+[TODO: address that "non-exclusively-androphilic" does exclude lesbians]
+
+But that was just my _guess_ at how to apply reductionism to describe the atmosphere of women-only spaces using lower-level criteria—and it was probably a _bad_ guess. (For one thing, [TODO: lesbians]) A woman who benefits from women-only spaces and knows more about psychology than me might say something different, and we should listen to _her_, not me. The "more than −1σ in both of these two dimensions" threshold was completely arbitrary; maybe she would prefer some other function. Maybe she [doesn't like the Big Five model](https://en.wikipedia.org/w/index.php?title=Big_Five_personality_traits&oldid=868909816#Critique). Maybe (in fact, almost certainly) multiple such women wouldn't _agree_ on the exact criteria.
+
+But even if such a committee of female psychologists _could_ agree on such criteria, I think most people would say that reorganizing the group as the "not-exclusively-androphilic people in our-favorite-hobby whose results on this-and-such personality battery match the following 1 KiB description ..." is not a particularly very appealing proposal.
+
+It would seem that in a world where psychological traits can't be cheaply, precisely, and verifiably measured, discrimination on the basis of [easily-observable traits that happen correlate with harder-to-measure traits that we actually care about](/2018/Feb/blegg-mode/) turns out to be a practical bright-line Schelling point for people to coordinate around.
+
+_Not_ an infinitely-thin, infinitely-bright line,[ref]As it is said: what about masculine women and feminine men (whose share of the population depends on where you set your sex-atypicality thresholds)? What about trans people (0.3%–[TODO] of the population, depending on how you define your categories and whose statistics you trust)? What about people with [5α-Reductase deficiency](https://en.wikipedia.org/wiki/5%CE%B1-Reductase_deficiency) or any of a dozen other specific intersex conditions?[/ref] but a line thin _enough_ and bright _enough_ that the forces of social evolution have coughed up some institutions and other cultural practices that take the line into account for _functional_ reasons.
+
+My goal in writing about this is certainly not to argue _for more sexism_—I'm looking forward to the postgender lesbian transhumanist future of Total Morphological Freedom as much as anyone else. (I already have my new name and outfits picked out!) If we can invent _new_ institutions and practices that serve more people more effectively, we should _do it_. But because I am a rationalist, because I cannot _unsee_ the cold, cisheteronormative logic of [Chesterson's fence](https://en.wikipedia.org/wiki/Wikipedia:Chesterton's_fence), I have to speak out when people are using clever word games to obfuscate the function of the existing fences. It's not that it's impossible to do better; it's that doing better isn't _trivial_.
+
+In response to the argument that women's restrooms function as safe havens that women can retreat to and exclude scary or threatening men from, Ozy writes:
+
+> I do not understand the relationship between this and psychological gender differences. It seems quite obvious that the relevant category here is "people who look like the vast majority of street harassers" versus "people who do not look like the vast majority of street harassers." The former group uncontroversially includes some trans women (closeted trans women) and some trans men (Buck Angel) and has nothing to do with psychology anyway. No matter how female-typical a trans man's psychology is, if he has muscles like Chris Hemsworth and a beard like a lumberjack, he belongs in the men's room.
+
+It has to do with _probabilistic predictions about_ psychology in a world where [male violence against females is _older than humanity itself_](https://en.wikipedia.org/w/index.php?title=Sexual_coercion&oldid=866576906). We could imagine an alternate universe designed by a loving God, where the people have the same physical forms as the women and men of our own world, but where rape and sexual harrassment and voyeurism are unknown, and in _that_ world, people with female bodies would have no particular reason to be wary of people with male bodies.[ref]Well, except for that _d_≈2.6 difference in muscle mass should a dispute escalate to physical fighting.[/ref]
+
+Certainly _most_ men are nice, civilized people who don't harrass women—and occasional Hemsworthlike, lumberjack-bearded androphilic trans men with a feminine personalities, present even _less_ of a threat. But when designing the social norms for a safe space for the modal cis woman, false positives (including someone who shouldn't be included) are probably going to be worse than false negatives (excluding someone who shouldn't be). If "Does this person look male?" is _easier to assess_ than "Does this person-of-whatever-sex look like a potential threat to my safety, comfort, and privacy?"—and possibly more importantly, is easier for third parties to _agree on_ when third parties are called in to enforce the rules—then the rule ends up being "no men" (or "no male-looking people").
+
+Which is _not_ necessarily a great rule!
+
+[***]
+
+[agree that everyone deserves a place to pee]
+
+Finally, Ozy makes an analogy between social gender and money. What constitutes money in a given social context is determined by collective agreement: money is whatever you can reliably expect everyone else to accept as payment. This isn't a circular definition (in the way that "money is whatever we agree is money" would be uninformative to an alien who didn't already have a referent for the word _money_), and people advocating for a _different_ money regime (like [late-19th century American bimetalists](https://en.wikipedia.org/w/index.php?title=Bimetallism&oldid=864176071#Political_debate) or contemporary cryptocurrency advocates) aren't making an epistemic _mistake_.
+
+I _really like_ this analogy! An important thing to note here is that while the form of money can vary widely across sociocultural contexts (from [shell beads](https://en.wikipedia.org/wiki/Wampum), to silver coins, to fiat paper currency, to database entries in a bank), not just any form will suffice to serve the functions of money: perishable goods like cheese can't function as a long-term store of value; non-fungible items that vary in quality in hard-to-measure ways (my bull might be younger or fitter than yours) can't function as a unit of account.