I am not interested in ideas that cannot in principle be proven or disproven. I am as capable as the next guy in believing in an idea that is not yet proven so long as it could in principle be proven or disproven.
In my chosen field of autism, I believe that the cause will turn out to be assortative mating of two hyper-systemizers. I believe this because we already have 3 pieces of the jig-saw: (1) that fathers of children with autism are more likely to work in the field of engineering (compared to fathers of children without autism); (2) that grandfathers of children with autism—on both sides of the family—were also more likely to work in the field of engineering (compared to grandfathers of children without autism); and (3) that both mothers and fathers of children with autism are super-fast at the embedded figures test, a task requiring analysis of patterns and rules. (Note that engineering is a chosen example because it involves strong systemizing. But other related scientific and technical fields [such as math or physics] would have been equally good examples to study).
We have had these three pieces of the jigsaw since 1997, published in the scientific literature. They do not yet prove the assortative mating theory. They simply point to it being highly likely. Direct tests of the theory are still needed. I will be the first to give up this idea if it is proven wrong, since I'm not in the business of holding onto wrong ideas. But I won't give up the idea simply because it will be unpopular to certain groups (such as those who want to believe that the cause of autism is purely environmental). I will hold onto the idea until it has been properly tested. Popperian science is about being able to let go of an idea when the evidence goes against it, but it is also about being able to hold onto an idea until the evidence has been collected, if you have enough reasons to believe it might be true.
The causes of autism are likely to be complex, including at the very least multiple genes interacting with environmental factors, but the assortative mating theory may describe some contributing factors.