A note on verb phrase complementation

Evidently, if VPC does not explain the ungrammaticality of (7) there is no explanation for the ungrammaticality of (i ib). It appears Rosenbaum's argument for NP domination in the complement structures dealt with here is: x IS DOMINATED BY NP IF IT CAN BE PSEUDO-CLEFTED. To bring this argument into disrepute we need to (a) establish that strings dominated by VP (according to Rosenbaum's analysis) can be pseudo-clefted; or (b) establish that what is pseudo-clefted is not an NP. (a) is discussed and (b) is not. The main point under (a) is that both prefer-type and condescend-type complements can be pseudo-clefted; if this is established, then obviously pseudo-clefting cannot be used to distinguish deep structure differences between them. The fact that she condescended to talk with us can be pseudo-clefted was overlooked by Rosenbaum because he took condescend as the most underlying form of the verb, that is, that form of the verb which has not undergone any transformational operations. (For further arguments, cf. JL 4 (I967) 9I Wagner, and Bowers, YL 4 (I967) 83-84.) We find, in fact, with Wagner and Bowers, that when to is inserted in the sentences that are deemed ungrammatical by Rosenbaum, they become perfectly acceptable pseudo-clefted sentences which appear to be paraphrases of their non-pseudo-clefted congeners. Thus