A Glorious Victory for KM!
Congratulation, fellow KMWorld readers! Humanity has finally caught up with us. Knowledge as we knew it is dead! Forgive my exaggeration. Nevertheless, there’s something true about it. After all, modern KM was enabled by the raw connectivity of the internet and was fully jolted into life by the endless affordances enabled by the web. This released document management and collaboration from the limitations of real-world physics. The draft on your desktop became a nubbin that could be opened to collaborators for comment and revision. It could be put through processes informal or rigorous, ending up in a densely linked online archive that could be mined and recalled with a few keystrokes.
Amazing! And it is now so perfectly normal that we take it for granted. But surely I’m describing a world in which knowledge is more alive than ever. So why would I declare it to be dead? For three reasons: First, there has, in fact, been a deep change in how we think about knowledge. Second, AI is taking this change to a new level. Third, this column’s opening paragraph is clickbait, which, if you’re still reading, apparently worked.
Changes in How We View Knowledge
I can see the changes in our idea of knowledge in my own life: I no longer care very much about knowledge, at least the way we thought about it for the millennia before KM became a commonplace business tool. That old sort of knowledge consisted of true statements that we have some justification for believing. The nature of the justifications varied by topic area and discipline—math, physics, legal, and how to keep a custard from curdling—and all have different standards and methods for establishing the truth of a statement.
That hasn’t changed, but some things have, as those in the KM world know. KM shows that knowledge does not consist merely of atoms of truth.
Rather, knowledge is a set of relationships among truths. Relational KM databases made that clear at the beginning. More sophisticated tools, such as knowledge graphs, demonstrate that those relationships are more complex—and thus richer—than we realized. Now that we can take advantage of those relationships, they become not external to the knowledge bits, but integral to them. Take a true statement out of its contextual relationship, and it loses all its value—and its intelligibility.
That linked network lets us look at knowledge from different viewpoints. KM in that contextualized sense has made knowledge multidimensional. In one context, a dip in your sales is a threat to your quarterly results, but in another, it’s a sign to make changes, often in multiple dimensions, such as innovation investment, belt tightening, or addressing new markets.
It also means that the only way you can ever pin anything down once and for all is the way we can pin a butterfly down: by driving a pin through its thorax. The refusal of today’s knowledge to be finally fixed in place means that it’s endlessly and unpredictably reusable.
Finally, our knowledge networks are unstable: New connections are discovered, old ones may be deprecated. The value of making those connections tempts us to make them even when they’re not entirely certain. For example, we’ll want to connect the dip in sales to factors that might be crucial but that are not provable or are just a hunch. Hence, in today’s rich knowledge networks, we can record the reliability of a connection in an item's metadata: Brainstorming, conjecturing, speculating, even being wrong are all essential forms of knowing.