Saturday 29 December 2007

The starting point

I've been thinking about knowledge and knowledge management intermittently for over ten years. As a computer and management consultant I started from the need to manage information and knowledge in organisations and from Peter Drucker's insight that 'knowledge work' was becoming central to the economy.

I believe there's a disconnect between the two. By knowledge work Drucker meant, roughly, work done by people with formal education at or above degree level. But most of what's written about 'knowledge management' (KM) has nothing to say about the kind of knowledge that is learnt in gaining a degree. I don't say that the KM literature is worthless - only that it's incomplete in an important respect.

My own thinking started from a definition and two propositions.

My definition

Knowledge is assertions that are based on evidence. Assertions don't become knowledge because they are widely believed or found in textbooks (or holy books) or endorsed by authority (whether political, organisational or religious) but because they are supported by logic and observation.

Notice that this is NOT the same as defining knowledge as "justified true belief". Truth is not part of my definition because it cannot be definitively known.

Scientific knowledge is the form of knowledge that most clearly exemplifies my definition and it's reasonable to seek illumination from the philosophy and practice of science. There is, of course, knowledge in other areas, eg history, marketing and business management.

Of course my definition begs the question as to how much evidence is needed to to convert an assertion into knowledge. I confess that I do not (yet) have an answer to this question of which I'm confident. My best current answer is to replace the opinion/knowledge dichotomy with a scale which has prejudice at one end and established knowledge at the other. This would be consistent with Russell's advice to 'give to each proposition that degree of belief justified by the evidence'.

Two propositions

A) Knowledge is most interesting and important where it is general. Thus Newton's laws of motion are more important than the orbital parameters of any one planet. And an effective customer segmentation scheme is more important than any one customer's spending pattern.

B) Much knowledge can be seen as a model of the domain concerned. Newton's laws of motion are a mathematical model and a great deal of science can be seen as models of various kinds.

Models, generally less precise and mathematical, are also common in management science and marketing.

Some knowledge, and this is especially true of tacit knowledge, is probably not models; it's certainly not explicit models. I'll return to this point in subsequent posts.

Implications
The processes of knowledge creation and verification are important parts of knowledge management. It's useful to see them as model building and testing.