This work proposes possible categories for analysing complexity in English text. Particular attention is paid to the following categories: the long subject construction, the nominal group, the appositional group, and Interpolation structures. These categories are analysed in relation to two dimensions of text: firstly, in relation to the individual sentence, and secondly, in relation to the larger discourse context of clause relations. (A clause relation is a binary relation between members. Each member consists of one or more clauses or sentences). The first dimension examines how the presence of these categories functions to interrupt the structure of the clause as a minimal grammatical form. It is assumed that the structure of the clause is basic to the structure of text and is central to the reading process, When the structure of the clause is interrupted excessively by these categories, it is predicted that there will be a consequent impairment of reading comprehension. The second dimension of complexity examines how the structural interruption of the individual sentence can entail the structural interruption of the clause relation. Syntactic complexity at this level of structure is increased when the minimal grammatical form of the clause relation (that is, the relation between two members) is delayed excessively by structural interruption. Thus, although the individual clause is the primary unit of description in this analysis of complexity, the analysis is not restricted to the single sentence. Moreover, the individual categories proposed are descriptive procedures which derive from the examination of texts as aspects of performance, not as aspects of competence. Finally, the notion of clause is put forward as a grammatical constraint which has important implications for our production of texts. The clause is fundamental to the structure of texts and provides the mechanism for its cohesion.