Wednesday, August 03, 2005

DI figures it out

The DI figured out how a typo gets into print:
[Times reporter Bumiller] apologized for the error saying it was "nothing more than a mistake madie in haste" and has assured us that the Times will run a correction in the print edition tomorrow. In explanation she indicated that she'd taken notes very quickly in shorthand and later under deadline pressure from editors inadvertently substituted "biblical" for "biological."

Whether it's a Freudian slip or not it is likely to be a comment that grows into an urban myth. So, it is with relief that we see the reporter and the paper taking steps to quickly correct the mnistake [sic].
I'm not big on mocking typos, but the Discovery Institute putting a typo in "mistake" in a post about a typo is just too amusing.

You'll recall that I guessed these were notes taken in shorthand. Here's why: Shorthand is a phonetic system. When you take notes in a phonetic shorthand, you lose vowels. So, blgcl to bblcl would be a rough approximation of the difference we're talking about.

This is the problem with inferring "design" without a proper model of the phenomenon under consideration. That's why DI's "mn" typo makes a lot of sense, when the "madie" typo at the top is less reasonable. "M" and "N" are right next to one another, while "I" isn't near "D" or "E."

When Dembski and his merry band make probability calculations, they almost always assume that all outcomes are equally likely. In fact, certain changes are more likely than others. Substituting or supplementing letters next to one another is more likely that a substitution or supplementation with vert distant letters on the keyboard.

If you were a Martian, or even if you grew up typing on a Dvorak keyboard, you'd never figure out why some typos are so common and others are rarer. You might start thinking that someone was hiding a code in those common errors. When you learned more about the QWERTY keyboard, you'd realize that a uniform probability distribution is wrong.

Similarly, a person typing randomly on a keyboard (a game I liked to play when I was young) will occasionally produce text that is readable, if not coherent. I was typing these things before I could read them, so the game was that I'd sit on my Dad's lap at his typewriter, pound out some lines, and he'd read them to me. I was very proud when I hit a string of words. The text I would type was obviously going to be concentrated in some areas of the keyboard.

Again, without knowing what the keyboard looked like, those strings of readable text might have seemed like they meant something, but it turns out they didn't. There was no intelligent design there, just random chance.