A recent development in the social sciences is to treat language as a medium of social action and as a topic of investigation, rather than as merely a vehicle of communication and a resource to be used in carrying out research. Traditionally, researchers have regarded language as a nuisance whose inexact properties must be remedied in designing a coding format or interview schedule and in fitting subjects' responses to the format or schedule. Now, under the influence of contemporary intellectual developments in ordinary language philosophy, speech act theory, ethnography of speaking, ethnomethodology, and sociolinguistics, students are exploring how language is deeply implicated in social activities and is therefore itself deserving of systematic research attention. Thus, actual discourse has begun to be studied in casual, everyday situations and in medical, educational, and legal settings as well. Of the institutional areas, probably legal ones have received the least attention from language scholars, but the lack is rapidly being repaired. Atkinson and Drew's Order in Court,' Bennett and Feldman's Reconstructing Reality in the Courtroom, 2 and O'Barr's Linguistic Evidence3 represent respectively the efforts of two sociologists, two political scientists, and an anthropologist, all of whom examine the structure of spoken discourse in courtroom trials. These books offer much that is interesting, if not provocative, for the student of language and for the practitioner of law. They do exhibit how inchoate this particular kind of research enterprise is: the researchers sometimes claim more than is warranted from their findings (O'Barr), they exaggerate the significance of the phenomenon described (Bennett and