Articles

Semantics aside, "knowledge" can be managed

Re: Working Knowledge: (Ubiquity, August 8, 2000)

T.D Wilson's comments on "Working Knowledge" seem to hinge on a semantic definition. The writer seems to be asserting that knowledge does not exist outside of a human brain. If it is outside of a brain, then it cannot be knowledge. This is a self-referential definition that almost forbids discussion of knowledge as an entity. If "Knowledge is what the knower knows," then, by this definition, only the specific knower has access to knowledge. I can't know what is in anyone else's brain and everything else that is not in a brain is not knowledge by definition. This leaves me to declaim that the only knowledge that actually exists is what I know -- a highly egocentric view, and surely not what the writer means!

The dictionary (Webster's) has many definitions of knowledge associated with the word "knowing"--the act of cognition that can really only be done by a human; but it also has a definition as "the body of truth, information and principles acquired by mankind (sic)." This is the definition used by people in knowledge management. As David Bohm pointed out in "Thought as a System" knowledge of any sort: knowledge in the act of being thought, knowledge not in the act of being thought, knowledge in a hardware device, or in a book is not the thing itself, it is a representation of the thing. This implies that knowledge can indeed be passively "represented." This usual definition is knowledge equals applicable data plus data structure (relationships) plus some context (some knowledge of use).

The whole idea of the written word is to transport knowledge from one brain to another, and we see these elements in most books. While it certainly can be argued (and has been through the centuries) that there is no knowledge without humans, I could give a good argument that this may not be true.

In the August "Communications of the ACM" ("The Business of Software" p.19) I proposed that there are actually five media in which knowledge can be stored: DNA, brains, hardware, books and software. It's not too big a step to view DNA as the "knowledge of how to make living things" that is stored in a somewhat executable form. Books contain knowledge stored in a non-executable form. Software exists only that it is executable knowledge, which is why we use it.

A key point, though, is that only a human brain can intentionally create knowledge. The other media can store it, but they cannot generate it. In software, for the first time in history, we have a knowledge-storage medium in which the knowledge is persistent, intentional, modifiable and executable. The old word for "Knowledge Management" was epistemology, and it really is a very old concept.

Considering knowledge-in-software, knowledge management is pretty much the same thing as systems design. True, we are somewhat reinventing and repackaging knowledge acquisition management, but reinventing and repackaging things is also an old concept. Personally, as society moves (has moved) to a knowledge-asset basis rather than a resource-asset basis, I think knowledge management, whatever you call it, is just about the most important thing that we can do.

� -- Phillip G. Armour


� 'Knowledge' vs. 'Information'

Re: Working Knowledge: (Ubiquity, August 8, 2000)

Pierre Berger comments: "I think that Davenport and Prusak have made a clear picture of what the general public and management in general intend by "knowledge." But I remain thirsty about more substantial findings, if this is supposed to be a research paper. It would probably be necessary to enter in a modeling and field-testing process of "knowledge." If possible quantitative modeling. A long and perhaps tedious time, for sure. But it would give a firmer ground to the humanistic thinking of the authors."

� I dispute that D&P have made any kind of "clear picture" of 'knowledge' --indeed, they totally confuse the situation. They slip in this article as in others they have written, from "knowledge" to "information" and back again. This is management consultancy rubbish and has nothing to do with research of any descriptions. As for the possibility of any kind of quantitative research into "knowledge" -- if anyone could do that then "artificial intelligence" would be a reality and, in spite of all the hype and all the money, no one has yet managed to determine what human intelligence is -- a necessary precursor to any artificial correlate.

� Lyda Woods comments: "In my experience, knowledge management has helped novices see patterns in information, data or troubleshooting that experts have helped to highlight. This was true of the expert system (a kind of knowledge base) I helped to develop two years ago. Our goal was to provide newer analysts with troubleshooting decision models that were based upon expert knowledge. The knowledge base served as a kind of extended mentoring system since experts could reach more people through the system than by talking with each individual and many of the questions were the same anyway."

� But this, again, is to confuse "knowledge" with "information" -- an expert system records what an expert is able to convey about his/her knowledge, i.e., information about the person's knowledge base -- and, necessarily, it must be partial, because it is impossible to convey everything one knows about anything. An expert system, therefore, records not knowledge, but information connected through decision rules -- which, themselves, may or may not be firmly grounded in scientific understanding; instead -- they may simply "work". Such systems may indeed be useful -- although the minimal extent of their use suggests that they are not as useful as their originators hope (what has happened, for example, to all of the medical-related expert systems of some years ago -- they appear not be used even in teaching, let alone practice) -- but they are based upon information, not knowledge.

� It is this slippery confusion of information and knowledge that detracts from the genuine advances that information management, computer science and computational logic have made in the organization of information over recent years and the advocates of "knowledge management" are simply self-seekers selling yet another "solution" to gullible senior management.

� -- T.D. Wilson