The Firm as a Distributed Knowledge System: A Constructionist Approach

There are two key questions mainstream management researchers have traditionally addressed in their studies of firms’ behaviour. First, in what direction should a firm channel its activities? And second, how should a firm be organized? The first is a question of strategy, the second of organization design. What are the typical assumptions behind these questions? What do they take for granted? First, that there is a quasi-optimum in (or at least, a good enough solution to) what a firm should pursue and how it should be organized. And second, that the quasi-optimum can be reached if all the necessary knowledge is possessed by strategists, if a system of preferences is already established, and if the relationship between means and ends is known (Mintzberg 1990: 180–7; Mintzberg, 1994: ch. 5). How could these ‘ifs’ be turned into certainties? Only if management researchers, through their studies of aggregates of firms, could identify patterns of behaviour which would then codify into ‘if, then’ propositional (or declarative) statements to be taken as valid under certain specified conditions (Tsoukas 1994a: 4; 1997b). As a result, practitioners would benefit by being able to base their policies on scientific knowledge (Ansoff 1991: 143, 146). Those policies would, ideally, also consist of ‘if, then’ rules (what Brown and Duguid 1991: 41 call ‘canonical practice’) which would be drawn upon by organizational members in their daily practices. The reader may have noticed that the preceding view of what traditional management research has been trying to achieve owes a great deal to Hayek’s formulation of what neoclassical economics tried to do (1945, 1982, 1989). For