There's a thing I've sometimes noticed happening in social science debates. In the spirit of How to Write Quickly While Maintaining Epistemic Rigor, let's consider an example.
A note: I'm going to pick a controversial example, of the causes of transsexuality. This is because for the past few years, I have been studying transsexuality to try to understand its causes better, and so it's an example I have thought deeply about and know a lot about. And because I've been hit by the problem described in this post.
But understandably, it's also something that a lot of people, particularly trans people and people involved in pro- or anti-trans politics, care a lot about. Given the centrality of the example, I think its reasonable to debate object-level matters of the example in the comments, but everyone on all sides should be careful about mindkilling.
Blanchard's autogynephilia theory
Why do trans women transition? As in, what is it that causes them to be transgender, rather than to live happily as men? While there is disagreement about this question, insofar as the trans community gives you an answer, it tends to be something like:
Your brain expects to have a certain kind of body, either male or female, and it expects to live in a specific gender role, either as a man or as a woman. Some people are born with brains of the opposite sex. This causes distress, and can be fixed by transitioning to live as the sex that matches their brain. And thus you get trans people.
This narrative has a great degree of flaws and inconsistencies, which I won't belabor here (though if you can't see any, feel free to ask in the comments, and I will reply). It's probably not surprising that the narrative is so flawed; it's not selected for truth, but instead for political convenience, because it allows an easy way to describe what it is like to be trans, and because the policies people tend to derive from it favor trans people.
But given the inaccuracy, we might wish for another, more accurate narrative. There are a million other idiosyncratic theories than this one, but there's probably one narrative that stands above all of the others in terms of scientific support, quality of derivation, and so on. It has many names, but I will call it the Blanchardian model, named after Ray Blanchard, a major figure in its development:
There are two different kinds of trans women, homosexual (exclusively attracted to men) and autogynephilic (have a sexual interest in being women). The homosexual ones are the ones who are closest to the classical transgender narrative, in that they tend to be very feminine, and this is likely a factor in why they transition. The other kind, autogynephilic transsexuals, are the most common kind. They have a sexual interest in being women, which may present in various ways, such as sexual fantasies about having a female body, transvestic fetishism, etc.. They are attracted to women, because autogynephilia is a sort of "inversion" or "miswiring" of their pre-existing attraction to women.
Autogynephilia itself can be quite broad. I think the instance people are most familiar with is transvestic fetishism, but according to a survey I did, the most common forms might involve sex as a woman (with either a man or a woman), or masturbation as a woman. A classical example of autogynephilia in internet porn is transgender transformations, erotic material depicting men who turn into women, but surprisingly I didn't hear much of that in my survey. (It should be noted that most autogynephiles don't transition, but instead live as non-transgender men.)
There's a huge ton of details, qualifications, etc., that can be added to this model; at times the model makes claims that seem unreasonable, but once you start learning the evidence for the claims, some of the unreasonable-looking claims start seeming more like hidden insights than like reasons to disbelieve. For brevity, though, I won’t go into the details in this post.
Critics are really bad
At this point, I've spent years studying the causes of transsexuality. At first, I was critical of the Blanchardian theory for various reasons. But I gradually learned things that made it seem more and more plausible. And one thing that struck me just as much as the raw evidence that was presented, was the quality of the arguments presented. Often, the researchers critical of Blanchardianism had plainly obvious rationalizations, misrepresented theory and evidence, and made extremely questionable inferences.
To give a fairly typical example, there is the Identity-Defense Model of Gender Variant Development, which claims that autogynephilia develops as a result of taboos around males who want to be female. This being in an attempt to downplay the correlation between autogynephilia and transsexuality by proposing that it is due to reverse causation.
But it relies strongly on the notion that taboo things become erotic, and while this is a popular idea, I've yet to see any convincing evidence for it, nor have the various people I've seen advocate for the theory been able to provide much. (Feel free to debate this in the comments.) It doesn't seem to make any sort of evolutionary sense, and it doesn't seem to me to check out in terms of commonsense correlations (the most popular erotic acts seem to be the least taboo, while highly taboo erotic acts like incest seem highly unpopular - and this doesn't seem to me to be reverse causality).
To give another example of bad critics, here is a quote from probably the most popular critique of autogynephilia, Julia Serano's Case Against Autogynephilia:
To put this sexualization in perspective, consider the following analogy: Many natal women have sexual fantasies about being raped (reviewed in Leitenberg & Henning, 1995). It is one thing to respectfully attempt to explore and understand such rape fantasies. It would be an entirely different thing to insist that there are two subtypes of women—those who have rape fantasies and those who do not; to use the label “autoraptophiles” when describing women who have such fantasies and to insist that they are primarily motivated by their desire to be raped; to include “autoraptophilia” as a modifier in the Diagnostic and Statistical Manual of Mental Disorders; and to encourage the lay public to actively distinguish between those women who are “autoraptophiles” and those who are not. Such actions would undoubtedly have a severe, negative impact on women (who are already routinely sexualized and marginalized in our culture).
While one can definitely discuss the political controversy that the concept of autogynephilia stirs up, Julia Serano's analogy doesn't seem to hold at all. The damage done by the "autoraptophilia" concept would presumably be in suggesting that women want to be raped, and so contributing to the likelihood of women getting raped. Which is bad because women very much want to avoid getting raped. Meanwhile, the analogous implication for trans women would be suggesting that they want to... be female? But that is hardly harmful for trans women, nor is it a unique implication of Blanchardianism.
If anything, Serano's own conception of autogynephilia, as indicative of a desire to be a woman, rather than being one component that adds up with other component to influence preferences, seems on a deeper mechanistic level closer to the theory that "autoraptophilic" women want to be raped.
Blanchardians are bad
So, due to a combination of empirical evidence for Blanchardianism (which I have not reviewed in this post), and the bad critics (mentioned previously), I ended up buying into the theory and using it as a basis for my research. After all, if the evidence supports it, and nobody can critique it well, surely it must be right. And I ended up becoming pretty notorious for it, being basically the resident Blanchardian in various rationalist-adjacent groups for years.
However, while there were quite a few insights I accumulated from Blanchardianism, there were also quite a few places where I found it to be stretching the truth more and more. One major case of this is the concept of erotic target location errors:
The way Blanchardians make sense of autogynephilia is as an inverted attraction to women. That is, under ordinary circumstances, males find women to be sexually attractive as partners. However, in a subset of males, ?some glitch? happens, and this makes them find it erotic to be women, where the things they would otherwise find erotic when done to women, become things they would find erotic to be done to themselves as women.
One major prediction of this model is that autogynephiles can't be gay. And for what seems to be complicated historical reasons that don't make sense, Blanchardians also insist that autogynephiles can't be bisexual. Instead, apparent homosexuality among autogynephiles is explained away with various theories, a major one being what is known as "pseudo-bisexuality/pseudo-androphilia" or "meta-attraction".
Meta-attraction is a phenomenon where autogynephiles have a secondary sexual interest in men, not due to attraction to men in the classical sense, but instead as a side-effect of their autogynephilia, limited to only exist with themselves as women. For instance in surveys I've seen a lot of straight men who fantasize about being women and having sex with men. The consensus theory from Blanchardians is that things like this explains all autogynephiles' sexual interest in men, and that none are attracted to men in the classical sense.
I bought into this and advocated for it for a while, partly because Blanchardianism had earned my trust, and partly because there seemed to be some evidence for it. However, my belief became more and more strained over time, as I gradually accumulated evidence that contradicted its predictions. This eventually culminated in a blog post, Meta-attraction cannot account for all autogynephiles' interest in men, arguing that the consensus was contradicted by the balance of evidence so far.
This led to... not much. While there was acknowledgement that my post presented nontrivial evidence (email response by Michael Bailey), it has still been continued to be taken as an axiom that autogynephiles cannot be attracted to men, without any good reason.
The poor response to this lead me to reevaluating my trust in the Blanchardian researchers, and reexamining a variety of issues I had bought into. It turns out that there's actually a lot of really questionable ideas that are just taken for granted by them! Sometimes they reach the level where I'd call them dishonest! This has led me to accidentally gradually build up a series of posts that call "my debunking of Blanchardianism" (in a "ha ha only serious" way).
Here's what I see as the dynamics in this and other situations:
You've got a group of politically motivated people, who advocate strongly for one narrative. This narrative is really bad, because it's not optimized for truth, it's optimized for political convenience.
But as a result, people find the original narrative unworkable, so they come up with their own counternarrative, which is somewhat more reality-based, but really not all that better; it ends up overly rigid and narrow, because it's main function is to cover the cases that don't fit well under the original narrative.
The people who are politically biased criticize the counternarrative. However, because they only really care about the political aspects of it, they don't focus on the parts where the counternarrative is weakest, and often don't even fully understand the counternarrative; instead they just focus on the aspects they find most politically objectionable.
The end result is that the counternarrative looks like it is well-supported by evidence, and has stood up to any critique you could throw at it. But really it's full of myths and it starts to break once you push it a bit on an axis orthogonal to the political controversy.
Often, the counternarrative may still have a nugget of truth, or a good point about some things. I would still absolutely apply ideas from Blanchardianism to analyzing transsexuality. But this sort of thing can demote the counternarrative from being the central part of the story to just being an obnoxiously polarizing piece of the puzzle.