Got into a conversation yesterday about autism, specifically what the actual clinical definition is. Apparently, there's a debate raging over whether "mild autism" exists. One group maintains that people who show some autistic symptoms, but not to the degree associated with Oliver Sachs level cases ( i.e. people who exhibit some of the difficulty in dealing with other people and some of the characteristic physical tics of the autistic) should be classified as "mildly autistic." The other group holds this is an incorrect claim. They claim that since the "mildly autistic can be cured through therapy (and sometimes, medication), their condition doesn't qualify as autism, since autism is incurable.

I think the only useful reply to someone making that sort of claim is a good solid pimp-slapping. Not because they're necessarily wrong -- I'm no expert in the field, and I'm certainly willing to believe that the differences between autism and mild autism outweigh the similarities -- but because that kind of argumentation is so entirely unproductive. We concern ourselves with X. We define X in a certain way. We therefore ignore thing Y which does not match our definition for X. Which is all well and good, except that our exclusion of Y, unless we have done some extra work in the interim, says nothing about X or Y in and of themselves, only about our definition of X. That is, we have said nothing new about Y by drawing a line that excludes it, and it's better not to go around acting as though we'd learned something. If your definition of autism stipulates incurability, then you can't make interesting points about curability -- you're only allowed to leverage this definition in order to speak about other attributes. It's an issue of circularity, of having enough equations for the number of unknowns you're working with.

I first felt the need to go off on this rant while reading an account of the endless back-and-forth in the world of Austen critics on the whole lesbianism issue, only here I think both sides are being equally fatuous. One camp wants to label Austen a lesbian, on the strength of her strong bonds to her female friends, the discourse of "love" and "affection" among her women, notes how strongly certain traits fit with our contemporary understanding of certain lesbian personae. The other camp, vehemently "defending" Austen, points to the general conditions of English culture and claims that such things are typical, that Austen is in no way a standout, that the modern understanding of "lesbian" is only applicable in a modern context. They're both correct, so far as they go, but in order to score a few political points, they each wind up making basically the same interpretive error. The story here is not Austen herself; the interesting questions all concern the nature of the society she lived in. It's horribly incorrect to label her a proto-feminist lesbian writer and then hold her up as a shining example of courage and ahead-of-her-time attitudes, but it's no less incorrect to think of her as straight (in modern terms) either. Having set up a system of definitions for dealing with Georgian English sexuality, you can't then assume that a term taken from that system -- "lesbian" or "straight" -- carries over into modern usage. You could, quite legitimately, supply a set of definitions that cross cultural and temporal boundaries, a set of standards useful for looking at different social contexts. But such a set is useless for scoring the kind of cheap political points the opposing camps of Austenites are looking for -- by bringing past and present within a consistent set of definitions, you rule out the ability to shiftily transfer into a normative realm that sits outside your definitions -- which is why they don't bother with the exercise.

Also matching this pattern are such classics as the ontological argument, synthetic ethics, and most replies to Turing's test, and classifying ketchup as a vegetable. To the extent that there is a question, one's definitions may clarify or obscure it, but they will not ever actually answer it.