Today we wrap up the Grenada story with Day 5, which wasn't that interesting, as it was a lot like Day 3, but it will include
what was technically Day 6, the trip home, which is sort of interesting.
Sometimes it's really crazy how many topics Douglas Hofstadter and I are both interested in. My latest reading is
"Metamagical Themas," a collection of columns he wrote in Scientific American from 1981-1983. The topics are so varied, and sometimes quite
unscientific, and the writing so freewheeling yet thoughtful, that it's hard to believe these really appeared in that magazine, and harder to believe
you'd see anything like it today, in any magazine. One column that struck my fancy was about the nature of nonsense, and the history of nonsense
writing. He didn't mention Finnegans Wake, because he feels there's a lot of meaning there, but he did cite a lot of other writing, including a few
really obscure things that I'm going to have to check out.
It got me thinking about what's easy and hard for people and computers. It's hard to
give computers all the creativity and common sense that we take for granted, so what might be hard for people and easy for computers, other
than the obvious, pure computation? I think nonsense might be another thing like this. It's very hard or maybe impossible for people to think truly
randomly. We always think in patterns, there are always connections between thoughts. Nonsense writers might be able to fool the reader by going
through many connections before writing down the next idea, thereby disguising the trail, but it's really quite hard to disguise it completely, and
not create another trail with all the disguising attempts.
I'm reminded of an exercise in the cryptology class I took, in which we were asked to write a sequence of 40 1's and 0's,
and try to make it as random as possible. We then ran algorithms that measured randomness, and of course found that most of our compositions were
obviously not random, because of our tendency to have thoughts like "hmm, that's a lot of 0's in a row, better get back to 1's" or vice versa.
This may not be very novel, because the whole disguising process, like the algorithms computers use to produce "random"
numbers, is perhaps just a lot more computation! But it's certainly a different kind of computation, because we have to sort of try not to think
too much, to avoid the patterns.