Abstract

Predication and specification assert themselves in the syntax of cleft sentences on a variety of planes. The distinction between predicational and specificational cleft sentences has been commonplace in the realm of pseudoclefts at least since Higgins’s seminal work; it is less familiar perhaps but equally significant in the domain of ‘it’-clefts (cf. esp. Declerck’s work). The examples in (1a,b) and (2a,b) illustrate the two readings.

(1) what John doesn’t eat is food for the dog a. PREDICATIONAL — ‘the food items John doesn’t eat are fed to the dog’ b. SPECIFICATIONAL — ‘John doesn’t eat the following: food for the dog’ (2) it was an interesting meeting that I went to last night a. PREDICATIONAL — ‘the meeting I went to last night was interesting’ b. SPECIFICATIONAL — ‘I went to the following last night: an interesting meeting’

Predication arguably underlies the syntax of both predicational and specificational cleft sentences. In predicational (1a), ‘food for the dog’ is straightforwardly predicated of ‘what John doesn’t eat’; in predicational (2a), ‘an interesting meeting’ is predicated of ‘it’, which here is used as a referential pronoun. In specificational (1b), ‘food for the dog’ serves as the subject of the underlying predicate ‘what John doesn’t eat’; and likewise, in specificational (2b), ‘an interesting meeting’ is the subject of a small clause, this time with ‘it’ being the predicate. In both (1b) and (2b), the underlying predicate inverts with its subject in the course of the derivation (à la Moro), forcing the presence of a form of the copula in contexts in which the copula is otherwise optional.

(3) I consider what John doesn’t eat (to be) food for the dog a. PREDICATIONAL — ‘to be’ is optional b. SPECIFICATIONAL — ‘to be’ is obligatory (4) I consider it (to be) an interesting subject that they are discussing tonight a. PREDICATIONAL — ‘to be’ is optional b. SPECIFICATIONAL — ‘to be’ is obligatory

The gross syntax of predicational and specificational cleft sentences is taken care of by the above. But several questions remain. One is what the semantics of ‘specificationality’ emanates from. Since specificational pseudoclefts are well known to oscillate in most cases between two alternative word orders (‘what John doesn’t eat is food for the dog’; ‘food for the dog is what John doesn’t eat’), it is unlikely that the semantics of ‘specificationality’ is uniquely the result of predicate raising to SpecIP. For specificational ‘it’ clefts, this question does not arise in the case of English — because, for reasons that remain to be properly understood, English predicate-’it’ must generally raise to the structural subject position (*’John was it who kissed Sue’ contrasts sharply with ‘it was John who kissed Sue’). But in languages like Dutch and German, the counterpart of ‘it’ in ‘it’ clefts can and sometimes must remain in situ, with the semantics of ‘specificationality’ arising nonetheless.

In addition to this fundamental question about the nature of ‘specificationality’, specificational ‘it’ clefts also raise the question of how to integrate the wh-/’that’ clause into the structure. With the predicate and subject positions of the copular clause being taken by ‘it’ and the focused constituent, respectively, I argue in detail that the wh-/’that’-clause is a null-headed relative clause in right-dislocated position — with the null-headed relative clause and the material preceding it entertaining a relationship of asyndetic specification mediated by a functional head. The null head of the relative clause is shown to have to be formally licensed via a strict concord relationship with the relative operator, and content-licensed via a strict concord relationship with the focused constituent. Both specificational ‘it’-clefts and specificational pseudoclefts feature a null-headed relative clause. In ‘it’ clefts, the null head is radically empty (i.e., devoid of both phonological and semantic features), hence dependent for its licensing on a concord relation with the relative operator and the focused XP.