The problem with all arguments about whether or not autism should be ‘cured’ is that the definition of ‘cure’ is extraordinarily poorly specified.

One definition is the cure searched for by those looking for a biomedical treatment for autism: a procedure which, if performed, causes an autistic person to become nonautistic. One could have a legitimate debate about biomedical treatments– the very real suffering undergone by many autistic people, compared to the feeling of many autistic people that eliminating their autism would be killing them and replacing them with a neurotypical with the same face– if it were not for the fact that arguing about the ethics of biomedical treatments for autism is like arguing about the ethics of unicorns farting rainbows. There is no single treatment that cures autism, and we’re not going to find one, and the currently available alleged ‘cures’ are unethical for lifting the hopes and emptying the wallets of desperate parents.

Another definition of ‘cure’ is developing a medication treatment for some symptoms of autism: for instance, going nonverbal, severe sensory sensitivities, epilepsy, the inability to eat more than two foods causing one to have nutritional deficiencies, etc. I am pretty sure that not even the Autistic Self-Advocacy Network objects to coming up with better treatments for epilepsy in autistic people.

The only point of contention here is what exactly we’re treating. Many autistic people would object to treatments for special interests, stimming, autistic ways of relating, or other odd but morally neutral traits. But I suspect the vast majority of treatments people would actually investigate would be completely uncontroversial for everyone.

A third definition of ‘cure’ is better referred to as ‘prevention’: the research into finding a way to prevent the existence of autistic people, most likely through abortion of children with autistic genes. The effectiveness of prevention depends on the etiology of autism. If autism is highly polygenic and environmentally influenced, the way that intelligence is, eliminating autism will be as difficult as only having high-IQ children– which is to say, very hard. This form of autism cure can be safely classified into the unicorns-farting-rainbows category.

On the other hand, it might turn out that autism is actually a bunch of different things. This isn’t actually that unlikely: some conditions, like fragile X and Rett syndrome, have symptoms that are extraordinarily similar to autism; people with Rett syndrome probably would be classified as just autistic if we didn’t know what caused Rett syndrome. The executive function and social responsiveness parts of the brain are quite easy to break; autism may very well be a catchall category for “your executive function and social responsiveness are borked and we don’t know why.”

In a very lucky universe, it would turn out that autism that tends to cause a high level of suffering would be a different condition from autism that tends to not cause a high level of suffering, and thus some autistic people would wind up aborted and Ozys would not.

However, I’m not sure that we’re in that universe. Consider two identical autistic people; the only difference is that one of them is an excellent programmer, and one is not. The excellent programmer gets a job at Google, where no one expects him to talk about anything other than programming. He mostly socializes among autistic and autistic-friendly people, making his difficulty at socializing with neurotypicals irrelevant. Although he can’t cook, drive, or clean, Grubhub, Uber and a maid service get those things covered; his very tolerant girlfriend takes care of the rest. He can afford to live in a nice house in a quiet area; his assistive tech– including a weighted blanket and a top-of-the-line set of noise-cancelling headphones– allows him to live a life without undue sensory stimulation.

The one who can’t program lives with his parents, because he is incapable of living independently. He is tremendously lonely, since no one particularly wants to talk to him, especially since he’s not great at maintaining conversations on any subject other than his special interest, tea flavors. His parents’ house is full of noise, from the cars driving by to the music the neighbors play on weekends; the stress causes him to go nonverbal and, eventually, to self-injure or to attack his parents. Desperate, his parents put him on anti-psychotics, which make him even more miserable and withdrawn, although at least less prone to attacking people. His parents worry about what will happen to him when they die, because he cannot take care of himself; he may wind up in an institution.

‘Having a special interest in programming’ is not exactly the sort of thing that’s genetically coded for. I suspect what will happen is that some forms of autism, like Rett syndrome, will be generally terrible, and others will have a mixed portfolio– some people will be miserable, while others will be better off autistic than they would be nonautistic. The treatment of the latter group is extraordinarily controversial.