Originally posted on 06-10-07:
Your car stopped running so you have it towed in. The garage calls you back in a few days and says it’s ready, and when you go to pick it up, the mechanic says, “It was broken and I fixed it. That will be $5,499.”
After you get your breath back, you ask, “What was wrong with it?” At which point he spins his computer monitor around so you can see, and starts going through the results of analyses of the chemical composition of all the fluids in your car, the metal alloys and plastics that its parts are made of, the measurements of wear in each of those parts, etc.
He actually doesn’t get very far with this before you stop him in exasperation and say that you aren’t really interested in this level of detail, and you’d like to know what parts he replaced or adjusted and why. He says that some object apparently punched a hole in your oil pan, the oil drained out, the cam chain wore from lack of lubrication until it broke, bending all the valves, which knocked chips out of the pistons, which scoured all the cylinders, so he replaced the engine.
You’re not happy, but this is a level of explanation that you can understand. His declaration that it was broken didn’t give enough detail, and his molecular analysis gave too much—you needed something in between.
Which brings up Douglas Hofstadter and his latest book, “I Am a Strange Loop,” which I mentioned a few days ago (weeks?). I found many interesting things in the book, but I think he neglects the importance of different levels of explanation, and their relevance to the kinds of questions we’re asking.
He wants to draw a parallel between thermodynamics and what he calls “thinkodynamics.”(p.34) In thermodynamics, it is possible to say many useful things about the behavior of gases without trying to detail all the collisions of atoms involved, which he summarizes as, “Statistical mechanics can be bypassed by talking at the level of thermodynamics.” The parallel is that in thinkodynamics, we can bypass talking about synaptic gaps and ion channels and talk about ideas.
I agree that much of our behavior can be understood by talking at the level of ideas, but the parallel with thermodynamics breaks down when we realize that the physicists ideas of pressure, temperature, etc. summarize “highly predictable regularities,” in the behavior of “invisible microscopic constituents.”(p.33) Human behavior at the level of ideas does not deal with highly predictable regularities—consider the vast number of meanings for an idea like “love.”
Real world example: Joe feels depressed at the idea of going to work, and calls in sick a lot. He goes to a psychiatrist who says, “Your neurotransmitters are out of balance—take this Prozac and come back in a month.” Joe takes the magic pill, his depression lifts and his attendance at work becomes exemplary.
Sue feels depressed about work, too, but she doesn’t like the idea of drugs so she goes to a talk therapist. They discuss her relationship with her supervisor, and the therapist suggests ways Sue might approach her boss. She talks to the supervisor, who turns out to be very understanding and alters her behavior toward Sue, whose depression lifts.
In both these examples, the idea of “depression” is involved, but the “highly predictable regularities” of thermodynamics are missing. As James Taylor sang in “Daddy’s all Gone,” “It’s not that simple.”
I think there are some very useful levels at which we can talk about human behavior, somewhere between neurotransmitters and complex ideas like “depression.” Looking at the way activity moves from one area of the brain to another is particularly useful. You may remember my talking on that level, and there’s more to come, but this short version is already getting pretty long—later.
Focused On A Different Level