-Psychology is very complicated; every human is their own unique snowflake, but it would be impossible to navigate the world using the "every human is their own unique _maximum-entropy_ snowflake; you can't make _any_ probabilistic inferences about someone's mind based on your experiences with other humans" theory. Even if someone were to _verbally endorse_ something like that—and at age sixteen, I might have—their brain is still going to go on to make predictions inferences about people's minds using _some_ algorithm whose details aren't available to introspection. Much of this predictive machinery is going to be instinct bequeathed by natural selection (because predicting the behavior of conspecifics was very useful in the environment of evolutionary adaptedness), but some of it is the cultural accumulation of people's attempts to organize their experience into categories, clusters, diagnoses, taxons. (The cluster-learning capability is _also_ bequeathed by natural selection, of course, but it's worth distinguishing more "learned" from more "innate" content.)
+Psychology is very complicated; every human is their own unique snowflake, but it would be impossible to navigate the world using the "every human is their own unique _maximum-entropy_ snowflake; you can't make _any_ probabilistic inferences about someone's mind based on your experiences with other humans" theory. Even if someone were to _verbally endorse_ something like that—and at age sixteen, I might have—their brain is still going to go on to make predictions about people's behavior using _some_ algorithm whose details aren't available to introspection. Much of this predictive machinery is going to be instinct bequeathed by natural selection (because predicting the behavior of conspecifics was very useful in the environment of evolutionary adaptedness), but some of it is the cultural accumulation of people's attempts to organize their experience into categories, clusters, diagnoses, taxons.[^cluster-learning-capability]
+
+[^cluster-learning-capability]: The cluster-learning capability is _also_ bequeathed by natural selection, of course, but it's worth distinguishing more "learned" from more "innate" content.
+
+There could be situations in psychology where a good theory (not a perfect theory, but as good as our theories about how to engineer bridges) would be described by (say) a 70-node [causal graph](https://www.lesswrong.com/posts/hzuSDMx7pd2uxFc5w/causal-diagrams-and-causal-models), but it turns out that some of [the more "important" variables in the graph happen to anti-correlate with each other](https://surveyanon.wordpress.com/2019/10/27/the-mathematical-consequences-of-a-toy-model-of-gender-transition/), such that stupid humans who don't know how to discover the correct 70-node graph, do manage to pattern-match their way to a two-type typology that actually is better, as a first approximation, than pretending not to have a theory. No one matches any particular clinical-profile stereotype _exactly_, but [the world makes more sense when you have language for theoretical abstractions](https://astralcodexten.substack.com/p/ontology-of-psychiatric-conditions) like ["comas"](https://slatestarcodex.com/2014/08/11/does-the-glasgow-coma-scale-exist-do-comas/) or "depression" or "bipolar disorder"—or "autogynephilia".[^lucky-simplification]
+
+[^lucky-simplification]: In some sense it's a matter of "luck" when the relevant structure in the world happens to simplify so much. For example, [friend of the blog](/tag/tailcalled/) Tailcalled argues that [there's no discrete typology for FtM](https://www.reddit.com/r/Blanchardianism/comments/jp9rmn/there_is_probably_no_ftm_typology/) as there is for the two types of MtF, because the various causes of gender problems in females vary more independently and aren't as stratified by age.