Some challenges:
There are a lot of benefits to scale; if you figure out some true thing and share it with others, then the advantage from that true thing can be multiplied by the number of people you share it with, which may be a lot of people. Further, as an individual you only have limited time and knowledge, which prevents you from figuring out all that many things, whereas a group of people cooperating could figure out many more things.
However, scale and cooperation are hard to achieve. You could expect everyone to check all of the evidence themselves, but that would be very expensive in terms of time (and sometimes also other things, like privacy), so you need people to trust each other. But there are a couple of reasons why trust is hard to achieve.
First, there's selection bias. The sorts of people who are likely to put a lot of effort into pushing some idea are also the sorts who are likely to be biased about that idea. And if they weren't biased to begin with, surely a lot have become biased just by being known as The People Who Share That Idea.
Secondly, there is enemy action. If you acknowledge innate racial differences in intelligence, for instance, then you make a lot of enemies who, if you become popular, will try to undermine your reputation through various means. These include telling others that you are biased without evidence, picking things you've said out of context, etc..
Even if you somehow figure out how to achieve trust, there are other challenges. In order to achieve scale, you need some platform to have that scale on. And here it's hard if your platform isn't mainstream so that it is accessible to people. But e.g. Discord will ban you if you hang out in a server that talks about race differences in intelligence (source: I've seen it happen to a person), and there are lots of people coordinating to try and make other platforms more strict in terms of what information they allow to be shared (often under the guise of "curbing misinformation" or "preventing hate"). This in effect interferes with the possibility of achieving scale and cooperation in telling the truth, which makes it harder to figure out the truth and less impactful once you do figure it out.
And then... What about if you succeed in telling the truth? I used to think that it would be fine, I'm personally liberal and I don't think race differences in intelligence should be used as justification for discrimination. But then there's the second part of this blog post, which criticizes the idea that you can just straightforwardly expect liberalism to work without group difference denial.
And heck, what about the "double taboos" - in circles that discuss race differences in intelligence and sex differences in personality, it is often taboo to discuss race differences in personality and sex differences in intelligence, so I have a hard time telling how much meat there is on these topics. But what if there is - what if black people are innately more criminal, and not due to some sort of intelligence->poverty->crime connection? What if women are 5 IQ points less intelligent than men? If these were widely accepted, or even if they were demoted from double taboo to single taboo, what impacts would this have? I don't know.
What do you see as the main problem with applying "That which can be destroyed by the truth should be" to liberalism? Surely people came to believe in liberalism for reasons that are in part empirical.
(One issue is that the lack of agreement on a common political framework could cause conflict-heavy coordination failure, but I'm not sure if that's what you see as the main problem.)
What do you see as the main problem with applying "That which can be destroyed by the truth should be" to liberalism?
Disclaimer: I haven't thought all of this through. As MTSW points out, "The idea that people are unequal in ways that matter is legitimately too horrifying to contemplate, so liberals deny the inequality, and conservatives deny that it matters." - I don't think I can think all of the consequences through myself, and I haven't seen anyone seriously discuss the potential outcomes of it.
But I think it may be useful to start with the fact that it matters how you select and present the "truths"; e.g. there is the whole Chinese Robber Fallacy, where you selectively present facts that makes a group look bad. Obviously this is not the sort of thing "that which can be destroyed by the truth should be" is meant to apply to; it's more meant to be for the cases of "that which can be destroyed by the whole truth should be", or something along those lines.
So bringing this around to group differences - one of the debates about group differences is what they tell you about individual traits. For instance, if women have much (d~1.2) worse mechanical ability than men, does this tell you things about any given woman's mechanical ability? On the one hand, the answer is clearly no/negligible; a group difference of 1.2 will (if I did my calculations correctly 😬) only reduce your range of uncertainty by 14%. You're still going to need to treat people as individuals and investigate their actual mechanical ability if you want to know how good they are. On the other hand, as MTSW points out:
Give people photographs of various women and men and ask them to judge how tall the people in the photos are, as Nelson et al. 1990 did, and people's guesses reflect both the photo-subjects' actual heights, but also (to a lesser degree) their sex. Unless you expect people to be perfect at assessing height from photographs (when they don't know how far away the cameraperson was standing, aren't "trigonometrically omniscient", &c.), this behavior is just correct: men really are taller than women on average, so P(true-height|apparent-height, sex) ≠ P(true-height|apparent-height) because of regression to the mean (and women and men regress to different means). But this all happens subconsciously: in the same study, when the authors tried height-matching the photographs (for every photo of a woman of a given height, there was another photo in the set of a man of the same height) and telling the participants about the height-matching and offering a cash reward to the best height-judge, more than half of the stereotyping effect remained. It would seem that people can't consciously readjust their learned priors in reaction to verbal instructions pertaining to an artificial context.
So men and women would regress to different means, which means that if your test of mechanical ability is imperfect, you would "rationally" discriminate against women. In fact it turns out that the usefulness of a stereotype for discrimination is highly nonlinearly related to the amount of information it provides/degree to which it reduces your uncertainty; as stereotypes that only reduce your range of uncertainty by marginal amounts are still highly useful. (When trying to select for someone who is high in some trait using a proxy, the degree to which you succeed in expectation will be proportional to the correlation r between the proxy and the trait - but the amount it reduces the uncertainty will be proportional to 1-sqrt(1-r^2), which is close to 0 for most of the range of r.)
(Sidenote - one point I've thought up a while ago, which seems underappreciated, is that how bad this is actually depends on some subtle causal relationships. In addition to there being a sex difference in mechanical ability, there is also a sex difference in mechanical interests. One theory goes that the sex difference in mechanical interests causes the sex difference in mechanical ability, because if you engage with mechanical stuff more, that gives you experience with it. Another theory goes that sex differences in mechanical abilities cause sex differences in mechanical interests, because it's not very fun to engage in some topic that you are bad at. These make somewhat different predictions because in the former case, interests screen off ability from sex, while in the latter case, ability screens off interests from sex; that implies that, in the former case, women and men who are equally interested in mechanical stuff regress to the same mean, while in the latter case, women regress to a lower mean than men even when equally interested. The former seems "better" in that the "women are worse at mechanical stuff" only applies to women who don't care about mechanical stuff anyway. Personally I suspect the former more accurately captures how it works, but I haven't seen it evaluated in detail.)
Anyway, what does this have to do with the Chinese Robber Fallacy? Well, if your test of mechanical ability wasn't imperfect - if you knew exactly how good everyone's mechanical ability was - then you would not be incentivized to rationally discriminate. So plausibly, equal treatment might be "destroyed by" knowledge of demographic truths, but restored by knowledge of individual truths. So truth about group differences may make group discrimination worse, while truth about individual differences may make group discrimination better.
But also, an unfortunate fact is that group differences are more legible and easier to measure. You can just select a bunch of people from each group, give them a relatively bad (but quick/cheap) test, and this will give you a fairly solid rank-ordering of the groups in the trait, because the noise cancels out. Then often group membership will be trivial to observe, as it is correlated with physical characteristics that are hard to hide, so you can apply this information about group differences as an ultra-cheap but highly inaccurate test of the trait value for individuals. Meanwhile if you wanted to know the trait value for individuals, you would have to test every single individual that you wanted to know the trait value for, and you would have to use a much more thorough but expensive test, etc..
So there's a Chinese Robber-style asymmetry, where the truths that increase group discrimination are inherently easier to obtain, spread, and apply than the truths that decrease group discrimination.
And then there's the issue of whether it's even desirable to know the individual-level truths. Truth and privacy inherently trade off against each other.
Surely people came to believe in liberalism for reasons that are in part empirical.
This is definitely a good point, but I don't know whether it solves it. As far as I know, the modern movements in support of gender and racial equality originate from all the way back when women and racial minorities were completely subjugated by white male authority. I feel fairly certain that one doesn't need group difference denial to see that this is wrong and immoral.
But this then seems to have birthed movements that have made group difference denial obligatory, which then seems to have increased the expectations for equality and the demands for policies to achieve it. If racial differences in intelligence don't exist, then not only should black people be allowed to do whatever white people are allowed to do, it's also indicative of hateful discrimination if black people achieve worse socioeconomic outcomes and therefore this discrimination must be stopped.
Presumably, acknowledging innate group differences isn't going to bring us all the way back to maximal oppression, and presumably it is going to cut down on some of the worst excesses of activism - but that still leaves an enormous range inbetween, of varying degrees of oppression, discimination, activism. Do I know where we would end up if we freely acknowledged innate group differences? I have to admit that I don't. Do I even know where we should end up? I might favor some sort of "treat people as individuals" liberalism, some go more egalitarian than me and suggest using the knowledge of innate group differences to achieve greater equality, and some go less egalitarian than me and want to cut down on immigration. And that's just the state currently, where there is much denial of innate group differences - what happens if the taboos start getting lifted?
(And what about backlash? If it's discovered that academics and activists have been lying about these topics for decades, what will the reaction to that be?)
(One issue is that the lack of agreement on a common political framework could cause conflict-heavy coordination failure, but I'm not sure if that's what you see as the main problem.)
Overall, I don't really have a main problem in mind. What I do have in mind is that I don't know what would happen. "The idea that people are unequal in ways that matter is legitimately too horrifying to contemplate, so liberals deny the inequality, and conservatives deny that it matters." - This is a topic that lacks investigation and discussion by smart people who acknowledge the inequality, care about the moral equality that has been achieved, and most importantly, care about the truth.
There is a lot to unpack in your comment.
It's true that you can be technically correct on the details while misleading others about the whole. Misleading people is bad and should be avoided. (Incidentally, news is misleading because of sample bias.)
It's also true that "being fair to the people you're judging" and "optimizing for your own predictive performance" are two different kinds of fairness. I explained how they constitute mathematically contradictory values in my book review of The Alignment Problem. Paul Graham wrote a good article on the subject too.
As far as I know, the modern movements in support of gender and racial equality originate from all the way back when women and racial minorities were completely subjugated by white male authority. I feel fairly certain that one doesn't need group difference denial to see that this is wrong and immoral.
But this then seems to have birthed movements that have made group difference denial obligatory, which then seems to have increased the expectations for equality and the demands for policies to achieve it.
The pendulum swings. The pendulum swings back.
Do I know where we would end up if we freely acknowledged innate group differences? I have to admit that I don't. Do I even know where we should end up?
I don't know where we'll end up either but the best place to end up is probably somewhere our epistemics are grounded in reality. Otherwise, how could we know we're in a good place? Empiricism and transparency have a great historical track record. I project this will continue.
(And what about backlash? If it's discovered that academics and activists have been lying about these topics for decades, what will the reaction to that be?)
In this situation, that which can be destroyed by the truth should be.
I think ultimately, I agree, hence why this was a side-point to my original comment, rather than being the main focus. The distortions today seem quite big and growing ever bigger, and it seems like the median peace/equality benefits are probably small enough to not be worth the cost. But I do think it is appropriate, within some group that understands the importance of figuring out the truth, to also speak up clearly about the potential downsides of it. (If nothing else, we should spend some time thinking of the potential negative tail of worst case scenarios.)
Potentially relevant to the discussion: I ran a poll among my twitter followers, and 50% did not want differences between people to be revealed, I assume because they thought it would make themselves or the world worse off. I don't know for sure how my twitter followers compare to the general population, but I suspect they would be more positive towards differences between people being revealed.
That which can be destroyed by abstract truths might also be abstractly true.
Only when you are dealing with claims which represent fully formalized intuitions does it apply that 'that which can be destroyed by the truth (is false and therefore) should be.'
Abstract imperatives like "don't be a dick" and "be cool to each other" are important to remember even if you have a very good formalization, because you basically never know if you've really formalized the full set of intuitions, or if you've only formalized some parts of the set of intuitions which e.g. "don't be a dick" and "be cool to each other" capture.
On the other hand, I am curious as to what would happen if we formalized intuitions about levels of abstraction in general.
I am very frank by nature. In my career my coworkers have often expressed a lot of respect for the fact that I will ask "tough questions", give thoughtful critical feedback, and argue well for idiosyncratic opinions. I almost never decline to say something that I think is true and interesting or useful because I think it will have negative repercussions on me. (Although if I think it might be hurtful to someone, or they might misinterpret it, I may try to optimize how I say it.)
However, I am also noticeably poor at understanding other people's nonverbal behavior. I am frequently in situations where other people (who I consider to be worth their weight in gold to me when performing this function) show me that I have been oblivious to how I was affecting other people. For example, I have had to learn a lot about how to criticize people in ways that they will be receptive to, and about the function of public praise. Perhaps not unrelatedly, I also have exceptionally low amounts of general anxiety and neuroticism. So basically the thing in the world that I am absolutely the worst at noticing and responding to is when someone is unhappy or angry with me in some way.
I think these go together. I never get negative feedback on my behavior, because I am too oblivious to notice the negative feedback which is constantly flying at me. So I do the "brave truthteller" thing by default and enjoy the conspicuous respect. I bet that most "brave truthteller" types are in a similar position. That's not to say that it's a bad policy -- I think it's just a high-variance policy -- but it's a lot easier to execute the policy when you don't suffer the bad beats.
The lack of recognition probably has to do with your own ability to keep things positive internally, so when you do talk about negative or depressing stuff, it has different type of emotional effect on you compared to most others. I don't have as many "friends" as a result of this, but I'm able to find friends who I think matter for areas that others may not care much about for this as well. It's really just a case of "even though you are reading this, it doesn't mean it was written for you."
I am willing to bet money that the average IQ of descendants of H1-B visa holders is (or will be) higher than the IQs of both the average American and the average Indian.
I don't think this necessarily implies ethnic differences in IQ. It could be that Americans and Indians have the same IQ distribution and higher-IQ Indians are more likely to immigrate.
I believe the point is that today's "higher-IQ Indians [...] more likely to immigrate" become tomorrow's Americans.
This is not a sleight-of-hand; Indian Americans (or Indian Britons, or Chinese Australians, or members of ethnicity X living in country Y) do constitute an ethnic group, in precisely the same way e.g. African Americans constitute an ethnic group.
This is because membership in these groups is decision-relevant, in a way that membership in broader groups such as "all Indians in the world" is not: e.g. when you are selecting from a pool of job applicants, you will in most cases be dealing with applicants who either (a) already live within the country, or (b) intend to move to the country--either of which subjects them to the selection effect induced by the H1-B visa process. And as it is in this context that "ethnic groups" (and moral questions surrounding the fair or unfair treatment thereof) are even a thing worth noticing to begin with, there is no sleight-of-hand in the original post.
Even in the complete absence of personal consequences, expressing unpopular opinions still brings disrepute on other opinions that are logically related or held by the same people. E.g., if hypothetically there were a surprisingly strong argument for murdering puppies, I would keep it to myself, because only people who care about surprisingly strong arguments would accept it, and others would hate them for it, impeding their ability to do all the less horrible and more important things that there are surprisingly strong arguments for.
In this case the principle that leaves the state of evidence undisturbed is to keep any argument for not murdering puppies to yourself as well, for otherwise you in expectation would create filtered evidence in favor of not murdering puppies.
This is analogous to trial preregistration, you just do the preregistration like an updateless agent, committing to act as if you've preregistered to speak publicly on any topic on which you are about to speak regardless of what it turns out you have to say on it. This either prompts you to say a socially costly thing (if you judge the preregistration a good deal) or to stay silent on a socially neutral or approved thing (if the preregistration doesn't look like a good deal).
Obviously the idea is not to never risk making enemies, but the future is to some extent a hostage negotiation, and, airy rhetoric aside, it's a bad idea to insult a hostage taker's mother, causing him to murder lots of hostages, even if she's genuinely a bad person who deserves to be called out.
You don't talk about because you want others to accept your position. You talk about it, so others have a chance to convince you to abandon that position, either for you to take theirs or something entirely different. How do you know that you've read everything to take up your position if you don't bother giving others who have put into their own time and thoughts into this a chance to present their arguments? But at the end of the day, we just gotta what we gotta do that makes us happy.
I can log onto a free blogging platform via Tor and no government in the world could deanonymize me.
How sure are you about this? I think maybe someone could deanonymize you, it would just take a lot of resources (pwning a bunch of exit nodes, and then correlating when your computer sends a thing with when the thing hits the exit node). Or they could analyze your writing and cross-reference that with other stuff you've written nymously.
see gwern for an interesting read on deanonymization techniques:
https://www.gwern.net/Death-Note-Anonymity
From what I know about darkweb drug dealers, the US government has never taken down an illicit drug rung by breaking Tor itself. But I'm not a specialist.
Writing analysis is a bigger threat, especially since the technology to deanonymize writing by cross-referencing it is advancing.
US government has never taken down an illicit drug rung by breaking Tor itself.
The policy of the US government is to reconstruct a plausible narrative of how they caught the offender if they use exploits like this so that they can continue to use them in future.
If you look at the number of dark markets that have used Tor and been shutdown by the feds, I don't think it's implausible that Tor is already compromised.
Fails to make a clear point; talks about the ability to publish in the modern world, then brushes over cancel culture, immigration, and gender differences. Needs to make a stronger argument and back it up with evidence.
Strong upvoted. I have always considered people who complained about the tiny and insignificant social consequences for heresy in the modern U.S. cowards, and this post explains why.
My impression is that we would be in relative heaven by now if this image realistically represented the thoughts of most people as it implicitly intends to. Most people would rather prevent legibility of comparison.
One line stood out to me in particular, and it stood out due to a misunderstanding I had which I suspect many other readers have as well. The line is "...there are significant heritable intelligence disparities between ethnic groups?"
At first read, this post implies that genes in different populations determine at least 50% of the variability in IQ, which is not true. I know your stance on saying the truth without regard to dumbing it down for audiences that shouldn't read it, but I think this is one point where you should explain more. I conjecture that a nontrivial amount of people reading that line also have the same confusion I had.
What is the confusion? That heritability is different from genetic heritability. This post enlightened me: https://bostonreview.net/science-nature-race/ned-block-race-genes-and-iq
An example from that article is that wearing earrings used to be highly heritable because you just had to look at whether they were female or male. As more people have started wearing earrings, the earring wearing trait has become less heritable. Similarly, IQ is not genetically determined and you see kids from low socioeconomic status way surpassing their parents IQ when adopted into higher socioeconomic status families.
On this particular point, while you are trying to succinctly say the truth, you are communicating a falsehood that needlessly alienates readers who have the same misunderstanding as me. I'm not refuting your point. I'm merely arguing that you should explain more here about a common confusion people will have when reading this.
[T]his is one point where you should explain more.
I will explain more. Total heritability of intelligence (in the US) might be as low as .40 (but probably isn't). Heritability of intelligence due to being in one particular genetic bucket must be strictly lower than total heritability of intelligence. "Significant" can be below 50%.
An example from that article is that wearing earrings used to be highly heritable because you just had to look at whether they were female or male. As more people have started wearing earrings, the earring wearing trait has become less heritable.
This is true. Exactly what "heritability" means is beyond the scope of this post but Dynomight has a great article on what heritability means, mathematically. (Including how a trait can be more than 100% heritable and how changes to society can change the heritability of a trait.)
IQ is not genetically determined
My claim is that IQ has a genetic component. I am not claiming that IQ is 100% genetically determined. Are you claiming IQ doesn't have a genetic component at all?
Thanks for the thoughtful reply, and the interesting read!
My claim is that IQ has a genetic component. I am not claiming that IQ is 100% genetically determined. Are you claiming IQ doesn't have a genetic component at all?
I'm not claiming that IQ has zero genetic component, but I am saying that it's not straightforward to conclude there are significant ethnic differences in IQ that are determined by genes.
To be specific, I'm arguing that IQ between ethnic groups in the US is likely much less than 50% determined by genes. Finding genes correlated with IQ doesn't imply genes play a direct causal role, and there are very strong explanations that don't involve genes such as socioeconomic status for example.
I'd wager around 0 to10% of the variation within normal IQ ranges is determined by genes for some cases, although that's speculation based on evidence. I can't find any rigorous scientific study of genes changing IQ (within normal ranges, as you can have genes that make the brain dysfunctional).
Heritability of intelligence due to being in one particular genetic bucket must be strictly lower than total heritability of intelligence. "Significant" can be below 50%.
Do you claim that heritability of intelligence due to being in one particular genetic bucket is closer to 50%? Or how much lower would you put it?
Society is not an epistemic community. Its purpose is not to find the truth. When will you "people who care about truth more than about how you look"1 get this? That not all true things should be said, at least in public. These things might hurt public order & morality, and people will ostracize and fire you for them.
1 note: right, I changed it. It's the same, though. Also, "Are you afraid to say that there are significant heritable intelligence disparities between ethnic groups? It's the obvious conclusion ..." violates our society's established standards of discourse.
Thank you for joining our community's collective endeavor to better approach the truth. Your sentence "When will you autistic types * get this?" is rude and non-collaborative. It violates our established standards of productive discourse. Is there a way you can rephrase your comment to be less unnecessarily insulting?
My favorite thing about living in the 21ˢᵗ century is that nobody can stop me from publishing whatever I want.
I mean this in the most literal sense. I can log onto a free blogging platform via Tor and no government in the world could deanonymize me. I might even do this already and there would be no way for you (or anybody else) to know.
Different countries have different levels of freedom. I have friends who live under authoritarian regimes who would literally be disappeared by the secret police if they declared their political views outright. They say "it's dangerous to express my views publicly" but they never say "I'm an independent thinker". The only people I know who claim they're scared to express independent thought live in the United States where freedom of the press is protected under the First Amendment of the Constitution.
People tell me they're worried of being cancelled by woke culture. I think this is just a convenient excuse for laziness and cowardice. What are you afraid of saying? That there are personality differences between men and women? Women consistently score higher than men in agreeableness and neroticism.
Are you afraid to say that there are significant heritable intelligence disparities between ethnic groups? It's the obvious conclusion if you think critically about US immigration policy.
"87.2% of Indian-American adults in 2010 were foreign-born[1]." US immigration policy makes it very hard for natives of India to immigrate into the United States. It's so hard to get an H1-B visa that a young hardworking ambitious Indian data scientist who reads this blog[2] told me about how working for a US company was "unrealistic". If most Indian-American families immigrated recently and recent Indian immigrants to the United States are smarter than average Indians then one of the following statements must be true.
Intelligence is more than 50% heritable. US immigration policy does not distinguish between heritable intelligence and non-heritable intelligence. If there is a difference in intelligence between these populations then at least 50% is likely to be heritable. I am willing to bet money that the average IQ of descendants of H1-B visa holders is (or will be) higher than the IQs of both the average American and the average Indian. I'd be happy to be proven wrong. I would prefer to live in a world without ethnic differences in IQ. But I do not believe that is the world I live in.
I would love to be proven wrong about the gender-personality connection too but that won't happen if I keep my conclusions to myself.
If you work for Cyberdyne then you should use an alias when you say bad things about Cyberdyne so you don't get fired. If you live under an authoritarian regime then you should behave appropriately. Neither excuse applies to most readers of this blog. Denying reality hurts social justice. Epistemic rationality is a prerequisite to instrumental rationality.
Source: Pew Research Center. ↩︎
People who read Less Wrong tend to be good engineers. ↩︎