How Should We Revise the Paratactic Theory?

That is to say, the surface form of indirect speech is itself paratactic and requires only the insertion of a period (‘a tiny orthographic change ... without semantic significance’, as Davidson put it) to be made explicit. If SF is true, then the ‘that’ of ‘says that’ is, grammatically speaking, a demonstrative, and the utterance to which it refers is just the reporter’s utterance of the words immediately following it.2 Early objectors to PT tended to attack LF directly, arguing that it did not yield a satisfactory account of the truth conditions of indirect speech reports. Defenders of PT usually found ways of answering, or at least of dodging, these objections and the theory survived and prospered. More recent critics, however, have concentrated their fire upon SF, and this flanking attack may ultimately prove more damaging for PT. The case against SF is strong. In an elegant 1986 paper, Segal and Speas marshalled a range of evidence, all tending to show that, from a grammatical point of view, the ‘that’ of ‘says that’ is a complementizer, not a demonstrative