One of the central problems of social theory has long been that of ho\v to link so-called 'macro' and 'micro' levels of social structure, social organisation and social process. Over the years, many social scientists have suggested that perhaps the macro/micro dichotomy is not the most helpful \vay in which to understand how the observable dimensions of social life in the here and no\v are linked to durable patterns \vhich lie beyond the control or the a\vareness of individuals (see for example Cicourel 1978; Collins 1981; Giddens 1982; Mehan 1987). Conceptualising social life in terms of a dichotomy implies that there are different types of data for each, equally observable (or not, as the case may be), and that, in addition, the linkages should be identifiable. And yet, empirical work fails to identify such types. Instead, Giddens, for example, proposes thinking in terms of structure and action, but in ways which allow structure and action to be in a dialectical relationship to each other, a relationship which itself occurs in the observable patterns of situated social action.