Wednesday, February 02, 2005

Failed design inferences

People are bad at discerning design. From Carl Sagan's Cosmos, p. 66:
Most curious is [Kepler's] view of the origin of the lunar craters, which make the moon, he says, "not dissimilar to the face of a boy disfigured by smallpox." He argued correctly that the craters are depressions rather than mounds. From his own observations he noted the ramparts surrounding many craters and the existence of central peaks. But he thought that their regular circular shape implied such a degree of order that only intelligent life could explain them. He did not realize that great rocks falling out of the sky would produce a local explosion, perfectly symmetric in all directions, that would carve out a circular cavity—the origin of the bulk of the craters on the moon and the other terrestrial planets. He deduced instead "the existence of some race rationally capable of constructing those hollows on the surface of the moon. This race must have many individuals, so that one group puts one hollow to use while another group constructs another hollow."
Three things jump out at me from this passage. First, people are bad at discerning design. The heuristic of complexity often turns out to be complexity relative to our experience. Our experience is limited, though, and so we find ourselves claiming that craters are designed by a race of intelligent moon-men. Second, this sort of design inference is an argument from ignorance. In this case, Kepler was unaware of meteorites, and so he imagined that little green men were responsible. Third, the current proponents of ignorance (the Discovery Institute and the rest of the ID crowd) will undoubtedly find themselves left behind by the march of science, just as Kepler's ideas about intelligently-designed craters have been.

Interestingly, Kepler's inference to the presence of a large number of limited designers is perhaps more justified than the DI's inference to a single omnipotent transcendent designer. At least we know that limited designers exist; we cannot fall back on any such knowledge to assist the DI's case.

UPDATE: followup post here.