Do we need to define knowledge in order to ‘do’ knowledge management’? The answer is probably not, and the answer may even be that defining knowledge is counter-productive to ‘doing’ KM. There is of also the question of whether we need to theorise ‘knowledge’ in order to theoretically inform our practices of knowledge management, but that is another matter, and will not concern us here.

Simply, the reason why we do not need to define knowledge is that epistemology is not necessarily the primary (or any) concern of the practice of KM. The concern is for ‘capacity for effective action’, not for epistemological analysis.

Snowden provides an interesting framework for KM, which although it is not explicitly concerned with epistemology, it does lead stakeholders on to epistemologically informed distinctions, of what he calls his four “contexts” but which might more usefully be called ‘states’ and specifically, ‘operational states’, in my own interpretation of the issue, as follows …

Operational states can be defined in terms of both perception and action. A complex state is one in which variables are self-organising and, therefore, the events are ordered but unpredictable. A prescriptive state is one in which the variables follow rigid external laws, and the events are predictable. (There is a subset of prescriptive states, routine states, in which events are not only predictable, but stable and well known and well established. And there is a fourth state, the chaotic or disordered state).

The question of how we might go about dealing with these four states, in practice, must be resolved, foremostly, at the resolution of variables, not events. And at the level of variables, the question is two-fold: epistemological and political.

Epistemologically, the issue is whether the variables are in principle at least self-organising, possibly self-reproducing, and possibly also maintain, themselves, an identity of self. Or on the other hand, is the behaviour of the variables determined entirely by laws external to them-‘selves’?

Politically, the issue is whether the variables are permitted to exercise ‘self-organisation’ at these three levels. ‘Political’ in this sense incorporates some of Bruno Latour’s recent definition of 'politics’, i.e. a politics which is inclusive of the animate and the inanimate (see here for an account and a critique of his position).

So the four states (complex, predictable, routine and chaotic) are a function of both epistemology and politics – or perception and action. In other words, on the one hand there is residual level of possible interaction between variables, and on the other hand there is the permissible, actual, level of interaction within a political context. Variables which are capable of self-organising may or may not be permitted to exercise such a capability, and this applies to variables (and thus events) within both ecological and social systems.

This provides a framework for the practical matter of ‘doing’ knowledge management.

To wit, you (individually or as a group of stakeholders) need to be aware of the four operational states, both in terms of what is possible in principle, and in terms of what is permissible in practice – within the political context of which you are a large or small part. And the ‘permissible’ is dynamic, it changes, and therefore, ‘you’ (individually or collectively) can shift variables and events from one ‘contextual quadrant’, or one of the four ‘operational states’ to another. You might find that difficult, but it depends on what you are ‘capable’ of – i.e. how much or how little you are prepared to permit or restrict the actions of variables within an event, and how far you are capable of ‘going’ to ensure what you want.