As a matter of ontology, on some topics it does. I can appreciate the position that sex, even social sex, is not one of those things (though I disagree with it), but denying the existence of intersubjective reality is not tenable.
Take language. We here are using words that have a roughly shared meaning. You can do empirical tests about what a word means, e.g. by telling a bunch of test subjects that the box on the right has some object they value and the box on the left has nothing and asking them to pick whichever box they want. If they consistently pick the box on the right, we’ve established an objective truth about the meaning of the words “the box on the right”. But that meaning is just a matter of agreement between people. It’s true, objectively true, solely by virtue of the fact that people agree about it.
Other intersubjective realities are political facts like laws and national borders, religious facts like “the Pope is the head of the Catholic church” and “the New Testament is the holy book of Christianity”, and economic facts like “a bitcoin costs more than $6000”. These claims are true, predictions we make about them will reveal that they have an objective reality, but nonetheless that reality is entirely dependent on what people believe. When people stopped believing that “bitcoin is worth more than $12000”, bitcoin stopped being worth more than $12000.
I’d argue that the social aspects of sex are like this. Someone is a woman in social situations if we all agree that they are a woman in social situations. If everyone in a room were independently asked to divide the room into men and women, and everyone put the transwoman in the women column, then she is a woman. That does break some implications from the statement “X is a woman”, e.g. that it entails “X has XX chromosomes”, but those claims are conceptually distinct and there’s no necessary implication. It isn’t the case that only those people who have XX chromosomes are considered to be women in social situations.
Where you make inferences from calling someone a “man”, you will often make more accurate inferences when you call trans men “men” than when you call them “women”, particularly when you weight inferences by relevance. That means calling transmen “men” is more effective than calling them “women”.
But you’re also admitting that you have no knowledge from which to derive your certainty.
I’m not sure of the ratio, but there’s clear evidence of cultures that have taken either approach (sometimes simultaneously). This again undermines any claim to necessary implication being a historical universal.
I appreciate this. But you’re simultaneously saying that despite all the ways in which someone born with a penis can have other characteristically female traits, the only characteristic we should consider is genital shape. That’s a weird thing to do in situations where other sex characteristics are much more salient and relevant. Why should generally inaccessible information be preferred over accessible information in contexts where the accessible information is more relevant anyway? (and to return your tone of good faith, I acknowledge that genital shape is more relevant in the context of gendered bathrooms, but I still don’t find it compelling)
A bit aside, but: loan words are, for the most part, not brought in by linguists but by immigration and second languages. “Shenanigans” came out of the north east not because of Harvard linguists, but because of a dense population of 1st and 2nd generation Irish Americans who grew up hearing their parents and grand parents using a word from their ancestral language with no adequate equivalent in English. No one said, “let’s make ‘Shenanigans’ a thing”, they just used a word that was already in their vocabulary, and continued using it when they adopted a second language.
In the domain people have been trying for a generation at least to introduce e.g. a gender neutral pronoun into English (ey, xe, ze, etc.). But instead, despite few people explicitly advocating for it and plenty of misguided pedants resisting it, “they” is becoming the accepted third person singular generic pronoun.
New words are coined for new concepts, or for concepts that don’t have good words yet, but more often (and particularly where we’re borrowing from another language, it’s a lot more likely to be organic use by polyglots and cultural transplants.
I am pretty sure that almost all transsexuals do in fact do this.
I think you’re overestimating this set of things. Look at the wiki article for causes of transsexuality, particularly the sections on brain structure and brain function. The neurological and psychological parts of transsexuals are in many ways closer to the sex they choose than the sex as determined by their genitals or chromosomes.
Let me use an analogy to show you the mistake I think you’re making. If I present myself as a Muslim, that does not in itself make me a Muslim. It’s possible to lie about being a Muslim, in the same way that you would be lying if you intended someone to believe your string of adjectives. Nonetheless, if I sincerely believe myself to be a Muslim, that is sufficient to make me a Muslim.
You’re offering as a reductio something that transsexuals aren’t doing and no one is defending here, so it doesn’t work as a reductio.
Who’s talking about assault? I thought we were just trading tales of unjustified fears.