Comment by KennyBlanken | Hacker News Reader

Comment by KennyBlanken | Hacker News Reader

Parent Full thread

It's emotionally charged because the author is Ashutosh Jogalekar, who in the mid 2010's got himself fired from Scientific American as a blogger by praising the theories of a scientific-racism book:

https://www.washingtonpost.com/lifestyle/style/whats-going-o...

> Jogalekar praised the book, saying it confirms the need to “recognize a strong genetic component to [social and cognitive] differences” among racial groups.

It should be noted that Jogalekar is a chemist.

What did the people whose work Wade cited for his book say?

> "Wade juxtaposes an incomplete and inaccurate account of our research on human genetic differences with speculation that recent natural selection has led to worldwide differences in I.Q. test results, political institutions and economic development. We reject Wade's implication that our findings substantiate his guesswork. They do not. We are in full agreement that there is no support from the field of population genetics for Wade's conjectures."

Instead of keeping his head down, Ashutosh followed up shortly thereafter with a blog falling over himself to find reasons to excuse or explain Feynman's rampant history of womanizing and general asshole-ism:

https://blogs.scientificamerican.com/the-curious-wavefunctio...

...and then we have Jogalekar defending EO Wilson, whose book Sociology, proposed (in 1975) that human behavior was determined by genetics, was widely criticized as sympathetic to eugenics. Sound familiar?

Ask yourself this: why does Jogalekar keep defending authors of works that tread around scientific racism and eugenics?

Just because you call it "scientific racism" and "eugenics", doesn't mean it's wrong or untrue, just that you don't like it.

It absolutely is wrong and untrue, though, in addition to being racist. There being significant IQ differences between ethnicities/"races" that are caused by genetics is a long debunked theory; just look at the scientific discourse around something like "The Bell Curve".

It's very politicized, that makes it hard to judge who is doing real science and who is advancing a political agenda.

One thing I wonder is what makes intelligence "special" so that there are no differences, as compared to other physical traits where you do see differences between ethnicities?

> It's very politicized, that makes it hard to judge who is doing real science and who is advancing a political agenda.

The science is clear on this question, and has been for decades; there never has been any scientifically well-supported theory showing genetic causes for observable differences of IQ between ethnicities.

> One thing I wonder is what makes intelligence "special" so that there are no differences, as compared to other physical traits where you do see differences between ethnicities?

The position isn't even that there are no differences in IQ between populations; the well-supported scientific consensus is

a) IQ differences between individuals of a population are larger than between populations in the first place, and

b) IQ differences between populations are largely not caused by genetics.

Note that neither of these statements contradict the heritability of IQ, which a statistical effect.

> a) IQ differences between individuals of a population are larger than between populations in the first place, and

What does this mean? I assume I am interpreting it wrong, because it just seems like nonsense.

The fact that IQ differences between individuals in a population is larger than the alleged IQ differences between populations seems totally obvious. Any given population probably contains some disabled people and also some geniuses, for, let's say, a 50-point spread. But it would be a meaningful population-level difference if the average Finn was 100 but the average Swede was 110. Or are there people out there seriously arguing that there are peoples out there totally consisting of super geniuses or extremely disabled people?

If you mean that there’s no evidence of a difference in iq distribution in different populations, I think it’s clearer just to say that.

One particular (I think well founded) worry about the hereditarian position is that if genetic differences between ethnicities are admitted, people start thinking about stuff like, say, an active "blackness" that drags people down, and that they'd start insisting that every black person has to be dumb or something like that. That people would stop really meeting people as individuals (woke blank slatism, of course, has that exact effect already, which is partly why it sucks).

It's somewhat hard to get people to dispassionatedly understand that all the hereditarian position being true would mean that what differed between populations would be the counts of people with a certain levels of ability, and that's it. Dunk a smart black man, a smart white man, a smart East Asian and a smart Ashkenazi in green paint and you have four smart green men unburdened with any mystical force acting on their mental faculties.

Another possibility would be fears of things like "eh, they're dumb anyway" to justify giving up on people with a certain skin color because they won't achieve highly anyway, which would be heinous.

"Well-supported scientific consensus" and "scientifically well-supported theory" are doing a lot of work when the loudest enforcers of the "consensus" consistently operate in dishonesty for strategic reasons, see e.g. the 2019 Cofnas paper: https://www.tandfonline.com/doi/pdf/10.1080/09515089.2019.16...

> there never has been any scientifically well-supported theory showing genetic causes for observable differences of IQ between ethnicities

I feel like this is the key to your argument, but I'm having a hard time understanding what you're trying to say. Theories don't show causes, they attempt to fit data into a mental model that can then be falsified or fail to be falsified.

Are you saying we should reject data because a satisfying theory hasn't been found yet? I don't agree with that principle. In my view, that seems analogous to rejecting the idea of gravity because we don't have a good theory of how it works, or that the data leading to the string theory hypothesis should be discarded because we can't test and falsify string theory.

And yet a recent study showed at least some is heritable. And this whole “between” populations nonsense doesn’t mean different populations have the same mean.

So some variation in IQ is caused by genetics, but to a very small degree compared to other factors. Am I interpreting (a) and (b) correctly?

Intelligence is very highly heritable, at least within populations.

Apparently yes, sound like the wage gap

  - differences between individuals of a population are larger than between populations in the first place
  - differences between populations are largely not caused by *sexism*

It makes no sense to posit that IQ differences between population have a different cause than IQ differences between individuals. Adoption raises the IQ by at most 5 points ! What environmental difference could have a much more dramatic effect than that (between population IQ difference are up to 15 pts) ?

> One thing I wonder is what makes intelligence "special" so that there are no differences, as compared to other physical traits where you do see differences between ethnicities?

IQ theory has the following problems:

1. Intelligence is not the same as height, as one of these is low-dimensional and mechanical, and the other is high-dimensional and very complex.

2. There are problems defining what intelligence is.

3. The "intelligence" as measured by IQ is problematic because it's engineered to follow a normal distribution. This then contradicts claims by IQ-ists that IQ predicts things which are very much non-normally distributed.

4. Many of the apparent successes of IQ theory can be explained by saying that low IQ relative to a surrounding population is indicative of a learning disability. This doesn't make higher-than-average IQ useful though.

5. People who score highly in IQ tests are interested in puzzles and seeing patterns in things. This is not necessarily a useful skill in real life, and may even be detrimental sometimes.

6. IQ-ists claim to be dispassionately following what the science says, but are generally incompetent at the fundamentals of science. For instance, they use measures of statistical association like Pearson Correlation Coefficient, which have been abandoned in all the hard sciences. These old-fashioned methods of statistical association are very brittle, and are easy to misuse. They appear to be ignorant of more modern and robust statistics. When their data has been checked using more modern methodology, their conclusions don't hold up.

7. IQ-ists claim to be dispassionately following what the science says, but their mistakes have serious consequences for people's lives. It can be used to justify racism and eugenics.

Good lord, listen to yourself.

> People who score highly in IQ tests are interested in puzzles and seeing patterns in things. This is not necessarily a useful skill in real life, and may even be detrimental sometimes.

Have you been outside in the last thirty years? "Real life" has been completely designed by people who are good at puzzles and patterns! Grocery store checkout machines, venture capital, social media, automobiles, the internet--all of these are puzzles and patterns. The better you are at puzzles and patterns, the better you'll be at "real life".

There is a difference between discussing the weaknesses of IQ theory and completely ignoring reality in favor of making sure you can't be accused of "racism or eugenics". You sound like someone in an authoritarian religious state making fervent declarations of whatever the parish is telling people to say. I agree that IQ does not exist, it is an abstract metric invented a while ago--but on the whole, people who score highly on IQ tests are going to have a much, much easier time succeeding in the modern puzzle-and-pattern reality.

Not disputing any of your other point but this one:

> 1. Intelligence is not the same as height, as one of these is low-dimensional and mechanical, and the other is high-dimensional and very complex.

Is way off. Height is influenced by a huge number of factors. It’s not just some simple genetic program.

I think you have their point a bit backwards: they're not saying the causes of height are simple, but rather that the output is simple to measure. If you measure 5 groups and one is 20% taller than the rest, that's pretty clear evidence of something. You get one simple number, one dimension.

Intelligence is much squishier because IQ is a suspect measure at best. If a group scores 20% higher on some intelligence test, is that because of genetics? Cultural focus on learning the specific skills measured by this test? A biased test? Completely different methods of testing being falsely compared? There's no real meterstick for intelligence like there (literally) is for height, so trying to extract a number you can throw into a statistical analysis is a lot more fraught, if it's possible at all.

Intelligence is absolutely trivial to measure. You just make a huge pile of tests of mental ability, the more the merrier. It won't be very sophisticated, but you have an intelligence test. Intelligence is so easy to measure in a rough way that people have accidentally built intelligence tests when they were trying not to.

A lot of the other critiques are at least somewhat suspect: Modern intelligence tests explicitly try to use tasks that you won't train yourself that much in in daily life and that try to eg. stress working memory capacity directly without testing for other skills. There's also evidence that training for some kind of skill like strategies for n-back numbers doesn't improve performance on similar tasks using letters, for example.

Yes, it is not as good of a measure as a tape measure is, everyone knows that. But saying it's suspect and not useful is kind of like saying computer benchmarking is worthless and won't tell a crap machine from a powerful one.

> they're not saying the causes of height are simple, but rather that the output is simple to measure.

By talking about the dimensionality, they are saying both.

I agree that measuring height is easier, but I disagree that it’s causally simpler, and that is frankly what matters.

Is someone taller because of genetics or nutrition? Is there a nutrition difference because of culture, or because of racism, or because of wealth or because of IQ? Etc. Etc.

> … if it's possible at all.

It sounds like you are someone who doesn’t believe intelligence can be measured.

That's fair. I think I actually remember reading somewhere that worldwide height is catching up with European height, and what was posited to be genetic differences was actually due to nutrition. Don't have a source at hand, though.

> It sounds like you are someone who doesn’t believe intelligence can be measured.

I'm certainly skeptical about the utility of IQ or any other test as a general measure of "intelligence", when what intelligence even means is an active matter of debate.

> I'm certainly skeptical about the utility of IQ or any other test as a general measure of "intelligence", when what intelligence even means is an active matter of debate.

That's partly because it's mythologized to hell and back, and people often turn to the lack of a definition given concisely in words as proof that we don't really know it.

But, g exists, we know its nature rather well, and being able to do hang on to that measure has enabled considerable detective work as to the physical underpinnings of it.

Most simply, I think treating it as a brain performance benchmark, just like you would for a computer, doesn't go very far wrong. It's variation in the general monkey brain blueprint we're all built from, and has physical correlates that imply it as a rather general measure of the brain's performance (it's eg. linked to faster average reaction times, which would make no sense for a measure of book-learning aptitude). So, more IQ = monkey with more CPU cycles and RAM, done. Zero mythology, zero romanticism, simple.

As one analogy, you could liken it to gravity: For the longest time we knew things fell down, we could measure that speed, etc. but we didn't know why things fell down. Then we made some theories, and then new ones, and then new ones. But we don't actually know the cause and the mechanism completely. We know mechanisms of how it acts and measures of how strongly, but stuff like a quantum mechanical theory of gravity is anyone's guess at this point.

All along that journey of discovery, we still knew a fundamental fact: Somehow things fell down. That observation is the anchor. In intelligence research, that observation is g.

> when what intelligence even means is an active matter of debate.

Is there anything in social science that isn’t an active matter of debate?

1 & 2 are very intertwined I think. Height just happens to be easily measured.

3. You mean because IQ is assumed to be normally distributed in a population? What are you thinking of when you say "This then contradicts claims by IQ-ists that IQ predicts things which are very much non-normally distributed."? I don't think IQ distributed normally means that you can't have non-normally distributed consequences in other areas?

I disprove you thusly: "IQ-ists"

Yes hmm I wonder what. I wonder if maybe there have been any events in the last century that would make people touchy about this.

> any events in the last century

Not sure which events specifically that you're referring to, but usually when somebody refers abstractly to "events in the last century", they mean the nazis. It's worth noting that the nazis themselves rejected IQ testing and theories as justifying "Jewish supremacy".

Actually I rewrote that comment to not refer to any specific events because I figured whichever one I chose someone would find a way to use it to invalidate the entire concern.

Anyway! The world isn't neatly divided into "pure Truth, delivered to us by Science" and "other things, which can be used to do harm."

This doesn't make potentially dangerous research subjects verboten or banned! But the people doing that research should have a solid understanding of who has an interest in their work, why, and what it may be used for. Having that understanding isn't "advancing a political agenda" I don't think.

> Having that understanding isn't "advancing a political agenda" I don't think.

You're reading a bit too much between the lines here. I didn't mean any specific side, I assume both sides engage in this so it's meant in a general sense.

Do you actually thing that the anti-IQ crowd is mad that the nazis (notoriously anti-IQ) lost WWII and that is the reason they are so touchy about this ?

As a Jew, I will tell you that the Nazis didn’t care about our IQ. Many Jews proudly used their brains to help Germany in WWI (Fritz Haber comes to mind) and yet arguably Germany lost because of the brain drain which led directly to our atomic bomb, among other things, that might have helped them.

They didn’t care about whether we could run a mile or win at athletics. It was ancient ethnic hatred only many centuries earlier based on religion.

Saying that different populations have different IQs could surely be a post how justification for that kind of hatred, but it could also be justification for amelioration of the non-genetic effects through social programs.

Right, and this is exactly my concern. Not that this sort of research will uncover new information that will make people form new prejudices. But that they will use this research to justify acting on the prejudices they already have.

I didn't mention nazis and they are just one example. If you pare away the context then sure, this could go either way, it could be used to reduce these divisions. But the history of scientific racism clearly shows another pattern.

And therefore we are all the same above the neck.

Among “races”? I think that’s debunked, but among populations?

What exactly is wrong and untrue, E. O. Wilson's work?

Reply on news.ycombinator.com