Tense use in direct and indirect speech in English

This paper sets out to revisit the problem of tense in English direct and indirect speech with past reporting clauses. The general point it makes is that in all instances of speech representation, the temporal reference point defined by the ‘original’ speech exchange described in the reporting clause (she said, he thought, etc.) should be incorporated into the analysis of tense. It creates a second, ‘surrogate’ temporal zero-point within the intensional domain of the represented speaker's utterances, alongside the real temporal zero-point of the actual speaker's discourse. Recognition of this intensional domain with its own time of utterance results in an analysis of tense in the reported clause of direct speech as ‘intensionally’ absolute, i.e. as directly related to the time of utterance not of the actual, but of the represented speaker. Backshifted tenses in indirect speech are likewise argued to be ‘intensionally’ relative, involving the represented speaker's surrogate temporal zero-point which is shifted to a past time of orientation from which relations are subsequently plotted. In addition, indirect speech is shown to occur with two different types of absolute tense use: intensionally absolute tenses similar to those in direct speech, and true absolute tenses in semantically factive examples, where the tense form is related directly to the actual, not the surrogate, time of utterance.