I.
From Identity, Personal Identity, and the Self by John Perry:
“There is something about practical things that knocks us off our philosophical high horses. Perhaps Heraclitus really thought he couldn’t step in the same river twice. Perhaps he even received tenure for that contribution to philosophy. But suppose some other ancient had claimed to have as much right as Heraclitus did to an ox Heraclitus had bought, on the grounds that since the animal had changed, it wasn’t the same one he had bought and so was up for grabs. Heraclitus would have quickly come up with some ersatz, watered-down version of identity of practical value for dealing with property rights, oxen, lyres, vineyards, and the like. And then he might have wondered if that watered-down vulgar sense of identity might be a considerably more valuable concept than a pure and philosophical sort of identity that nothing has.
Okay, but I can think of something worse than that.
Imagine Heraclitus as a cattle rustler in the Old West. Every time a rancher catches him at his nefarious business, he patiently explains to them that identity doesn’t exist, and therefore the same argument against private property as made above. Flummoxed, they’re unable to think of a response before he rides off into the sunset.
But then when Heraclitus himself needs the concept of stable personal identity for something – maybe he wants to deposit his ill-gotten gains in the bank with certainty that the banker will give it back to him next time he shows up to withdraw it, or maybe he wants to bribe the sheriff to ignore his activities for the next while – all of a sudden Heraclitus is willing to tolerate the watered-down vulgar sense of identity like everyone else.
(actually, I can think of something even worse than that, which is a TV western based on this premise, where a roving band of pre-Socratic desperadoes terrorizes Texas. The climax is no doubt when the hero strides onto Main Street, revolver in hand, saying “There’s a new sheriff in town.” And Parmenides gruffly responds “No, I’m pretty sure that’s impossible.”)
At its best, philosophy is a revolutionary pursuit that dissolves our common-sense intuitions and exposes the possibility of much deeper structures behind them. One can respond by becoming a saint or madman, or by becoming a pragmatist who is willing to continue to participate in human society while also understanding its theoretical limitations. Both are respectable career paths.
The problem is when someone chooses to apply philosophical rigor selectively.
Heraclitus could drown in his deeper understanding of personal identity and become a holy madman, eschewing material things and taking no care for the morrow because he does not believe there is any consistent self to experience it. Or he could engage with it from afar, becoming a wise scholar who participating in earthly affairs while drawing equanimity from the realization that there is a sense in which all his accomplishments will be impermanent.
But if he only applies his new theory when he wants other people’s cows, then we have a problem. Philosophical rigor, usually a virtue, has been debased to an isolated demand for rigor in cases where it benefits Heraclitus.
A fair use of philosophical rigor would prevent both Heraclitus and his victims from owning property, and thus either collapse under its own impracticality or usher in a revolutionary new form of economic thinking. An isolated demand for philosophical rigor, applied by Heraclitus to other people but never the other way around, would merely give Heraclitus an unfair advantage in the existing system.
II.
A while ago I wrote a post called Military Strikes Are An Extremely Cheap Way To Help Foreigners which was a response to a Matt Yglesias post called the opposite. Yglesias was opposed to “humanitarian” military intervention (think the air strikes on ISIS going on right now, justified under the cause of preventing a genocide) and his argument was that this was extremely cost-ineffective compared to just giving the money to GiveWell’s top-rated charity – at the time he was writing, malaria prevention.
I argued he was wrong about his numbers. But I also argued he was unfairly making an isolated demand for philosophical rigor.
Once you learn about utilitarianism and effective charity, you can become the holy madman, donating every cent you have beyond what is strictly necessary to survive and hold down a job to whatever the top rated charity is.
Or you can become the worldly scholar, continuing to fritter away your money on things like “hot water” and “food other than gruel” but appreciating the effective-utilitarian perspective and trying to make a few particularly important concessions to it.
Or you can use it to steal other people’s cows. This is what I accused Matt Yglesias of doing. Presumably there are lots of government programs Yglesias supports – I suggested PBS – and he would never dream of demanding that we defund them in the hopes of donating the money to malaria prevention. But if for political reasons he doesn’t support air strikes, suddenly that plan has to justify itself according to rigorous criteria that no government program that exists could possibly pass.
Government spending seems to be a particularly fertile case for this problem. I remember hearing some conservatives complain: sex education in public schools is an outrage, because my tax dollars are going to support something I believe is morally wrong.
This is, I guess, a demand for ethical rigor. That no one should ever be forced to pay for something they don’t like. Apply it consistently, and conservatives shouldn’t have to pay for sex ed, liberals shouldn’t have to pay for wars, and libertarians shouldn’t have to pay for anything, except maybe a $9.99 tax bill yearly to support the police and a minimal court system.
Applied consistently, you become the holy madman demanding either total anarchy or some kind of weird system of tax earmarks which would actually be pretty fun to think about. Or the worldly scholar with a strong appreciation for libertarian ideas who needs a really strong foundational justification for spending government money on things that a lot of people oppose.
Applied inconsistently, you’re just stealing cows again, coming up with a clever argument against the programs you don’t like while defending the ones you do.
III.
But this is the sort of uncouth behavior we expect of political partisans. What about science?
Suppose there are scientists on both sides of a controversial issue – for example, economists studying the minimum wage. One team that supports a minimum wage comes up with a pretty good study showing with p < 0.05 that minimum wages help the economy in some relevant way. The Science Czar (of course we have a science czar! We're not monsters!) notes that p < 0.05 is really a shoddy criterion that can prove anything and they should come back when they have p < 0.01. I have a huge amount of sympathy with the Science Czar on this one, by the way.
Soooo the team of economists spends another five years doing another study and finds with p < 0.01 that the minimum wage helps the economy in some important way. The Science Czar notes that their study was correlational only, and that correlational studies suck. We really can't show that minimum wages are any good without a randomized controlled trial. Luckily, the governments of every country in the world are totally game for splitting their countries in half and instituting different economic regimes in each part for ten years, so after a decade it comes out that in the randomized controlled trial the minimum wage helped the economy with p < 0.01. The Science Czar worries about publication bias. What if there were a lot of other teams who got all the countries in the world to split in half and institute different wage policies in each of the two territories for one decade, but they weren't published because their results weren't interesting enough? Everything the Science Czar has said so far makes perfect sense and he is to be commended for his rigor and commitment to the job. Science is really hard and even tiny methodological mistakes can in principle invalidate an entire field.
But now suppose that a team shows that, in a sample of six restaurants in Podunk Ohio, there was a nonsignificant trend towards the minimum wage making things a little worse.
And the Science Czar says: awesome! That solves that debate, minimum wage is bad, let’s move on to investigating nominal GDP targeting.
Now it looks like the Science Czar is just a jerk who’s really against minimum wage. All his knowledge of the standards of scientific rigor are going not towards bettering science, but toward worsering science. He’s not trying to create a revolutionary new scientific regime, he’s taking pot shots.
I see this a lot in medicine. Someone jumps on a new study showing the selenium or chromium or plutonium or whatever cures cancer. It is brought up that no, really, the medical community has investigated this sort of thing before, and it has always been found that it doesn’t.
“Well, maybe the medical community wasn’t investigating it the right way! Maybe the investigators were biased! Maybe they didn’t randomize right! Maybe they used a population unusually susceptible to cancer-getting! Ninety percent of medical studies are wrong! Those twenty experiments showing a lack of effect could be total bunk!”
Yes, maybe these things happened in each of the twenty studies that disagree with you.
Or maybe they happened in the one contrarian study you are getting so excited about.
IV.
The unholy combination of isolated demands for philosophical rigor and isolated demands for scientific rigor is isolated demands for mathematical-statistical-conceptual rigor, ie the sort of thing this blog has been talking about all week.
I have already been made fun of for how many different things I am metaphorically comparing IQ to – speed, blood pressure, comas – so I guess it can’t hurt to add another example I only thought of today. How about crime? It’s usually measured by crime rate – a made-up statistic that combines subfactors like arson (maybe higher when fire insurance pays out better), property damage (maybe higher during periods of ethnic tension and frequent riots) and theft (maybe higher when income inequality is worse). There is assumed to be a General Factor Of Crime (presumably caused by things like poor policing, dark alleys, broken families, et cetera) but I would be extremely surprised if anyone had ever proven Beyond A Shadow Of A Doubt that the factor analysis works out here.
When Cosma Shalizi says he’s not sure about the factor analysis in IQ, I have no quarrel with him, because Cosma Shalizi’s response to everything in the world is to glare at it for not being sufficiently statistically rigorous.
But when other people are totally happy to talk about speed and blood pressure and comas and the crime rate, and then suddenly switch to a position that we can’t talk about IQ at all unless we have a perfect factor-analytical proof of its obeying certain statistical rules, then I worry they’re just out to steal cows.
Likewise, if someone were to just never acknowledge any sorts of groups of objects except those that could be statistically proven to fall out into absolutely separate clusters in which variance within each cluster is less than variance between clusters, well, at least they would be fun to talk to at dinner parties.
But when people never even begin to question the idea of different cultures but make exacting demands of anyone before they can talk about different races – even though the two ideas are statistically isomorphic – then I think they’re just out to steal cows.
So this is another technique for avoiding Eulering – is your interlocutor equally willing to apply their complex mathematical argument to everything else.
I think if I hadn’t known anything about Bayesian probability, I would have examined the McGrews’ Bayesian argument for the Gospels by seeing if it applied equally well to Mormonism, the control group for Christianity.
V.
The old man stamped his boot in the red dirt, kicking up a tiny cloud of dust. “There’s a new sheriff in town,” he told them.
“No, I’m pretty sure that’s impossible,” says Parmenides. “There’s no such thing as change, only the appearance thereof.”
“Well then,” says the old man, “I reckon you won’t mind the false illusion of your surroundings appearing to change into a jail cell.” And he took out his six-shooter and held it steady.
“Hold on,” said Thales. “We don’t want any trouble here. All is water, so all we did was steal a little bit of water from people. We can give you some water back, and everything will be even, right?” He gestured to a watering trough for horses on the side of the street, which was full of the stuff.
“Just so long as you don’t mind being sprayed with some very hard water from my squirt gun,” the old man answered, and the six-shooter was pointed at the Milesian now.
“Ha!” said Zeno of Elea. “You don’t scare us. In order to hit Thales, your bullet would have to get halfway to him, then half of the remaining distance, and so on. But that would require an infinite number of steps, therefore it is impossible.”
“Sorry,” said the old man, “I couldn’t hear you because it’s logically impossible for the sound waves encoding your speech to reach my ears.”
“We’re not even the same people as the guys who stole those cattle!” said Heraclitus. “Personal identity is an illusion!”
“Then you won’t mind coming to the courthouse with me,” replied the old man “to help the judge imprison some other people who look just like you.”
The last of them, the tall one, said nothing. He just raised his revolver in a fluid motion and shot at the old man.
The old man saw it coming and jumped out of the way. The air was briefly full of bullets. Bang! Thales went down! Bang bang! Heraclitus! Bang bang! Parmenides and Zeno. Bang bang bang! The old man was hit in the arm, but still standing. Bang bang bang bang…
It was just the old man and the tall one now. The tall one picked up his gun and fired. Nothing happened. Out of bullets.
The old man smiled wryly, his six-shooter still in his hand.
“I know what you’re thinking. You’re thinking – did he fire six shots, or only five? Well, you’ve got to ask yourself a question – do you feel lucky? Well, do you, punk?”
The tall one didn’t budge. “Man is the measure of all things,” said Protagoras. “If I believe you fired six shots, then by my personal epistemic standards, you fired six shots.”
The old man didn’t say anything.
“You see,” the Sophist continued. “Out of all of them, I alone was truly consistent. They all came up with clever theories, then abandoned them whenever it conflicted with their self-interest. I was more honest. I just said at the beginning that my self-interest determined truth, and so never suffered any temptation to depart from my position.”
The old man took off the bandana covering his face. “Man may be the measure of all things. But I’ve taken your measure, Protagoras, and found it wanting.”
“Socrates?!” the Sophist gasped.
“The only truly consistent people are the dead, Protagoras,” he said – and squeezed the trigger.
s/Protagoras”/Protagoras,”/ and also, oh my god, this may be the best essay to come out of the rationalist community yet 😀
I’m surprised you didn’t call out Eliezer, who applies the scientific principle to everything except for his favorite QM interpretation, where Bayesian inference obviously trumps the “Eld Science”.
Does it? As far as I can see WMI is simply assuming that physics keeps working the same way even if we can’t check. Would the Copenhagen interpretation from the point of view of an observer totally isolated from the rest of the universe be any different from the WMI?
I assume that can’t be what the grandparent is talking about, because I care little for Bayesianism but MWI remains obviously true, and I see no connection between the two.
I’m afraid it is. Shminux/Shmi Nux has a hornet in eir bonnet about this particular issue.
I have a question to those who think that MWI is obviously true, or that the sentence “MWI is true” actually means anything at all.
Do you also believe it makes sense to wonder which gauge fixing condition is true in electrodynamics? Say someone on the Internet argued that the Lorenz gauge is the obviously true one, and denounced physicists for not having exposed the Coulomb gauge for the sham it is. Would that person be making any sense? If not, what is the difference? And which of the arguments EY uses to defend his position as non-meaningless, from chocolate cakes in the rings of Saturn to spaceships outside of the observable universe, cannot be equally as well used to defend non-meaninglessness of the Lorenz gauge religion?
I am also surprised by this claim.
Bro do you even Bayes? Elly doesn’t “apply the scientific method to everything except QM”, he applies Bayes to everything, and the “scientific principle” is a lossy, dumbed-down version of Bayes with a bunch of social rituals sprinkled on top to make it work as a social practice. Bayes in the lambda calculus to science’s Java.
Exactly. Eliezer thinks of the social process of science as an pragmatic approximation to Bayesian reasoning, much like a man might calculate the physics of a bullet using Newtonian mechanics even though he knows that Einsteinian mechanics are more accurate. Most of the time they will both give the same answer (Reverend Bayes isn’t going to show up tomorrow and tell you that the theory evolution is false), but in edge cases it should be possible for a Bayesian to outperform idealized science. The correct interpretation of quantum mechanics is supposedly one such instance, in which science has been too slow and remained carefully agnostic on an issue which a Bayesian reasoner would declare to be solved.
Which would turn out to be really funny if we find some other, deeper physics lurking behind quantum mechanics, which would not admit an MWI.
the “scientific principle” is a lossy, dumbed-down version of Bayes with a bunch of social rituals sprinkled on top to make it work as a social practice.
The modern scientific method is actually an improvement over what you call “Bayes”, made to correct for the fact that humans are fallible. If humans were free of biases and not prone to seeing patterns where there are none, we wouldn’t need the concept of falsification. After all, an experiment which has already been performed gives us exactly the same amount of information as one which will be performed in the future. But the way actual human brains work, it is much easier to derive the results of yesterday’s experiment from your theory, than to predict the results from tomorrow. If experimenters never made mistakes and never squinted at their apparatus until it showed what they knew had to be correct, then we could also do away with the requirement of reproducibility.
What you refer to as “Bayes” had actually been the standard way of thinking about the scientific method in the past. A scientist was supposed to look at all available experimental facts and derive a theory from them by a process called “induction”. It didn’t always work so well and that led Karl Popper to formulating the scientific method as we know it today.
Yeah. I didn’t say anything in disagreement with this.
Speaking of which, the most purified case of Eulering is Eliezer’s use of Solomonoff induction to support his beliefs on metaphysical topics, from the (non)existence of God to the interpretations of Quantum Mechanics.
Everyone generally agrees that simpler explanations are better. The problem is, for some people God is the simplest explanation in the world, and for others quite the opposite, and ditto for the MWI. That’s where EY triumphantly exclaims, “Ha! But thanks to Solomonoff induction, we can precisely state what is really the simplest.”
I happen to know what Solomoff induction means. It is a method of assigning Turing machines (or weighted combinations of Turing machines) to strings of bits, which can be thought of as one way to formalize the intuition of a “simple explanation”. There is nothing at all in it which could possibly have any relevance to the existence of God or the interpretations of QM. He might as well be saying, “(a+b^n)/n = x, therefore, MWI is correct”.
I don’t know anything about Solomonoff Induction beyond how Eliezer has explained it. But if Eliezer is explaining it wrong, then the thing Eliezer was explaining, the thing that’s not Solomonoff Induction, that is the correct answer to the question of why God isn’t simple.
Like the argument that the simplest explanation for everything is always “The woman down the street is a witch, she did it” makes it obvious that our common-sense notions of simplicity don’t work, and that we need to switch to simplicity in a more formal and mathematical sense.
The way Eliezer describes Solomonoff Induction seems to me like a perfect example pointing at what it would mean to have a more formal and mathematical sense of simplicity. If his particular example doesn’t work, it seems obvious to me that something else along those lines would.
A priest can say “God is perfectly simple,” and it sounds common-sensically correct. But everyone agrees that it is very hard to program a computer to perfectly simulate God.
Likewise, it is about as easy to get a computer to print the string “universe” one billion times as it is to make it print the string once, which means that just having more of something doesn’t raise its “complexity”.
I don’t agree that it is very hard to program a computer to perfectly simulate God. Since God doesn’t do anything, it doesn’t seem hard at all. Note this is not a flippant answer. This is the crux of the problem. The reason we are even talking about this is that Christianity (MWI) makes no predictions which would allow us to test it experimentally. If the Catechism of the Catholic Church stated that every time holy water is electrolyzed an angel with the flaming sword appears, then instead of talking about Solomonoff induction we would be out buying batteries and salt.
Let me go over the argument. I’m going to describe it without math, so it will be a simplified version, but I don’t think it affects the validity of the argument.
Say we are looking for a sequence of bits of letters or whatever. We can consider all computer programs that output this sequence. We pick the shortest of these programs. For example, if our sequence is the first 100k digits of pi, then the winning program will probably be one that uses some formula for generating digits of pi, and not one with hardcoded PRINT “3.141592…”. We nominate that program as the simplest possible explanation for our data[1].
Now let’s say we’ve encoded all our observations of the world as a sequence of bits. Then, it would make sense to call the winner program the simplest model of the world that agrees with all observations. We could also call it the most Occam-razor compliant scientific theory.
We don’t require that the candidate programs stop after outputting our sequence. If we’re looking at “ABC”, then a program which outputs “ABCDEF” also qualifies. In the example with digits of pi, it seems likely that the winning program would be one that just goes on forever outputting digits of pi, not one that stops after the first 100k. Similarly, our Occam-approved scientific theory program would probably not stop after reproducing all the observations we had, it would also predict future observations, and it would probably be very good at it.
(Please let me know if your/Eliezer’s understanding differs from what I’ve written above. I think I managed to capture the essentials.)
This is all very interesting and useful for things like AI research. But I don’t see how on Earth it could be relevant for metaphysical questions. Imagine that by some magic I can give you the shortest computer program for any sequence you give me. Imagine you actually encode everything you’ve ever observed in your life (or maybe everything that human scientists ever observed in history) in a sequence of bytes and that I gave you the shortest computer program for that sequence. How are you going to use it to decide whether God exists? You can look at its output, since once again, Christianity makes no experimental predictions that would differ from what your atheist physics teacher would tell you. Do you hope to inspect the source code to see whether there is a concept of God encoded somewhere in it? I can tell you what you’ll see when you look at that file: indecipherable gibberish. Which is the same thing you see after looking at any computer program that has been machine-generated (and a fair number of human-written ones too). Open the binary of your web browser in a hex editor–is there a concept of God encoded in this file? How about the interpretations of Quantum Mechanics?
[1] In the actual formulation we don’t completely throw away programs that are longer than the winner or programs that printed something different than our sequence, we just assign lower probability weights to them depending on how much longer and how much different, but I don’t think it’s important for this argument.
@Anonymous: I think that your point would be clear if instead of talking about Christianity you talked about deism. Christianity is broad, but generally speaking, Christians make falsifiable predictions (about the efficacy of prayer, for example), then make excuses when they’re falsified. Of course, push a theist far enough, and they often retreat towards deism, but the thing that makes no predictions at all is deism itself.
Anonymous,
I found your response very clear and helpful. But I don’t see why the MWI advocate won’t just respond “my computer program is shorter!” The two computer programs don’t just encode the Schrödinger equation and some Hilbert space mathematics; they need some way to extract from all this math, actual predictions of our experience. Conceivably, that is where the interpretations of QM will differ.
@Anonymous: Have you read Why Won’t God Heal Amputees? or “Religion’s Claim to be Non-Disprovable”? They make a pretty good case that the standard Christian world-view does naturally imply certain predictions, and that crypto-deism is simply the position theists retreat to when those predictions are shown to be false.
@Paul Torek: Thanks for the kind words.
What would the MWI advocate mean by “my computer program”? A hypothetical computer program implemented by a human programmer who believes in MWI? And that such program would supposedly be shorter than one implemented by a programmer who just learned about the state vector reduction and never heard about MWI?
As long as we’re talking about computer programs which reproduce or predict observations, I don’t see why the version implemented by the MWI advocate should be any different from the one written by someone who’s more into Copenhagen, or who doesn’t care. The sequence of steps you need to perform to calculate the results of any given experiment is exactly the same no matter how you choose to think about it. When physics students take a QM exam, the calculations they perform are the same regardless of whether they believe MWI or not.
@Toby Bartels, jaimeastorga2000: I was specifically thinking about the Roman Catholic doctrine, which I think I understand reasonably well. To the best of my knowledge it does not make any empirically tested predictions, related to prayer or anything else. But I’m happy to talk about deism instead, I don’t think it matters for the point I was trying to make.
On the contrary, it’s Bayesianism that he applies to everything except for quantum, where he reverts to frequentism.
And I’m surprised no one has brought up Eliezer’s “polyamory” in a discussion of him and ulterior motives for philosophizing.
Regarding the Matt Yglesias thing, my steelmanned interpretation is that if your opponent holds Position A on Grounds X, you hold Position Not-A on Grounds Not-X, and you believe that Position A is untenable on Grounds X, it is reasonable to make the argument towards your opponent that they should abandon Position A because it’s contrary to X-ism, and this wouldn’t be an isolated demand for rigor because you’re not an X-ist anyway. So, here, Yglesias can say that you should oppose air strikes on utilitarian grounds, and make that argument because he believes it, though his real reason for opposing air strikes is on non-utilitarian grounds. If you say, “Why don’t you cut PBS and donate the money to charity?”, he can respond, “I’m not the utilitarian – you are. I just think that utilitarianism supports my position on this issue, so if you’re a utilitarian, you should oppose air strikes.”
This isn’t actually what Yglesias is doing. He’s arguing for selective application of a system that doesn’t have any coherence when applied selectively. The argument he made would be sensible if there was a comparative choice between expenditure on malaria and expenditure on strikes, but there is no such comparative. It isn’t true that the money spent on strikes would otherwise be spent on malaria. The only way that money would be spent on malaria is in a utilitarian system (where the funding for strikes would come from something else with very little value).
This is why opposing an argument for government expenditure with claims like ‘X is more cost-efficient’ is absurd, because it’s an attempt to smuggle in the premise that no expenditure at all is justified unless it’s literally the most cost-efficient expenditure.
It’s another matter that utilitarianism is actually an absurd system, which no on really believes in. Extreme selfishness makes a lot more sense and actually remains consistent with observed human behaviour, yet for some reason people continue to speak as though ‘utilitarianism’ was some sort of model world-view through which all action should be considered (though none of them are willing to extend it to it’s conclusions).
Even if no individual was a utilitarian, it would still have value as a way for the society of selfish individuals to plan their next move. In fact, that’s how our society is modeled already, isn’t it? Neither the market nor democracy is necessarily tilted toward utilitarianism – they’re just ways for people to coordinate their enlightened self interest. Accepting utilitarianism as the ‘correct’ moral calculus just means choosing “cooperate” in the prisoner’s dilemma that is existence.
I don’t think the conservatives arguing for airstrikes are our claim to be utilitarians.
No, but some progressive arguing for air strikes may be utilitarians or at least be to utilitarian arguments on this issue.
Don’t we have enough ways to dismiss positions without engaging with them? I know all debates are etc but it seems like people in this sphere get themselves in more trouble by going meta than by spending time learning and understanding details. That was always the danger of learning about biases. I guess it’s fine for keeping yourself from getting persuaded too easily, as long as you remember the thing about reversed stupidity, and don’t act surprised when people dismiss your position on meta grounds like “didn’t engage.”
Speaking of which, it seems like you’re either really insisting on missing the point about causality, or just really interested in arguing against the worst/least-interesting versions of “[concept] isn’t that real/useful.” It’s like you admit that there are people who take a consistent position, but that since many people are inconsistent, you can ignore them? [Maybe there are two ways of looking at it: either the consistent people are ideologues, so it’s not evidence when they say a thing; or the inconsistent people are just shopping for good arguments for the opinions they already hold, so it’s not evidence against the argument applied consistently. Oh no, opposing meta arguments, what shall we do…]
Oh, and analogies in social science are misleading enough without calling them isomorphisms. I know what you meant, but that gave me some rhetoric-chills.
This comment is looking pretty mean. I really liked that last bit!
I think we have quite enough debates about specifics on this blog, much deeper than most people get. Meta-ideas are a great way of stepping back and figuring out why we are having particular debates and what we hope to gain from them.
WTF. Where does Scott say this?
Since “ideologues” is just a pejorative – not an actual argument – I still don’t see why consistency would be a bad thing.
I agree, so why dismiss Shalizi and other consistently disagreeing commenters? This particular principle isn’t very good for finding out what is true, just for making sure you don’t take certain arguments seriously. It seems particularly risky for questions about social science and informal discourse, where good “consistency” doesn’t look anything like abstract first-principles reasoning because things really are complex, analogies are always imperfect, etc.
Re: going meta, yes, like I said, it’s a tradeoff. I’m making a particular claim about where we stand with respect to it. You’ve now disagreed. I could keep listing all the ways going meta predictably goes bad, and you could keep listing all the things it’s good for, but how do we actually resolve that disagreement but by evaluating the object claims?
Which of the following are you claiming?:
1 Going meta displaces object level arguments. And object level arguments are better.
2 Going meta decreases the quality of subsequent object level arguments.
I read you as claiming 1, but now it looks like you might be claiming 2.
Regarding 1, going meta complements having object level debates. We can have plenty of object level debates and a few meta level debates (its not like we have meta level debates very often on this blog). So I reacted with annoyance when it looked like you were saying something like : “why are you having meta level debates when you should be having object level debates?”.
But if you are claiming 2, then I think that this particular idea does get at truth, but not the truth of particular claims: rather the truth of why people choose to evaluate claims in the way that they do. This is very important, because it is often impossible to resolve a debate by keeping it on the object level. Often this leads to privileging the question.
Here is an example. Suppose groups A and B are at war. Supporters of B point out to many questionable war tactics being used by A, citing death statistics and creating a lot of debate about whether the specific actions performed by group A were justified. Meanwhile Group B’s tactics are just as bad. One might very reasonably try to shift the debate to the meta level of what sorts of norms all participants of war should adhere to. Supporters of group B will then accuse these people of trying to dodge the object level issue that group A’s tactics have been really really horrible. Which may very well be true, but the question: “Has group A’s tactics been horrible?” though perfectly object level, is a nevertheless a privileged question.
Identifying a flaw in an argument is engaging with it. Identifying a pattern of flaws helps identify arguments that exhibit that flaw.
OK, but “some people apply this argument inconsistently” isn’t a flaw in an argument, particularly if you have people maintaining principled consistency (Shalizi) or people maintaining consistency by adjusting their application of an argument based on detailed understanding of different situations. It could be evidence about the argument, but that’s what my comment in square brackets was about.
How about crime? It’s usually measured by crime rate – a made-up statistic that combines subfactors like arson (maybe higher when fire insurance pays out better), property damage (maybe higher during periods of ethnic tension and frequent riots) and theft (maybe higher when income inequality is worse). There is assumed to be a General Factor Of Crime (presumably caused by things like poor policing, dark alleys, broken families, et cetera) but I would be extremely surprised if anyone had ever proven Beyond A Shadow Of A Doubt that the factor analysis works out here.
Ha, so in the UK it is known that the reported crime rate given by the police is a pretty bad measure of crime, because the police do things like record several crimes as one for example, so that their stats look better. Instead the British Crime Survey (which is now the England and Wales crime survey) records people’s experience of crime, and is considered a much better measure, being a nice randomly conducted survey.
So in other words while I think everyone would agree that the crime rate is measuring crime to some extent, if I wanted to use it to, say, argue that crime had dropped or risen I would be misleading people. Another excellent metaphor for IQ!
As long as the ability and the willingness of the police to fudge the numbers remains constant, it’s still useful for relative measures, such as whether crime is increasing or decreasing.
Which, admittedly, is unlikely. From one mayor to the next, or one chief of police to the next, there could be more or less pressure to keep crime rates low.
I apologize for the off-topicness, but there are two Tom Swifties from your recent Twitter spree (which I loved) that I didn’t quite get:
“”The medication cured my autism but also made me gain weight,” Tom said fatalistically.” (I can see “fat” at the beginning, but I can’t pull anything out of the rest of the word…)
“”I got a job producing another season of Lassie,” Tom said moronically.” (Similarly, I can see what’s probably a “more” at the beginning, but nothing else, unless it’s “More on Nic(kelodeon)”, which doesn’t seem like a very strong Lassie association to me?)
Any help?
“Allistic” being the antonym of “autistic”. No idea about the Lassie one though!
More-on-a-collie?
*groans.*
We shouldn’t get too ad hominem about inconsistent philosophers. It is a fact that there are certain things (such as philosophical skepticism) which humans cannot bring themselves to believe on an emotional level. This doesn’t change the fact that there are legitimate arguments not refuted by the inconsistency of their proponents.
My inner monologue upon reaching and proceeding to read Part V: “oh my God did he actually do it, please let him have done it, please let him have done it, HE DID IT, HE DID IT, AAAAAAA yes yesssss”
Uh…the rest of this post was also excellent, and described very useful reasoning techniques in an illuminating way.
But if the cattle-rustler is a member of a historically-disadvantaged group, surely it is in the service of justice to allow them to steal your cows?
If you acquired the cows within a social system whose history includes institutional inequality, is your claim to “own” the cows even legitimate to begin with?
And what about the cows’ human rights, which are violated when you herd them and slaughter them and etc.
Assuming Protagoras had a sixshooter like Socrates did, when Socrates squeezes the trigger it’s just gonna go “click”, proving that, in fact, Sophism is the one true philosophy.
You’re forgetting about the first shot: “He just raised his revolver in a fluid motion and shot at the old man.” Also, there’s two bangs before Heraclitus, so one of those could have been from Heraclitus, Parmenides or Zeno…. Perhaps they cancel out. It’s still not looking good for Socrates.
Ah, but this was the Platonic Ideal of Guns, which, obviously, is never unloaded because that would be less than ideal.
Kickstarter to make this western a web series!
Seriously… how would you feel if I took this idea and made a YouTube series with sock puppets?
Holy crap I really really need to get a job.
do it well and that can BE a job.
“But when other people are totally happy to talk about speed and blood pressure and comas and the crime rate, and then suddenly switch to a position that we can’t talk about IQ at all unless we have a perfect factor-analytical proof of its obeying certain statistical rules, then I worry they’re just out to steal cows.”
Huh. When people start talking about IQ, then I worry that they’re out to justify some extremely shitty and anti-humanistic policies. Historically, that’s what’s happened. Count the number of people who ever pointed to their high IQ and said “My high IQ ought to give me special obligations”, divide by the number of people who ever said “My high IQ means I deserve the power and status I’ve got, and ideally I should have more of it”, and you get a sadly small number.
We can tolerate some vague and shaky models. Often a vague and shaky model is better than nothing. But when poor models start getting abused, that’s when it’s worth making a fuss about them.
There’s a particularly common way of abusing this kind of model, and that is the description vs. explanation error. You take something that is at heart a description of a phenomenon (for instance the DSM criteria for alcohol abuse), and use it as an explanation of the very same phenomenon. Why is he an alcoholic? Because he does these things. Why does he do these things? Because he’s an alcoholic.
Is the shaky model called IQ abused this way? Yes. It’s at heart a description of stuff some people are good at. But its original purpose, and still the way it’s used to this day by its fans, is to explain personal success, or argue that personal success is somehow your birthright.
But maybe we can just ignore the people who use it that way, and use – correctly, as a description, not an explanation – it for something useful? Sure, as soon as you tell me what that useful thing is. (Something that wouldn’t be better served by Shalizi’s “Arete” quotient, too!). When we know what you’re planning to use it for, we can maybe suggest better models, too.
Is the shaky model called crime rate abused in this way? No. People rarely say people commit crimes because there’s have a high crime rate. We all know it’s a description, not an explanation. The shaky model called crime rate might be used to explain other things, though (like migration), or you might use other things (things not implicit in its definition) to try explain the crime rate. Imperfect though it is, it has uses. And people are perfectly willing to switch to more specific models if the use case warrants it.
I’ve seen people use the crime rate to “demonstrate” the inferiority of certain minorities far, far more often than IQ.
But maybe we can just ignore the people who use it that way, and use – correctly, as a description, not an explanation – it for something useful? Sure, as soon as you tell me what that useful thing is.
Sure. Here are two things: determining job aptitude, and learning how to better educate children of different ability levels. Both are more or less functionally illegal right now, because they involve “disparate impact” on members of races with a lower mean IQ (in the latter case, when schoolteachers sort children by ability to focus different methods on different IQ levels).
100% of the jobs I’ve applied to required an IQ test in the form of HR questions, programming questions,etc. in the ‘technical’ part of the interview.
When I did statistical work for school districts (5 years or so ago), 100% of them had programs called things like TAG (talented and gifted) or magnet schools,etc. Accelerated programs for gifted students. All but 1 had vocational programs for less high performing students.
You say these things are “functionally illegal” but they are actually ubiquitous.
With respect to IQ tests, I primarily had this court case in mind: http://en.wikipedia.org/wiki/Griggs_v._Duke_Power_Co
I’m not a legal expert, but from what I understand most employers avoid IQ tests because of the possibility of being penalized under this precedent.
Are you saying that you actually took an IQ test in your job applications, or that you were asked questions correct answers to which correlate with IQ? Employers of course use various tests to determine who to hire, and success at many of those tests will correlate with IQ. The SAT correlates with IQ, but people don’t say that college admissions are decided on the basis of an IQ test.
Two points with respect to ability sorting. First, from what I understand programs for gifted students as well as special education programs are much more common than sorting within, say, 75-125 IQ children. There’s much hand-wringing over the racial disparities even in those, though, and from what I understand these programs, as well as more fine-grained ones, are more common in less racially diverse schools, because there’s less danger of the groupings “failing to reflect” the demographics of the school, thus opening the school up to a lawsuit. What were the racial demographics of the schools you worked for like?
My source for this information is mainly Education Realist’s blog, for instance this post:
http://educationrealist.wordpress.com/2013/11/21/unstructured-musings-on-choice/ . ER is a schoolteacher who I have no reason to believe is lying when he says that schools risk lawsuits “if they suspend too many black or Hispanic students” or “if they group kids by ability and the racial demographics are unrepresentative of the school community.” I’ve also seen news stories about such lawsuits in the past, but I’m afraid I don’t have any citations on hand.
Edit: the above only applies to public schools, in case that was not clear from context.
It’s not at all clear what is the legal status of IQ tests for hiring. Griggs and its codification into law seemed to ban them. It also banned diplomas, but everyone uses diplomas in hiring and no one gets in trouble (though there are recently rumblings).
The supreme court recently ruled that tests cannot be used for firemen, even tests of knowledge that firemen need to know to do their job. There is some suspicion that the courts just think that there’s no book-learning in firefighting but they might be more lenient on technical jobs.
But technical companies heed their lawyers’ advice not to use formal written tests as they pass the magic 50 employee line below which all is fair. Instead, as you say, they use oral exams in interviews. It’s mysterious why people think that this is more likely to be approved by the courts, but no one sues on these grounds. Perhaps the purpose is to avoid having a concrete instrument that can be blamed in court, rather than a procedure. Or perhaps the purposes is to avoid leaving records. But not keeping records also means that it’s hard for the company to learn about the quality of the test and to improve it over time. Or even to learn about validity of the same question across different examiners.
Gifted and talented programs are usually excused on the grounds of being supplemental: they are a minor part of the curriculum. Perhaps more importantly they affect very few students, as Troy says. Tracking goes in and out of fashion for many reasons. Probably education theorists’ thoughts on racial disparities and the possible impact on the IQ gap are a big part, but I don’t think legal concerns are a big part. Maybe they are becoming more important. For an example of the myriad factors that go into this, my elementary school had tracking for math and english in a couple of grades, but eventually stopped because the other schools in the district were jealous. It was killed because everyone loved it. Also, parents at my school were prevented from using the school on the weekend for supplemental classes for the same reason.
Magnet schools for the gifted are rare and usually have racial quotas (a great solution, but illegal for private employers). Most magnet schools are topical or just different. Really, they are a way of signalling that the parents care enough and are organized enough to apply. Also, Sailer thinks from his experience that they just lie about their admissions procedures.
There are legal actions against schools based on racial disparities in punishment, but this is a recent development. (which is why I think that legal action against tracking is recent if true)
One thing about disparate impact law and IQ tests that’s worth keeping in mind: even if a company is within its rights to give the tests because they’re relevant to the employee’s work, no company wants to be the one to take that case to the Supreme Court and see an article in Slate titled “Guess Which Company Is Fighting Tooth And Nail To Save Its Racist Employment Exam?” and then if they win their name goes in the name of the decision, and the decision becomes a byword for racism and every election the Left talks about how we have to do something about it. Like imagine instead of railing against Roe v. Wade, the Right was constantly going on about Coca-Cola Corporation v. Wade? You’d see a lot more Pepsi-drinkers in the red states. And you might still lose, and then you’re the company that fought for its right to be racist and lost.
It’s easier to give in and implement defacto racial quotas than to deal with all that nonsense.
Anoynmous: Imagine that in the limit IQ tests work perfectly. Employers would then hire people starting from the highest IQs going down. At some score (at least in some professions) they would run out of employees and you’d find that, oh, everyone with an IQ 105 is employed and everyone with an IQ 100 is unemployed 100% of the time. The exact cutoff could fluctuate depending on job availability, but the effect would still be that people below a certain point would find their chance of finding a job reduced all out of proportion to the IQ point spread.
The result would be a permanent underclass who can never get a job because their IQ is too low.
In other words, the fact that employers have to use inaccurate measures, can’t improve the tests, and have different results for different examiners is a feature, not a bug. They need to use tests whose results have a lot of noise so that people with a lower IQ only have a somewhat lower chance of getting a job without becoming chronically unemployed.
let’s ban resumes
Jiro:
Suppose that you have a population with, say, 100 people with an IQ of 100, and 50 people with an IQ of 110. Employers prefer to employ people with higher IQs, and there are 100 job slots. In World A, employers are allowed to use super-accurate IQ tests to select their employees, so all 50 people with an IQ of 110 have jobs, as well as 50 of the people with an IQ of 100. In World B, employers aren’t allowed to use IQ tests, so there’s some noise – 10 additional 100-IQers get jobs that they wouldn’t have been able to get otherwise, and 10 110-IQers fail to get jobs. Do you think World B is better than World A? If so, why? If the same number of people are employed and unemployed in both worlds, as it is here, why would it matter how unemployment is distributed across IQs?
If anything, World A is better than World B, because employers want higher-IQ workers for a reason, probably because they’re more productive.
Anonymous: Thanks for the helpful info.
Jiro: Low IQ people are not helped by being given jobs they cannot do well. They lower the productivity of the company, and are liable to get fired, leaving them worse off than before.
Thankfully, not all jobs require high IQs. If you’re concerned about the employability of lower IQ workers, you should be concerned with things like mechanization of manual labor, not with keeping companies from using IQ tests, which (as blacktrance points out) companies will find a way to get around, just less efficiently.
@Jiro
Your assertion that employers would hire from highest IQ down seems unfounded: http://abcnews.go.com/US/story?id=95836 .
I would expect employers to prefer a range of IQs, like the police department in the story, with both higher and lower IQs being less ideal for most types of jobs.
blacktrance:
“Suppose that you have a population with, say, 100 people with an IQ of 100, and 50 people with an IQ of 110.”
The problem happens because employers will select the people from high to low. If a substantial subgroup has exactly the same IQ, employers won’t be able to select them from high to low. And if this subgroup is the only one that exists after the employers have used up the rest of the applicants’ the problem won’t happen at all.
In other words, if IQ scores actually were distributed like that, you are correct that there would be no problem and selecting by IQ without noise would be better than selecting with noise. But IQ scores are not distributed like that.
Troy:
“Low IQ people are not helped by being given jobs they cannot do well.”
Their chance of getting a job can be affected all out of proportion to the difference in performance, if they fall past the cutoff. You could have someone with 1% poorer performance and a 100% less chance of getting the job. Such people are indeed “helped by being given jobs they cannot do well”, since “cannot do well” may only mean a miniscule difference.
Jiro:
Your hiring from highest to lowest IQ is how it would work if wages were fixed beforehand and independent of IQ.
What would happen if employers started to use IQ tests is that intelligent people would get paid more than less intelligent people for the “same” work, and any job that could actually be done the same way by stupid people would be done by stupid people because they get paid less.
Jiro:
The example can be modified to accommodate that objection. Suppose instead there is a population of 150 people, 2 people with an IQ of 150, 2 with an IQ of 149, 2 with an IQ of 148, and so on, and 100 job slots, as before. In World A, where IQ tests are legal and highly accurate, employers hire the top 100 workers by IQ, so the hundred people with an IQ above 100 all have jobs, and the 50 people with lower IQs are unemployed. In World B, IQ tests are illegal, but there are still the same 100 job slots, so some people with higher IQs fail to get jobs, and some people with lower IQs succeed, but in net, employment is the same. Would you say that World B would be better than World A? If so, why? The same number of people are employed in both – why is high-IQ people being unemployed less bad?
blacktrance: Yes, world B would be much better. In world A, if the IQ test is accurate, the person who scores 149 will score 149 every time, and for every employer. If employers start hiring from the highest to the lowest IQs, not only will some people not get jobs, but the *same* people won’t get jobs for the rest of their life. Whenever an employer expands his company he’lll be hiring from the top of the pool, and whenever someone loses a job because the company contracts, he’ll go back into the pool at the top, so the result will be that regular churn just causes the same people to be rehired at other employers while the people below the cutoff never get hired.
In scenario B, when someone loses a job, he goes back into the pool, but since the measure is so noisy, employers have some chance of hiring anyone in the pool. So while the percentage of joblessness in scenarios A and B may always be the same, this is spread over the whole pool in scenario B–for instance, scenario A may have 10% of the population never able to get jobs, ever, while scenario B may have, oh, 12% unemployed at one time, but exactly which 12% it is shifts.
I would think that scenario B would then be better than scenario A even if scenario B has 12% unemployment (because the employees who are hired are on the average dumber and companies don’t do as well) while scenario A only has 10% unemployment. Having X% of the population permanently unemployed is worse than having X% or even X+1% unemployed but having it be different people over time.
Jiro:
Oh, don’t worry. The racialists are just going to suggest some humiliating, paternalistic make-work program, complete with intrusive moralistic demands on the subjects’ lifestyle – perhaps in the vein of those Catholic pris… err, convents for sinful women.
We already have populations of permanently unemployed: the elderly, the disabled, children [ who are employable but legally unemployed (unless you count school as employment)], the mentally ill and so on. We already have provisions to provide for people in this situation. So why couldn’t this work for the iq cutoff too?
Economic growth tends to raise the standard of living across the board, although admittedly for the wealthy first. I wouldn’t want to try to predict growth gains from optimal iq sorted employment but it seems like it’s plausible the benefits would outweigh the costs of adding x more permanently unemployed people.
Misha: Even if you believe that perpetual unemployment is acceptable under this system, I have yet to see anyone who proposes employer IQ testing spontaneously mention that doing so won’t work unless you can accept having people with lower IQ perpetually unemployed. That’s a pretty big thing to leave out when selling people on the idea of employer IQ testing. It’s like giving someone some ice cream and saying “this tastes great”, while leaving out that the ice cream is opossum-flavored and that it tastes great to you because you like the taste of opossums.
@Troy: have you ever applied for a job? Yes, their aren’t official IQ tests, but there are often dozens of brain teaser type questions that are unrelated to job performance. These are in oral, and not written tests, but they are obviously in the mold of IQ tests. There are even interview books loaded with ‘microsoft interview questions’ and things like that. What the law technically says and how companies actually behave are vastly different (for an obvious second example. If you removed all disparate impact laws, I don’t think hiring practices would change much at all (although maybe for firefighters, which seem to be one area where people actually sue and win for some reason).
As far as school programs go, talented and gifted programs are not often supplemental, they are often entire separate curriculums that happen to have share a school building (with magnet schools, they even have their own buildings.) Most of them have a separate teaching staff.
As I said, I’ve worked with several large city school districts all over the country, and every single one had accelerated tracks for gifted students, and all but 1 had vocational programs to split off the low performing students from the rest of highschool students. You could argue that maybe more tracks would be better, or that tracks should split off earlier/later but it seems silly to argue tracking doesn’t exist.
In several districts I worked in (for instance, Dallas, Texas) there was often defacto self-segregation. There were latino schools, black schools, white schools,etc (Lincoln high was primarily black and the highschool with a confederate colonel for a mascot was primarily black). In order to lessen the obvious segregation issue, one of the IB (a high achieving/gifted) highschool program was slapped into the middle of a latino highschool, which lead to a situation with a 90% white gifted program in the middle of a 95% latino ‘normal’ highschool program. No one is worried about getting sued.
Most decisions as to curriculums and what not are made by school boards staffed with members who have literally no qualifications (education, law or otherwise), or made at a state legislature level. Lawsuits are really not a consideration.
Its possible that people worry about suspending/disciplining students disproportionately. But given the defacto segregation of schools, it seems unlikely (of course the all black school suspends all kids, etc).
Regardless- tracking exists in school districts all over the US, and test designed to measure intelligence exist in large companies all over the US. Stating they are functionally illegal is probably not even hyperbole, its just wrong.
The gifted program for my county was ‘science and technology’ — which, in practice, meant nothing more than that all the gifted-program students were forced to take woodshop, but they called it engineering. Would have been useful had the shop teacher not been senile, nearly blind in his one remaining eye, and totally incapable of teaching. (Most of the work involved memorizing worksheets while he retreated to his office to do something that involved making repetitive thumping noises. There were no tools in his office, but there was a computer with an internet connection.)
There were maybe three hundred students in it, and it served a third of the county — but the other two schools hosting gifted programs had a terrible reputation, which was probably justified. (One is in a 92% black suburb; the other is in an 87% black suburb. The one I went to is in a town that’s only 48% black.)
Most of the smart parents in the area probably sent their children to private school — for good reason, since the public school is a dead end. I only knew two people there who didn’t end up in the state college, and they were both black. Also, Sergey fucking Brin went there, and guess where he went to college.
The place was a shithole, run and staffed by incompetents. The county school board is a third-world joke, and the only reason the school wasn’t as much of a disaster as every single other one in the county is that the former principal never listened to them or anyone. Then he got promoted, continued to never listen to anyone, and within two years the school system was in a lot of trouble with the ACLU. (This was partially my fault, but it probably would have happened anyway.)
Tracking nominally exists here, but it doesn’t do much good. It probably can’t do any good, given that it would have to operate within the constraints imposed by the fact that local politics are third-world as hell and will remain so for as long as democracy exists here, but it could be made better by, for example, having one magnet program for the entire county and putting it in a separate school.
On the other side of the Potomac, they have a school like that. But the civil rights types are crawling up their ass.
Will, my experience with school districts is extremely different from yours. Vocational programs are quite rare, though they are coming back from a nadir maybe 20 years ago.
IB programs are like magnet schools, admission by parental decision, not by test score. Most magnet schools are a secret code that this is the place for ambitious parents to send their children to avoid the parents that are clueless. They are about race and especially class segregation. Smart kids whose parents are not clued into the system do not get recruited.
IB schools are much better because they are up front that they are about academic effort. Smart kids are much more likely to figure out that this is the place for them than with magnet schools.
But magnet schools often have enrollment limited and seats filled by lottery. Entering the lottery on time is a way of keeping out undesirables.
The difference between firefighters and microsoft is that government hiring is subject to strict bureaucratization. They have to have objective written tests to prove that they aren’t giving jobs to their friends. Whereas microsoft thinks the continual use of ad hoc IQ tests is worth the risk of nepotism. Google has institutionalized nepotism.
Yes, everyone tries to recreate IQ tests in hiring, but they do an awful job with their ad hoc tests. Similarly, parents try to modify school systems to produce segregation. They do a much worse job of emulating tracking because they aren’t even trying. Of course, if there were IQ based tracking, the parents might be trying to undermine it, too, because it is not exactly what they want.
Aimed at the latest Anon- I imagine experiences are different in district to district. And my sample is not random, its large, somewhat wealthy districts.
My point is only that tracking is not illegal, tracks exist in lots of places (look at the parent post I was responding to).
I’m not defending the US school system in any way. The biggest take away I had in working with school boards is that even in very wealthy districts, the school boards are basically groups of incompetents.
Similarly, regards to interviews and IQ tests, I’m merely pointing out that companies are clearly not at all afraid of using pseudo-IQ tests to hire people.
We can argue about proper implementations, but I think we both agree neither different tracks, nor pseudo-IQ interview tests are illegal.
The circularity dissolves when you realize that “he’s an alcoholic” can point to both object-level behavior (he drinks half a bottle of Jim Beam a day) and its unobserved but inferred causes (complicated psychology and biochemistry that boils down to habituation of various kinds). The two are linked closely enough that we can get away with using the same phrase for them, but their referents are not identical. “He’s an alcoholic[1] because he’s an alcoholic[2]” is not a tautological statement, even if it’s too ambiguous for people to use it much in the wild.
This sort of pattern is common wherever complex hard-to-measure variables influence object-level results.
There’s also the fact that “X because Y” is commonly understood as shorthand for “I believe X because Y”. Given the right context, it’s perfectly reasonable to respond to the question “Why is John McCain a Republican?” with “He has an R after his name. See? McCain R-AZ”.
This is, like, just straight-up factually inaccurate. IQ’s original purpose was to predict personal success, or specifically to predict which folks would be best able to turn material advantages such as high-level schooling into personal success, so that we could maximize societal flourishing.
The question of desert is properly orthogonal to factual questions like the role of IQ. However, it seems to me easy to argue that success that comes about because you were lucky enough to have an innately high IQ is undeserved. It seems harder to argue that success gained by working harder than everyone else when everyone is exactly as intelligent as everyone else is undeserved. Properly viewed, I feel like a belief in a strong role for IQ is a point in favor of redistributionism.
Properly viewed, I feel like a belief in a strong role for IQ is a point in favor of redistributionism.
I think it could be as well, but it depends on one’s ethical premises. Certain types of anti-humanist might band together with their high-IQ cronies and say, “Screw you guys.”
The science and the politics should be separated, of course, but that is as clear a demonstration of the phrase “easier said than done” that one could ever see.
Well, I mean, they could, but I don’t see how admitting the strong role of IQ makes things any worse. Hypocrites gonna hippo, you know? Hypocrites will always find some means of justifying banding together to screw us guys, we shouldn’t sacrifice the truth just to close off one specific avenue to them. If they’re not robbing you with Social Darwinism they’ll just rob you with Social Justice. In every age, in every place, the deeds of men remain the same.
…we shouldn’t sacrifice the truth just to close off one specific avenue to them…
Hell no. But as we venture down the avenue of objectivity it’s worth thinking about which ethical outlets it leads us to. (People might only follow if they’re sure that it goes to a decent part of town.) I agree with you – and Gregory Clark – that determinist conclusions could strengthen and not harm the case for social solidarity, but it could be a powerful tool for people who maintain that the successful have no obligations towards the masses. “Society is unfair” has more evident collective implications than “life is unfair”.
“In every age, in every place, the deeds of men remain the same.”
Did you just quote Sigrid Undset? 🙂
But anyway, I agree on that. No need to let social darwinists rob us of a useful tool – but that is a lot more relevant for the statistical methods they invented than for IQ. I’m questioning whether IQ is a useful tool.
@Harald K: It’s from an anime called Legend of the Galactic Heroes.
They don’t though.
In reality the deeds of man(or people) are extreamly sensitive to the social context social Norms systems of laws and systems of incentives.
The behavior of people in Finland Denmark and to a lesser extent the US are wildly differnt than the behavior of people in Russia South Africa and Vietnam. And the first 3 profit greatly from that difference.
Seems like the next obvious step is to come up with a Hard Work Quotient to predict which people are innately more likely to work hard, and then use that to redistribute away from them.
Good news! Conscientiousness is heritable too…
I wonder how heritable empathy is?
http://scholar.google.com/scholar?q=heritable%20empathy looks like it’s been investigated before, but I’m not familiar with the various scales and attempts to quantify empathy, so dunno.
It would be desirable, but tests for things other than IQ are gameable by smart people, if they know they are being tested and know what the test is for (being smart people, they usually do). So your HQ would measure some mix of hard work+intelligence+sociopathy.
Yes, but if the test is properly designed, gaming it would be hard work ….
If you count Binet’s test as the first, then I could grant that it had a decent purpose: to judge children’s needs for special assistance in the French school system. But it was Terman, not Binet, who adapted it and claimed it as a test of “general intelligence”. And yes, by then it was already deeply about justifying eugenics and (white) racial supremacy. The Pioneer fund was established in 1937, to this day they fund publications arguing for such things.
About desserts, and whether high IQ shouldn’t give you duties instead of privileges: It does seems easy to argue, right? That’s why I mentioned it, because that argument is conspicuously never made. Herrnstein and Murray did not make it. I don’t know of any other Pioneer fund recipients who’ve made it.
But anyway, what does it matter? IQ is a model. Models should be evaluated on how well they answer the questions you ask them. The original questions “who should be allowed to breed?” and “who should rule?” have gone out of vogue, but IQ tests were a pretty rotten way to answer them anyway. Maybe there are other questions the IQ model can anwer?
How about the question “who are most likely to be successful in life?” for that, it may be a good model, but one incorporating beauty, family wealth and physical performance as well would almost certainly be better (and if Shalizi is to be believed on the math of factor analysis, you will find a “g”-like factor underlying those too).
Is it “who are most likely to perform better at [specific task]”? In that case you’re better off with a model measuring performance at the specific task.
Is the question “How does the brain work, physically?” in that case the model of IQ has not led to any insights.
But as I said, I’m not ruling out there’s some question where this old tool invented to justify injustice could still be useful. Especially since there’s a good deal of data existing for it already. There’s just no need to be an apologist for it in general.
When has anyone said IQ should be the only factor you should ever look at? Yes, if you’re modeling life success you want to include things like physical fitness and family wealth; but you also want to include IQ. Your predictive abilities may be harmed by leaving out beauty, but they will surely be harmed much more by leaving out IQ. (My guess is that family wealth and IQ are going to be of roughly equal importance to your model.)
I’m sorry, but the name “Pioneer Fund” doesn’t trigger any tinfoil hat paranoia in me. Just knowing that they support a position does not seem to me like evidence against that position.
It’s instance 2 of argument from my opponent believes something.
BTW, you’ve responded to a mention of the word “desert” by changing it to “desserts”. I don’t want to come off as a Grammar Nazi, but I thought you should know that “desert” is in fact the correct spellings.
Why, I believe you’re arguing against intelligence as a concept! This is much more interesting than the usual anti-IQ arguments.
I could venture my view on this, but I serve the commentariat better by presenting the question than my answer. Why do we care about intelligence, this strange category broader than any specific task yet narrowed to exclude wealth, beauty, strength, and practical and cultural knowledge? (The neoreactionary voice in my head is gleefully saying “IQ tests don’t, as often charged, reward belonging to white culture, but they’d be better if they did.)
What useful question is answered by knowing someone’s intelligence, that is not better answered by a narrower or a broader measure?
Remind me of http://lesswrong.com/lw/kk/why_are_individual_iq_differences_ok/
This is wonderful thing about evaluating factual claims on their own merit. When you don’t mix politics with facts, you end up with interesting ideas and conclusions. I’m highly suspicious of people who say “But if you believe X, then you have to believe that the government should Y!”.
pure awesomeness, and this is coming from someone who really doesn’t have a very high opinion of philosophy, or western mythos.
You missed an opportunity for syllepsis.
Is there research on the extent to which people tend to overestimate real group differences?
Does the idea of IQ tend to discourage people from trying to find out how to improve it?
Every now and then (when appropriate, obviously), I ask people IRL what they think African American IQ is. These are generally college educated people who understand statistics well enough to be familiar with normal distributions, and I preface it by telling them that the British are used as the reference population, defined to have a mean IQ of 100 and a standard deviation of 15. (Estimate it yourself, by the way, before you get on to the next paragraph.)
The answer I get is generally ’95,’ and they are shocked when I tell them the true answer is 85. I have yet to hear anyone underestimate it.
If anything, I would expect it to have the opposite effect- now there’s a number you can track to see if you’re doing any better, rather than just subjective guessing of “I suppose I feel a little sharper on piracetam.”
Consider the hypothetical world where it were possible to easily adjust IQ upwards. Much like a diet drug or exercise plan would advertise that you could drop twenty pounds, a nootropic or mental regimen would advertise that you could gain twenty IQ points. Many nootropics and mental regimens do advertise themselves this way; they simply fail to deliver. It’s much easier to find a drug or lifestyle that makes people gain weight than lose it- similarly, it is much easier to find a drug or lifestyle that makes people lose intelligence, rather than gain it.
>The answer I get is generally ’95,’ and they are shocked when I tell them the true answer is 85
Is it still this high? I saw several review papers in the mid 2000s, starting with work by Flynn, that suggested the gap had closed by about ~5-6 points between the 70s and 2000. I think this had pushed the younger cohort differences to down around 10ish points.
From about 1.1 to 0.9 standard deviations. I’m going to stick with the round number 1. here’s the paper
The answer is subject to a large number of decisions, such as which test to use or what age to measure at. here’s a graph of several tests.
“Every now and then (when appropriate, obviously), I ask people IRL what they think African American IQ is”
The lack of the word “average”, while possibly a simple oversight on your part, suggests that you may be engaging in a worrisome reification of the concept “African American”.
And as for people not underestimating black IQ, clearly underestimating is more fraught than overestimating.
That story was delightful.
I’m curious – when I read the first part, the first thing I thought of was utilitarianism, and the second thing I thought of was natural law theory. Catholics are quite happy to use things not in perfect accordance with their natural ends … except when people are having sex in a way that they don’t like. Would anyone else mind sharing what came to their mind?
Sort of related – I’ve been thinking a bit about the relationship between bravery debates, policy debates and wars, and I’ve drawn some conclusions. For most people policy debates are like wars – and the art of war is destroying your enemies while getting other people to either join your side or sympathize with your side. This is very difficult to pull off because destroying one’s opponents is a hostile act. So people try to selectively point out all of the unethical tactics being used by their opponents, all of the losses incurred on their side, while ignoring the suffering of their opponents or the unethical tactics being used by their own side. I came to this conclusion while looking at how people talk about the Israel conflict. A supporter of Palestine will cite statistics of how many people Israel killed. And a supporter of Israel will cite how many people the Hamas have killed (or how many rockets the Hamas fired at Israel). Bravery debates are just a specific instance of this. Bravery debates are a way of saying “my opponents are using unethical tactics – in particular they are trying to silence opposition”. But I think it is part of a more general pattern of trying to paint your opponents as the bad guys.
Humans are hypocrites, so when anyone uses some “system” to explain their demands, it is a safe bet that they would oppose the same system when used against them. Finding good examples simply means finding people who argue by some “system”… as opposed to admitting openly they acted selfishly (but even those will usually argue that it is somehow okay when they act selfishly, but not so okay if someone else does), acting without explanation, or using such chaotic explanations that there really is no system.
So it’s mostly a question of which “systems” we are irritated by because our opponents use them against us. I could give a few examples, but that would be almost the same as giving a list of my opponents, so… uhm… I guess I’ll abstain for now.
How do you distinguish “they’re pointing to differences as a form of bravery debate” and “they’re pointing to differences because the differences are actually significant”? Is there some way other than “well, they’re pointing to important differences if I agree with them”?
That itself can be used as an isolated demand for rigor. When someone points out the differences between two sides of a dispute, and you don’t like that, call it a bravery debate–but don’t apply similar criteria to calling things bravery debates elsewhere in life. (Supporters of homeopathy and supporters of conventional medicine can both cite things that support their worldview. Of course, they may cite different things or things that occur in different quantities, but that is true for Hamas versus Israel too. Can homeopathy versus conventional medicine be best understood as a bravery debate?)
Why would you think that? I see nothing in the conservatives’ reasoning that would make me think they are advocating a meta-level rule like that. It seems like a pretty obvious example of an object-level rule; “sex-education is horrible, and I am particularly horrified that some of my money goes to support it.” If I oppose letting serial killers murder children for their own amusement, that doesn’t mean I am advocating a meta-level rule of not letting people indulge in their desires; it just means I don’t like children being killed.
“Sex education is horrible and we shouldn’t do it” is applying the meta-level rule “don’t do horrible things.”
“I find sex education horrible, so I shouldn’t have to pay for it, but other people can pay for it if they want to” applies the meta-level rule “people shouldn’t have to pay for things they find horrible,” which consistently leads to me not having to pay for the military, prisons, the police, or the NSA.
No, this is not the conservative position, or meta-level rule. Rather, the rule is:
“Sex education is horrible, so it’s outrageous people should be forced to pay for it. But other people can pay for it privately if they want to.”
The meta-level rule is:
“Government compulsion is only OK for things that are not just good in one person’s opinion, but overwhelmingly good (or conversely, to suppress things that are not just bad, but bad beyond redemption). There’s a broad swathe of things which we may like or don’t like, but it doesn’t follow the government should do much about. That’s the meaning of limited government, and indeed freedom.”
There’s no contradiction between “You should have to pay for the military, but you should get to choose on sex education.” What is missing from your post (and even more so from OldCrow’s below), is the concept that not everything is a matter for government, which is at the heart of the conservative position.
In this case, there’s also the added element of “stop trying to force your values onto my children and making me pay for the privilege.”
“is the concept that not everything is a matter for government, which is at the heart of the conservative position.”
I’m not sure by “the conservative position”, you mean general conservative ideology, or this particular position, but in either case you are wrong. Conservatism advocates substantial government involvement in general, and in the case of sex education, they are opposed to it even in the case of private actors; it’s just that the public funding of public schools gives them a further basis on which to attack it. Keep in mind that it was once illegal to send information about contraceptives through the mail.
Yes, you can always take an object-level rule and generalize it into one of many plausible meta-level rules which contain it as a particular instance. That does not mean it is a good idea to model people who are clearly thinking on the object-level as really thinking of whatever meta-level rule you think they obviously follow.
I am reminded of Kant’s first formulation of the categorical imperative, and the numerous ways in which a moral maxim can be constructed to generalize behavior in any situation.
Except in this case, if someone says “I shouldn’t be forced to pay for things I don’t support” or “We shouldn’t pay attention to correlational studies that don’t reveal the underlying cause,” they’ve explicitly brought the meta-level rule into their argument. They are thinking at the object level but arguing at the meta level. In that case it’s appropriate to point out that this probably isn’t their real objection, then try and move the argument to the object level.
<3 anyone who recognizes Kant's dumb stupid moronic ridiculous crazy idiocy
Edit – Dammit, this is what I get for being a long-winded asshole. Just read Ozy’s post.
I think Scott is referring to an (anecdotally) common argument made by conservatives who don’t like sex education, that really rests on the use of “because”. The idea being that if I believe something is morally wrong, it is unjust to use my tax dollars to support it, which is a sufficient reason to get rid of it even though you disagree with me.
This is one of those arguments that probably looks like a straw man until you run into it in the wild.
It’s one thing to argue that sex education is evil and convince enough people that you get it banned. Similarly, enough people think murder is evil that we send the cops after serial killers. But the murder analogy, applying the same meta-reasoning as the sex education example, is more properly “If I don’t have a problem with murder, the government shouldn’t be able to take my money and use it to fund the police force.”
The point is that the anti-sex-ed conservative would never listen to someone who said this – it’s obviously stupid. The only reason they use the argument for sex education is that it happens to be a convenient response.
Although it might be interesting to see what happened if we actually funded policies in proportion to their popular support. Probably disastrous, but interesting.
I think the reasoning is actually closer to that found in anti-Ten Commandment monument cases. There, the meta-reasoning is that government funds shouldn’t be used to promote beliefs that are only held by part of the polity (even if, in this case, it’s probably a substantial majority of the polity).
The argument used (successfully) in those cases is usually also phrased along the lines of “tax dollars shouldn’t be used to promote an ideology I disagree with,” which is basically the same argument conservatives are making about the (specific kind of) sex ed being promoted to their kids.
That’s a totally sensible argument against sex ed. I just don’t think it’s the one Scott was referring to. There are people who actually make the naïve sounds-like-a-strawman version I posted. Not because they’re dumb, but because they will latch on to an argument solely because it supports their position. And if you use the word ‘meta-reasoning’ they will look at you funny.
I did not intend to imply that this has any bearing on whether conservative positions on sex ed make sense. I was just trying to clarify the description of a particular mistake, which I think Scott is justified as using as an example of poor reasoning.
Logically justified, not necessarily rhetorically justified.
*insert additional caveats that I forgot to add
I think Scott is mis-stating the actual argument, or more accurately, misunderstanding it because the people making it are using shorthand rather than the precise technical terms that would be appropriate for a meta-level discussion.
I’ve imagined it as a check-off on one’s tax return. I expect we’d be drowning in advertising from government agencies and programs.
That seems like an obvious enough failure mode that there’d probably be limits on it. I do, however, think we’d be drowning in advertising from PACs and other politically involved non-government actors.
And if the person/group designing the checklist is clever, they’ll have a ton of influence. Which programs get grouped together, the names used to refer to each program, the ordering of the list… I wouldn’t be surprised if people check off more boxes near the beginning, so doing things alphabetically could lead to some weird naming conventions. Maybe they add ‘bundles’, so you can simultaneously support a bunch of military or redistribution programs – even more options for manipulation.
I wrote a short post about this on LW a couple months ago, as it happens.
(Short version: the opportunities for corruption are bad, but the basic inefficiencies of the system are probably worse. Programs don’t need money in proportion to their sexiness.)
They’re not making an argument they expect you to listen to. They’re providing a declaration of war.
(long time lurker who enjoys this blog and finds it very interesting, first time poster)
I’ve often imagined having a tax form that looked like a United Way donation sheet, only with your donation amount already filled in. You still get to earmark where your personal money goes.
I know it wouldn’t work, but I really wish it could. Heck, we could almost do away with politicians and just vote directly on the budgets 🙂
Sometimes you do see conservative rhetoric specifically involving the use of public money to find conduct that is deeply antithetical to their religious belief, and that doesn’t seem to be entirely limited to the “object-level” dispute over the merits. But I think in that case it’s too glib to extrapolate the point to be that nobody should contribute tax money toward programs that they oppose morally. I think a little more sympathetic attention to this argument would involve something like the recognition that our society didn’t spring into existence overnight last night, and that policies encroaching on a large segment of the population’a moral views should get special recognition where the moral views in question form part of the historical bedrock from which our society has evolved. You’d have to spell all this out with greater care and spend time making it more precise, but I do think it’s really unfair to sort of dismiss the whole thing in a way that suggests that it’s proponents never considered that some people oppose everything the government does and that you can’t realistically have an individual opt-out from contributing to tax-funded programs. I’m not even saying that the conservative view is correct for present purposes, just that the cavalier dismissal is way too simplistic to do the job.
Best,
CT
But by saying “I am particularly horrified that some of my money goes to support it”, they are appealing to a meta-rule that forcing people to pay for things they don’t like is wrong.
Just in case you ever need it (it helped me so it might help someone else) a control group for Mormons is the [Stragites](http://en.wikipedia.org/wiki/Strangites) — one of the splinter groups after Joseph Smith’s death that had new plates and witnesses.
The Mormon analogy is better than your LessWrong rebuttal to the McGrews, I think. I’d argue that we have ample disconfirmation of Mormonism that we don’t have for Christianity. For example, the archaeological record does not support any of the claims about the Americas made in the Book of Mormon; and Egyptologists have shown that Smith’s translation of the “Book of Abraham,” supposedly written in the same language as the plates from which he translated the Book of Mormon, bore not the slightest resemblance to the original.
The existence of an analogy between the Christian apostles and the Mormon witnesses could still be problematic, inasmuch as the latter is supposedly a case of multiple witnesses in similar circumstances testifying to the miraculous. But even if there is an analogy here, it’s open to the Christian apologist to say that this testimony is evidence for Mormonism; it just happens that we’ve got a lot of evidence against Mormonism as well.
I’m not convinced that the analogy is that strong — in particular, the testimony of the eight seems unremarkable on the hypothesis of fraud — but I haven’t read enough about the 11 witnesses to yet be confident on this point.
(long time lurker who enjoys this blog and finds it very interesting, first time poster)
First of all, the idea of a pre-Socratic Western is awesome and it strongly reminds me of Dungeon & Discourse.
Second, and I hope I’m not going offtopic…maybe it’s just me, but although no one has mentioned it I see a strong parallel to arguments about Free will and punishment for crime.
I’m thinking about – and I hope I’m not oversimplifying – the idea that
1. There is no free will.
2. With no free will, we should not hold people responsible for their actions.
3. Thus, we should not punish crime.
It seems to me that this is incoherent since it implies that ‘we’ have the free will to NOT punish people.
Basically, it seems inconsistent to me that the criminal has no free will but the judge, jury and justice system do.
I can think of partial counter arguments or flaws in this line of thought – for one, the people who argue for ‘no free will->no punishment’ wouldn’t have the will NOT to, but I do not find them very strong.
One could also argue that some people have free will and some do not – I suppose that is sort of implied by the insanity defense, in a way – but this is not an assumption I’ve ever seen in this kind of discussion (at least, not openly), where (lack of) free will is assumed to be some kind of universal.
(again, I do hope this is not an oversimplification of the matter, I do understand there are multiple interpretations of what free will or its lack means – it’s just that it seemed to me to be a field of philosopical discussion which is not very coherent so I thought I’d comment on it. )
This all assumes ought implies can. Determinists do not necessarily maintain that position.
You are of course correct – and I do know there are determinists who are ok with punishment justifying it in ways other than culpability-, but I’m not sure it applies in this specific case I’m talking about.
After all, shouldn’t the people saying we *ought* to stop doing something (punishing) believe we *can* do it if they are doing so as part of suggesting an actual policy change? What would the point of making the argument in a social (as in part of a wider society with a law system)setting, as some determinists do?
Again, of course, it’s not like they have the *choice* not to make the argument, if they are themselves determined!
And yet it still seems to me like a lack of rigor one could ‘accuse’ them of. I do see it as a less clear-cut case than the examples Scott provided, though.
An argument can be conceptualized as an input to the deterministic process of thinking. Exposure to good arguments about what “should” be can change minds and thus change the decisions people will make. It’s all predetermined, but going through the process is still necessary, since individuals don’t know which direction it’s predetermined until they’ve evaluated the argument.
Yes, the step from 1 to 2 seems incoherent. You could rewrite it into sanity, but I think that still leaves the step from 2 to 3 being also incoherent. Consider:
1. Cliffs have no free will.
2. With no free will, it makes no sense to hold cliffs responsible for their actions (such as dropping rocks on people).
3. Therefore, we should not punish cliffs (such as by detaining and restricting their freedom of movement).
And yet: People keep applying rock netting to cliffs.
There are other things than free will and desert that argue for punishment: incentive structures and danger containment.
Once you rewrite it to be coherent it becomes a very different argument though.
You are also right that there are pro-punishment arguments that, using utilitarian reasoning, do not requrie ‘blaming’ people at all.
Focusing on ‘danger containment’ you could even justify not calling it punishment – after all, with no moral judgement, it wouldn’t be THAT different from quarantining someone who has an extremely dangerous infectious disease.
That fits with your cliff example (although I doubt people blame cliffs!).
So my OP was really just referring to a quite specific line of thought that I’ve seen quite a few times – though I suspect (or at least, hope) that it’s not actually a popular position in ‘serious’ philosophical circles.
You are exhibiting a lack of comprehension about what the word “punishment” means. “Punishment” means action taken for the deterrent value. Applying netting is not at all motivated by deterrent value.
The right answer to your no-free-will example is John Perry’s:
By analogy, we just need to threaten putting the offending philosopher into a society with no deterrence for any crimes. They will quickly come up with some ersatz, watered-down vulgar version of a justice system that uses deterrence, and seek its protection. They might not call that deterrence “punishment”, but the rest of us will (hat tip: RCF). We might even insist that deterrence was the core of punishment all along, and that the philosopher’s metaphysical “free will” theory was beside the point. Note that the heavyweight metaphysics was equally beside the point whether the philosopher was against it (as in your example) or for it.
While I agree isolated demands for rigor are often selfishly motivated, sometimes they’re what you get from a rational Bayesian with the right sort of priors. While people should probably be held to a higher standard than this when making public arguments, I don’t think isolated demands for rigor are necessarily a bad thing, when private or individual.
So the Robin Williams episode has launched a thousand comments that really funny people are really depressed.
Really insightful, bright people are also more likely to be depressed.
This essay has really got me worried about SA.
Aw, Ozy would know. (A random SO of a random person might not, but certainly Ozy would. Right?)
Are the first two statements actually verified? All the research I’ve been linked to has failed to withstand preliminary sniffing.
And, if I may ask, what are SA and SO?
In this context, I’m guessing SA == Scott Alexander and SO == significant other.
Wait, why this essay? Because it’s insightful, bright, and really funny?
I think you may have just delivered the darkest underhanded compliment I’ve ever witnessed.
Also, SA has mentioned having experienced depression.
So there’s this tendency in my social surroundings to consider a certain ethnicity of people (Bongs, they’re called, for reasons unrelated to substance abuse) as more regionalist and backward. A lot of people think it’s because of some property of the Bongs or their culture. I think all of it is explained by the fact that Bongs form a majority of our community and so less rich/western people from that state turn up. And people have trouble accepting such a statement.
Now, an important part of the argument of my hypothesis is that there’s absolutely no reason to believe that Bongs are, on average, more regionalist than people of other ethnicities. I could tell them that this is because cultures* come in clusters with more internal variance than between-culture variance (before I read your posts I would have taken a hundred words and the condept of an effective theory to explain that, so thanks for that), and endure the blank stare thereby incurred. Or, I could tell them why what they call Bong culture is entirely arbitrary and it’s stupid to expect reality to follow your arbitrarily drawn lines (okay, the blank stares continue – let’s assume someone more comprehensible was explaining; it seems plausible that they’re much more comprehensible explaining the second thing than the first).
What I’m getting at is this: whether considering the clusterisation real is useful or not depends on the use it’s being put to, and people are more likely to call it out as a ‘false idea’ if they see particular clusterisations regularly being stretched beyond reasonable bounds – and you do have to admit that race in the western world (and also culture in these parts) was taken that way.
So yes, these people are being selectively rigorous, but… for a reason, you know.
As another example, suppose you’re a depressive whose job involves no math whatsoever and your brain throws up two hypotheses: a) math is hard for you, and b) you’re inherently bad with people of romantic and sexual interest. And it also provides much anecdotal evidence.
Obviously you’re going to be much more rigorous about the second hypothesis. Why? Because you realise that there could be a lot more at play in the existence of that hypothesis and you correctly suspect anything in that area. Note that this rigorousness, while epistemological, also incorporates other concerns about the costs of believing the second hypothesis if untrue.
I think that people, implicitly, are doing that exact calculation in all SJ writing. I’m sure that if I went around looking I could find excerpts where it’s explicit. And, to counteract that bias where you’re allowed to have arbitrarily many undeclared negative findings, I’ll wager that that calculation is made explicit in the king of all intros to feminism: http://badassdigest.com/2013/11/14/we-need-to-change-how-we-talk-about-rape/
And, as you yourself, have pointed out many a time, SJ is more often about social warfare than truth per se. Like bending a misshapen spoon tooo far the other way to rightify it.
That said, I agree with your discomfort with all this (despite liberally using selective rigour in life). Especially the IQ thing, which doesn’t seem to be particularly misused – leading me to suspect that people are trying to escape the implications of believing in the implications of the existence of intelligence.
*I’ve always found culture to be a weird idea, but that’s probably part of living in India.
Wow, for someone no good at explaining stuff, you do a damn good job.
And while we’re talking about low-dimensional scales for an arbitrarily large number of factors, let’s not forget money.
Alternate ending line “The only sufficient rigor is rigor mortis.”
Maybe this incarnation of the joke doesn’t need it, but I suggest the British spelling:
Reading these posts, I have thought and continue to think that you’re knocking on, if not an open door, then at least one that’s ajar. Simply put, many people look at what has been done (and what continues to be done) in the name of the concept “race,” and conclude that no, we really can’t have anything like disinterested scientific inquiry about it. The need to prevent future injuries–both macro-level atrocities and daily microagressions–is just too urgent. Hence, nothing whatsoever should be conceded intellectually to racists, including the concept of “race” itself, which should be eliminated from discourse via social norms (not legal ones! calm down!).
Unfortunately, most of them (leftists, I mean) don’t see the problem clearly enough to pose it this way. Rather, confirmation bias leads them to one or more theories (e.g. Lewontin) that race isn’t real. But I’ve come to believe that the esoteric, scorched-earth position on race is much more intellectually consistent. I’m on the fence about whether I agree with it. But when I scan my Twitter feed and see the latest news from Ferguson, it seems pretty damned reasonable, even if it is irrational.
First, this sounds to me as though it’s trying to make reductio ad hitlerum become a standard form of reasoning. And that’s terrible.
Second, I’m reminded of what happened with ‘HBD’ (Human Biodiversity): some people coined a new term and very soon other people shouted “crypto-racist”. The scorched earth position seems as though it would run into similar difficulty. Supposing that you were successful in suppressing all discussion of “race”, someone is going to rediscover or reinvent the concept under another name for use in discussing related real phenomena, and poof, we’re back on the euphemism treadmill.
Third, the people saying “too white” (some in high places) are going to be loath to give up their linguistic tools and concepts. If most leftists don’t see the problem clearly enough, as you say, I think that’s hardly the biggest obstacle; many leftists are invested in the concept of race as a bludgeon to wield against their enemies.
The question isn’t whether it’s terrible, the question is whether the terribleness can be justified because it makes racism literally impossible to think. This is basically a fait accompli in left-leaning academic disciplines like sociology and anthropology—disciplines whose influence filters out through the media, think tanks, etc. That influence used to be predominantly racist, and now it is thoroughly leftist and anti-racist. It’s hard not to cheer that development no matter what your politics are.
Also, when you’re weighing the scales of terribleness, note that almost no one is proposing to instate taboos that extend beyond race and gender (though the latter is a little more complicated). You can say whatever you like about climate change, antibiotic-resistant bacteria, black holes, quantum mechanics, etc., without the anti-racist police interfering. The essence of scientific inquiry is not threatened. To pretend otherwise, one has to engage in some extremely tendentious slippery-slope arguments.
No problem here. You yourself point out that leftists aren’t the least bit confused by this sort of thing.
No, these are mostly the same people who have given up race as a concept, and the apparent contradiction is easily resolved. When they say “too white,” they mean “benefiting from race qua social fiction”—and having a perspective that is conditioned by, yet oblivious to, the benefits of white privilege. To deny that race is real is not to deny that race as a social category has real effects—it’s just the opposite. So the fact that white privilege becomes a “bludgeon” (as you put it) to wield against the privileged—well, that’s not an inconsistency; it’s just politics.
In conclusion, let me repeat that I’m half-playing devil’s advocate… but in this case I happen to have a lot of sympathy for the devil.
It’s hard not to cheer that development no matter what your politics are, because the people whose politics it aligns with are going to be watching and making damn sure you cheer.
It’s hard not to cheer that development no matter what your politics are, because politics that disagree with me sufficiently are not really politics, but simply evil.
It’s hard not to cheer that development no matter what your politics are, because the thought of not cheering that development has been made literally impossible to think, and it’s hard not to cheer that development no matter what your politics are.
I’m pretty sure its impossible to make racism literally impossible to think.
Sorry. Despite being an avowed grammar Nazi, I ended up abusing “literally” in textbook fashion. Of course, it is possible to be, e.g., a contemporary sociologist and simultaneously a racist (or sexist). But were you to express those racist opinions as part of your work, they would be impossible to fit within the accepted theoretical framework that all of your colleagues would be using. Any one of your peers would be able to apply the accepted concepts in a paint-by-numbers way, and you would come out looking ridiculous.
I’m trying to say that this state of affairs has an ambiguous value, because it’s undeniable that these disciplines are much less racist than they used to be.
It’s basically racist now to say that the West had anything going for it other than a superior willingness to rape and steal when it kicked off the industrial revolution. Never mind racial realism, even referencing purely cultural superiorities like scientific rationalism and capital markets can be shouted down as “racist.”
I’ve come to agree with those who say that anti-racist is code for anti-white. Whatever value it once had has been subsumed entirely by its use as a weapon, to the point that I’m willing to oppose anti-racism on principle despite once thinking of myself as anti-racist myself. (Until I was informed that wanting to judge people by individual merits is actually racist.)
People arguing that you can’t say X about cultural superiority in the West because people will shout ‘racist’ must not have been actually reading any of the academic literature produced over the past decade.
I mean you’ll certainly be called a racist by some people, overtly or covertly, but your research will get a ton of play nonetheless. Basically the entire Great Divergence area of history is filled with cultural superiority books and articles. Gregory Clark still is well-respected and widely published. The idea that you can’t say or research these things just doesn’t mesh with reality, no matter how much you’d like to believe that your beliefs are all suppressed by some liberal conspiracy.
Also, it’s perfectly respectable — even incentivized — to talk about cultural superiority as long as it’s aimed at certain people. Josh Barro published an article called “There Are Two Americas And One Is Better Than The Other”.
I was talking to a mainstream progressive relative a while back. Mentioned the Bush-era United States of Canada vs. Jesusland maps, so she asked, “which one would you live in?” I said I’d take the one that didn’t call for the destruction of my people (and she heard about my firsthand experiences with genocidal rhetoric when it happened), to which she replied, “but you’re not a Bible-thumping—”
“No, I know a lot of people in Texas, and none of them are like that.”
“In Austin?”
“No, they don’t like Austin.”
“Well… they must not be women. If they were, they’d hate it, since it’s almost impossible to get an abortion in Texas.”
(I later mentioned Switzerland as an example of a country that deals well with cultural divisions, and got the reply, “what, the place that didn’t let women vote until 1970?”)
@nydwracu
Okay, now you’re making up an asymmetry that just isn’t there. At the time of the USofC v. Jesusland map, plenty of conservatives were making fun of Kerry for being a “Massachusetts liberal,” as if being from Massachusetts was something shameful in itself. And calls for genocide pretty clearly fall at least as much on the right side of the political divide. I would say much more so, in fact. How many liberals’ cars have you seen with bumper stickers advertising their hunting permit for conservatives?
On race, I agree that there is an asymmetry between what you can get away with saying without opprobrium about whites v. minorities. On ideology, I call bullshit.
Pretty much every autobiographical comment you make comes off as extremely filtered through confirmation bias. You’ve apparently never met a liberal who wasn’t extraordinarily intolerant nor a conservative who wasn’t a bastion of tolerance. Suffice it to say that other people have had… contrasting experiences.
The relevant difference isn’t an abstract crypto-deontological one; it’s in the presence of the universalizing impulse among liberals. (And neocons.)
How many conservatives care about what liberals do in Massachusetts, as long as it doesn’t filter back to them? How many liberals care about what conservatives do in Texas, as long as it doesn’t filter back to them?
I’ve seen the hunting-permit stuff. I don’t really care about that. I’ve heard about people getting nasty looks out in Western Maryland for prog-signaling bumper stickers, and I don’t care about that either. “This particular piece of land is ours, we’re going to do things our way, and if you don’t like it, get out” is a fundamentally different thing than “everyone should be like us”.
The contrasting governmental pipe-dreams of conservatism and liberalism — an independent CSA vs. a world government — make the distinction clear.
How many conservatives care about what liberals do in Massachusetts, as long as it doesn’t filter back to them? How many liberals care about what conservatives do in Texas, as long as it doesn’t filter back to them?
Again, this is just totally at odds with the actual reaction of out-of-state conservatives to, for example, Massachusetts becoming the first state to legalize gay marriage. Which is not to say that conservatives weren’t worried about a domino effect as well, but if you actually go back and look at the rhetoric they were very much opposed to gay marriage being legal anywhere.
And, seriously, the CSA is a terrible example for your point. “States’ rights” was always a load of bullshit. The pro-slavery forces were absolutely livid at abolitionist and then Republican refusal to use the coercive power of the Federal Government to enforce the Fugitive Slave Act; you can easily verify this by reading the declarations of secession.
nydwracu
“No, I know a lot of people in Texas, and none of them are like that.”
The only reason homosexual activity isn’t a felony in Texas is because the federal courts stepped in. Texas is rife with SOCAS violations, and being an open atheist is quite dangerous. Either you live in a liberal bubble, or your threshold for being a “Bible-thumper” is incredibly high.
[delete me]
Hi Thom,
As I’ve urged before in this and other threads, willful ignorance of facts about racial variations hurts underprivileged minorities more than anyone else. Would you deny medical knowledge about different susceptibility to disease or different effects of certain drugs among different ethnic and racial groups? Are Native Americans made better off by our being shocked, shocked, when someone suggests that their susceptibility to alcoholism might be genetic, thus squelching any research into how to help with this? Are poor countries around the world made better off when we assume that humans are a blank slate, and so that any policies that work well in Scandinavia can be made to work equally well in Zambia? Are African American children better off when we refuse to discuss the black-white IQ gap and so prevent schoolteachers from separating children out by ability (since that would lead to more blacks in the lower ability groups than higher ability groups, and so ipso facto be proof of racism on the part of schoolteachers) so as to teach children with similar abilities more effectively?
The only people taboos about race “help” are well-off white progressives by serving as a bludgeon for them to browbeat conservatives with. (I, for one, do not see the near total domination of academia by progressives today as a good thing.)
Except we already have a lot of the non-IQ-based interventions you suggest. This discourse is not preventing people from investigating whether black people are more susceptible to sickle cell anemia, or whether Native Americans or other groups are more susceptible to alcoholism etc. We already do that. And we base policy on that. There’s no problem there.
The problem with race-based IQ research is that we have a really, really long history of using (in hindsight terribly flawed) science to ‘prove’ that certain races are dumb, and then using that to do terrible things to them on various scales of terribleness, starting with slavery, progressing through Jim Crow and de jure segregation to de facto segregation, ghetto formation and more.
And while your ideal of using this to help people learn better sounds terrific on its face, every bit of recent history suggests that we would instead simply under-fund the stupid groups, because there are more black people in it. In fact, we’d probably see calls to separate the schools (after all, why do the dumb kids and smart kids need to be in the same institution?), and then magically the dumb schools would primarily be located in poor neighborhoods and would receive significantly less funding. And then we’d complain about inner-city schools, see an outflow of white students to better-funded private schools, leaving the poor, black folks behind.
Because that’s exactly what we’ve been doing for decades.
This is also why ‘isolated demands for rigor’ when discussing IQ make a lot of sense: we have repeatedly shown that we cannot be trusted with this knowledge. Until we can, we should probably stay away.
Hi Menelaos,
Here’s the position that Thom was playing devil’s advocate for, in his own words: “Simply put, many people look at what has been done (and what continues to be done) in the name of the concept “race,” and conclude that no, we really can’t have anything like disinterested scientific inquiry about it. … Hence, nothing whatsoever should be conceded intellectually to racists, including the concept of “race” itself, which should be eliminated from discourse via social norms.”
In other words, this position advocates never using the concept of “race” [or perhaps it only advocates this when the concept is understood as a genetic category], not just not using it when talking about IQ. And race-based medicine is controversial, although less controversial than race and IQ research. For example, according to this article by Nicholas Wade, the New England Journal of Medicine’s official position is that “race is so poorly defined that it is not a reliable biological concept and should be banished, if possible, from scientific vocabulary” (http://www.nytimes.com/2002/07/30/science/race-is-seen-as-real-guide-to-track-roots-of-disease.html).
And while your ideal of using this to help people learn better sounds terrific on its face, every bit of recent history suggests that we would instead simply under-fund the stupid groups, because there are more black people in it. In fact, we’d probably see calls to separate the schools (after all, why do the dumb kids and smart kids need to be in the same institution?), and then magically the dumb schools would primarily be located in poor neighborhoods and would receive significantly less funding. And then we’d complain about inner-city schools, see an outflow of white students to better-funded private schools, leaving the poor, black folks behind.
Come on. Teachers and education policy-makers are not neoreactionaries. This is a ridiculous slippery slope argument, filled with wild hypotheticals. The fact is that right now our education practices are clearly sub-optimal, and that there’s a very obvious way to improve them. Indeed, the sub-optimality I’m pointing to explains some of the inequities you’re concerned about. To the extent that underfunding of poor or black schools is a problem (we would disagree, no doubt, on the extent to which this explains the problems present in those schools), it’s largely due to the kinds of stupid policies we get when ignoring IQ: like basing funding on absolute test scores rather than test scores controlled for variables like race, or the average improvement on test scores over the course of a year. In other words, we treat children like blank slates, assume that disparate outcomes are the fault of the teachers and administrators, and punish the latter accordingly by lowering funding. A sensible education policy which recognized the importance of IQ would not have this result.
Note too that I’m not calling for some general education policy requiring ability sorting, which would undoubtedly be an overly centralized, overly simplified, bureaucratic mess. I’m just calling for teachers and administrators to be allowed to organize their classes and schools by ability without fear of lawsuit, and for the pros and cons of ability sorting to be freely discussed in, say, education school.
Basically, I’m asking you to trust America’s teachers. 🙂
Troy,
Your point on the New England Journal of Medicine is well-taken, but it’s a point easily sidestepped: talking about genetic sub-groups instead of race is easily and often done within medicine. The idea that there is real race-based medicine we could be doing but aren’t isn’t really grounded in what’s actually happening in research. You can find a ton of research on how various genetic clusters are susceptible to various diseases, and race is often used where appropriate to communicate that. While ‘race is not real’ is a slogan, people are not insane enough to use it to deny useful medical research. Unless this ends up happening, this is a non-issue for me.
As for ability sorting, again I point to recent history to suggest that “we should trust that these people will get it right” is not something we should be doing. Because when we’ve done that, the results have been pretty consistently, pretty terrible. There’s very little reason to believe that teachers will be any more immune than non-teachers.
I do find your remark that “teachers and education policy-makers are not reactionaries” telling. The key point to racism is that you do not need to be a neoreactionary to perpetuate racist policy. You don’t need to believe in the superiority of whiteness to perpetuate its societal position as such. The fact that black schools are often underfunded isn’t due to neoreactionaries who hate black people. It’s due to a range of different things tied to how we treat poverty, and how we consciously or subconsciously treat black people. To suggest that it’ll be okay because teachers will do the best they can and that that will prevent these problems is not how society currently works.
Lest you think this is a theoretical point, we have a real-world example. Dutch schools use ability sorting, with an end-of-primary school test coupled with a primary-school-issued advice leading to sorting into various secondary school levels. The advice takes precedence over the test score. Sounds good, right? Ability sorting, teachers are given the leeway to do that sorting to the best of their abilities, there’s an ‘objective’ component to it to help guide them, and I’m sure they’re all fine people.
Except ‘immigrant’ children (includes at least second-generation immigrants) are consistently given the advice to be sorted into a lower-level school than non-immigrant children with identical test scores. Pretty strong evidence of conscious or subconscious bias at work. And that’s before we even get to some potentially problematic designs of such tests, which would only exacerbate this issue.
Meneleos: Your Dutch story sounds like an excellent argument that tests should be used without any input by humans who can introduce bias into the system, which is exactly the reason I am so pro-IQ test.
Your point on the New England Journal of Medicine is well-taken, but it’s a point easily sidestepped: talking about genetic sub-groups instead of race is easily and often done within medicine.
Sure, and we can do the same thing with IQ. In all of these areas it’s possible to functionally talk about race and X without actually mentioning race.
The fact that black schools are often underfunded isn’t due to neoreactionaries who hate black people. It’s due to a range of different things tied to how we treat poverty, and how we consciously or subconsciously treat black people.
Here at least we agree; I just think it’s false assumptions about cognitive equality that lead to misappropriation of funds!
Edit, on Dutch teachers: I think Scott’s point is a good one (that your anecdote seems to be an argument for, and not against, IQ tests). I admit to being more skeptical than Scott, though, that the teachers are biased here. I’m not familiar with the situation and so can’t say with any confidence whether there are legitimate reasons for this pattern, but I can think of many reasons teachers might have for giving advice that didn’t directly mirror the tests: for example, language difficulties, behavioral problems, and suspicion of cheating. I don’t know if any of those are the case here; but I tend to give teachers the benefit of the doubt in these kinds of cases, at least until I know more of the story. I think that decisions based purely on IQ tests without any human input, like any centralized decision-making process, are too apt to ignore contextual factors (contextual factors which might, of course, not be evenly distributed among all demographic groups).
Scott:
That assumes that the initial design of the test by humans doesn’t already introduce bias. Even so, I generally don’t have an issue with IQ tests, but I do have issues with the policies I’ve seen resulting from the use of IQ tests when related to population groups rather than individuals. I’m not arguing against the accuracy of testing, I’m arguing that we’ve repeatedly shown that we will use testing to justify terrible things when it comes to race (or gender for that matter).
Troy:
It’s not just that we can talk about race without calling it race, it’s that there’s often no need to talk about race when talking about genetic population groups. Even so, medical research has not shunned identifying medical problems based on modern conceptions of race alone. To pretend that this is not happening is to invent a problem.
As for your defense of the Dutch example, think of it this way: is this more likely to be an outcropping of a phenomenon we know exists (subconscious racist biases), or is it the result of a certain group of people displaying a variety of symptoms at a much higher level than another group of people, despite identical test results which actually test for several of those issues?
I can think of several reasons why those results might not be due to bias, but I cannot think of any reasons that are more likely than bias.
As for your defense of the Dutch example, think of it this way: is this more likely to be an outcropping of a phenomenon we know exists (subconscious racist biases), or is it the result of a certain group of people displaying a variety of symptoms at a much higher level than another group of people, despite identical test results which actually test for several of those issues?
Not knowing much about the situation the second probability is hard to calculate. But I suspect that our priors for teachers being affected by subconscious racist biases are different. I don’t at all deny that people rely on racial stereotypes, but the evidence I’ve seen suggests that it’s more common when people have little other information to go on, which would presumably not be the case for teachers who have these children in class (although I suppose this depends on things like class sizes and who’s making the recommendations).
Familiarity can help reduce the impact of unconscious biases in specific cases, but they certainly don’t eliminate them. This pdf is an okay introduction to some of the ways this works.
Go back to your reasons: you suggested language (tested for), cheating (if you cheat (and get caught) you don’t pass), and psychological issues. The only one not tested for is not something you’d expect to see more of among immigrants. Which to me suggests that you’re looking for reasons to explain away the obvious answer of bias, rather than looking at the evidence and taking the most likely answer.
Which is not uncommon when we talk about racism. The need to justify, rationalize and explain away racial bias is overwhelming among the privileged class. You see it all the time, even in cases where we know these issues are due to racism.
@Menelaos:
Do you have a source for your account of that Dutch school bias (immigrants given advice to apply for a lower level schools than whites with a same test score)? This sounds very relevant to my interests, but googling fails me, most likely because I can’t think of the right search terms.
Only Dutch sources, which I don’t think will be overly useful for you. Here’s the main report I’m basing this on: http://www.amsterdam.nl/publish/pages/91307/rapportdepositievanallochtoneleerlingenbijdeoverstapvanbasisnaarvoortgezetonderwijs.pdf
To be fair, a later report suggest that this difference disappeared in multivariate analysis when corrected for behavioral problems observed by the teacher — but that uncritically accepts the teacher’s evaluation as accurate, which is the problem to begin with.
(http://www.onderwijsinspectie.nl/binaries/content/assets/Actueel_publicaties/2011/Onderadvisering+van+allochtone+leerlingen.pdf)
In addition, that report ignores the aspect of polling of student experiences in in secondary school. Those polls backup the original report, not the multivariate analysis.
Thanks! Google Translate manages Dutch-to-English translation quite well, so these links were very helpful.
The need to prevent a small subset of actions, which you identify completely with “future injuries”, from happening to the small subset of people who identify with factions you’re allied with.
Too many IQ-related posts so I’ll just post this here.
We now have high-res brain scanning technology. We’ve identified important structural components of the brain from the neuron to the modular scale. We’ve recently identified the 100-neuron “learning unit” in the neocortex. Does it bug anybody else that nobody is suggesting where g fits in with neurology?
This would seem to be extremely important. For one thing, if we understood where IQ/g come from, we could target interventions during development to ensure the absolute best possible outcomes, rather than relying on ad hoc experiments – “I guess breastfeeding and iodine are important? Maybe learn an instrument?” Additionally, if we really understood it, we could design drugs or even training methods that could actually boost effective intelligence during adulthood, at least in certain regards.
Is there research into this that I’m just not aware of? Quite possible since I’m totally ignorant, I’m more remarking on the fact that I don’t see it being discussed here at all.
Starting point: http://scholar.google.com/scholar?q=fluid%20intelligence%20fmri One of the more interesting neural correlates seems to be global connectivity, eg http://www.ploscompbiol.org/article/info%3Adoi%2F10.1371%2Fjournal.pcbi.1000395 or http://www.jneurosci.org/content/32/26/8988.long
Thanks.
Gwern, what are your current estimates of the impact of N-back training on children’s long-term Gf scores?
It just really *seems* like N-back training should work.
edit: To clarify, I’ve read your FAQ, just wondered what your current opinion on this particular topic was.
Very low. Aside from the problems with n-back, the Buschkuel lab has been hammering away at various flavors of WM training for children for over a decade now, and at best they’ve been able to show small increases in some kinds of grades when focusing on children with low WM. Combine that with the notorious fadeout and hollow IQ gains of programs like Headstart…
Some problems are easier than others. Genetic tests are pretty good at predicting milk production in cows, but people have no idea what the particular variations mean. IQ is going to enter that phase real soon now.
The climax is no doubt when the hero strides onto Main Street, revolver in hand, saying “There’s a new sheriff in town.” And Parmenides gruffly responds “No, I’m pretty sure that’s impossible.”
Dude, that is fuckin hilarious!
Frankly, I think you’ve missed Yglesias’s point.
I don’t think he’s saying that the objective of all policy should be to maximize global utility (he might think that, but that’s neither here nor there as far as his article goes). He’s saying that insofar as airstrikes are foreign humanitarian policy, they are very cost ineffective.
There are policies other than foreign humanitarian policies. Obviously, there are domestic policies focused on the USA too. There are also foreign strategic policies, which do not have humanitarian objectives. Then, at an even broader level, there’s the balance between public and private sectors and general — how much should we tax/spend overall?
Yglesias’s argument is not about the appropriate proportion of resources going toward domestic policies, foreign humanitarian policies, and foreign strategic policies. Rather, his argument is about the efficiency of policies within the realm of foreign humanitarian policies.
You might say that’s being selective by leaving whole other realms of policy off the table, but 1) I think he makes his assumptions that he’s making a limited case within the realm of foreign humanitarian policy reasonably clear, and 2) I think that it’s often reasonable to talk about the best deployment of resources within some limited domain. I think it sometimes makes sense, e.g., to talk about whether it’s a good idea for limited education resources to go toward, e.g., higher teacher salaries vs. facilities repair, without always bringing in “well, what if we cut money from the Joint Strike Fighter program to increase education spending overall? What if we cut education spending and teacher’s salaries to increase Social Security?”
I’ve been mulling it over, and I’m no longer convinced that Matt Yglesias’ argument is a selective demand for rigor. Consider that one might view things through a framework like this:
1. There are certain functions that it is right and proper remit for a national government to expend funds on.
2. Saving lives of people who are not citizens of that government is a legitimate expense, and it is proper to apportion some fraction of the national budget to accomplish that.
3. The national government also has other legitimate priorities among its own citizens, and it is also proper to apportion some fraction of the national budget to domestic issues.
4. Within the bucket assigned to help foreigners, “prevent malaria” is a more efficient means of saving lives than “drop bombs,” and thus we should prefer the former to the latter.
5. It does not follow from this that we should reapportion funds between the “help citizens” and “help foreigners” buckets, because the national government still has other legitimate priorities.