It's important to notice that, contrary to Murray's self-presentation of being only concerned with facts and not causality, this _is_ implicitly making causal claims: it's just that it's specifically making claims about the "crime → imprisonment" and "intelligence → success" links in the [causal graph](https://www.lesswrong.com/posts/hzuSDMx7pd2uxFc5w/causal-diagrams-and-causal-models), while remaining silent about the equations underlying the "genes → intelligence ← environment" and "genes → intelligence ← environment" v-structures. But this is a technicality. (If someone with the requisite power and influence were trying to found a discipline of Actual Social Science in our world, one of the obvious first steps would be to hire [Judea Pearl](http://bayes.cs.ucla.edu/BOOK-2K/) as a technical reviewer for Murray.) Murray's point is just that even if the trait differences were solely caused by environmental factors like poverty or the cultural legacy of slavery, the differences would still be real and still show up in statistics produced by procedurally fair and non-racist institutions; the strawman of "And the differences are 100% genetic" is expressly not implied. (Not that "And this looks likely to be somewhere between 40–80% genetic" would be more than 40–80% less unpalatable.)
-The problem is that this causality-blindness—or not-quite-coherently _attempted_ causality-blindness—is profoundly unsatisfying. The [difference between a causal model and a statistical model](https://academic.oup.com/ije/article/45/6/1895/2999350) is _about_ how the system would respond to interventions; [this isn't something you can dodge](/2021/Feb/you-are-right-and-i-was-wrong-reply-to-tailcalled-on-causality/) if you care about policy and possibilities, rather than just summarizing static facts about a static world. Compiling the stats and arguing that the stats aren't just lies is an important service, and educational to those who aren't already familiar with the stats, but no sophisticated advocate of structural-racism theories is going to have their worldview substantially altered by this book: it's very easy to concede that the trait differences Murray documents are real (not simply made up by lying teacher and cops), but that they would vanish in a sufficiently anti-racist Society, and Murray is _explicilty_ (and reasonably, given how difficult it is) declaring that question out-of-scope. Overall, my impression of _Facing Reality_ is favorable but restrained: I keep finding myself agreeing with Murray "as far as it goes", but thinking that it just doesn't go as far as Murray seems to suggest.
+The problem is that this causality-blindness—or not-quite-coherently _attempted_ causality-blindness—is profoundly unsatisfying. The [difference between a causal model and a statistical model](https://academic.oup.com/ije/article/45/6/1895/2999350) is _about_ how the system would respond to interventions; [this isn't something you can dodge](/2021/Feb/you-are-right-and-i-was-wrong-reply-to-tailcalled-on-causality/) if you care about policy and possibilities, rather than just summarizing static facts about a static world. Compiling the stats and arguing that the stats aren't just lies is an important service, and educational to those who aren't already familiar with the stats, but no sophisticated advocate of structural-racism theories is going to have their worldview substantially altered by this book: it's very easy to concede that the trait differences Murray documents are real (not simply made up by lying teachers and cops), but that they would vanish in a sufficiently anti-racist Society, and Murray is _explicitly_ (and perhaps reasonably, given how difficult it is) declaring that question out-of-scope. Overall, my impression of _Facing Reality_ is favorable but restrained: I keep finding myself agreeing with Murray "as far as it goes", but thinking that it just doesn't go as far as Murray seems to suggest.
After this overview, let me summarize the content of this pretty-short-for-a-book in the form of a pretty-long-for-a-blog-post. After introducing our topic, Chapter 2 covers the stats on American demographics. At present, the country is about 60% white, 18% Latino, 13% black, and 6% Asian, but the, um, black-and-white framing of American racial discourse makes more sense in this historical context that there were a lot fewer Latinos and Asians before a [1965 immigration reform](https://en.wikipedia.org/wiki/Immigration_and_Nationality_Act_of_1965): in 1960, the figures were 87% white, 11% black. Big cities have become much more multiracial, whereas smaller cities and towns remain either monoracially white or biracial (the two races being white/black in the South, or white/Latino in the southwest and southern California).
Murray briefly addresses two popular (but mutually in tension) classes of objection: that the gaps will vanish with better (more equitable) education policy, and that the tests are biased. The response to the we-can-fix-it objection is basically, "We tried that and it didn't work": a lot of money and effort has been poured into attempts to narrow the racial achievement gap over the past thirty years of its stability, but we just don't know of any interventions that lastingly increase cognitive ability for _anyone_. The response to the tests-are-biased objection is basically, "We checked for that and it doesn't work": psychometricians have all sorts of technical measures to verify that their tests are doing what they think, but beyond that, if the tests were biased, you would expect them to underpredict school and workplace performance, and that's not what we see: for example, black people get worse college grades than their SAT and ACT scores would predict, not better.
-Perhaps you could argue both that the tests are racist and that teachers are _even more_ racist?—but you'd want to be specific about what alternative metric of "merit" you think they're being racist _with respect to_. The thing about tests is, a vast space of "objective" procedures can come up with a number, without giving anyone a reason to care about that particular number; if you care about group rank orderings, you _could_ come up with a measurement that gets the the group rank ordering you want. (Black people have more melanin than white people, on average! People with more letters in their name take longer to say their name out loud, on average! Cats do better than humans on a test of scratching, on average! "Hispanic" comes before "White" in alphabetical order!) The problem is that it looks like the "cognitive ability" thing that psychometricians are trying to measure is actually pretty robust abstraction that summarizes variation _in individuals_ that people care about (like the ability to master a profession); if it were just a tool of racial oppression, it's hard to see why it would work so well _within_ groups.
+Perhaps you could argue both that the tests are racist and that teachers are _even more_ racist?—but you'd want to be specific about what alternative metric of "merit" you think they're being racist _with respect to_. The thing about tests is, a vast space of "objective" procedures can come up with a number, without giving anyone a reason to care about that particular number; if you care about group rank orderings, you _could_ come up with a measurement that gets the the group rank ordering you want. (Black people have more melanin than white people, on average! People with more letters in their name take longer to say their name out loud, on average! Cats do better than humans on a test of scratching, on average! "Hispanic" comes before "White" in alphabetical order!) The problem is that it looks like the "cognitive ability" thing that psychometricians are trying to measure is actually a pretty robust abstraction that summarizes variation _in individuals_ that people care about (like the ability to master a profession); if it were just a tool of racial oppression, it's hard to see why it would work so well _within_ groups.
+
+[TODO: mention https://en.wikipedia.org/wiki/Black_Intelligence_Test_of_Cultural_Homogeneity ]
In the chapter on violent crime, Murray presents a table of black/white and Latino/white ratios of arrests for violent crimes in thirteen cities for which data was available. The median black/white ratio was 9.0 (that is, 9 black people arrested for violent crimes per 1 white person so arrested) and the median Latino/white ratio was 2.4.
[TODO: I have paperwork from being in psych prison—yes, I know, they call it a "hospital"—that says I self-presented due to thoughts of sucide. That isn't true.]
-Notice how casually Murray mentions the decision of what charges are to be filed as "a main bargaining chip in a plea bargain negotiation"! As a description of the system as it actually exists, this is perfectly accurate, but it seems important to notice that the entire concept of plea bargaining is a perversion of justice. One would have hoped for a system that proportionately punishes people for the specific crimes that they _actually did_, in order to disincentivize crime. Instead, we have the Orwellian nightmare of a system that says, "We think you're guilty of something, but it'll be easier for you if you confess to being guilty of something less bad, [and swear under oath that no one threatened you to confess](http://benjaminrosshoffman.com/can-crimes-be-discussed-literally/)."
+Notice how casually Murray mentions the decision of what charges are to be filed as "a main bargaining chip in a plea bargain negotiation"! As a description of the system as it actually exists, this is perfectly accurate, but it seems important to notice that the entire concept of plea bargaining is a perversion of justice. One would have hoped for a system that proportionately punishes people for the specific crimes that they _actually did_, in order to disincentivize crime. Instead, we have the Orwellian nightmare of a system that says, "We think you're guilty of something, but it'll be easier for you if you confess to being guilty of something less bad, [and swear under oath that no one threatened you to confess](http://benjaminrosshoffman.com/can-crimes-be-discussed-literally/) (!!)."
In calling the current system an Orwellian nightmare, I'm _not_ saying I personally know how to do better. (If abolishing the police would just result in anarchy and mob justice, that would just be a different kind of nightmare.) I'm trying to highlight how the statistics output by the actually-existing "justice" and "education" systems need to be understood as data about what's happening _within_ the current power structure managed by these systems, and shouldn't be naïvely seen as solely reflecting an independently existing reality of "education" and "crime." If there were no schools, people would still learn things (people still _do_ learn things outside of school); if there were no law enforcement, people would still take advantage of each other (people still _do_ take advantage of each other, outside the reaches of the law).
[TODO: if police use drug crimes as a plea bargaining chip because they're easier to prosecute, that may be a pragmatic adaptation to it being hard to prove real crimes, but it also sets up a situation where the police as an entity are preying on communities that use drugs]
-[TODO: summarize chapter on first-order effects of IQ]
+[TODO: credentialism (I'm not writing this for school, and you're not reading it for school, either)]
+
+[TODO: "But I digress" transition sentence]
+
+In the chapters on first-order effects, Murray lays out some of the immediate and knowable consequences of the group trait differences that have just been argued to exist at all. You might think that groups would have approximately the same IQs _within_ a profession, but this turns out not to be true: the gaps are either the same or larger! The reason probably differs by job: for jobs that most people are capable of, like being a janitor, employers value non-cognitive traits like conscientiousness more, and the IQ gap from the general population "carries through" to the profession.
+
+For jobs that usually require credentials, another explanation takes prominence. For decades, colleges have been applying pretty aggressive affirmative action policies, admitting black and Latino students with much weaker scores than white or Asian candidates: as far as admission to elite colleges goes, being black is an advantage equivalent to about 180 points on the SAT. That has ripple effects: elite colleges "snap up" the most talented non-Asian minorities, who then tend to come in lower in the class rankings—and this _within_-institution racial inequality perpetuates itself at less-elite institutions (which are playing the same game with the students who didn't get admitted by their first choice), and down the pipeline to graduate school and the professions. If schools applied the same standards to everyone, you wouldn't see this distortionary effect where the same credential means different things based on race—Murray notes that the U.S. military, which makes heavy use of aptitude testing, has been strikingly successful in this regard—but that would come at the cost of lopsided diversity numbers. For exceptionally cognitively demanding jobs, the competition for which H.R. or the admissions office are not given the power to meddle in, we already see this anyway: as [Larry Summers (in)famously observed](https://en.wikipedia.org/wiki/Lawrence_Summers#Differences_between_the_sexes), small group differences in mean or variance result in much larger differences at the extreme tails of the distribution, such that there are just many _fewer_ non-Asian minorities who can perform at the level of elite CEOs and professors.
-[TODO: summarize chapter on first-order effects of crime]
+As to crime, Murray argues that the consequences of criminality differences are a bigger deal in big cities where segregated higher-crime neighborhoods form, in contrast to smaller towns and cities that are more racially integrated simply because there aren't enough minorities to form an enclave. Businesses and real estate developers are less willing to invest in high-crime areas, and with a higher ambient threat level, police officers in those areas are incentivized to be more jumpy about the use of force.
Murray wraps up with a chapter on "If We Don't Face Reality." The facts of IQ and crime differences don't imply any particular policy, but Murray wants researchers to at least be able to control for IQ as an independent variable, and for the targets of our ongoing Cultural Revolution to offer a little more resistance. Moreover, Murray contends, identity politics is an existential threat to the American system: it was one thing when just minorities thought of themselves as collective interest groups, but if the white working class picks up the same playbook, then the ideal of individualism will be truly lost within the tides of ethnic conflict. Murray identifies eliminating government-sponsored affirmative action as politically impossible, "not within our grasp", but that a partial solution would be for those on the center-left and center-right to reaffirm the American creed and the goal of equality before the law.
-To me, this seems ... pretty naïve? I _also_ feel a lot of affection and loyalty to the previous ideological regime, but what made equality before the law such an effective marketing promise was the unstated premise that it would lead to, you know, _actual_ equality. If that's _empirically not true_, and people _aren't_ actually colorblind—if people notice ethnicity as a cluster in high-dimensional configuration space and tend to care more about people "like us"—then the classic ideal of individualist egalitarianism
+As a child of the previous ideological regime, I'm strongly in favor of this! Unfortunately, I am not feeling optimistic about the American creed's prospects. Murray notes that the ideal of individualism is unnatural—we evolved to be loyal to our ingroup and distrust outsiders. A dominant group serving its own interests at the expense of others is the natural form of government; the American experiment—to the extent it was ever real—was the exception that required careful cultivation. But part of what made equality before the law such an effective marketing promise was the unstated premise that it would lead to, you know, _actual_ equality. If that's empirically not true—if people don't _believe_ that it's true—what happens to our Society?
+
+I don't know.
+
+----
+
+... and that's the book review that I would _prefer_ to write. A social-science review of a social-science book, for social-science nerds, in a world that wasn't _about to end_. Let me explain.
-That bell can't be unrung.
+When scholars like Murray write about "intelligence", they're talking about a summary of the differences _between_ humans: we can measure how well different humans perform at various verbal or spatial or mathematical thinking-tasks, and it turns out that, on average, people who are good at one thinking-task also tend to be good at others. Graph all the test scores on an appropriately high-dimensional plot, and [the longest axis of the hyperellipsoid](/2020/Apr/book-review-human-diversity/#the-length-of-a-hyperellipsoid) represents "general intelligence"—the dimension of human variation that we recognize as "smart" _vs._ "dumb."
------
+But this particular dick-measuring contest takes place in the context of a human civilization; it doesn't tell us very much about "intelligence" as a natural phenomenon—the capacity of an agent to achieve goals across a variety of environments. Maybe some humans read better than others, but from the standpoint of eternity, reading itself is a _recent_ cultural practice [(invented only 3500 years ago)](https://en.wikipedia.org/wiki/Literacy#Prehistoric_and_ancient_literacy) that piggybacks off of natural language capabilities that _all_ developmentally normal humans share. Cats and crows and octopuses _do_ have "intelligence"—various cognitive abilities that let them integrate sensory information into a model of their environment, allocate attention, execute motor plans to seek prey or avoid predators, _&c._, but you can't give them a [Stanford–Binet IQ test](https://en.wikipedia.org/wiki/Stanford%E2%80%93Binet_Intelligence_Scales), which was designed around the _specific_ set of abilties that humans have in common. But, in principle, humans aren't special.
-... and that's the book review that I would _prefer_ to write. A social-science review of a social-science book, for social-science nerds, in a world that wasn't _about to end_.
+And yet—it seems like humans _are_ special, in some ways. Of all the creatures on [the tree of life](https://en.wikipedia.org/wiki/Tree_of_life_(biology)), our lineage "took over the world" in the sense that if humans want a resource that cats or crows or octopuses are using, the nonhuman side of the ensuing conflict is predictably going to lose. (To the extent that we don't usually think of ourselves as engaging in a "conflict". Animals aren't _enemies_; they're just in the way.) This is not because humans are stronger or have sharper teeth than other creatures, but because of something about our "intelligence" in the natural-phenomenon sense, not the IQ test variation sense. It's not even necessarily about _individual_ human intelligence being a particularly formidable force: given no tools and no friends, and confronted by a hungry lion at ten paces, it doesn't seem easy to survive by thinking of some incredibly clever plan. If you had a gun, you could shoot the lion, but [no one individual knows how to make a gun](https://en.wikipedia.org/wiki/I,_Pencil) all by themselves, starting from nothing.
+Rather, the power of humanity over the rest of the biosphere has to do with our species having evolved a suite of cognitive capabilities adequate to support the accumulation of culture and technology. If you already had a spear, you might be able to think of of some incredibly clever plan to get a slightly sharper spear—which everyone in your tribe could imitate. And so on up the tech tree.
+
+[TODO: ↓ clunky prose probably needs a rewrite after better outlining]
+
+You can think of events on Earth before the rise of human civilization as mostly being shaped by evolution by natural selection: new complex ordered phenomena arose as the product of genetic mutations that allowed their bearers to survive and reproduce, thereby increasing the frequency of the mutation. Natural selection is a form of _optimization_: the accumulation of beneficial mutations creates functionality that looks "designed" for the purposes of reproduction, because they were selected for existence on that basis.
+
+But _after_ the rise of civilization, biological evolution stopped being the dominant force shaping planetary events, just because cultural evolution runs on a faster timescale. If some other species were on the evolutionary path towards developing the capabilities that would eventually result in them developing their own civilization, it basically "wouldn't matter".
+
+To the extent that our civilization is better for us to live in than the state of nature, it's because civilization is the product of the cumulative optimization of humans trying to acheive their goals: vast, complex infrastructure and economies look "designed" to cater to human needs—supermarkets to feed us, hospitals to heal us, cars and airplanes to take us where we want to go—because we selected them for existence on that basis.
[What I actually care about: the intelligence explosion; the alignment problem might be solvable by von Neumann clones; a civilization at our tech level with a more functional state religion that wasn't afraid of genetics would be able to produce them faster than we will; this is really bad guys]
[the Georgetown law professor thing: "I end up having this angst every semester that a lot of my lower ones are Blacks."—this is actually a very straightforward consequence of affirmative action (even if you don't believe in a genetic gap); it should be possible to reason about
https://www.insidehighered.com/news/2021/03/12/georgetown-terminates-law-professor-reprehensible-comments-about-black-students
-]
+
+
"rotation" of the class axis, as people sort by cognitive ability directly rather than race??
For the Singularity https://www.overcomingbias.com/2021/05/the-artificial-life-taboo-in-biology.html
+eugenicists treated blindness as hereditary!!!!
+https://www.datasecretslox.com/index.php/topic,4082.0.html
+
From my notebook circa November 2007—
I think I'm coming to terms with the fact that I'm male all the way down, no matter what I say or do or even feel. This is hard to deal with because my <<policies are based on the fact that>> I'm special or the Difference is trivial. <<Whereas all the testing says not really.>> (But I've _always_ been <<inherently gloomy>>)
> I don't want to hear about black thugs raping white women, but I'm going to hear it soon if you don't stop talking about white cops killing black kids
Glenn Loury on Tucker https://www.youtube.com/watch?v=wuPdPKHG18k
+https://www.cold-takes.com/why-ai-alignment-could-be-hard-with-modern-deep-learning/
+
https://johnmcwhorter.substack.com/p/why-charles-murrays-new-book-is-his
Coleman Hughes: it's possible to push back against Kendism without
https://www.researchgate.net/publication/354010004_Genetic_Ancestry_and_General_Cognitive_Ability_in_a_Sample_of_American_Youths
+only 14% approval for IQ enhancement https://www.pewresearch.org/science/2020/12/10/biotechnology-research-viewed-with-caution-globally-but-most-support-gene-editing-for-babies-to-treat-disease/
+
"Human capital mediates natural selection in contemporary humans"
https://ueaeco.github.io/working-papers/papers/ueaeco/UEA-ECO-21-02_Updated.pdf
I feel a lot of sympathy for the defendant
]
+
+[TODO: compare Great Oxidization Event and its timescale to the Industrial Revolution??]
+
+"Genetic and environmental contributions to IQ in adoptive and biological families with 30-year-old offspring"
+https://www.sciencedirect.com/science/article/abs/pii/S0160289621000635
+
+Modernity selects against those capable of maintaining it
+https://twitter.com/CovfefeAnon/status/1441997339309314049
+
+https://www.nytimes.com/2021/10/08/nyregion/gifted-talented-nyc-schools.html
+
+https://whyevolutionistrue.com/2021/10/09/the-art-institute-of-chicago-fires-all-122-of-its-unpaid-and-volunteer-docents-because-they-arent-sufficiently-diverse/
+
+NCAA Task Force Recommends Students No Longer Submit SAT, ACT Scores To Promote 'Racial Equity'
+https://archive.ph/QerlW
+
+> The American Medical Association has just released "Advancing Health Equity: A Guide to Language, Narrative and Concepts," a strange document that calls for doctors to insert progressive politics into even plain statements of fact.
+https://twitter.com/jessesingal/status/1454468272011743239
+https://www.ama-assn.org/system/files/ama-aamc-equity-guide.pdf
+
+https://www.latimes.com/california/story/2021-11-18/uc-slams-door-on-sat-and-all-standardized-admissions-tests
+
+https://slate.com/human-interest/2021/11/overstepping-auntie-relationship-parenting-advice.html
projects / Ultimately_Untrue_Thought.git / blobdiff