The concept of structure
Social Theory and Modern Sociology (Polity Press, Oxford, 1986)
Interpreting agency within the context of its duration helps link the notion of action to those concepts—that is, structure, institutions and so on—which have been so important to objectivist social scientists. To see how these connections might be made, we have however to consider afresh the concept of structure. Among English-speaking social scientists, the concept of structure has ordinarily been a received notion. In contrast, for example, to the concept of function, the idea of structure has received remarkably little discussion. Why should this be so? The reason is probably that most English-speaking social scientists have a clear idea of how the concept of structure should be understood. When they talk of structure, or of 'the structural properties of institutions', they have in mind a sort of visual analogy. They see the structural properties of institutions as like the girders of a building, or the anatomy of a body. Structure consists of the patterns or relationships observable in a diversity of social contexts. Now this notion of structure needs to be examined just as closely as the idea of action.
In the traditions of structuralism and post-structuralism, the concept of structure is used in a fashion quite divergent from that characteristic of Anglo-Saxon social science and philosophy. The easiest way to indicate this is by reference to Saussure's classic discussion oft he structural qualities of language. Structural features of language do not exist as patterns situated in time and space, like patterns of social relationships; they consist of relations of absences and presences embedded in the instantiation of language, in speech or in texts.' Structure here presumes the idea of an absent totality. To understand a sentence which a speaker utters means knowing an enormous range of rules and strategies of a syntactical and semantical kind, which are not contained with the speech act, but are nevertheless necessary to either understand it or to produce it. It is such a notion of structure (as an absent totality) which I hold to be important as a concept for the social sciences as a whole and basic to the notion of duality of structure. The problem with conceptualizing structure as a set of relations of 'presences', is that structure then appears as a constraint which is 'external' to action.
If we conceive of structure in this fashion, it is not surprising that action appears to be limited by structural constraints which have essentially nothing to do with it. For structures limit behaviour, although it may be that within those limits—so one would have to presume—the agent is capable of acting freely. According to the notion of the duality of structure, by contrast, structure is not as such external to human action, and is not identified solely with constraint. Structure is both the medium and the outcome of the human activities which it recursively organizes. Institutions, or large-scale societies, have structural properties in virtue of the continuity of the actions of their component members. But those members of society are only able to carry out their day-to-day activities in virtue of their capability of instantiating those structural properties.