Many in Knowledge
Management (KM) prefer to avoid defining its meaning. Their view is
that definition is a sterile, time wasting pastime contributing little
or nothing to the real work of KM. My view is different. It is that
definition is an important preliminary step on the road to specifying
one's cognitive map of knowledge processing and KM, and, ultimately, to
developing quality models useful for developing KM solutions. I also
think that arguments over definition are not fruitless arguments, but
important exchanges about what is a good starting point for developing
a cognitive map of KM.
The purpose of a definition is not to provide necessary and sufficient
conditions for its use. Instead, its purpose is to answer a question
like: “What do you mean by Knowledge Management?” with a short,
incomplete answer that:
- allows the questioner to infer something more of
the cognitive map (or conceptual map, or semantic network) of the
target of the question; and
- facilitates the beginning of further communication, and perhaps learning relative to that cognitive map.
Figure 1 illustrates the idea of a cognitive map. Nodes represent
concepts. Edges are relationships. Single headed arrows are asymmetric
relationships. Double-headed arrows are bi-directional associations.
Edges may be weighted between 0 and 1.00, or, alternatively, weighted
with words such as “few,” “more,” “somewhat, and signed (+, -) for
increase or decrease in the target node following a change in the
source node. Since qualitative concepts can be
represented in a cognitive map, it should be clear that the idea can be
used to represent fuzzy relationships, as well as logically crisp ones.
The cognitive map idea is therefore not limited to mathematical or
precise logical relationships but can also accommodate less exacting
formulations of relations between concepts.
Figure 7.1 A Cognitive Map
The light yellow ellipse highlights the area of
definition. The gold ellipse does the same for specification. The
orange ellipse the same for a measurement model. The pink area
represents the full model. The green arrow represents a feedback
relationship. Measures are experiential concepts. Abstractions are what
we measure. The abstraction/measure boundary is crossed by measurement
rules. A pattern of measurement rules, abstract relationships and
associations among measures defines a measurement model.
Keeping the complex and comprehensive pattern of the whole cognitive
map in mind, and the small area represented by the definition, I will
characterize the definition as the "elevator speech" (the 30 second
expression of the
idea, See Geoffrey Moore, Crossing the Chasm, pp. 159-162)
representing, however imperfectly, the cognitive map. That is, when
communicating with others about any term. You can:
- Refuse to explain it;
Specify it; or
Construct a cognitive map of it.
Which would you rather do in response to a basic question from someone
either at the beginning of a conversation or at a briefing? Do nothing?
Give the “elevator speech?” Give the five-minute overview? Or give the
And if there's disagreement over a specific definition there are a
number of good reasons why that might be the case, other than mere love
of philosophical disputation.
- First, The definition may not provide enough of the definer’s cognitive map to evaluate their statements using the concept.
- Second, the definition may not distinguish the concept from other concepts.
Third, the definition may redefine the term beyond common usage in a
manner that promotes confusion in communication. (This is a frequent
occurrence due to the desire of communicators to acquire the "halo
effect" of certain terms for their frequently different concepts.)
those disagreeing may forecast that bad model development will result
(wasted time and effort) from the starting place for model construction
provided by a particular definition. And
So, once again, why bother to define? To save time in responding to a
questioner, to create a basis for further communication with others,
and last, to specify a cost-effective starting place for further
specification, measurement, and modeling.
Fifth, the "definition" may be clearly be false.
(Most of this post is excerpted from my Enterprise Information Portals and Knowledge Management, Burlington, MA: KMCI Press/Butterworth-Heinemann, 2003, pp. 105-107)