In Features of naturalness in conversation, Martin Warren aims to define conversation by means of a description of nine compulsory features that set conversation apart from other types of spoken discourse. Consisting of multiple sources, determination of discourse coherence, language as doing, cooperation, unfolding, open-endedness, artifacts, inexplicitness, and shared responsibility, these nine features are based in part on John Sinclair’s (lecture at the University of Malaya, 1983) seven “discourse principles.” All of Sinclair’s principles are included in Warren’s set of “features of naturalness,” along with two original contributions of Warren’s: inexplicitness and shared responsibility.
With the professed ambition of documenting naturalness without the complication of elicitation or experimentation, Warren recorded conversations in public places without the speakers’ knowledge. He did, he claims, later inform speakers about the recordings and sought speakers’ approval to use the recordings for research. Using capital letters for monikers and without ethnographic detail, Warren’s transcriptions omit names and defining characteristics.
After an introduction that focuses on past studies aimed at describing conversation, Warren explains and defends his method of data collection. In the conversation snippets he labels “extracts,” Warren provides the talk’s location and participants (usually defined solely by a letter and the sex of the speaker). These examples are dispersed evenly throughout the next nine chapters, each of which concentrates on one of Warren’s nine features. The original data are paired with insight from other sources to illuminate the nature of each feature. The volume concludes by noting that, while accorded nearly equal focus in the volume, the nine features do not necessarily carry equal weight in interaction, and act together as distinct and singular phenomena. Because of the volume’s limitations in terms of exploring the relative status of each feature, Warren suggests further research. This research, he adds, might eventually result in a comprehensive model of conversation as a distinct and definable subset of spoken discourse.
