. This yields two possible LFs for English depending on whether [Wh+nominal] or wh alone is interpreted as a variable. It is therefore natural for me to propose that number of questions understood corresponds to the number of nominal segments at LF.