One ring to bind them

One of the most profound concepts that I came across late this year was of the 'organic circuit'. In The Metaphysical Club Menand writes on John Dewey:

... it assumes that the parts are prior to the whole, when in fact it is the whole that makes the parts what they are...

Meaning that divisions are artificial, there are no real discrete units, there is simply an organic whole. According to Dewey, we do not know, so that we can do, but doing leads to knowing. Knowledge in not the result of experience, any more than a response is the result of a stimulus; knowledge is experience itself in one of its manifestations.

As is usually the case, the same concept once discovered pops up more often and much sooner than expected in other places. At a different level was Jeff Hawkins's On Intelligence which presents a new framework for intelligence - ideas on how to build a human brain. Even after all these years AI did not produce a device that can perform the basic task that a three-year old can perform with ease. Bigger, faster is not going to get us there when the neocortex, the size and dimension of a dinner napkin still beats anything out there in telling a cat apart from a dog.

The key is to first understand how the brain actually does it. Neuroscience suffers from the the other problem - too much detail, no overarching theory. Echoing Dewey, we need to study the organic whole and not just the parts in isolation. Society is not a sum of individuals, but an aggregate of interacting individuals. The brain too works as an organic whole and there is no seeing that leads to perception that results in action, but all these acting together. No wonder that 80% of the connections in the brain are feedback connections.

There are always analyses, the breaking down in various ways. What about the synthesis? the putting together. Is that a tacit assumption that we know how to put together what we have pulled apart? If the market is any judge are there jobs for synthesists?

No comments: