This section discusses wh-movement in wh-questions. Section 11.3.1.1 begins with a discussion of wh-questions such as given in (119), in which the wh-phrase is moved into the initial position of its own clause. We will show that this movement is semantically motivated by the need to form operator-variable chains in the sense of predicate calculus. This leads to the formula ?x (Peter is reading x), in which ?x is the question operator and x the variable.

Other issues that will be discussed in Section 11.3.1.1 are the obligatoriness of wh-movement in wh-questions, pied piping (the fact that wh-movement sometimes also affects non-interrogative material that is part of the clausal constituent containing the wh-element), and stranding (the fact that wh-movement sometimes does not affect such non-interrogative material).

Section 11.3.1.2 continues with a discussion of so-called long wh-movement, i.e. cases in which a wh-phrase is extracted from an embedded clause, as in (121a). There are several restrictions on this kind of wh-extraction. For example, while long wh-movement is perfectly acceptable from object clauses selected by a verb of saying, it leads to a degraded result when the object clause is selected by a factive verb such as betreurento regret; cf. the contrast between the examples in (121a&b). Furthermore, long wh-movement is completely impossible from adverbial clauses such as (121c). We will discuss a number of factors that may affect the acceptability of this kind of long wh-movement.

Note that examples like (121c) cannot be saved by pied piping of the adverbial clause: sentence (122a) is infelicitous as a wh-question, although the same sentence is perfectly acceptable (with a different intonation contour) if wat is interpreted as an existential pronoun, as in (122b).

The unacceptability of both (121c) and (122a) thus shows that in certain syntactic environments a wh-element may be inaccessible to wh-movement. As a result, certain questions that can be easily expressed by an operator-variable chain in predicate calculus cannot be formulated as wh-questions in natural language. Section 11.3.1.3 briefly discusses a number of such syntactic environments in which a wh-element is inaccessible to wh-movement, often referred to as islands for movement.

Section 11.3.1.4 concludes with a discussion of so-called multiple wh-questions, questions like those in (123) with more than one wh-phrase. We will discuss the semantics of such constructions, as well as a number of restrictions on their form.