In English, there is a clear distinction between direct and indirect speech. In direct speech, the original speaker’s exact words are reproduced, without any change whatsoever, as in John said yesterday, ‘I’ll send you the data tomorrow‘. Note in particular that in this example the pronoun / refers to the original speaker, i.e. John; the adverb tomorrow is interpreted from the viewpoint of the original speaker’s deictic centre, i.e. the time reference of tomorrow is the day after John’s utterance. In indirect speech, two obvious changes take place. The first, which is optional, is that the speaker’s original wording may be changed, as long as the same content is expressed as in the original utterance; in the example above one might, for instance, replace leave by depart, as in John said that he would depart today.

More important is the shift in deictic centre. Apart from tense, all the other elements sensitive in the speaker’s original utterance to his deictic centre are shifted to correspond to the deictic centre of the person reporting this utterance. Thus, in the example above, in indirect speech we find he for I of the original, and today for tomorrow of the original. It is important to note that what is relevant in going from direct to indirect speech is the shift in deictic centre, and not only any mechanical procedure for replacing one set of forms by another.

In English, clearly there is a (possible) change in the tense of the verb in the shift from direct to indirect speech, as we can see in the replacement of will leave (with future time reference) by would leave in the examples just cited. At first sight, it might seem that this follows from the general change of deictic centre already observed for other deictics. With this particular ex­ample, this would indeed be compatible with the meaning of the future in the past would leave: a reference point is established in the past (namely, the time of John’s utterance), and the time reference of the situation located (namely, the time of John’s departure) is subsequent to this reference point. Below, however, we shall see that this is not the correct analysis of tense in indirect speech in English, since there is a different analysis which gives correct pre­dictions in a number of instances where the shift of deictic centre analysis gives the wrong prediction for use of tense in indirect speech; it just happens that in the example discussed both analyses give the same prediction.

It should be noted that the interaction of the rule requiring shift of deictics other than tense and the absence of any shift in tense, between direct and indirect speech, can lead to apparent conflicts between time reference of the verb and of adverbials.

The English sequence of tenses rale is subject to one interesting modifi­cation. Even when the main verb is in a past tense, it is possible (though not obligatory) to avoid invoking the shift to past sequence in the subordinate clause, provided that the content of the indirect speech still has validity. Thus, I can report John’s actual words I am ill, spoken in the past, either as John said that he was ill or as John said that he is ill. In the former version, no commitment is made as to whether John’s (actual or alleged) illness is a state continuing up to the present. In the second version, however, it is necessarily the case that I am reporting a (real or imaginary) illness which I believe still has relevance. One could use this version, for instance, if someone asked why John is not at work today, in which case the question makes clear that, whatever answer is given, it must have current validity. The same failure to apply the sequence of tenses rule can be seen with other tenses in the subor­dinate clause, e.g. John said that he would/will leave tomorrow; one cannot, of course, have John said that he will leave yesterday, where the time refer­ence of the subordinate clause clearly has past time reference, not continu­ing validity. A more complex example is reporting I will leave before Jane returns, as said by John in the past. It is possible to leave both subordinate verbs in the non-past, i.e. John said he will leave before Jane returns, with the implication that John’s leaving and Jane’s return are possible future events. It is also possible to leave just the last verb in the non-past, i.e. John said that he would leave before Jane returns, which leaves open the possibility that John has already left, but implies that Jane has not yet returned – this version could naturally be followed by... and he has/did, which cannot be appended to John said that he will leave before Jane returns. The version with both verbs shifted to past sequence, i.e. John said that he would leave before Jane returned, is the only version possible if John has in fact already left and Jane has already returned – this version could be followed by... and he has/did.

The fact that one can retain non-past tenses in sentences like John said that he is ill, where the content of the reported speech has continuing valid­ity, might seem to be an argument in favour of the deictic centre approach to tense in indirect speech in English: given that the illness is reported as holding at the present moment, the present tense is the obvious tense to use. However, adoption of the deictic centre analysis does not account for the range of data found here, in particular since it is still possible to use the past tense (i.e. John said that he was ill), even if John’s illness still has present time validity (although, of course, in this case the present validity is not made explicit).

In stating the sequence of tenses rule, we noted that a verb in a non-past tense must be placed by the corresponding past tense when the main verb is a past tense (subject to the option of not doing so when the situation referred to by the subordinate verb has continuing validity). What happens, however, if the verb in direct speech is already in the past tense? Here, English allows two possibilities. Thus, if one wants to report John said, ‘I arrived on Friday‘ in indirect speech, either one can replace the Simple Past by the Past Perfect to give John said that he had arrived on Friday, or one can simply leave the verb in the Simple Past John said that he arrived on Friday. It should be noted that the two versions seem to be distinguished stylistically, the version with the Past Perfect being more literary, the version with the Simple Past more colloquial. This suggests that there are in fact two slightly different versions of the sequence of tenses rule, either of which may be applied. In the first variant, a tense in direct speech must be put into the corresponding past tense, so that a past tense in direct speech simply remains in the past, given that the past corresponds to itself across the non-past/past dichotomy. In the second variant, a tense in direct speech must be put into the past tense expressing one added degree of anteriority. For most tenses in direct speech, the two rules have the same effect, but for a past tense in direct speech they produce, respectively, the Simple Past and the Past Perfect. In fact, the sec­ond given variant must be supplemented by a restriction “provided such a tense exists”. Thus, if the Past Perfect occurs already in direct speech, it simply remains in indirect speech, as in reporting I had seen her before yes­terday as John said that he had seen her before yesterday, since English does not have any tense that would express anteriority to the Past Perfect.