Wednesday, December 08, 2004

On Intelligence by Hawkins

I picked up Jeff Dawkins brilliant book, On Intelligence this morning and couldn't put it down. I love the fact that MIT, the pre-eminent bastion of AI turned down his application for admission because Hawkins had the common sense to suggest that the creation of true AI would come from a study of the human brain. There is much in this to remind us that the scientific community can be extraordinarily obtuse in regards to what an outsider might consider as obvious. I am somewhat perplexed at this point, although perhaps Hawkins will remedy my perplexity, with his contention that what he calls invariant representation or the creation of patterns out of sensory input should have little correspondence to the Aristotelian/Platonic notion of form. Indeed, he refers to the idea of eternal forms as "loopy" to our modern way of thinking, although he does feel that Plato was dead-on in terms of his focus. I say this because there is a certain kind of reductionism at work even in Hawkin's notion of what constitutes mind. The idea that the neocortex might be engaged both in pattern recognition and predictions about those patterns really doesn't bring us much closer to understanding how it actually does this than the older theories he critizes. Somebody or something recognizes patterns. When you think about it why should the brain create an invariant representation at all, unless there were something in reality--something like Aristotle's form--to prompt it to do so?

That something or observer to my mind is the deeper identity of the soul, of which, intellection in Aristotle's theory of epistemology (how you know anything) was just one of the five other powers (growth, movement, sensitivity, intellect and volition.) Aristotle's brilliant answer as to how knowing occurred is to my mind still one of the best. The conversion of the phantasm is very similar to what Dawkins is referring to as invariant representation. The only real difference is that Aristotle ties all the pieces together. In Aristotle's theory the passive intellect, much like hot wax imprints the form of whatever thing the sense perceptions are registering and the agent intellect converts that form to a phantasm, or image that matches sensory data with the invariant pattern of the form. Using Hawkin's lingo, with Aristotle's, we might say that as the eyes, for example, register the thing we call a chair, the "accidental" sensory data flows through the neocortex and the pattern recognition occurs when the invariant representation of the form is compared to the current input. However, the neocortex could never compare patterns by itself--SOMETHING has to do the comparing and this is what seems to be missing so far from Hawkins and is not missing from Aristotle. That something is not like viewing an infinity of images of yourself in two mirrors and confusing the reflection of those images with your identity. Those two images don't create your identity, they merely reflect it--you recognize that you are gazing at multiple images of yourself. In other words, you know the difference between the image and yourself because you already know the what, the form of yourself, the invariant representation of something that has substance--that which stands under the image. This recalls the Christian notion of being made in the image and likeness of God--you have substance, life force, soul--whatever you want to call it because your identity originates from the Divine Mind. (Indeed, much of How to Manage Your DICK: Redirect Sexual Energy and Discover Your More Spiritually Evolved, Enlightened Self was about the relationship between the higher self's knowledge of itself and how sexual and other energy impacts that knowing.)

I still think that there is much to be said for Plato's notion that all knowledge is remembering. Our ultimate identities and abilities and seem to originate from something closer to Aurobindo's concept of force/consciousness than the endless regression that seems to characterize much of the thinking about how AI works.. How does force become consciousness? What are the dynamics of recognition? This was something that Dad was fascinated with towards the end of this life and rightly so. Understanding the relationship between energy and consciousness is far more important than trying to figure out how things work without a Maker or Unmoved Mover. Tearing apart the brain is like tearing apart a computer to figure out how the computer made itself. You can't do it. You gotta find the builder of the hardware and the programmer of the software.

I don't want to sound like I'm poo-pooing Hawkins, he is really onto something--a guy not afraid to venture into uncharted waters and challenge some of the intuitive paradigms that may in fact not be true. Now if he would just take on global warming...

No comments: