Argument clauses are special in that they allow wh-extraction under specific conditions. The examples in (323) show that the extracted wh-phrase may be either an argument of the embedded verb or an adjunct. The traces ti refer to the original position of the moved wh-phrases in the embedded clauses.

The fact that wh-phrases can be extracted from argument clauses is rather special as this is categorically prohibited from adjunct clauses. The examples in (324), for instance, show that conditional clauses are strong (absolute) islands for wh-movement; they block wh-extraction of both arguments and adjuncts.

There are good reasons for assuming that the wh-phrases in (323) are not moved in one fell swoop into their target position but that this involves an additional movement step via the initial position of the embedded clause; cf. Section 11.3 and Chomsky (1973/1977). This is indicated in the structures in (325), in which the traces refer to the positions that the moved phrase has occupied during the derivation; the CP/TP structure of clauses assumed here is discussed in Section 9.1.

The main reason for assuming that the wh-phrases are moved via the initial position of the embedded clause is that this immediately accounts for the unacceptability of examples like (326b&c); since the clause-initial position of the embedded clause is already occupied by the subject pronoun wie'who', wh-extraction of the object/adjunct must apply in one fell swoop and this is not allowed. Note that (326c) is acceptable when the adverb wanneer'when' is construed as a modifier of the matrix predicate, but the reading intended here is the one in which it modifies the embedded predicate dit boek kopen'buy this book', as is indicated by the traces.

Although more can be said about the restrictions on wh -movement (see Section 11.3.1), the above suffices for the main topic in this subsection: which verbs may function as bridge verbs, that is, allow wh -extraction from their argument clauses? For example, whereas verbs of saying/thinking normally license wh -extraction from their complement clause, verbs of manner of speech such as fluisteren 'to whisper' normally do not; this observation is attributed by Erteschik-Shir (2006), to an unpublished paper by Janet Dean (Fodor) from 1967.

The examples in (328) show that irrealis verbs expressing a hope or a wish constitute another set of verbs that readily license wh-extraction in Dutch; cf. Haider (2010:108) for the same observation for those varieties of German that allow wh-extraction from embedded dass-clauses.

Factive verbs like betreuren'to regret', on the other hand, systematically seem to block wh-extraction from their complement clause given that examples like (329b-d) are generally judged to be unacceptable; see, e.g., Hoeksema (2006:147).

      Examples like (329b&c), in which an argument is extracted seem degraded but are often considered to be better than examples such as (329d), in which an adjunct is extracted. This so-called argument-adjunct asymmetry is often attributed to the referential status of arguments; see Rizzi (1990). That referential status is relevant is clear from the fact that argument extraction becomes even better when the argument is discourse-linked, that is, when the hearer is asked to pick some entity or set of entities from some presupposed list. Although there is considerable variation in acceptability judgments on examples of this type, some speakers even seem to consider them fully acceptable; see, e.g., Zwart (2011:209) for cases of object extraction. If acceptable, the examples in (330) show that factive clauses are not strong, but weak (selective) islands for wh-extraction.

      All in all, the list of bridge verbs seems to be relatively small. Hoeksema (2006) collected a sample of such verbs selected from written sources published after 1780. The complete collection consists of 963 tokens and 110 types. Most types have a very low frequency: nearly fifty types occur only once. Restricting ourselves to types occurring minimally six times, we get the result in Table (331). Bridge verbs are not only relevant for wh-questions but also for relative clause and topicalization constructions.

Since Hoeksema does not give his list of attested examples, we searched the internet with the string [Wat V[±past] je dat'what do/did you V that ...?' in order to check whether the verbs in Table (331) occur in the relevant type of wh-question. The three cases for which we could not find such examples are marked by a number sign; their use may be restricted to relative clause or topicalization constructions; see example (322b), for instance, which was taken from Hoeksema (2006). This leaves us with no more then twenty verbs that are regularly used as bridge verbs in wh-questions, and there is in fact only one verb, denken'to think', that is really frequent (>300 tokens). Another relatively frequent bridge verb is the irrealis verb willen'to want' (>100), but all other verbs are relatively infrequent (<100). Other corpus-based research has revealed a similar preference for the verb denken and, to a lesser extent, willen; cf. Verhagen (2005:119ff.) and Schippers (2012:105).

Wh-extraction has given rise to two main lines of research, which Erteschik-Shir (2006) refers to as, respectively, the structural and the semantic approach. We will start with arguments in favor of the structural approach, according to which bridge verbs have some special syntactic property that makes their complement clauses transparent for wh-movement. For example, Erteschik-Shir mentions that verbs of manner of speech such as fluisteren'to whisper' differ from verbs of speaking and thinking in that they can occur without a propositional clause, and she suggests on the basis of this that embedded clauses co-occurring with verbs of manner of speech may have a syntactic status different from embedded clauses that co-occur with verbs of speaking and thinking.

More support for the structural approach comes from languages like English and German. In English, the set of bridge verbs seems to coincide more or less with the set of verbs allowing that-deletion in embedded declarative clauses, an observation again attributed by Erteschik-Shir (2006) to Janet Dean (Fodor). Verbs of speaking/thinking, for example, allow that-deletion while factive verbs like to regret do not (although it is not too hard to find cases on the internet). Since Dutch does not allow dat-deletion in embedded declarative clauses, we cannot provide similar evidence on the basis of this language.

As for German, Haider (1985:55) claims that most bridge verbs trigger embedded verb-second in the German varieties that have it (the Northern varieties as well as standard German), although irrealis verbs like mögen'to like' are an exception to this general rule; cf. Haider (2010:124, fn. 25). The examples in (334) show that wh-extraction even requires the embedded finite verb to be in second position in those varieties that do not allow wh-extraction from embedded declarative dass-clauses; wh-extraction in dialects not allowing a set-up such as in (334a) obligatorily triggers verb-second, as in (334b). Since Dutch does not have this form of embedded verb-second, we cannot provide similar evidence on the basis of this language.

In short, arguments in favor of the structural approach to wh-extraction emphasize that bridge verbs obligatorily take a complement clause and that, in some languages at least, such verbs may affect the form of these clauses by licensing complementizer deletion or embedded verb-second. The semantic approach, which originates in Erteschik-Shir's (1973) seminal work, emphasizes the effect of information structure on wh-extraction. The generalization is that wh-extraction is possible only from clauses which are focused, that is, which express the new information of the clause. This immediately accounts for the fact that wh-movement is normally not possible from factive clauses given that these present propositions the truth of which is presupposed by the speaker; see the discussion in Section 5.1.2.3. It may, however, also account for the contrast between the primeless and primed examples in (335b&c). That we are dealing with a so-called weak island is clear from the fact that the two primed examples exhibit the argument-adjunct asymmetry discussed in Subsection I. For completeness' sake, it should be noted that the intended reading of the two (c)-examples is the one in which the adverbial wanneer'when' modifies the embedded clause; matrix scope is not intended here.

Erteschik-Shir (1973:95ff.) shows that adding negation to the matrix clause has the effect of defocusing the embedded clause; whereas example (335a) without negation can readily be used to introduce the proposition expressed by the embedded clause in the domain of discourse, example (335b) with negation is most naturally interpreted as the denial of the presupposed truth of the embedded proposition. This means that (335b), but not (335b'), is in accordance with the generalization that wh-extraction requires the embedded clause to be part of the focus of the clause.
      Since in general the addition of information to the matrix clause makes it more likely that the embedded clause is defocused, the generalization predicts that this may also have a degrading effect on wh-extraction. This may account for the contrast between the examples in (327), repeated here as (336). The verb fluisteren'to whisper' is more informative than the verb zeggen'to say' since it includes a manner component: Jan expressed his assertion in a low voice. In fact, adding a manner adverb like zachtjes'softly' in (336c) seems to have a similar degrading effect on wh-extraction, a fact that seems to have escaped attention in the literature so far.

Erteschik-Shir's generalization is completely in line with the findings in Verhagen (2005:124ff.): on the basis of his corpus research mentioned at the end of Subsection I, he claims that attested cases of wh-extraction differ only minimally from what he assumes to be the basic template, which he takes to be the one given in (337). He further claims that "invented examples of wh-extractions are judged worse to the degree that they deviate more from the [..] pattern [in (337)]".

The nature of the evidence reviewed above suggests to us that each of the two approaches has something different to contribute to the description of wh-extraction. The structural approaches may be correct in claiming that wh-extraction is subject to certain structural conditions, for example, that the embedded clause is a complement of the verb in the matrix clause. The semantic approaches, on the other hand, may be correct in postulating additional conditions on the use of wh-extraction constructions, for example, that the embedded clause is the focus of attention and therefore constitutes the new information of the clause, and that as a consequence the semantic contribution of the matrix clause must be limited.