This book sets itself the ambitious aim of assessing the relative merits of two fundamentally different approaches to the analysis of discourse: “grammatical research” and “conversational analysis.” The latter, derived from the work of ethnomethodologists in the 1960s and 1970s (Garfinkel 1967), is concerned with the analysis of the principles involved in such phenomena as turn-taking in conversation (Sacks, Schlegloff, and Jefferson 1974); the former is not a specific theory nor a single approach, but appears to cover standard linguistic methods in the description of discourse, focusing on features of language itself and their role. The evaluation of the relative value of these approaches is potentially a rather wide-ranging enterprise; however, the book is in fact clearly focused, and is firmly grounded in the practical analysis of data, which gives its findings an authority that could not be achieved by more abstract and theoretical reasoning alone. These findings could certainly be seen as a vindication of the “grammatical” approach at the expense of the “conversational”; however, it is not a polemical demonstration of its superiority, but rather a balanced and sympathetic assessment. Using data from recorded conversations, the author examines three marginal linguistic features that might be said to have a conversational role, and concludes that this role can be seen as linguistic, as opposed to purely conversational. Since linguistic structure itself can thus be demonstrated to have interactive relevance, the author concludes that, in effect, at least some of the claims made for conversational analysis are to be rejected. Whether an issue of this general and ultimately theoretical kind can really be decided in this practical way may be open to question, as it seems to depend more on the overall approach to the phenomenon than on the analysis of data, but there can be no doubt that the detailed analysis of specific features of conversation from this perspective makes the book a valuable contribution to the study of the linguistics of conversation, regardless of how we evaluate its theoretical conclusions.