Last year, Ray Sims compiled a discriminating list of 62 KM definitions (and counting). I think 62 definitions is a great start. As far as I'm concerned, the more the better—and much better than just one definition of knowledge management that tries to please everybody.
Best of all, many of these definitions are very different from how I see KM. For example:
"…the process of improving the job performance of knowledge workers by eliminating relevant ignorance and inability as quickly and inexpensively as possible…"
Why are multiple, bad definitions of KM better than one definition, even if it's really good?
"I find definitions to be extremely limiting. They become permanent representations of the thing being defined and leave no room for interpretation, adaptation or evolution over time," writes Adrian Ward (via Sims), who prefers to make distinctions. "Distinctions allow something to take on different guises in different contexts. They allow you to accumulate a list of those manifestations and characteristics over time. They also allow you to include what something is not in addition to what it is, to set boundaries around the thing being described. They provide people with a much richer sense of meaning and understanding. In short, something moves from being a dead set of words to being alive in the mind of the receiver."
I've come to believe that ambiguity and abstraction are among the most useful gadgets in the KM and organizational learning toolkits. You are forced to adapt and apply, rather than simply accept and adopt. A single definition assumes clarity and agreement, even when this is not the case. It assumes understanding where there may be confusion. No matter how precise someone tries to be, differences in expectations, experience, context, culture and language will always cause people to hear things differently. And as long as that's happening, you might as well turn this diversity to your advantage.
In short, it's the difference between knowledge and learning.
I prefer multiple definitions because clarity and agreement cannot be assumed. Meaning must be negotiated and confirmed. Even if it's only a temporary agreement or working definition for the task at hand, that represents a position triangulated from the multiple points of view of all participants. You'll have a much better commitment to success if consensus was earned rather than enforced.
If the definitions are mutually exclusive, that's even better. Paradox and contradiction drive us think deeper, and ultimately reframe the problem or question, like Zen koans.
I will say there was one definition of KM that I really loved:
"The mistaken idea that what is in people's heads (knowledge) is fundamentally the same stuff as can be documented in words, pictures, charts, etc (information). This underestimates the unique and essential value-adding role of people, who make things happen by applying skills, experience, reason, intuition, passion, and decision to information. You can't bottle this stuff."
Although I love "You can't bottle this stuff," none of these supplant my favorite definition, which isn't even one of mine. Not surprisingly, I have several definitions of my own, including this irreverent column. In 2002, I created another working definition for destinationKM.com (though I might not use it today):
Knowledge management refers to strategies and structures for maximizing the return on intellectual and information resources. KM depends on both cultural and technological processes of creation, collection, sharing, recombination and reuse. The goal is to create new value by improving the efficiency and effectiveness of individual and collaborative knowledge work while increasing innovation and sharpening decision-making.
Another effort, Tangient's knowledge transfer wiki, assembled 27 definitions knowledge management. Surprisingly, I didn't see a single definition of knowledge there. Not one.
Now, zero (0) definitions of knowledge is very Zen. But I don't know how far you can get in discussing KM if you aren't willing to speculate about K.