Xiang (2021) notes the following puzzle: plural wh-questions involving certain collective predicates are predicted to carry a uniqueness presupposition (Dayal 1996), yet intuitively they don't (cf. Gentile and Schwarz 2020). She proposes that such questions have 'higher-order readings' (Spector 2007, 2008), and crucially that they have answers naming boolean conjunctions. We show that recourse to higher-order question readings is mistaken: Xiang's puzzle should be solved with higher-order plurality, and we provide empirical justification for this approach, mirroring for questions the recent findings for declaratives by Buccola, Kuhn, and Nicolas (2021).