+What are the reasons a male-to-female transition might seem like a good idea to someone? _Why_ would a male be interested in undergoing medical interventions to resemble a female and live socially as a woman? I see three prominent reasons, depicted as the parents of the "transition" node in a graph.
+
+First and most obviously, femininity: if you happen to be a male with unusually female-typical psychological traits, you might fit into the social world better as a woman rather than as an anomalously effeminate man.
+
+Second—second is hard to quickly explain if you're not already familiar with the phenomenon, but basically, autogynephilia is very obviously a real thing; [I wrote about my experiences with it in a previous post](/2021/May/sexual-dimorphism-in-the-sequences-in-relation-to-my-gender-problems/). Crucially, autogynephilic identification with the _idea_ of being female, is distinct from naturally feminine behavior, of which other people [know it when they see it](/2022/May/gaydar-jamming/).
+
+Third—various cultural factors. You can't be trans if your culture doesn't have a concept of "being trans", and the concepts [and incentives](/2017/Dec/lesser-known-demand-curves/) that your culture offers, make a difference as to how you turn out. Many people who think of themselves as trans women in today's culture, could very well be "the same" as people who thought of themselves as drag queens or occasional cross-dressers 10 or 20 or 30 years ago. (Either "the same" in terms of underlying dispositions, or, in many cases, just literally the same people.)
+
+If there are multiple non-mutually-exclusive reasons why transitioning might seem like a good idea to someone, then the decision of whether to transition could take the form of a liability–threshold model: males transition if the _sum_ of their levels of femininity, autogynephilia, and culture-related-trans-disposition exceed some threshold (given some sensible scheme for quantifying and adding (!) these traits).
+
+You might ask: okay, but then where do the two types come from? This graph is just illustrating (conjectured) cause-and-effect relationships, but if we were actually to flesh it out as a complete Bayesian network, there would be additional data that quantitatively specifies what (probability distribution over) values each node takes conditional on the values of its parents. When I claim that Blanchard–Bailey–Lawrence's two-type taxonomy is a useful approximation for this causal model, I'm claiming that the distribution represented by this Bayesian network (if we had the complete network) could also be approximated a two-cluster model: _most_ people high in the "femininity" factor will be low in the "autogynephilia" factor and _vice versa_, such that you can buy decent predictive accuracy by casually speaking as if there were two discrete "types".
+
+Why? It has to do with the parents of femininity and autogynephilia in the graph. Suppose that gay men are more feminine than straight men, and autogynephilia is the result of being straight plus having an "erotic target location error", in which men who are attracted to something (in this case, women), are also attracted to the idea of _being_ that thing.
+
+Then the value of the sexual-orientation node is pushing the values of its children in _opposite_ directions: gay males are more feminine and less autogynephilic, and straight males are less feminine and more autogynephilic, leading to two broadly different etiological trajectories by which transition might seem like a good idea to someone, even while it's not that the two types have nothing in common. For example, this model predicts that among autogynephilic males, those who transition are going to be selected for higher levels of femininity compared to those who don't transition—and in that aspect, their stories are going to have _something_ in common with their androphilic sisters, even if the latter are broadly _more_ feminine.
+
+(Of course, it's also the case that the component factors in a liability-threshold model would negatively correlate among the population past a threshold, due to the effect of conditioning on a collider, as in the famous Berkson's paradox. But I'm claiming that the degree of bimodality induced by the effects of sexual orientation is substantially greater than that accounted for by the conditioning-on-a-collider effect.)
+
+An advantage of this kind of _probabilistic_ model is that it gives us a _causal_ account of the broad trends we see, while also not being too "brittle" in the face of a complex world. The threshold graphical model explains why the two-type taxonomy looks so compelling as a first approximation, without immediately collapsing the moment we meet a relatively unusual individual who doesn't seem to quite fit the strictest interpretation of the classical two-type taxonomy. For example, when we meet a trans woman who's not very feminine _and_ has no history of autogynephilia, we can predict that in her case, there were probably unusually intense cultural factors (_e.g._, internalized misandry) making transition seem like a salient option (and therefore that her analogue in previous generations wouldn't have been transsexual), instead of predicting that she doesn't exist. (It's possible that what Blanchard–Bailey–Lawrence conceived of as a androphilic _vs._ autogynephilic taxonomy, may be better thought of as an androphilic _vs._ not-otherwise-specified taxonomy, if it's not easy to disambiguate autogynephilia from all other possible reasons for not-overtly-feminine males to show up at the gender clinic.)
+
+Care must be taken to avoid abusing the probabilistic nature of the model to make excuses to avoid falsification. The theory that can explain everything _with equal probability_, explains nothing: if you find yourself saying, "Oh, this case is an exception" too _often_, you do need to revise your theory. But a "small" number of "exceptions" can actually be fine: a theory that says a coin is biased to come up Heads 80% of the time, isn't falsified by a single Tails (and is in fact _confirmed_ if that Tails happens 20% of the time).