ABSTRACT
This chapter discusses the application of discourse analyses to the assessment of oral verbal communicative behaviour of traumatically brain injured adults. The discourse analysis procedure begins with the elicitation of a discourse sample, ideally five sentences or more in length. L. E. Nicholas and R. H. Brookshire have noted that a valid discourse sample should consist of a minimum of 200 words. There are numerous measures that can be generated at the sentence level once the transcribed discourse sample has been distributed into a basic unit such as T-units. Grammatical aspects of discourse have been studied in several recent investigations. The traumatic brain injury (TBI) subjects' greater impairment in maintaining global, as opposed to local, coherence suggests that their disordered discourse results from impaired macro-organisational abilities more than from disrupted meaning relationships between contiguous concepts. A variety of productivity measures have been applied to TBI speakers and may be categorised into three groups: total output, speaking time and mazes.
