Section 10.3.1 has discussed the clausal comparative correlative construction, illustrated in (505a). Although verb-second does not apply in this construction, we have seen that the second clause should be regarded as a main clause; Den Dikken (2003a/2005/2009) analyzes the first clause as a free relative clause that is adjoined to this main clause, as shown in representation (505b).

Here we focus on the internal structure of the main clause of the construction and try to answer the question whether wh-movement is involved in its derivation. A first argument for answering this question in the affirmative is that the comparative part hoe/des te beter precedes the subject of the clause (here the pronoun hetit), and that it also precedes the complementizer datthat, if present.

An affirmative answer would further predict that the comparative phrase hoe/des te minder should also be able to undergo long wh-movement, as in the examples in (507).

Since the data include many variables (such as the possibility of using a complementizer, which will be ignored in the following, and others not mentioned here), and the acceptability status of the relevant examples is not always immediately clear, Den Dikken sent a questionnaire to 17 professional syntacticians. For the examples in (506) without the complementizer dat, Den Dikken found that (506a) was accepted by all informants, while (506b) was accepted by most informants as a less preferred option; this is as expected, and can be seen as reflecting the fact that the use of des te is part of the formal register of the language. What is important for our present purpose is that example (507a) was considered perfectly acceptable by 11 informants, while the judgments of the remaining informants ranged from marked (4 cases) to unacceptable (2 cases). The approval rate for (507b) was much lower: 6 informants found this example perfectly acceptable, 6 rejected it, while the remaining 5 reported an intermediate status. We can conclude that wh-movement is clearly involved in the derivation of the main clause in correlative constructions when the correlative marker is hoehow. This is less clear in the case of the marker des te, and indeed Den Dikken (2009) argues for a different analysis in this case, for which we refer to the article.