DSM - Dumb Stupid Manual

From a Crooked Timber post,

I appreciate that "the population with a diagnosis of autism" is diverse, but I can't see that the present article gets us much further, accepting as it does that there is a single condition that is autism. 
Given, as we appear to agree, that there is little necessarily in common between the people who have the diagnosis, the most probable explanations would seem to be that the name covers several conditions with overlapping symptoms or that the definition is terminally vague and just consists of the constituent elements of human behavior prefaced by the words "too much" or "too little".
It's certainly true that in the US at least the DSM-IV definition makes diversity inevitable, involving as it does a Chinese menu approach to diagnosis - "(I) A total of six (or more) items from (A), (B), and (C), with at least two from (A), and one each from (B) and (C)" - which leads by perms and combs to thousands of different presentations. This diagnostic approach would tend to lead to the conditions-with-overlapping-symptoms problem; the vaguer DSM-V approach approaches the second "too much of things"  problem.
At the very least, talk about 'the autisms'; ideally, drop the term altogether and talk in terms of actual describable problems of communication, behavior, and the like. 'Autism' is incurable; problems with communication suggest trials of AAC.

Incidentally, do American Chinese restaurants still have menus like that?  I've never seen one in my visits.