It must be the times. Go to any Knowledge Management (KM) professional
meeting. Read any KM Journal or popular magazine. Join any KM
newsgroup. The story is the same. Except for those pesky KMCI types and
a few of their friends, no one seems to be interested in practices,
methods, or theory about evaluating knowledge claims. Now, given the
general and long-standing interest in decision support, and the fact
that KM is often justified as improving it, isn't this
passing strange?
The problem of knowledge claim evaluation is a decision problem itself.
It is the problem of selecting the best among competing knowledge
claims, and the problem exists whether or not one thinks that knowledge
is a type of belief network, or whether it is a type of semantic
network. For even if one thinks that semantic networks are only
information, one should still care very much about the relative quality of
information and its relationship to knowledge (viewed as belief) and
should, therefore, select among competing knowledge claims by deciding
which one has the highest quality. So, why aren't KM professionals
concerned about Knowledge Claim Evaluation?
Explanation1: "The Old Knowledge Management" and Knowledge Sharing
"The Old Knowledge Management" is about Knowledge Sharing. Its value
propositions are better decision support, higher job productivity and
performance, and capture of knowledge assets that would otherwise leave
the organization. The Old Knowledge Management is not about making new
knowledge, problem solving or innovation. So why should it be concerned
with Knowledge Claim Evaluation, the sub-process that allows us to
decide what is knowledge and what is "just information"?
Even if one thinks there's some truth to this explanation, the Old
Knowledge Management has now been supplemented by a concern for
knowledge making and innovation. "
Second Generation Knowledge
Management" (SGKM) has arrived and KM is concerned with much more than
knowledge sharing as a visit to any of the major periodicals and news
magazines in the field will attest. Yet the appearance of SGKM has
hardly increased the concern for Knowledge Claim Evaluation within the
mainstream of KM.
Explanation 2: The Belief that Knowledge Making in Business is a
Practical Activity and Includes No Time or Resources for Knowledge
Claim Evaluation
I've heard from some that Knowledge Claim Evaluation is not very
important in KM, because it is a business activity, process, or
discipline, not a science. The implication, of course, is that science
uses Knowledge Claim Evaluation because of its deliberative, exacting,
theoretical, and precise character, while business with its much more
imprecise and action-oriented practical reasoning just can't afford the
time and effort that the deliberative approach to knowledge making
requires.
This line of reasoning, if it represents a widespread attitude in KM,
may provide a part of the reason why there is so little concern about
Knowledge Claim Evaluation in KM. As I explained in
All Life Is Problem
Solving, however, all of our non-routine knowledge making, and that
means whether in science or business, or any other area of
organizational or human behavior, involves problem recognition,
formulating tentative solutions, and error elimination.
In
organizations we do perform Knowledge Claim Evaluation. It is how we
attempt to eliminate errors in our knowledge claims.
The only important
questions are whether we do so with full awareness of what we are
doing, and whether our practices produce knowledge claims that are
effective in raising the quality of our business process performance or
not.
Explanation 3: The Belief that Knowledge Claim Evaluation Is Based On Authority
Knowledge Claim Evaluation is not of great concern to KM because, since
knowledge claims cannot be justified as true through evaluation, there
are only three theories of evaluation that count in organizations
anyway: (1) what managers think, (2) what experts think, and (3) what
one's community thinks. In all three cases, some form of authority:
managerial, expert, or community consensus, "justifies" our knowledge
claims.
This is another view that may explain why Knowledge Claim Evaluation is
not of greater interest to KM. If only authority can justify our
knowledge claims, the issue of how we ought to select among knowledge
claims is of no importance. We have no choice. What we select is
determined by various authorities, by politics and not any rational
procedure.
Explanation 4: The Belief That Knowledge Is Socially Constructed,
Determined By Social And Cultural Background, and Unaffected By Reality
Social constructivism, an epistemological theory held by many in the
social sciences, holds that reality as well as our knowledge of it is
socially constructed and that such knowledge constructions are
unaffected by an independently existing reality. Social constructivism
often goes along with two other beliefs. First, the distinction between
objective and subjective knowledge is meaningless because all knowledge
is a function of our social and cultural context and can only be
justified relative to that context. And second, such justification can
only be provided by community consensus, since only it reliably
reflects the influence of social and cultural context on our knowledge.
Since community consensus is the only legitimate basis of knowledge,
explanation 4 partly agrees with explanation 3. It holds that Knowledge
Claim Evaluation is a simple matter of determining whether a knowledge
claim network is backed by a community consensus. So we need not spend
our time worrying about effective methods of Knowledge Claim
Evaluation. All we need do is see to it that knowledge is effectively
shared so that the community is informed. Then we just need to wait for
consensus to emerge.
Of course, the problem with this reasoning begins with reality. Reality
is
not socially constructed. Our knowledge of it is
certainly mediated by our social networks, along with our psychological
predispositions, and biological heritage, but it is also influenced by
reality itself, which exists apart from our social construction of it.
Since reality, and our knowledge of it, are it least partly
independent, the issues of the correspondence of our knowledge claims
with reality, i.e. of their truth, and of which of a competing set of
knowledge claims is closest to the truth, need to be faced. And since
neither correspondence to reality, nor closeness of approach to the
truth can be measured directly, facing these issues means facing the
issue of how we can effectively evaluate our knowledge claims.
We know enough about knowledge claim evaluation through the centuries
to know that
it is not effective to use any form of authority,
including community consensus as a criterion for evaluating knowledge
claims. Knowledge claims cannot be validated by community consensus,
but rather should be continuously tested and evaluated in order to
eliminate error.
The Job Ahead
Whether KM's lack of concern about Knowledge Claim Evaluation is due to
the idea that only knowledge sharing is important, or to the idea that
business is imprecise and neither needs nor has the time nor the
resources for it, or to the idea that Knowledge Claim Evaluation in
business can only be based on authority, or, to the idea that reality
and our knowledge of it are both socially constructed, or all four, or
some other reasons that haven't occurred to me, is interesting, but, in
the end, secondary. What is important, is that this lack of concern
means that KM has not been doing anything to enhance the key
sub-process in the Knowledge Life Cycle, Knowledge Claim Evaluation. It
hasn't been doing anything to distinguish among knowledge claims
according to their quality, which also implies that it hasn't been
doing anything to distinguish objective knowledge from information, or
to measure the success of knowledge claim evaluation in producing
effective knowledge.
Knowledge Claim Evaluation is not ignored in every field of business
application today. Not even in fields that are closely related to
Knowledge Management. Knowledge Discovery in Databases and Data Mining
(KDD) has, since its inception in the 1990s, considered validating
models an important activity, and continues to produce useful research
on validation criteria that are applied in model estimation. But KDD
has had little effect on KM, perhaps because its orientation toward
using formal reasoning in development of its own perspectives is
foreign to most KM practitioners.
The job ahead is to develop methods of Knowledge Claim Evaluation that
will enable knowledge workers to do a better job of selecting among
competing knowledge claims. In our book,
Key Issues In the New
Knowledge Management, KMCI Press/Butterworth-Heinemann, 2003, Chapter
5, Mark McElroy and I have begun this process by outlining a theory of
fair comparison and two formal approaches to measuring "truthlikeness".
But this is just the first little bit of work in an area that requires
substantial effort.