This paper proposes a new account of the interpretation of tenses in English complement clauses by extending and revising Hans Reichenbach’s tense theory. English embedded tenses are not always interpreted relative to the matrix tense or the speech time, as Reichenbach originally proposed, and their interpretations are often affected by the speaker’s information and context. The seemingly-inconsistent interpretations of embedded tenses can be explained in a systematic way under the assumption that the reference time of the matrix clause is interpreted as the reference time of the embedded clause or its speech time. It is argued that the two reference times of a complex clause are not identical when the reference time of the embedded clause is determined by pragmatic factors such as previous discourse and conventional implicatures.