Genders Discrimination

Categories in your head

In physical reality, light with wavelength between 380 and 700 nm passes through your eyeball and activates cells in your retina. In your mind, you perceive colors. This makes it tempting to think that your perception of color is simple a wavelength-detector, but that is not the case.

Sometimes the same exact light waves produce differently-perceived colors. Sometimes you perceive a color whose wavelength is missing. Sometimes you perceive a color that has no associated wavelength at all. And sometimes you eat a mushroom and perceive all the colors at once with your eyes closed. The wavelengths are out there, but the color is only in your mind a is a feature of your mind, not of reality.

This is true of all your perceptions, including things like your emotions, sense of self, and the very reality of objects.

An important feature of perception is categorization — you learn concepts for grouping similar perceptions together. Your categories then affect your immediate perception, downplaying differences within categories and amplifying differences between them. Modern Westerners see the rainbow as made up of 7 distinct colors — up from 3-4 in medieval times and 5 in the 18th century. The rainbow itself, of course, is a continuous spectrum that could be broken up into any number of bands.

A general principle of developing perception is: more discrimination is better. You can better make and enjoy visual art better if you can tell pumpkin orange from safety orange from tiger orange, as opposed to seeing them as the same rainbow stripe. The movie Inside Out taught you that there are only 5 emotions, but you can actually increase the range and granularity of your perception of emotion to understand others and manage your own feelings more skillfully. The more attention you pay to a domain, the more fruitful distinctions in perception you can make.

Queer whales

My friend points out: just because out there are two clusters of biological sex defined by gamete size, doesn’t mean that people’s perception of gender is identical to those. What my friend calls people’s “gendering faculty” is similar to their perception and discrimination of colors, and has a lot to do with the particularities of the society they inhabit — for example whether men in that society grow their hair long or wear kilts.

To continue with the color analogy: it’s basically impossible to make yourself see red as yellow, but it’s not that hard to learn to discriminate and enjoy a hundred subtle shades of orange.

In a famous essay on categories and gender, Scott Alexander explained that it’s not wrong to call a whale a fish if what you care about is where to hunt them and not phylogenetics. Since there’s no ground truth of what a “fish” is, we should call whales fish (or call schizophrenics emperors, or men — women) if it makes life easier for others for social reasons.

People pointed out in response: although the categorization schemata in our brains aren’t the same as categories in reality, we rely on them to track and predict reality. Subverting your categories isn’t costless. It makes inhabiting reality harder. We should figure out how to get the social benefits after accepting reality, not by denying it.

Learning about both the whale’s habitat and physiology shouldn’t make you confused or ambivalent about whether the whale is a fish or mammal. It should increase the number of categories you have to sort animals into: fish (in sea, lay eggs), whale-like (in sea, nurse young), land mammals (on land, nurse young). Adding categories makes you smarter about all animals. Once you learn how whales queer the fish-mammal binary, it doesn’t make sense to keep insisting on that binary.

Let a thousand genders bloom

Today, the phrases “trans men are men” and “I’m a genderqueer femme-presenting bisexual they/them” are associated with the same political tribe and position on gender ideology. This is a result of both strange memetic forces and the necessity of presenting a unified political front on the national level. But philosophically, these two phrases are in contradiction.

As a positive statement, “trans men are men” strongly suggests that there are only two genders and every person is more or less a central example of one or the other — the only question is whether this determination is made on the basis of biology or dress or self-identification. But whichever basis you choose, this binary is contradicted either by intersex people, or drag queens, or detransitioners, or genderqueer bisexual they/thems.

As a normative statement, “trans men are men” can mean “trans men should be treated as men” or “trans men should be perceived as men”. But it’s always at least somewhat the latter, both because trans people (like all people) care a lot about how they’re perceived gender-wise and also because, on a deep level, perception and action are the same thing. On a collective level, asking people to change their behavior by forcing them to flip their perception of gender is a dubious proposition.

On the other hand, treating gender as a multivariate socially-mediated cloud of perceptions that people can learn to discriminate myriad categories in has a lot of benefits.

Everyone already does this all the time. We don’t use our gendering faculty to predict someone’s chromosomes but to predict things like their sexual behavior and interests, and to that end little kids and very old people are also to a large extent a gender onto themselves. We make different assumptions about a slim, flamboyantly-dressed, effeminate man and a bushy-bearded pudgy man in overalls. Reifying more of those categories with common names will make everyone less confused about everyone else’s gender expression. It will also give people more levers to shape other’s perception of them. For example, signaling which sex you’re attracted to via a particular style of flamboyant dress you wear, and thus affecting how they’re treated.

There’s a difference of course between my individual perception of a dozen genders and things that require social agreement, like bathrooms. But if society can go from 3 rainbow colors to 7 we can probably add at least a few genders as well. As soon as new categories become common knowledge people will shape themselves to match them, the way most cis people wear socially-gendered clothing today to be perceived as their desired gender.

Again, when we see someone walking into the women’s restroom and decide if it’s socially acceptable or not we go by their appearance and demeanor, we don’t run to measure their gamete size. Restroom policy should follow from a more accurate common understanding of gender, gender categorization shouldn’t follow from the fact that most buildings today have exactly two types of restrooms.

I don’t want to get too far in the weeds of politics and public norms. If some trans people want to fight against having to conform to any genders whether binary or more granular, or if they want to fight directly for restroom privileges based on self-identification regardless of perception, that’s their fight.

My blog is about personal epistemology. Personally, I think I’ve gained a more accurate picture of the world when I started perceiving “rationalist-adjacent female-attired somewhat-passing demisexual trans women who like hanging out with each other and write great Rust code” as their own gender. I make different predictions about what they like, how they’re seen by others, and how good they are at programming that I would about either men or women — more accurate predictions. I don’t know where it puts me politically on the great contemporary trans debate, but I’m in favor of more genders discrimination.

10 thoughts on “Genders Discrimination

  1. I genuinely don’t get the main point you’re trying to make. Could you clarify it?

    Why would it be more useful for anybody to use more than, let’s say, 5 main gender categories: cis man, cis woman, trans man, trans woman, and “other” (with some inner distinctions if somebody’s very curious)? Sexual orientation and gender expression are obviously closely related but separate categories.
    Why would anybody think of the mentioned “rationalist-adjacent female-attired somewhat-passing demisexual trans women who like hanging out with each other and are often obsessed with some technical topic” as a pure gender category? It describes a specific type of person, characterized by a few different categories – gender identity (trans woman), sexual orientation (demisexual), gender expression (female-attired), and interests (rationality, hanging out, technical topics).


    1. My take-away from Jacob is that it can be helpful to see the shades of gradation in things like gender, even if we don’t take them as absolute, in the way we learn to see shades of “red” even though “red” is nothing but a category we made up. More divisions = more gradation = a more fine-grained perspective = the possibility of better predictions. If you just labeled colors “light” or “dark”, you would be worse at e.g. distinguishing that dangerous lion over there from that harmless fox (or whatever).

      I actually think everything, not just gender, is like this. Everything is a spectrum, but categories are still useful.
      (I wrote about it here:


  2. Jacob, do you have any good (anec)data on the feasibility and stability of closed polycules involving 1 cis straight man and 2 cis women (usually bi)?

    Could it be a viable choice for men for whom the emotional or sexual involvement of their female partner(s) with anybody other than them and another cis woman is a dealbreaker? Let’s assume this is not a modifiable/negotiable preference.


  3. I think I’ve gained a more accurate picture of the world when I started perceiving “rationalist-adjacent female-attired somewhat-passing demisexual trans women who like hanging out with each other and are often obsessed with some technical topic” as their own gender.

    You’re the hero we don’t deserve


  4. Last week I caught myself saying “I bought three guns this month” when in fact I had only purchased two handguns and an AR-15 lower-receiver. The reason I lumped the lower receiver with the two handguns, is because as far as the American government is concerned, buying a AR lower receiver is buying a firearm, and is therefore subject to the same paperwork that a full rifle would be subjected to. But my local gun store only charges one flat fee to fill out the paperwork for up to three firearms. So it made fiscal sense to bundle my “firearm” purchases together, and then once handgun1+handgun2+lower_reciever were bundled in my shopping cart, they became bundled in my mind, at least for a moment.
    As I’m sure you’ve deduced by now, I’m a trans girl. I don’t categorize myself as female all the time. I avoid women’s bathrooms, because I don’t like making women uncomfortable and I worry about the risks of making women’s spaces accessible to perverts. But I did sign up for the “ladies only” firearms class at the at that gun store, because “ladies” in this context clearly means “people who’s gender does not meet the criteria of a traditional gun owner and thus have fewer opportunities to learn about guns” and I do think that category applies to me.
    P.S. LT deserved her gold medal. Sorrynotsorry.

    Liked by 2 people

  5. Leaving for a moment aside the object level of gender and sex.

    At the metalevel, the whole categories were made for men and whatever the cluster structure of thingspace was meant to say (sorry, i fail to be able to parse yudkowsky’s writing) is one of the subjects where “rationalists” seem to be dangerously close to scientific antirealism, negating the existence of natural kinds.

    Natural kinds are not just a matter of perception. If they were, then the “special sciences” could not be understood in a realist way. And well, the purpose of science is not to make prediction but to understand the structure of reality.

    (Also, objects do exists independently of perception :p)


    1. Reality exists independently of perception. The division of that reality into objects is at least partly a matter of convention, and likewise the classification of those objects into kinds is partly a matter of convention.

      My body is an object. But does it include the air in my lungs? The blood in my veins? How about all the microorganisms in my gut? What if I have a false tooth, an artificial leg, an artificial hip joint, a stent helping an artery stay open, a transplanted kidney?

      Presumably you’d like cats to form a natural kind, and dogs another. As long as we consider only currently-living things, that works pretty well, but we can trace a path from any cat to any dog where each step goes to a parent or offspring organism and there are no big leaps anywhere. Where are the boundaries of those natural kinds? I say they’re fuzzy and that if we insist on drawing lines, or quantifying the fuzziness, there’s some freedom in exactly how we do it.

      For any sufficiently tightly specified purpose, there might be an optimal way to draw these boundaries. But different people on different occasions have different purposes, and different purposes will lead to different optimal boundaries.

      None of this means that there aren’t real differences between cats and dogs, or between me and not-me. Just that the details of the delineation are somewhat up for grabs. (And in some contexts the best lines to draw might be in surprising places; for instance, consider Dawkins’s “extended phenotype” notion which suggests that for some evolutionary-biological purposes we should consider a beaver’s dam to be part of the beaver, because what “the beaver” is from the gene’s-eye perspective is the thing that enables those genes to survive and reproduce or not, and building dams is part of that just as building limbs is.)

      Liked by 1 person

      1. The fact that the boundaries are fuzzy is not a proof that a kind doesn’t exist. Emergence of natural laws always involves a degree of fuzziness, the further you move up away from physics and chemistry the fuzzier, nevertheless they still describe real features of the world. Note that I don’t claim that all our scientific categories are natural kinds, just that the scientific process can in principle capture them with better and better approximation, and also more or less quantify (or at least identify) that they are fuzzy

        It is true that there is also a degree of convention, but the degree of convention depends on the categorization itself. My problem is that the discussion in the post is a motte and bailey (oh the irony!). It starts by saying that all categories are conventional to a degree (fine) and then uses this to conclude that because of this there is no problem in abandoning for all practical purposes a natural kind (one which is almost not fuzzy at all) in favour of a completely conventional category.

        I am not saying that we should not use the concept of gender. But it needs to be justified way better than this. I agree that our language doesn’t need to only contain natural kinds, but we do need to be very careful in distinguish which of them are and which are not.


        1. I’m not sure what the claim that “a natural kind exists” really amounts to. Suppose we agree that (1) there are a variety of ways to draw a boundary with things-generally-considered-fish on one side and things-generally-not-considered-fish on the other, and (2) that although you can draw that boundary in various different ways, almost everyone will want some concept in their toolbox that draws some such boundary, but that you say “fish form a natural kind” and I say “fish do not form a natural kind”. What do we actually disagree about?

          I’m guessing the answer is that what you mean by “natural kind” is pretty much the same as #2, so my hypothetical position is incoherent. But so far as I can see, the categories-were-made-for-man approach (or more specifically the position taken by Scott in the piece with that title) doesn’t deny #2, either in general or for the particular instance of sex and/or gender. So maybe my guess is wrong and you mean something else by “natural kind”, but in that case I’m not sure quite what.

          Maybe you could be more concrete. (A) You suggested that the categories-were-made-for-man approach is “dangerously close to scientific antirealism”. What, concretely, is the danger you anticipate — what specific bad things do you think may ensue — and why is the term “scientific antirealism” appropriate? (B) You worry that the OP here concludes that “there is no problem in abandoning for all practical purposes a natural kind” — I assume you mean “men” or “women” or something of the sort. What specific problems do you think the OP is overlooking or denying? What specific bad things will happen if more people, more of the time, work with finer-grained gender subdivisions of the sort celebrated by the OP? And (C) What makes you think OP is proposing to abandon a natural kind, as opposed to supplementing it with finer-grained less-binary classifications when those turn out to be more useful?


  6. In one of Eliezer Yudkowsky’s (and other’s) recent fiction projects, I encountered the word “gendertrope” and feel like it unlocked part of the ‘puzzle’ of what transgenderism (and other gender non-conformism) is (to my understanding anyways).

    Sadly, I think that, even with a better understanding, the broader social/political problem seems, if anything, less tractable – there’s no tangible, even-somewhat-‘immediately’ accessible reality to which anyone can appeal.

    I remain confused as to whether the insistence/demand that others (e.g. everyone) recognize one’s self-identified ‘gendertrope’ as not only ‘valid’ but ‘true’ is reasonable. I worry that the relatively ‘militant’ participants, on any side of the conflict, will continue to act as if their victory is only possible with the destruction of any and all opposition.

    Liked by 1 person

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s