First note: After I install the extension it takes me to a page that says,

Fair enough, but is that a crux for you, or for Zack?

If you knew there wasn't a slippy slope here, would this matter?

Or to say it differently: we can unload some-to-most of the content of the word woman (however much of it doesn't apply to transwomen) onto the word "cis-woman", and call it a day. The "woman" category becomes proportionally less useful, but it's mostly fine because we still have the expressiveness to say everything we might want to say. 

In the skeptic's view, if you're not going to change the kid's diet on the basis of the second part, you shouldn't social transition the kid on the basis of the first part.

I think I probably would change the kid's diet?? Or at least talk with them further about it, and if their preference was robust, help them change their diet.

But if the grown-ups have been trained to believe that "trans kids know who they are"—if they're emotionally eager at the prospect of having a transgender child, or fearful of the damage they might do by not affirming—they might selectively attend to confirming evidence that the child "is trans", selectively ignore contrary evidence that the child "is cis", and end up reinforcing a cross-sex identity that would not have existed if not for their belief in it—a belief that the same people raising the same child ten years ago wouldn't have held. (A September 2013 article in The Atlantic by the father of a male child with stereotypically feminine interests was titled "My Son Wears Dresses; Get Over It", not "My Daughter Is Trans; Get Over It".)

Wow. This is a horrifying thought.

Under recent historical conditions in the West, these kids were mostly "pre-gay" rather than trans. (The stereotype about lesbians being masculine and gay men being feminine is, like most stereotypes, basically true: sex-atypical childhood behavior between gay and straight adults has been meta-analyzed at Cohen's d ≈ 1.31 standard deviations for men and d ≈ 0.96 for women.) A solid majority of children diagnosed with gender dysphoria ended up growing out of it by puberty. In the culture of the current year, it seemed likely that a lot of those kids would instead get affirmed into a cross-sex identity at a young age, even though most of them would have otherwise (under a "watchful waiting" protocol) grown up to be ordinary gay men and lesbians.

What made this shift in norms crazy, in my view, was not just that transitioning younger children is a dubious treatment decision, but that it's a dubious treatment decision that was being made on the basis of the obvious falsehood that "trans" was one thing: the cultural phenomenon of "trans kids" was being used to legitimize trans adults, even though a supermajority of trans adults were in the late-onset taxon and therefore had never resembled these HSTS-taxon kids. That is: pre-gay kids in our Society are being sterilized in order to affirm the narcissistic delusions[29] of guys like me.

I definitely want to think more about this, and my views are provisional.

But if this basic story is true, it sure changes my attitude towards childhood gender-transitions!

I was skeptical of the claim that no one was "really" being kept ignorant. If you're sufficiently clever and careful and you remember how language worked when Airstrip One was still Britain, then you can still think, internally, and express yourself as best you can in Newspeak. But a culture in which Newspeak is mandatory, and all of Oceania's best philosophers have clever arguments for why Newspeak doesn't distort people's beliefs doesn't seem like a culture that could solve AI alignment.

Hm. Is it a crux for you if language retains the categories of "transwoman" and "cis woman" in addition to (now corrupted, in your view) general category of "woman"?

I guess not, but I'm not totally sure what your reason for why not would be.

...or maybe you're mainly like "it's fucked up that this particular empirical question propagated so far back into our epistemology that it caused Scott and Eliezer to get a general philosophical question wrong."

That does seem to me like the most concerning thing about this whole situation, if that is indeed what happened. 

He asked for a specific example. ("Trans women are women, therefore trans women have uteruses" being a bad example, because no one was claiming that.) I quoted an article from the The Nation: "There is another argument against allowing trans athletes to compete with cis-gender athletes that suggests that their presence hurts cis-women and cis-girls. But this line of thought doesn't acknowledge that trans women are in fact women." Scott agreed that this was stupid and wrong and a natural consequence of letting people use language the way he was suggesting (!).

I wonder if the crux here is that Scott keeps thinking of the question as "what words should we use to describe things" and not "what internal categories should I use"?

Like, I could imagine thinking "It's not really a problem / not that bad to say that transwomen are women, because I happen to have the category of "transwomen" and so can keep track of the ways in which transwomen, on average, are different from cis women. Given that I'll be able to track the details of the world one way or the other, it's a pragmatic question of whether we should call transwomen women, and it seems like it's an overall pretty good choice on utilitarian grounds."

messy evolved animal brains don't track probability and utility separately the way a cleanly-designed AI could.

Side-note: a cleanly designed AI could do this, but it isn't obvious to me that this is actually the optimal design choice. Insofar as the agent is ultimately optimizing for utility, you might want epistemology to be shaped according considerations of valence (relevance to goals) up and down the stack. You pay attention to, and form concepts about, things in proportion to their utility-relevance.

It might seem like a little thing of no significance—requiring "I" statements is commonplace in therapy groups and corporate sensitivity training—but this little thing coming from Eliezer Yudkowsky setting guidelines for an explicitly "rationalist" space made a pattern click. If everyone is forced to only make claims about their map ("I think", "I feel") and not make claims about the territory (which could be construed to call other people's maps into question and thereby threaten them, because disagreement is disrespect), that's great for reducing social conflict but not for the kind of collective information processing that accomplishes cognitive work,[21] like good literary criticism. A rationalist space needs to be able to talk about the territory.

I strongly disagree with the bolded text. It often helps a lot to say phrases like "on my model" or "as I see it", because it emphasizes the difference between my map and the territory, even though I'm implicitly/explicitly claiming that my map models the territory.

This is helpful for a bunch of human psychological reasons, but one is that humans often feel social pressure to overwrite their own models or impressions with their received model of someone speaking confidently. In most parts of the world, stating something with confidence is not just a claim about truth values to be disputed, it's a social bid (sometimes a social threat) for others to treat what we're saying as true. That's very bad for collective epistemology!

Many of us rationalist-types (especially, in practice, males) have a social aura of confidence that fucks with other people's epistemology. (I've been on both sides of that dynamic, depending on who I'm talking to.)

By making these sorts of declarations where we emphasize that our maps are not the territory, we make space for others to have their own differing impressions and views, which means that I am able to learn more from them

Now, it's totally fine to say "well, people shouldn't be like that." But we already knew we were dealing with corrupted hardware riddled with biases. The question is how can we make use of the faculties that we actually have at our disposable to cobble together effective epistemic processes (individual and collective) anyway.

And it turns out that being straightforward about what you think is true at a content level (not dissembling), while also adopting practices and norms that attend to people's emotional and social experience works better than ignoring the social dimension, and trying to just focus on the content.

...Or that's my understanding anyway. 

See for instance:

Load More