Interesting discussion going on at the actkm group. A few days ago, Cory Banks proposed collaboration on a KM Body of Knowledge (BOK). Last night (eastern US time), and yesterday, the discussion intensified about what was involved in creating such a BOK, and, of course about what a Body of Knowledge is. For the ongoing discussion, I recommend that you access the actkm.org list serv and follow it. Here I’ll just state my view about what a Body of Knowledge is and then point out why the so-called Bodies of Knowledge one finds in the IT world are really, from my point of view, just Bodies of Information, or, if you prefer, Bodies of Content, but not Bodies of Knowledge.
I’ll begin with linguistic content, and with the idea that, in general, some, but, not all, of it is knowledge, and that since such content asserts something, it is non-random which makes it all information as well as content. Now, let’s assume we have a Body of Content, or what is the same thing, a body of linguistic information, then how should we distinguish bodies of linguistic information from bodies of linguistic knowledge, when both of these are comprised of content that makes assertions about the world?
Now one could take the position that there is no difference between information content and knowledge content, because there really is no linguistic knowledge, and that knowledge is really mental in character and refers to beliefs. I can understand this position well enough, but reject it because it rejects the idea of objective knowledge since mental knowledge cannot be shared or tested intersubjectively. Apart from this however, such a position suggests that the idea of a “Body of Knowledge” is an impossibility. Since this discussion is predicated on the assumption that there are bodies of knowledge, the idea that knowledge is only mental in character is another subject for another day. For today we are assuming that Bodies of Knowledge are possible and are discussing the question of how they differ from Bodies of Information.
Next, one could also take the position that Bodies of Knowledge are different from Bodies of Information in that the contributors to Bodies of Knowledge believe they are contributing true information to the body, whereas the contributors to Bodies of Information aren’t concerned with the truth or falsity of information. This position however, rests the identification of a Body of Knowledge on the intent of the contributors of content to a repository. Of course, such an intent will be very difficult to measure or divine in practical applications. So, if we accept this position, we are basically saying that while there may be a difference between bodies of knowledge and information, there is no measurable difference that has practical implications.
Next, we could look for differences in the bodies based on their differing content, and say that linguistic knowledge content is richer than mere information content in some important way. This is the position I take. Specifically, I think that bodies of linguistic knowledge are comprised of knowledge claims supplemented by other necessary information, whereas bodies of linguistic information don’t have such additional information. “Knowledge Claims” are statements we make about the world, or about what’s valuable, or what’s right, or what’s valid or invalid. They can be expressed in single statements or in networks of statements, But, what is the nature of the additional information that together with knowledge claims defines linguistic knowledge?
I think this additional information is comprised of “meta-claims.” “Meta-claims” are claims about claims. They’re important because the “track record” of performance of our claims, and of their survival in the face of criticisms, tests, and evaluations, and of their relevance to the problems they are supposed to solve, is made up of meta-claims. Knowledge claims assert what we think is true, or right, or efficacious, or valid. But such an assertion doesn’t make them “knowledge.” It only makes them “information.” Objective knowledge also requires the track record of performance and survival in the face of evaluation perspectives and/or criteria and competing alternative claims. So, in my view, Bodies of Knowledge are comprised of both knowledge claims and meta-claims about the performance of these claims. Bodies of Information, or Content need not include meta-claims about performance, and that is the difference between them and Bodies of Knowledge.
Very often, the phrase “Body of Knowledge” is applied to knowledge claims “established” in various fields of science and engineering. These “bodies” however, are not codified in any single knowledge base, and both the knowledge claims and meta-claims involved are distributed across many information/content sources. Nevertheless, scientific bodies of knowledge certainly fit the idea of objective knowledge I’ve just offered since both knowledge claims and meta-claims are part of scientific bodies of knowledge, and without the presence of the meta-claims we would not have scientific bodies of knowledge.
Moving to “Bodies of Knowledge” found in the IT world, I don’t view these as “real” BOKs, because they don’t provide the rich meta-claim context necessary for a knowledge base. They may provide the assertion of a best practice or a lesson learned, or a popular and useful technique, but they don’t provide the story of the knowledge claims that are part of a BOK. That story needs to include the problems giving rise to new knowledge, development of alternative solutions if that occurred, the process of evaluating solutions including any “safe-fail” experiments that may have been performed to test new knowledge, and the record of the new knowledge in practice. That is, it must include all the knowledge claims and meta-claims necessary to see the knowledge claims in their full evaluative context, so that humans using the BOK can evaluate its claims anew.
In some fields of business, you’ll find these kinds of stories told and stored for future reference: specifically in the Quality Management field, many organizations, Toyota provides a good example, cultivate meticulous reporting including description of problems, root cause analysis, development of solutions, experiments to test solutions, implementation, and analysis of gaps between expectations and actual results. Knowledge Bases of such reports are Bodies of Knowledge in the sense I’ve defined the term. But when people normally talk about BOKs they’re not talking about the full context of knowledge claims, but only about what has worked in the past and is considered “best” or “learned.” This sort of BOK is really a “Body of Content” or “information, because its users must begin its evaluation from scratch with no capability to access the track record of the claims in the so-called BOK.