That's my line! ---- "The ontology suggested by our new communication protocol differs from the ontology suggested by the old protocol in ways that increase the message length of concepts relevant to your values; sucks to be you" is at least showing your vanquished political opponents a modicum of respect for what they were! Playing dumb with smug you-are-ontologically-confused act is really quite unnecessary! ---- There's a cognitive-science-of-actually-existing-English-speakers thing that you keep refusing to acknowledge, that's still relevant for actually-existing-English-speakers even if it's not how you would set things up if you were designing a sane language from scratch. I flat-out disbelieve your claim that you don't know how to write a different viewpoint as a sympathetic character (you're far too good of a writer for that claim to be remotely plausible), but let me help you out; here's a short story idea— ----- A cis woman is testifying in court about a brutal rape that horrifically traumatized her. The rapist has since transitioned. "And then—" says the witness, reliving those awful moments, "and then, he took his erect penis—" "Her erect penis," says the judge. "What?" says the witness. "You will refer to the defendant with the correct pronoun, or I'll hold you in contempt of court." "And then she took her—" the witness breaks down crying. "I'm sorry, Your Honor; I can't do it. I'm under oath; I have to tell the story the way my brain represents it. In my mental universe, the person who raped me was a man—an adult human male—and in my native dialect of English, men are referred to with he/him pronouns. If I say 'she', it doesn't feel true to the memory in my head; it feels analogous to lying, even if it's not technically a lie because pronouns aren't facts." "At age 13 I was programming on LambdaMOO where people had their choice of exotic pronouns and nobody thought anything of it," says the judge. "Denied." "O-okay. So then—then sh-she ..." the witness breaks down crying again. "Your Honor, I can't. Look, I know the court allows witnesses to testify in non-English languages with the help of a translator. Can't you treat my testimony like that? Let me say what I experienced in the language that seems true to me, even if the court conducts its business under a different language convention?" "You're in contempt," says the judge. "Baliff! Take her away!" ---- Not a sympathetic viewpoint character? Not even a little bit? ----- Thanks for asking! My answer is that language choices and medical-procedure choices are correlated. If I'm deciding whether to transition, "transition" isn't an atomic action; it's a bundle of many complementary interventions, "(1) HRT (2) ask everyone to switch pronouns (3) ask everyone to use a new name (if 'It's short for Zacqueline' won't fly), &c." In principle, you could pick some interventions but not others (I in fact tried 5 months of HRT only), but the pronouns part specifically is usually part of a bundled "Transition? Yes/No" decision. If I'm considering the bundled decision, in order to make the best choice, I need good information about the costs and benefits. For example, one concern I might have is, "Gee, what if I don't physically 'pass'—is that going to make my social life more awkward, if people don't instinctively use the correct pronoun for me?" I should be more eager to transition (on the margin) if the real-world answer to that question is "No, that won't be a problem", than if the real-world answer is "Yes, that will actually be awkward." If all the high-status people I trust say, "No, that shouldn't be a problem because pronouns don't have truth conditions", but it turns out that the high-status people are wrong and the real-world answer is "Yes, it will be a problem" then I might make a worse decision! ------ That was really confusing, because I had spent the previous decade assuming that my thing wasn't the same thing as being "actually" trans. (22,000 word work-in-progress draft blog post about this: http://unremediatedgender.space/drafts/sexual-dimorphism-in-the-sequences-in-relation-to-my-gender-problems.html ) And so I started talked talking to people in your robot cult about the autogynephilia/two-types hypothesis, which seemed like a scientific question that was hugely relevant to a lot of our people, such that—if it were true (and it's obviously a better fit than the atomic 'gender identity' theory), and if this so-called "rationality" community isn't fraudulent (which it is, but I didn't know that then)—people around here should be interested in getting the right answer and, if necessary, correcting the state of public knowledge. One thing that was weird about those conversations is that they kept getting _derailed_ on some variation of, "Well, the word 'woman' doesn't necessarily mean that; have you read Scott Alexander's 'The Categories Were Made for Man, Not Man for the Categories'"? That really wasn't what I was trying to talk about—I thought I was trying to talk about an empirical thesis in psychology, not a philosophy-of-language issue! But because I believed in the power of good-faith debate, I took the bait. Psychology is a complicated empricial science—no matter how "obvious" I think something is, I could always just be wrong, in the real world, and not just as a matter of modesty. But this redefining-words mind game? That was clear-cut. (We had a whole Sequence on this!) At least that part of the argument, I knew I could win. So I wrote a couple of blog posts explaining the problem: http://unremediatedgender.space/2018/Feb/the-categories-were-made-for-man-to-make-predictions/ http://unremediatedgender.space/2018/Apr/reply-to-the-unit-of-caring-on-adult-human-females/ That would have been the end of the line of argument, until your "Hill of Meaning in Defense of Validity" Twitter proclamation, which seemed to me to be very blatantly and heavily optimized to lend public support to the "there's no rationality problem with redefining concepts to make people happier" position, even if every individual sentence permitted a true interpretation. > remember the tweet that set this whole thing off - at least as an initial spark In my mental universe, that was NOT the initial spark!! In my mental universe, the initial spark was the "over 50% probability at this point that at least 20% of the ones with penises are actually women" (https://www.facebook.com/yudkowsky/posts/10154078468809228) thing. In 2016, I was trying to talk to people about the autogynephilia/two-types hypothesis, which was an empirical claim in psychology. I was not expecting to have to go on a philosophy-of-language crusade! But I noticed that my attempts to talk about psychology kept getting _derailed_ on some variation of, "Well, the word 'woman' doesn't necessarily mean that; have you read Scott Alexander's 'The Categories Were Made for Man, Not Man for the Categories'?" So, because I believed in the power of good faith debate, I took the bait, and wrote two blog posts explaining the problem in February 2018 (http://unremediatedgender.space/2018/Feb/the-categories-were-made-for-man-to-make-predictions/) and April 2018 (http://unremediatedgender.space/2018/Apr/reply-to-the-unit-of-caring-on-adult-human-females/). That would have been the end of that line of argument, if your "Hill of Meaning in Defense of Validity" proclamation ----- Without making any claims about conscious "intent" (whatever that means), can you see why it looks to me like your public statements from 2016 and 2018 are better predicted by a "Eliezer Yudkowsky's public statements are part of an optimization process whose target includes tricking autogynephilic males into cutting their dicks off" model, rather than a "Eliezer Yudkowsky's public statements are optimized for neutrally teaching rationality" model? (Where, again, "Part of an optimization process" isn't about individual human intent, it's about what outcomes the system as a whole steers reality towards, robustly to perturbations: if the response to "20% actually women" being argued against is to immediately pivot to "Pronouns aren't lies"—and if people who have intuitions about English singular pronouns bearing sex-category information are told that "their feelings don't get to control everybody's language protocol", but Scott Alexander's post advocating that people's feelings should control everybody's language protocol, doesn't get that response—that pattern of selective argumentation and silence by high-status influential people has the same effects whether the underlying algorithm is Kolmogorov complicity or conscious scheming.) (As usual, Policy Debates Should Not Appear One-Sided—obviously, if you have a sufficiently large robot cult, then no matter what you say, some of your followers are going to model you as participating in an optimization process that is hostile to them, and throw a fit—in explaining the forces and threats that I see on _my_ chessboard, I'm not claiming that my chessboard, like, actually matters, what with the lightcone at stake.)