It will be argued that cleft and pseudo-cleft sentences are monosentential constructions rather than ordinary copular sentences. The main reason for the proposal are the well-known connectedness effects. The three noun phrase types (R-expressions, pronouns, anaphora) behave as if clefting had not applied. In addition, in a traditional analysis, where (pseudo-) cleft sentences are analyzed as copular sentences with an embedded relative clause, these constructions categorically display violations of Ross' 'influencer constraint'. All information about temporal and modal qualification for the interpretation of the pseudo-cleft as a whole comes from inside the apparent subordinate wh-clause.
I will argue that simplex, cleft and pseudo-cleft sentences are related transformationally. Focusing of a constituent forces obligatory movement of this constituent into the specifier of a focus phrase within the splitted CP-layer. In English, this position is dominated by a topic phrase which is headed by the copula, which has been moved there. The spec position is occupied by an expletive. The complement of the focus phrase is an expression which contains a variable (trace of the focused constituent) and whose semantics parallel that of a question. Pseudo-clefts are derived by one more transformational step: topicalization of the complement of Foc° into the spec position of TopP (it-replacement). That this movement is topicalization is suggested by several, language-specific indicators (topic marking, intonation...). This analysis is a direct implementation of Carlson's idea (1983) to treat pseudo-clefts as self-answering questions. Further evidence (apart from connectivity) will be provided showing that the traditional, biclausal analysis is wrong, whereas the proposal at hand is superior: semantic restriction on the post-copular focus position, case pattern and multiplicity in questions and relative clauses.
