A couple nights ago while I was falling asleep I had a vision of a humor research experiment in progress. A large audience was in a theater, watching a comedy movie. Each member of the audience was being videotaped by a camera mounted to the back of the seat in front of him/her. Someone was going to go through each tape later and for every “laugh point” in the movie, make a note of whether the person either 1) laughed with open mouth, 2) laughed with closed mouth, 3) smiled or the like, 4) did nothing, or 5) made a demonstration of displeasure. Then all this data was going to be analyzed by computer, looking for jokes that divided the audience in similar ways, or individuals who divided the jokes in similar ways, etc., and finding the degrees of correlation between such things. From this they would (I dreamed) come up with a set of 7 or 8 basic variables that defined the space on to which humor could be mapped. Then they’d come up with a sense-of-humor profile for each individual in terms of those variables. Then they’d correlate the profiles with survey questions that each person would have filled out…and would thereby be able to quantify and/or discredit assumptions like “slapstick is blue-collar humor.”

In the light of day this seems to me, unlike most things I think of when I’m falling asleep, perfectly reasonable – to the point that I imagine it, or something like it, must actually have been done at some point. The demographic bit at the end would certainly be valuable to the entertainment industry – more than Nielsen-type ratings, which don’t differentiate between enjoyment and compulsive/pleasureless viewing, which I imagine accounts for a good chunk of the US entertainment market. The suits might say that they don’t really care about the difference, but obviously they should.

So anyway, if anyone can find me a link to the results of this experiment, I’d appreciate it. All I could find was this statistical sense-of-humor analyzer, which, after forcing me to rate what must have been 40 or 50 jokes, finally admitted defeat and told me that it could find NO jokes in its database to recommend to me. I think it was right about that.

This is much more impressive at a similar task. But if you want to stump it, you will, so go easy. I always feel proud of it when it does anything right. Just now I was thrilled that it got “catalog” right on question 17. Catalog!