A Proper Treatment of the Sequence-of-tense Phenomenon in English

Abstract

My purpose in this paper is to provide an appropriate account of the sequence-of-tense (SOT) phenomenon in English. I address some problems with the previous studies of the sequence of tense such as Ladusaw (1977) and Ogihara (1992). Ladusaw's (1977) claim that the matrix past tense is copied onto the based-generated present tense in the complement clause runs into trouble when it deals with present-under-past sentences. Ogihara's proposal that the embedded tense is deleted under its identity to the matrix tense is not sufficient to account for sentences where the non-stative predicates occur in the complement clause. I propose the Tense Co-indexing rule which roughly says that the embedded tense is co-indexed with the matrix tense under identity. This rule is intended to denote a precedence or simultaneous relation between the matrix tense and the embedded tense. I also propose a SOT rule that says the embedded tense is deleted when bound by the matrix tense. The Tense Co-indexing rule triggers the application of the SOT to a given sentence. The analysis which I propose here takes account of those sentences that are not properly treated by Ladusaw (1997) and Ogihara (1992).