The problems of artificial intelligence show interesting things about intelligence. One of its longstanding problems is how hard it is to program machines to detect triviality and significance. Program a computer to maneuver obstacles in a room and open up a door, and it’s very difficult to have the computer not freeze if you take electrical tape and make a black “X” on the door. Have a computer watch a person open a jar, and it’ll copy everything the person does – scratching his face, pausing to mumble, etc. This is the principle behind the “prove you’re not a spambot” tests we have to take on comboxes.
AI might solve all these problems -it might have already solved them. The only point is that detecting significance and triviality is crucial to learning, and so an account of learning as mere induction from experience or deduction from general ideas will not be adequate. Both of these things exist in a context where some significance and triviality has been determined in advance. So what is the power to determine this? To call it an intuition is empty – to intuit something is just to see it, and so all we’ve said is that we see significance by seeing it.
This eye for significance and triviality is what Peirce called abductive reasoning, but he was just as quick to compare to an instinct for the right answer. Steven Pinker points out that the process seems to involve an ability to put ourselves in the mind of another – when we see someone open up a jar we can tell what is significant and what isn’t because we can see the action through their eyes. This seems right – significance and triviality are inseparable from how a thing is supposed to be, and to speak of some arrangement being as it is supposed to be involves teleology and intention. A blind force or a random mutation can be lucky, but it can’t meaningfully be said to accomplish what was supposed to happen. Pinker is right that we learn to open a jar by putting ourselves in the mind of the opener, but we learn about nature in just the same way. Understanding the making of a silk flower involves the same sort of thought as understanding the making of a real one.
Thus, we get the first theology suggested by the thought of Steven Pinker.