In most cases, a wh-question calls for an answer that names an entity in the set denoted by the extension of the wh-complement. However, evidence from questions with necessity modals and questions with collective predicates argues that sometimes a wh-question must be interpreted with a higher-order reading, in which this question calls for an answer that names a generalized quantifier. This paper investigates the distribution and compositional derivation of higher-order readings of wh-questions. First, I argue that the generalized quantifiers that can serve as semantic answers to wh-questions must be homogeneously positive. Next, on the distribution of higher-order readings, I observe that questions in which the wh-complement is singular-marked or numeral-modified can be answered by elided disjunctions but not by conjunctions. I further present two ways to account for this disjunction–conjunction asymmetry. In the uniform account, these questions admit disjunctions because disjunctions (but not conjunctions) may satisfy the atomicity requirement of singular-marking and the cardinality requirement of numeral modification. In the reconstruction account, the wh-complement is syntactically reconstructed, which gives rise to local uniqueness and yields a contradiction for conjunctive answers.

在大多数情况下，WH -question呼吁答案名称在集合实体表示由延长WH -complement。但是，来自具有必要模态的问题和具有集体谓词的问题的证据认为，有时必须以较高阶的解释来解释wh-问题，在该问题中，该问题要求给出一个命名为广义量词的答案。本文研究了wh问题的高阶读数的分布和成分推导。首先，我认为广义量词可以作为wh的语义答案-问题必须是同质的。接下来，关于高阶读数的分布，我观察到wh补码是单数标记或数字修饰的问题可以通过消除析取法来回答，而不能通过合词来回答。我进一步提出了两种方法来解释这种析取-析取不对称性。在统一说明中，这些问题允许析取，因为析取（但不是合词）可以满足单数标记的原子性要求和数字修饰的基数要求。在重建过程中，wh补语在语法上进行了重建，这引起局部唯一性，并导致了合取答案的矛盾。