The Mapping Fallacy

Charles Murtaugh reports, and links to other reports, that the similarity score between human and chimpanzee DNA is down from 98.5% to 95% or so.

Robert Musil muses that maybe the chimp’s “junk” DNA overlaps the human’s “working” DNA or vice versa; Paul Orwin sets him straight, convincingly. With all due respect to the real scientists in this discussion, like Murtaugh and Orwin, I think everyone is missing the point.

People care about the similarity of human/ape DNA because they are in the grip of what I will call, since there is no term in logic that I know of to describe it, the mapping fallacy. The mapping fallacy is the belief that similar inputs to a function generate similar outputs. That depends on the function. For f(x) = x + 1, sure. For f(x) = x³, a qualified yes, as long as you hold x down to a reasonable size. For certain complex functions, useful in cryptography, not at all: they generate a completely random-looking output for any change in the input, no matter how small.

To return to biology: our input, x, is the genetic sequence; our output, f(x), is the resulting life-form. We know x produces f(x), but we have no idea how. Maybe two radically different genetic sequences produce the same life-forms. Maybe two scarcely distinguishable sequences produce completely different ones. The point is, we don’t know, because we don’t know what sort of function, or mapping, we’re dealing with here. And until we do know, no one should much care whether human and gorilla DNA are 95% similar or 98.5% or 99.44% similar.

Mapping fallacy examples:

• Stephen Wise, the animal rights guy du jour, arguing that chimpanzees should have rights because we have 98.3% (apparently a compromise figure) of our DNA in common.
• Jared Diamond, a distinguished biologist who ought to know better, arguing that the concept of race is absurd because there are larger genetic differences within than among races. (I agree with Diamond about race but that doesn’t make his argument any more respectable.)
• The BBC story I linked above, arguing that the percentage dispute is important because “…the relatively small difference between human and chimp genomes would offer insights into the gene differences that might render humans more vulnerable to disease.”

(Update: The Man Without Qualities comments further, complaning mildly that Orkin did not exactly “set him straight,” conceding, on the contrary, that the study under consideration does not distinguish between “useful” and “junk” DNA. Fair enough. Now if he would only link to my post, instead of my email address, we could all go home happy.)