Your evidence was your own sense (not a paper, not even a survey) that people think differently when writing in a different programming language.

I refer not to my own sense so much as what is more or less universally acknowledged by influential thinkers in that field. That doesn't preclude the culture being wrong, but I do put Paul Graham on approximately the same level as Pinker, for example.

While Pinker is an extremely good populariser and writes some engaging accounts that are based off real science, I've actually been bitten by taking his word on faith too much before. He has a tendency to present things as established fact when they are far from universally agreed upon in the field and may not even be the majority position. The example that I'm thinking of primarily is what he writes about fear instincts, regarding to what extent fear of snakes (for example) is learned vs instinctive. His presentation of what has been determined by primate studies is, shall we say, one sided at best.