People often wonder if a computer could ever have a sense of humor. This question seems natural to those who think of humor as a pleasant but unnecessary luxury. But I'll argue quite the opposite — that humor has a practical and possibly essential function in how we learn.
When we learn in a serious context, the result is to change connections among ordinary agents. But when we learn in a humorous context, the principal result is to change the connections that involve our censors and suppressors.
In other words, my theory is that humor is involved with how our censors learn; it is mainly involved with negative thinking, though people rarely realize this. Why use such a distinct and peculiar medium as humor for this purpose? Because we must make a sharp distinction between our positive, action-oriented memories and the negative, inhibitory memories embodied in our censors.
Positive memory-agents must learn which mental states are desirable. Negative memory-agents must learn which mental states are undesirable.
Because these two types of learning required different processes, it was natural to evolve social signals to communicate that distinction. When people do things that we regard as good, we speak to them in encouraging tones — and this switches on their positive learning machinery. However, when people do things we consider stupid or wrong, we then complain in scornful tones or laugh derisively; this switches on their negative learning machinery. I suspect that scolding and laughing have somewhat different effects: scolding tends to produce suppressors, but laughing tends to produce censors. Accordingly, the effect of derisive humor is somewhat more likely to disrupt our present activity. This is because the process of constructing a censor deprives us of the use of our temporary memories, which must be frozen to maintain the records of our recent states of mind.
Suppressors merely need to learn which mental states are undesirable. Censors must remember and learn which mental states were undesirable.
To see why humor is so often concerned with prohibition, consider that our most productive forms of thought are just the ones most subject to mistakes. We can make fewer errors by confining ourselves to cautious, logical reasoning, but we'll also discover fewer new ideas. More can be gained by using metaphors and analogies, even though they are often defective and misleading. I think this is why so many jokes are based on recognizing inappropriate comparisons. Why, by the way, do we so rarely recognize the negative character of humor itself? Perhaps it has a funny side effect: while shutting off those censored thoughts, our censors also shut off thoughts about themselves — and make themselves invisible.
This solves Freud's problem about nonsense jokes. The taboos that grow within social communities can be learned only from other people. But when it comes to intellectual mistakes, a child needs no helpful friend to scold it when a tower falls, when it puts a spoon in its ear, or thinks a thought that sets its mind into a fruitless and confusing loop. In other words, we can detect many of our own intellectual failures all by ourselves. Freud's theory of jokes was based on the idea that censors suppress thoughts that would be considered naughty by those to whom we are attached. He must simply have overlooked the fact that ineffectual reasoning is equally naughty — and therefore equally funny — in the sense that it, too, ought to be suppressed. There is no need for our censors to distinguish between social incompetence and intellectual stupidity.