Lee Romero

Seth Earley on the Fractal Nature of Knowledge

A few weeks back, I was asked by Stan Garfield via email about how I might go about measuring if “knowledge specialization” is increasing – it was a question originally raised by Arnold Kling and Arnold had the hypothesis that increasing knowledge specialization in organizations was making management of those organizations more difficult.

Seth Earley was included on the email thread as well, and, while I replied (only on email – I didn’t post my reply here, though I could if anyone’s interested), I was sure Seth would have some good insights about how to go about grappling with the question.

Yesterday, Seth posted his reply on his blog, which I think highlight a good point about the initial theory – that even trying to analyze the level of specialization in knowledge is tricky because knowledge is fractal – no matter how detailed a look you take at it, there are always levels of detail below that.  To quote Seth:

[Knowledge] “is endlessly complex and classification depends on scale and perspective. It’s not a matter of “there should be more categories… “; there are more. It simply depends on where you look and your perspective.”

In my own reply, I had a vague feeling of unease about the idea of measuring increased knowledge specialization but did not think through what it meant, I tried to come up with ways one might try to discern a hypothetical increase in knowledge specialization. I’m glad to see Seth managed to more concisely crystalize the vague unease I had with the question.

I also really liked Seth’s summarization:

“The bottom line is that economic value is created not by understanding where all the knowledge is and micromanaging activities, but by providing broad constraints on targets, problems to solve, competitive differentiation, values, and resources and then creating the right circumstances that allow teams of people to focus knowledge and expertise on solving problems. Knowledge classifications are part of the tools for communicating value and telling the organization when trial and error has produced something that can be reused and applied to solving other problems.”