Looks like I was precisely wrong


I counted the number of words in each paragraph for some writings, selected more or less at random, and fitted a t-distribution centred at zero (well, really a half-t distribution) using scipy (which totally rocks).

Now obviously this is very preliminary, but:

Temple Grandin - Thinking in Pictures: sigma = 128, df = 22889374

Cory Doctorow - Down and Out in the Magic Kingdom: sigma = 41, df = 8.3

Jane Austen - Pride and Prejudice: sigma = 51, df = 2.7

Mark Twain - Huckleberry Finn: sigma = 28, df = 1.3

Temple Grandin is autistic. Her curve didn't even look like it had a tail. Cory, well, it would be a fair bet he has some leaning that way. Not very many data points, but it's exactly the opposite of what I expected, damnit!!

To add insult to injury, it appears that autistic people have more white matter than normal (and, as the linked paper notes, are often "frankly macrocephalic"), but the corpus callosum is no larger, most of the increase being near the grey matter on the surface of the brain. Again, exactly the opposite of what I would have expected.

Well, at least reality is being consistently contrarian, perhaps even measurably so.

Addendum: Temple Grandin's curve may even be log-normal, a somewhat related distribution. If so, her brain has crossed some critical threshhold and is working in some qualitatively different way. That would be quite exceedingly weird. Hmm. ... further addendum er, scratch that, because what we are dealing with here is the Levy stable distributions, extremal form, to which this data looks like it will all fit perfectly, no qualitative differences. Levy stable distributions are the distributions of ultimate evil, and the secret rulers of the universe. They're big brothers of the normal distribution. They laugh at the puny central limit theorem and tell it to go eat infinite variance. They assimilate outliers without blinking. And, rather annoyingly, they wouldn't be caught dead expressing themselves in an analytic form.