The Two-Type Taxonomy Is a Useful Approximation for a More Detailed Causal Model

A lot of people tend to balk when first hearing about the two-type taxonomy of male-to-female transsexualism. What, one scoffs, you're saying all trans women are exactly one of these two things? It seems at once both too simple and too specific.

In some ways, it's a fair complaint! Psychology is 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" theory. In order to compress our observations of the world we see, we end up distilling our observations into categories, clusters, diagnoses, taxons: no one matches any particular clinical-profile stereotype exactly, but the world makes more sense when you have language for theoretical abstractions like "comas" or "depression" or "borderline personality disorder"—or "autogynephilia".

Concepts and theories are good to the extent that they can "pay for" their complexity by making more accurate predictions. How much more complexity is worth how much more accuracy? Arguably, it depends! General relativity has superseded Newtonian classical mechanics as the ultimate theory of how gravity works, but if you're not dealing with velocities approaching the speed of light, Newton still makes very good predictions: it's pretty reasonable to still talk about Newtonian gravitation being "true" if it makes the math easier on you, and the more complicated math doesn't give appreciably different answers to the problems you're interested in.

Moreover, if relativity hasn't been invented yet, it makes sense to stick with Newtonian gravity as the best theory you have so far, even if there are a few anomalies like the precession of Mercury that it struggles to explain.

The same general principles of reasoning apply to psychological theories, even though psychology is a much more difficult subject matter and our available theories are correspondingly much poorer and vaguer. There's no way to make precise quantitative predictions about a human's behavior the way we can about the movements of the planets, but we still know some things about humans, which get expressed as high-level generalities that nevertheless admit many exceptions: if you don't have the complicated true theory that would account for everything, then simple theories plus noise are better than pretending not to have a theory. As you learn more, you can try to pin down a more complicated theory that explains some of the anomalies that looked like "noise" to the simpler theory.

What does this look like for psychological theories? In the crudest form, when we notice a pattern of traits that tend to go together, we give it a name. Sometimes people go through cycles of elevated arousal and hyperactivity, punctuated by pits of depression. After seeing the same distinctive patterns in many such cases, doctors decided to reify it as a diagnosis, "bipolar disorder".

If we notice further patterns within the group of cases that make up a category, we can spit it up into sub-categories: for example, a diagnosis of bipolar I requires a full-blown manic episode, but hypomania and a major depressive episode qualify one for bipolar II.

Is the two-type typology of bipolar disorder a good theory? Are bipolar I and bipolar II "really" different conditions, or slightly different presentations of "the same" condition, part of a "bipolar spectrum" along with cyclothymia? In our current state of knowledge, this is debatable, but if our understanding of the etiology of bipolar disorder were to advance, and we were to find evidence that that bipolar I has a different underlying causal structure from bipolar II with decision-relevant consequences (like responding to different treatments), that would support a policy of thinking and talking about them as mostly separate things—even while they have enough in common to call them both kinds of "bipolar". The simple high-level category ("bipolar disorder") is a useful approximation in the absence of knowing the sub-category (bipolar I vs. II), and the subcategory is a useful approximation in the absence of knowing the patient's detailed case history.

With a sufficiently detailed causal story, you could even dispense with the high-level categories altogether and directly talk about the consequences of different neurotransmitter counts or whatever—but lacking that supreme precise knowledge, it's useful to sum over the details into high-level categories, and meaningful to debate whether a one-type or two-type taxonomy is a better statistical fit to the underlying reality whose full details we don't know.

In the case of male-to-female transsexualism, we notice a pattern where androphilic and non-androphilic trans women seem to be different from each other—not just in their sexuality, but also in their age of dysphoria onset, interests, and personality.

This claim is most famously associated with the work of Ray Blanchard, J. Michael Bailey, and Anne Lawrence, who argue that there are two discrete types of male-to-female transsexualism: an autogynephilic type (basically, men who love women and want to become what they love), and an androphilic/homosexual type (basically, the extreme right tail of feminine gay men).

But many authors have noticed the same bimodal clustering of traits under various names, while disagreeing about the underlying causality. Veale, Clarke, and Lomax attribute the differences to whether defense mechanisms are used to suppress a gender-variant identity. Anne Vitale identifies distinct groups (Group One and Group Three, in her terminology), but hypothesizes that the difference is due to degree of prenatal androgenization. Julia Serano concedes that "the correlations that Blanchard and other researchers prior to him described generally hold true", but denies their causal or taxonometric significance.

Is a two type typology of male-to-female transsexualism a good theory? Is it "really" two different conditions (following Blanchard et al.), or slightly different presentations of "the same" condition (following e.g. Veale et al.)?

When the question is posed that way—if I have to choose between a one-type and a two-type theory—then I think the two-type theory is superior. But I also think we can do better and say more about the underlying causal structure that the simple two-types story is approximating, and hopefully explain anomalous cases that look like "noise" to the simple theory.

In the language of causal graphs (where the arrows point from cause to effect), here's what I think is going on:

Let me explain.

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. 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.

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 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 trans women 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 the case 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 feminine androphiles form a distinct cluster, but 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).

At this point, you might ask: okay, but why do I believe this? Anyone can name some variables and sketch a directed graph between them. Why should you believe this particular graph is true?

Ultimately, the reader cannot abdicate responsibility to think it through and decide for herself ... but it seems to me that all six arrows in the graph are things that we separately have a pretty large weight of evidence for, either in published scientific studies, or just informally looking at the world.

The femininity→transition arrow is obvious. The sexual orientation→femininity arrow (representing the fact that gay men are more feminine than straight men), besides being stereotypical folk knowledge, has also been extensively documented, for example by Lippa and by Bailey and Zucker. Evidence for the "v-structure" between sexual orientation, erotic target location erroneousness, and autogynephilia has been documented by Anne Lawrence: furries and amputee-wannabes who want to emulate the objects of their attraction, "look like" "the same thing" as autogynephiles, but pointed at a less conventional erotic target than women. The autogynephilia–transition concordance has been documented by many authors, and I claim the direction of causality is obvious. (If you want to argue that it goes the other way—that some underlying "gender identity" causes both autogynephilia and, separately, the desire to transition, then why does it usually not work that way for androphiles?) The cultural-factors→transition arrow is obvious if you haven't been living under a rock for the last decade.

This has been a qualitative summary of my current thinking. I'm very bullish on thinking in graphical models rather than discrete taxons being the way to go, but it would be a lot more work to pin down all these claims more rigorously—or, to the extent that my graph is wrong, to figure out the correct (or, a more correct, less wrong) graph instead.

(Thanks to the immortal Tailcalled for discussion.)