This talk investigates the following three readings of complex questions: (a) pair-list readings of multi-wh questions (which student voted for which candidate?), (b) pair-list readings of questions with a universal quantifier (which candidate did every student vote for?), and (c) choice readings of questions with an existential quantifier (which candidate did one of the students vote for?). To derive these readings, we can either unify (a-b) given that they both are pair-list readings, or unify (b-c) given that they both involve quantifying-into question. I argue that the two pair-list readings have a crucial semantic difference and hence must be derived differently: (b) is subject to domain exhaustivity and presupposes that every student voted for some candidate, while (a) isn’t (cf. Dayal 2002). I further offer a function-based approach to derive each of these readings compositionally (Xiang 2016: chap. 5-6). Compared with earlier function-based approaches (Dayal 1996, Chierchia 1993, a.o.), my approach predicts the contrast between (a) and (b) w.r.t. domain exhaustivity, and manages to treat quantifying-into question uniformly as regular quantification. If time permits, I will also review the family-of-questions approach (Fox 2012, a.o.), and discuss the quantificational variability effects in sentences like John mostly knows which student voted for which candidate.