My name sounds a lot like Scott Aaronson’s and I get confused for him a lot. I try to encourage this confusion, since it can only increase people’s opinion of me. So let me propose a tool for investigating morality through algorithmic systems very similar to Aaronson’s recent post on eigenmorality.
I say we use DW-Nominate.
I wrote about DW-Nominate before. It’s the tool political scientists use to calculate Congresspeople’s position on the political spectrum. Whenever you hear an alarmed-sounding voice on a black-and-white attack ad say something like “Senator Schmendrick is the third most liberal senator in Congress,” chances are they used DW-Nominate to calculate “third most liberal”.
The system is beautifully elegant. They take a Congressperson’s votes on all the issues and compare them to other Congresspeople’s votes to find blocs of Congresspeople who tend to vote together. Then they do factor analysis stuff to see how many dimensions of “similar voting” there are. They end up with something that looks a lot like the traditional left-right dimension and occasionally a Northern-US-vs.-Southern-US-dimension that doesn’t always matter that much. Then they use Congresspeople with multi-decade careers to bridge the gap between current Congresses and past Congresses, and use dead Congresspeople with multi-decade careers to bridge the gap between past Congresses and even-further-in-the-past Congresses, so that they can compare any Congressperson, living or dead, to any other Congressperson, living or dead. They also get the opportunity to evaluate bills as liberal or conservative, based on whether liberal or conservative Congresspeople support them.
And the neat thing about it is that at no point did they enter into the system that it was supposed to give “left” vs. “right” – or even that it was supposed to come out with only one major grouping. It could have found that actually Democrats and Republicans vote much the same, but men always vote with other men and women with other women, or that the real difference was a religious worldview versus a secular worldview or whatever. Instead they found that our notion of left and right emerges naturally from the data, even if you’re not looking for it, and that this transcends party lines – ie some Democrats are further left than other Democrats, and this has consistent effects. If you really wanted, you could use this to rate whether, say, cutting carbon emissions vs. gun control is a more truly leftist cause, by seeing which bills get more heavily supported by leftists.
And I wonder what would happen if you tried DW-Nominate with moral decisions.
Now there’s a very boring interpretation of that proposal, which is that we hand a hundred people a hundred different multiple-choice questions on moral dilemmas, and then use factor analysis stuff to see if we can divide them into groups. I bet we’d come out with something a lot like “utilitarians vs. deontologists vs. virtue ethics”, or maybe something more like “religious vs. secular” or maybe “people who use all Haidtian foundations vs. people who just use care and fairness”. Actually, forget what I said before, this would already be a quite interesting thing to do and somebody should do it.
But I would be much more interested in a (much harder) naturalistic experiment. What if we took the real decisions people engage in? I’m not even talking about obviously morally charged decisions like whether to have an abortion, I’m talking about things from “what college major should I have?” all the way down to “do I drink alcohol at age 17?” to “do I call my parents tonight like I promised I would, even though I’m very tired?”
A lot of people have to go through very similar decisions, which allows DW-Nominate style ranking. There would be some wiggle room in deciding which two decisions were equivalent (is the person who decides not to call loving parents making the same decision as the person who decides not to call abusive parents?), but let’s say we get a panel of raters to decide among themselves which decisions are equivalent and throw out any they can’t agree upon. This isn’t meant to be Pure And Objective here, only statistically useful.
In the same way that ten thousand Congressional votes, suitably analyzed, naturally group people into two categories that look to our trained eyes like Left and Right, would ten thousand little life decisions, suitably analyzed, naturally group people into two categories that look to our trained eyes like Good and Bad?
If so, it would be pretty easy to tell who the best person was, in the same way we can identify the most liberal member of Congress. We could give them a nice little award. Even better, it would be pretty easy to tell which option on each decision is more moral, for the same reason DW-Nominate can tell us that supporting gun control is more liberal than opposing it.
Suppose that we learned that one factor that naturally fell out of the data included giving money to the poor, supporting one’s aging parents, never committing violent crimes, avoiding ethnic slurs, conserving water and electricity, helping one’s friends when they were in trouble, and everything else we traditionally think of as good moral choices. And suppose this factor was heavily, heavily associated with being pro-life, to the same degree that the “liberal” factor in DW-Nominate is heavily associated with gun control. Would this provide some evidence in the debate over abortion? I’m not sure, but it would sure get me thinking long and hard about it.
I mean, we would probably also find some really silly things . Like that our moral factor loads on not getting tattoos of flaming skulls, ie the decision to get a tattoo of a flaming skull clusters with lots of immoral decisions. Presumably we would want to be able to say that getting a flaming skull tattoo is not itself immoral, but is correlated with immorality. But then we might as well say the same thing about being pro-life. Indeed, maybe everything religious will end out correlated with morality for religious reasons. We’d probably have to sort through this and fight a bunch of interminable correlation vs. causation debates.
But then there are areas where this could really shine.
I think this might solve a problem that Aaronson thought was unsolveable in his proposed algorithms. He said that in a world that was completely backwards – for example Nazi Germany – where everybody thought right was wrong and wrong was right, any moral sorting algorithm will give backwards results because it has to start with majority opinion in some sense. His example was that a PageRank style algorithm where people-believed-to-be-moral are the ones people-believed-to-be-moral believe are moral would fail, because most Nazis would believe that the high-ranking Nazi authorities were moral, and then the circle would complete with the high-ranking Nazi authorities getting to determine who the moral people were.
I think DW-Nominate might go part of the way solving that problem. Consider three different things we might find if we DW-Nominated Nazi Germany:
1. There is a General Factor of Morality, which includes giving to the poor, caring for your aged parents, cooperating with your neighbors, et cetera. People high in this General Factor of Morality are much more likely to oppose Nazi policies and hide Jews in their attics.
2. There is a General Factor of Morality, which includes giving to the poor, caring for your aged parents, cooperating with your neighbors, et cetera. People high in this General Factor of Morality are no more likely to hide Jews than anyone else, or maybe less likely to hide Jews.
3. There are multiple dimensions of morality. One dimension is something like “prosocial in-group patriotism” and captures things like paying your taxes on time, going without luxuries in order to help the war effort, and sending nice care packages to the troops. Another dimension is something like “willingness to go against consensus when it’s the right thing to do” and would include whistleblowing against corruption and being a passive resister to unjust wars. Hiding Jews in your attic might be negatively correlated with the first factor but positively correlated with the second factor. Universally beloved things like giving to the poor and caring for your aged parents might load about equally on both factors, or be a third factor, or whatever.
If Hypothesis 1 were true, that would be super interesting. It would suggest there’s something kind of objective about morality. Also, we could make it do work. Like we could go around to the Nazis, and say “Look, you agree that helping the poor is moral, right? And caring for your aged parents? Well, now that we’ve established what morality is, we have bad news for you. You don’t have it. Moral people are much more likely to oppose you. So stop doing what you’re doing.” This might actually work. Or if it didn’t, then when World War II ended and everyone agreed they should have listened to the General Factor Of Morality, then maybe after ten or twenty iterations of this people would start listening eventually.
If Hypothesis 2 were true, that would also be super interesting, albeit disappointing. It would mean that morality probably isn’t very objective, and that our moral positions are a lot closer to random than we want to believe. If being moral in every other way we can think of had minimal correlation with being moral in the particular way of saving Jews from the Nazis, it would mean that there was no consistent basis to morality and it was just a hodgepodge of popular positions. Or that if there was a philosophically consistent basis, it has little to do with how it’s practiced in the real world.
If Hypothesis 3 were true, that would be very boring, but possibly still worthwhile. Like we could have debates on whether Factor I Morality is more important than Factor II morality, and what to do when they contradict each other, and these debates would probably be more interesting than our current more vague debates on things like “what do you do when your duty and your moral intuitions conflict?”
I don’t really have some grand plan for how this could be used to solve everything or how a utopia could be created around it (although now that I mention it, if we can easily identify the most moral people in a population, they would make good candidates for judges and other high officials, though perhaps not legislators or executives).
I just think it would be fun to study.
I imagine this study could be massively confounded by akrasia and conscientiousness.
For instance, suppose that Hypothesis 1 is true, but for some reason conscientiousness is correlated with having muddled, contradictory, middle-of-the ground beliefs in morality. The result is that the greatest amount of moral behavior would be correlated with having average, status-quo-biased moral beliefs.
Or maybe Hypothesis 1 is correct, but akrasia prevents people who are high in General Factor from doing things that go against the status quo. In this case Hypothesis 2 would seem to be correct at first glance, and Hypothesis 1 would only be revealed to be correct if you found some way of controlling for akrasia.
Or maybe Hypothesis 3 is correct, but certain dimensions overcome akrasia more easily than others. This would result in certain dimensions appearing far more important than they actually are if you fail to control for akrasia.
I feel like from an outside view someone who does lots of morally good actions really is more moral (and as such more suited to be a societal leader etc.) than someone who has extremely accurate moral intuitions but never acts on them.
Yeah. It’s less important that Huck Finn be able to formulate the principle that it’s wrong to betray people, especially your friends and benefactors, than that he does not, in fact, betray someone, who is his friend.
@lmm
Not necessarily. You need to overcome a lot less akrasia to command other people to do something moral than you do to do something moral yourself. Compare how much taxpayer money Congress spends on welfare programs vs. how much of their salaries they donate to welfare programs.
Let’s say I wanted to help someone do this study, because I found it VERY INTERESTING and RELEVANT TO MY INTERESTS. As a computer programmer and database guy who isn’t vetted in Academia, what’s the easiest way for me to contribute / help kick off something like this?
My first thought was that people post a lot of information on their social media accounts. Also, people seem to love applications that generate statistics based off their accounts.
So maybe create an app that e.g. requests access to a user’s Facebook data (due to various privacy concerns, probably best if it just asks for their public data) and then extracts data from that. E.g. places of study, events visited, websites “liked” or shared, and a lot of other information could be gathered from there and transformed into a format where you could run some suitable algorithm on it. For extra coverage, offer the option to link accounts on other social media services to it.
Then you generate a report of the user, indicating which of the things in their data correlate with which clusters, and give the app some sexy name like “ARE YOU GOOD OR EVIL” to get people to use it.
I’m not sure of the extent to which you’d actually manage to classify people’s moral decisions this way, but you could at least classify their political affiliations. Maybe give some set of questions about moral dilemmas for all users to answer, though that would only give the answers they think they ought to give rather than the things they actually do. (If you could convince the people at YourMorals.org to offer their users the option to share data, that could be really interesting.)
I would say find morally important questions on the General Social Survey and use those as proof of concept first. Google “general social survey” and you should find the info you need.
If you mean how to get it published, I dunno. Ask James Miller or one of the other academics here.
“If Hypothesis 1 were true, that would be super interesting. It would suggest there’s something kind of objective about morality.”
It would mean that there are clusters of beliefs about morality. It would say nothing about whether people’s beliefs about morality are correct or incorrect. Using people’s beliefs about morality to find out stuff about morality is only as useful as using people’s beliefs about physics to find out stuff about physics.
It would be interesting to see which clusters are bigger than others, and what they rank on.
But the difference between physics and morality is that if you want to test, say, whether a lighter object falls faster than a heavier one, all you need is a couple bowling balls of various weights and a nice tower. And it might help if you can have a few guards to keep people away from the drop zone while you’re testing.
But the point is, physics can be tested.
Morality can’t. Short of sending a probe into the afterlife and seeing Anubis weighing the hearts of the dead against the Feather of Justice, we don’t know what morality is.
We don’t have consensus on what “good policy” is and we can see people’s ideas of good policy clustering in some pretty interesting ways. It would be really really interesting, and I wanna see it happen.
You’re assuming the existence of objective morality, which you haven’t proven. To my eyes morality looks subjective, a lot more similar to left-right politics than to physics.
Objective morality doesn’t have to be mind-independent, as is commonly assumed. See moral constructivism, particularly Hobbesian constructivism. In this conception, people can be mistaken about moral truths, without any mind-independent stone-tablet morality.
If you want to do meta-ethics, the interesting question is: if there is no such thing as a moral fact, why does everyone find it 100% intuitive to reason as if there is such a thing? Why do we all think the words “good” and “evil” have objective or intersubjective meanings (cognitivism: moral statements pick out specific subsets of concept-space) if they’re actually just expressing our vague, fuzzy, irrational feelings (emotivism)? At the same time, given cognitivism, and given a lack of an error-theory, how the hell do we locate the correct morality, and why does it seem to be only partially consistent over time and space (realism, even if humans-only realism)?
To me the obvious parallel is attractiveness. Which feels as though it’s a fact about someone, but actually is only partially consistent between people and times. You can even see our senses for the two as being similarly evolutionarily formed. I could say similar things about fun. So the fact that morality subjectively seems like something where there are objective facts doesn’t seem particularly surprising or in need of explaining.
OTOH you can use people’s beliefs about grammar to find out stuff about grammar (though the naive way of doing so isn’t always reliable); who says that morality is more like physics than like grammar in this respect?
The comparison of morality to grammar is INCREDIBLE!
“Using people’s beliefs about morality to find out stuff about morality is only as useful as using people’s beliefs about physics to find out stuff about physics.”
… having first singled out the most credible experts on the topic. So, a pretty good idea, then?
Say I want to know the relative mass of quarks. I don’t have access to a particle accelerator; I don’t even really understand how one would go about designing an experiment to test this. Heck, I’m a little confused about what the question means.
But I can Google up the opinions of physicists, who have demonstrated their understanding of physics based on unambiguous, easily tested questions like gravity.
“Using people’s beliefs about morality to find out stuff about morality is only as useful as using people’s beliefs about physics to find out stuff about physics.”
This works *today* because the people who’s beliefs about physics count are drawing from several hundred years of very successful experience. IDK if it would have worked as well in the time of Francis Bacon.
It’s an open question whether we can ever get to the same point with moral questions (and I’m betting most people here would argue that we can’t).
So if caring for your aged parents correlates with opposing Nazis, this proves that Nazis are evil. If caring for your aged parents correlates with supporting Nazis, this proves that correlation with caring for your aged parents doesn’t actually mean anything vis-a-vis morality.
Somehow I feel like you are committing a rationalist sin here and failing to treat Nazism with scrupulous fairness.
I’m about half convinced that in fifty years or so we’ll be threatening people with the prospect of going to Nazi Germany after they die.
The implicit symmetry-breaking assumption is that in non-Nazi cultures caring for aged parents is definitely correlated with opposing Nazis.
Is that actually true?
Quite possibly not.
Funny thing. http://www.dailystormer.com/craig-cobb-denied-the-right-to-care-for-his-aged-mother-because-he-is-an-evil-racist/
Sometimes I think you should be the blogger and SA the commenter, but I don’t have any longer-form samples from you besides anime crap I don’t care about.
As someone who does care about anime and admires many of suntzuanime’s comments here, I’m very glad to have now learned that suntzuanime has an anime blog.
Trolling is easier than it looks. But thank you for your kind words.
Oh HEY, another anime blog to follow!
I wanna be offended, but you’re kind of right. Let it be officially recorded that suntzuanime is added to the list of people who are welcome to post non-mind-killing articles here with approval if they email me about it.
Also, is suntzuanime the same guy as @admittedlyhuman, or do they just have very similar avatars?
Are you actually using the avatars as large evidence? They’re just both closely cropped portraits of anime girls. They don’t even have the same color hair: one is brown, other red. Nor do they have the same hairstyle, though maybe you can’t tell that from these crops. Maybe you can’t tell anime characters apart, but google can:
admittedlyhuman; suntzuanime.
Obviously STA likes anime. It is somewhat informative that AF likes it enough to use an icon. But AF’s icon is chosen for the expression, while STA’s is a much more bland positive expression.
On the one hand, I would tend to agree with you that the avatars really don’t have much in common beyond being anime girl faces. On the other hand, that is in fact me, so either “anime girl face” is already enough bits to make a correct judgment, or there are similarities there that Scott has picked up on and you and I have not.
I assumed that Scott was really using evidence from the writings. Maybe the avatars raised the hypothesis to his attention, perhaps irrationally, but my guess was that he wouldn’t be asking the question if he hadn’t considered the further evidence. But confirmation bias makes that a tricky inference.
If we assume that you have a twitter account that interacts with Scott, we don’t need many bits to identify it. The question is how much probability we should put on that hypothesis. Actually, it probably makes sense to go in the other direction: Someone who interacts a lot with Scott on twitter probably comments on his blog. But it wouldn’t be surprising if “Fakey Fakeson” is commenting under a vague pseudonym without a personalized avatar.
No, if caring for your parents correlates with opposing Nazis then this suggests that whether you care for your parents may be a good measure of how moral you are. We’re not trying to determine whether Nazis are evil, we’re trying to determine whether certain machine learning algorithms can capture morality.
I immediately thought the same thing, when I noticed that he did not include any potential results that would compel him to consider the moral correctness of killing all the Jews. If you’re going to expect a Nazi to give up on that because of your algorithm, then symmetry is required, or else you already know that the argument cannot be used against you and so your presentation of it is purely partisan.
I would assume that patriotism-type-morality among Germans would be highly correlated with supporting the Nazis, and patriotism-type-morality among Brits, Americans, Soviets, etc would be highly correlated with opposing the Nazis. If I only surveyed Germans and found a moral correlation between Naziness and morality, I would assume that was what was going on. If I surveyed all countries, and found a moral correlation between Naziness and morality even in Allied nations, then it might be worth bringing to attention the hypothesis that we’ve shown Nazism is moral.
The obvious solution: check for correlation between patriotism in Allied countries and support for native fascist/far-right populist groups, like the B.U.F. or the 1930s’ KKK or people like Father Coughlin. (Those are all dissimilar in some ways, though.)
Edit: well, duh, that’s what you mean by “Naziness and morality even in Allied nations”, right?
the reason this was such a terrible example is that patriotism is tied to filial piety. Not only that, but since Nazis are horrible racists who believe in genes, they would be especially concerned for their particular kin.
As always, there is a relevant xkcd:
http://xkcd.com/451/
Amusingly, your proposal is used as an example of “imposter” sociology there. (Can using logarithms for heat dissipation be far behind?)
There’s also another relevant xkcd–a graph showing the DW-NOMINATE scores of Congress over time. http://xkcd.com/1127/
I suspect this pretty clearly demonstrates that DW-Nominate doesn’t really match L-R intuitions over a sufficient large domain (in this case temporal.)
DW-Nominate may be better than existing alternatives, but it has definite flaws. Most obviously, it’s based on floor votes. So any issue that never makes it to a floor vote won’t be captured, even if there are significant ideological correlations with that issue.
Political tactics also make difference. Everyone knows the “right-left” axis. Politicians competing for votes will introduce bills which tend to sharpen that difference for campaigning reasons. This will exaggerate the difference between politicians, and an increase in this behavior will be difficult to distinguish from actual greater polarization of the electorate or of elected politicians.
What if the hypothetical test found out that the average Nazi was actually more moral than the average American today?
Before you jump on me I’m not talking about the top-level sociopaths like Goering or Hitler, but the balding older Brownshirt or the young woman donating hair to make tank treads for the Wehrmacht and singing ‘Deutschland Uber Alles’ with her kids. The type of ordinary person we occasionally see in more sophisticated movies as part of a morality play about ‘the banality of evil.’
After all, it’s pretty well established that conservatives tend to win handily on these morality/cooperation tests and while cladistically National Socialism and Italian Fascism were progressive their core supporters were overwhelmingly the conservative middle classes of their nations. It’s hard to say for sure that being a Party member by itself would do that much to change their score, especially since most of the atrocities were committed out of the public eye.
Your shibboleths are interfering with your protestations! But yeah, empirically I would expect the same result. (Nazi leaders were largely conservative petit bourgeois as well, albeit probably atypical ones in some ways, and “aging brownshirts” would be more likely to be lower class and probably atypically hooliganish ones.)
Doesn’t that broadly fall under the category of Hypothesis 2?
I think suntzuanime articulated it better than I did, but the problem with hypothesis 2 is that it seems to approach the whole idea of the correlation test backwards; if our measures of morality don’t sort Nazi’s into the right bin, then we need to go back and pull out bits until it does regardless of how well they predict otherwise.
It should be pretty uncontroversial to say that Nazi crimes like T4 or the holocaust or the mass murder of slavs on the eastern front were atrocities on a scale unimaginable before the 20th century. A system which fails to label people knowingly complicit in them as Immoral has a severe flaw. But there were millions of Nazis, and the overwhelming majority spent those eight years living more-or-less mundane lives. Demonizing them might be useful to keep people from making the same choices in the future, but it’s actively counterproductive when developing an objective measurement tool.
This “cladistics” nonsense is getting out of hand. Pope Francis with his distributism might be the most cladistically Reactionary political celebrity today, going all the way to Aquinas – and so what? How useful is it to keep that fact in mind when talking about him specifically?
le pragmatism face
Actually, knowing the legacy of Catholic economic thought and the theology behind it is extraordinarily useful to understanding people and movements which espouse it. Knowing whether someone’s ideology is based in Chesterton or Marx tells you a lot about it’s ultimate direction.
At the least, it saves you from looking like an idiot by pattern-matching distributist rhetoric to socialism and denouncing the pope as being “to the left of Pol Pot” (not to diss Jim, really insightful dude, just a bit too quick on the trigger). On the surface Francis does push a lot of similar prescriptions to a generic social democrat but the society which would result from following his philosophy to the end is of more or less the exact opposite nature.
Plus, on a practical level you have to admit that in terms of pushing worker ownership and ending debt slavery the guys at Mondragon or a usury-free Catholic credit union have a much more appealing public face than Spanish guerillas. From a Critical Theory / intersectionalist view sure they’re part of ‘the Establishment’ TM, but if you’re supposed to be such a humanistic materialist then surely supporting the most accessible and least bloody path to a goal should be more important than having the right ideological shibboleths.
I believe this is precisely what Multiheaded means when he criticizes cladistics. The old Aristotelian classification that says bats are birds is bad if you want to know what species are related to each other but good if you want to know what can fly. Knowing intellectual history is useful in many ways and fun to boot, too, of course, though whether things like “ultimate direction” are a primarily phenetic or cladistic matter – and whether cladistic assumptions are particularly meaningful in the presence of high rates of horizontal gene transfer – are a separate question.
@Oligospony
Touche, that’s not really an area of disagreement I suppose.
It just frustrates me endlessly to see destructive modernist ideas like Fascism and Capitalism standing in as the face of reaction, when genuinely traditionalist societies like the Holy Roman Empire have existed in one form or another pretty much since the neolithic revolution. Reaction isn’t some untried experimental idea; civilized people have lived according to reactionary principles for millennia, with brief interludes like ours coming only at the final collapse of their empires.
@Not an Apologist
You’re betraying your name. Societies like the Holy Roman Empire have existed since the neoliberal revolution, and people have revolted against them every chance we got.
@Eli @NaA
You both sound like nice and sincere people, but I think you’re both very wrong. I direct you to Corey Robin’s argument that Reaction as an ideology is inherently revolutionary and a mirror of Jacobinism. The medieval society in particular had very little in common with how certain ideologies paint it – the violence, the gender roles, the political processes, the urban culture, the role of religion are all reinvented and stripped of contradictions in the supposed process of recovery. This is true even of such a clearly well-intentioned thinker as Chesterton!
Well, I think part of the implicit association for me is that neoreaction pattern-matches to the ideas of secular right-wing intellectuals from a century ago, who mostly ended up supporting fascism precisely (though not exclusively) because it would save capitalism, and who if they broke with fascism did so for reasons that were basically incorrect (such as Spengler’s complaint that Hitler wasn’t sufficiently bellicose) or superficial. (To be fair this cluster also didn’t jump on board with Nazi racial thought either.)
How about compiling a big database of moral views and Implicit Association Test results? Sure, there’d be loads of confounders; like if republicans turned out to have more racial bias on average we probably wouldn’t want to conclude that all republican policies are racist. But if a relevant bias turned out to correlate with opinion on an issue after controlling for everything, that would tell us something interesting.
I think we can probably assume that racism is correlated with right-leaning political views. But I’m sure we could find something equally “evil” correlated with left-leaning political views; why single out racism?
Just by way of example. I thought that was clear, but if you prefer I can try to neutralize the text.
IATs are pretty racism-specific; if they were intended as a general example of “tests for moral behaviour” (and I don’t think all things are so easily tested) that wasn’t clear to me.
Ah, I see the confusion. IATs are already used for things other than racism, though admittedly usually still identity biases that SJ deals with. But what the test actually measures is association of categories. So while the most famous test uses ‘black’ and ‘white’, you could do ‘urban’ and ‘rural’, ‘blue collar’ and ‘white collar’, ’emotional’ and ‘rational’, ‘modern’ and ‘old’, just about any pair you can think of.
Now that I think about it, just creating these alternate IATs would be fascinating in itself even without the correlating with morality part.
Edit: But even if there’s a reason I’m missing that IATs really are only good for things like racism, it would still be a valuable to isolate if there’s anything conservatives believe *because of* racism. An objective test to measure any particular leftist-correlated vice would be similarly valuable.
The other problem is data show IATs are not actually reliable or valid for all people; only stupid young people for the most part (regardless of racism, and stupidity etc old people just have too slow of reaction times that they fail to be reliably measured too. Non-stupid young people regardless of views and what they say they support politically just have such high and effective performance that distinctions can’t be made)
Someanon, do you have links? I consider myself a non-stupid young person, and the IATs I’ve taken seemed pretty effective.
Why are you sure that we could find something equally evil that correlates with leftist views controling for other factors?
My intuition is that we wouldn’t find anything close.
There are only so many obviously awful things happening at once in developed countries. Racism is pretty up there. The other similarly bad things, support for policies that can responsibly be expected to kill lots of people I would also expect to correlate right.
But what negative issue which itself is not under dispute exists that you could even suspect that it might correlate left?
Red Terror, for once. Attacking certain kinds of culture. (Hell, even Orwell was viciously anti-Catholic.) Covering for systemically failed bureaucracy as long as it maintains ideological appearances. (We would counter that cheering for well-working rightist bureaucracy is worse due to it having a malicious purpose.)
Contempt for the white working class, especially the rural white working class.
Yeah, we lefties can be scumbags too, sometimes.
A very tangentially related question that I’ve been pondering: why is it that pro-state leftists and anti-state rightists tend towards modernism (cosmopolitanism, both kinds of materialism, train boners, deductive arguments about how to produce the most widgets,) while anti-state leftists and pro-state rightists tend towards romanticism (localism, mysticism, love of nature, beating the shit out of each other at punk concerts?) The Pournelle chart usefully illustrates, but does not explain, this (if it’s a real phenomenon, which I believe it is, even if all three dimensions are fuzzy and presumably there are only two real dimensions, my guess being that most of the fake dimension is probably statism.)
Consider the economic issues / social issues 2 x 2 chart. Typically the four corners of this side are liberal, conservative, libertarian, statist. So each of your four groups is a side. Pro-state leftists and anti-state rightists are the two economic issues sides. Anti-state leftists and pro-state rightist are the two social issues side.
Your first lists of traits sound very economic. Your second list of traits sound very social.
So I would unravel it as: “Is the economy more important, or is it ordinary human interactions?” In the first case, modernity is the greatest invention in the history of man (and if you’re on the left, we need a big state, and the right, a small one.) In the second case, we should have stayed together in tiny villages (and if you’re on the left, the state should get out of the way of our free-loving communes, and if you’re on the right, the state should protect our traditional villages from deviants.)
Wow, that’s really elegant and obvious-in-retrospect. I’ll have to ponder it for a bit to consider if it fully explains what I’m thinking of, but I feel real stupid for not having thought of it myself. Thanks!
There definitely exists a pretty big romanticist anti-state rightist cluster in ideaspace: http://en.wikipedia.org/wiki/Paleoconservatism
Huh. I’m the kind of leftist who doesn’t have a definite position on whether the state is always good or always bad, or even usually so. I tend towards some cosmopolitanism, pro-historical-materialism, anti-consumerism (because saving for being able to make whatever life choices I want is a superior investment), yes-train-boner, yes-deductive-arguments… and also pro-localism, and just generally in favor of absolutely everything about punk concerts.
I am once again confused.
You shouldn’t be, since I mean to invoke gradients rather than binaries. 🙂
Here is a book by a contrarian leftist cheering very enthusiastically for consumerism and commercialized culture. Sounds like a fun low-hanging idea to entertain, at least.
(One instant attraction for me is that Livingston attacks that awful fucking greyface buzzkill, Christopher Motherfucking Lasch. Fuck his shit, I tried reading him and loathed his every word SO. FUCKING. MUCH!!! You can probably chalk up some of that to my gender stuff.)
Then I should clarify: I’m anti-consumerist in the sense that I personally prioritize saving over spending (in many matters), and believe it is more-or-less objectively rational to do so. I don’t believe in the Protestant Work Ethic or any such shite for a second, as I was raised Jewish: it’s about as foreign to me as Christmas, something “they” do that “we” don’t.
So I’m all in favor, say, of everyone going to get ice-cream together, as long as it doesn’t mean someone will go into debt or miss a rent payment, and that everyone is being responsible about their future, etc.
Which makes actual social outings really depressing, as people who didn’t major in computers have a really hard time making enough money to be financially responsible right now :-(.
The thing about DW-Nominate is that the political spectrum that it detects is not exogenous. The political views of politicians are affected by the system they participate in. If we used DW-Nominate in a hypothetical Muslim democracy, it’s plausible that it would detect a Sunni-Shiite dimension as the first dimension if those are the two voting blocs. You would have “moderate Sunnis” who occasionally side with Shiites and vice versa, and be able to detect those policies which are super-Sunni. Conversely, if we had a system of sortition (and it had been established before the party system, so there weren’t holdover effects from the current system in which people import the left-right spectrum into their identity), it’s plausible that the first dimension wouldn’t clearly represent a left-right distinction, or that the higher dimensions explain significantly more variance. So DW-Nominate doesn’t necessarily say something fundamental about the spectrum of human beliefs, but rather it reflects our political structure.
As for using Nominate on moral decisions, it would definitely be very interesting, but what it detects would depend on the source of input data. If crime statistics were used, then it would detect law-abidingness as the first dimension. If you used “every decision made by a person” as the first dimension, who knows what it would detect. Maybe it would detect IQ, so it could be correlated with another random thing.
The late psychologist Hans Eysenk had a grad. student study the personality of people in various political parties. Courageously, she joined both the British Union of Fascists and the Communist Party at the same time so she could get access to some subjects. She found that they both had very similar personalities with sociopathic tendencies and fantasies.
In regards to Not an Apologist‘s question about the “balding older Brownshirt or the young woman donating hair to make tank treads for the Wehrmacht and singing ‘Deutschland Uber Alles’ with her kids,” Eysenk himself, as a teenager, had a girlfriend who was in the Hitler Maidens – even though he made no secret of his hatred of the Nazis and had to leave Germany to get into a university because of it. So Eysenk obviously didn’t think she had any great moral deficiency.
As to your question to the Nazis:
This would definitely not work. The Nazi would simply counter that under them the state organises many charitable works that also encourage solidarity among the people. These actions properly belong to the state, he would argue, and by doing these things off their own initiative and not in solidarity with the state, those people are harming the unity of the state and therefore acting immorally.
Please pick a slightly different name to comment under, as I’ve been using this one for some time already.
Note that at a certain point membership became compulsory, so this may be completely meaningless.
This is in the period just before and after the Nazis came to power. He felt that she was completely blind to what the Nazis were really about. Maybe she saw them as a force for order in the chaos and violence of the Weimar Republic.
In The True Believer Eric Hoffer hypothesized that communists and facists were recruiting from the same pool of potential supporters, and that their antagonism was due to competition over this scarce resource.
I don’t think this idea will work, because I strongly suspect that people who tend to act morally are also prone to wanting to signal morality using status signals of good behavior.
IOW, people who don’t cheat on their spouses or their taxes (pro-social behavior) are also probably more likely to recycle and donate to african hunger relief. (signalling that may or may not actually benefit anyone.)
How would this method distinguish between truly pro-social behavior, and mere status-signalling?
It would capture all behaviors and still rate them, but I think a lot of comments overlooked that this is only speculative and not something feasible at all. To really do this as suggested (the original text implies as much) would require monitoring people like 24/7 with invisible robots. It’s only a hypothetical.
A much more interesting discussion to have, that could get a lot of comments just because it’s on a general subject everyone here keeps talking about, is what we learn from the fact that NeoReaction is basically a control group for the LW/Singularity movement. Compared to society at large both have all the same population structure characteristics in several significant ways.
I would definitely say LW and Singularity are two similar-but-distinct ideological subsets of the “techno-geek” social grouping commonly found on Hacker News, or in the founding population of Reddit, or even on Slashdot.
My biases and knowledge together lead me to reason that neoreaction is obviously such complete bullshit that neoreactionary views can be used as a litmus test of basic morality and rationality within the techno-geek sphere, who are already more likely to espouse things like evolution and global-warming as tribally-evinced beliefs, whether or not they understand the relevant science on even a high-school level. Mind, the other thing my mind throws out as an obvious litmus test is belief in a Kurzweilian Singularity that goes right by default rather than ruining half the planet as in a Charles Stross novel, or even tiling the solar system in $SOMETHING as in Eliezer Yudkowsky’s worst nightmares.
What I’m wondering is to what degree neoreactionaries can be used as a control group for measuring the presence or absence of an effect in other parts of the techno-geekosphere.
Give Yudkowsky some credit; his worst nightmares are certainly even worse than that.
Ok, fine, Multiheaded: Singularity goes wrong and tiles the entire galactic supercluster in $SOMETHING.
(Keep in mind that “FAI goes wrong and creates everlasting Hell” is not one of Eliezer’s nightmares, as he has repeatedly pointed out that the directional difference between Heaven and Hell is not just a matter of having a single sign-bit hit by a cosmic ray, so that’s just not going to happen at all.
I believe him on this, because none of my FAI goal-systems would allow for such a flaw either, so I’m pretty reasonably sure the final design simply won’t allow that fault to happen!)
Factor analysis and principle components analysis seem to provide similar results, and have a large number of practical uses. I wish it was something they gave us an overview of in introductory statistics, because it’s not that difficult to apply and is hugely useful. The TruthCoin project you’ve mentioned before uses PCA to determine outcomes, so people’s votes are weighted by their coherence with the largest principle component.
On an unrelated note, I think this link might be of interest to the occasional “niceness in activism” themes you discuss. There is, of course, a chance you’ve cited it before and I’ve missed it.
http://onlinelibrary.wiley.com/doi/10.1002/ejsp.1983/abstract
Wait, so are college majors able to be sort of morally ranked? Or was that just being listed as a more general choice?
It seems like it wouldn’t be too hard to design a fairly simple questionnaire that asks people what their decisions were. And if it is a short questionnaire, it shouldn’t be too hard to get a hundred responses. Even a small number of responses should give you an idea of whether this is worth pursuing.
This seems like the sort of thing sociology majors must do for class all the time. And, y’know, you wrote the LessWrong survey, so you might not be a bad person to write the questionnaire. Getting 100 strangers to take it is outside my expertise, but I wouldn’t think it could be that terribly difficult to find someone willing and able – don’t sociology majors do this sort of thing all the time? 🙂
This proposal is virtue ethics in disguise.
No, it is empirical virtue ethics. It is to virtue ethics what cost benefit analysis is to utilitarianism.
The semi-supervised version, where you gave it a few handpicked examples of good people and a few handpicked examples of bad people and nucleated the clusters around them, would be virtue ethics for sure. But the unsupervised version Scott is talking about doesn’t have moral exemplars, it just has giant correlation matrices, so I’m not sure it’s fair to call it virtue ethics.
It reminds me more of Coherent Extrapolated Volition, except more open about its willingness to say “you’re part of the Bad Person cluster so your volition doesn’t count”.
But isn’t this arguing in circles? Taking your example of gun control as a “leftist” cause – why is it leftist?
“Because leftists vote for it!”
And how do you identify a leftist?
“Because they vote for things like gun control!”
I wouldn’t call myself a leftist (by American, or even by certain Irish values of the term) and I’d be all for gun control (I don’t like the things).
“Leftist” is just a label. Nothing changes if we reverse the labels. The point is to notice that people who vote for gun control also vote for abortion rights or whatever. We’re identifying clusters and dimensions; there’s no privileged orientation of the axes.
But fifty years ago the Black Panthers were openly bearing arms as a political statement, and Governor Ronald Reagan was so pissed off by this that he introduced a gun control law in California… Hell, the NRA used to agitate for keeping American gun culture responsible and respectable. There has been a 180 degree flip on the issue in the American political spectrum since the defeat of the Black Power movements and the rise of angry-white-man identity politics.
Here is a tiny modern (but I repeat myself) leftist party in America speaking out against gun control.
Uh-huh? And how does that fact change if we start calling the left “right” and the right “left”…?
But how is the pro-gun left any less committed to a broadly leftist worldview simply by being pro-gun? Would you predict any correlation between a leftist group being pro-gun and it being to the right of similar leftists on other issues (rather than having more emphasis on civil libertarianism, or going for an American patriotism appeal, etc)? For example, the CPUSA is firmly anti-gun, yet they’re basically milquetoast Obama-enabling liberals at this point.
Yes, but then the labels are used to classify the data, and the data is used to go “So liberals and conservatives are not just made that way by nurture, they are naturally different, maybe even down to a genetic level via evolutionary pressure!” and we get a lot of blah about “people who are independent” versus “people who are herd-minded” and the likes.
What was that last survey about values versus conservativism and lliberalism, where it was noted that ‘conservatives’ placed more weight on group consensus, and there was a lot of gloating on both sides of the polictical divide: those who considered themselves ‘liberals’ or ‘leftists’ or ‘Democrats’ or ‘Big Endians’ were spraining their wrists patting themselves on the back for not being ‘herd-minded’ and ‘sheeple’ but having the courage and intelligence and forward-thinkingness not to cling on to old discredited out-moded traditions; those who liked to think of themselves as ‘right’ or ‘conservative’ or clear blue water types did likewise about not being selfish, self-absorbed individualists but rather concerned with the common good.
I mean, why not call the different groups “Sand” and “Stones” or “Shells” and “Twigs”? Because ‘left’ and ‘right’ may only be labels, but they are labels with a lot of baggage attached (to mangle a metaphor).
Instead they found that our notion of left and right emerges naturally from the data
See, that is my problem with this right there. How do we judge what the data is “left” or “right” or “Burgundian” versus “Huns”?
Congressperson Burgundia is elected on a platform of “Vote for me and I will push through legislation that severely restricts those dangerous potential weapons, pitchforks.” Congressperson Hunstantia is elected on a platform of “Vote for me and I will oppose such pernicious legislation and defend your gods-given right to have a pitchfork in your hand and a pike in the thatch”.
Our congresspersons, once elected, get a chance to vote for a bill called “The (we couldn’t work out a catchy acronym like DREAM or a tugging at the heartstrings name like ‘cute little girl’s law’ for this so we’re just calling it the) Farmers You’ll Have Someone’s Eye Out With That Bill” which will permit licensed government inspectors to make sure all your pitchforks, slash hooks, sickles, dehorning knives and the like comply with health and safety laws.
Is this a ‘conservative’ or ‘liberal’ bill? Which way do you think our respective congresspersons will vote? Does it switch from being a ‘left’ to a ‘right’ bill if Burgundia votes against it (maybe a manufacturer of scythes got a lobbyist to slip the congressperson a bung once they were elected)?
I mean, the Labour Party in our own current coalition government got into power on a raft of promises of standing up to the Troika and the bank bondholders and protecting the social contract and all the rest of it, but once in power have gone along with the austerity budgets of the main centre-right partner. Yet, while they’re happy to throw their socialist principles under the bus in order to cling on to power (they’re going to get hammered in the next general election as they’ve already gotten a hammering in the recent local and European elections, so they’re determined to hang on until the bitter end), they also are quite happy to instigate such campaigns as a push for same-sex marriage/marriage equality (depending how you want to call it) because it’s a soft target. It won’t cost them anything in either support or money, it will permit them to keep up their claim to being socially progressive, it’s an easy ‘left’ or ‘liberal’ cause with no real cost to them.
By which I mean, on real left-wing goals involving monetary and budgetary targets, they roll over happily for the right-wing party. But they’ll produce enough ‘voting left’ legislation (or at least talk about it, like our Education Minister who is very occupied with getting religion out of schools but not so much with ‘once the kids have been educated, what’s there for them except emigration?’ or ‘how about access to third level education for low income families?’) to keep up the image of “we’re a left not a right party”.
Two examples of similar projects that might help: Scotch single malt whisky is traditionally characterised as from one of five (ish) regions. Analysis going backwards from flavour descriptions of large numbers of whiskys produces clusters which actually quite closely match the traditional regions. There’s probably a circularity problem here because people may well describe the whisky in a particular way because they know it’s from a region, but you’d expect use of extensive collections of reviews to correctly identify some real flavours with real descriptive power.
Similar stuff to DW-Nominate is widely applied in psychology, medicine and testing, as Item Response Theory. Here, an underlying trait is derived circularly from the interaction between students and questions (or patients and questionnaires), learning both how good the students are and how hard the questions are from just looking at correctness of responses. The results are entirely relative, and if you don’t specify an orientation (such as “more correct answers should produce a positive increase in the underlying trait), the scale and location on the scale will be arbitrary, but the relationship between the items/questions and the measured traits will always “work”: the better students will all be at one end of the scale, and the harder questions will be at the same end of the scale. When you measure disparate groups over time (such as with DW-Nominate), the arbitrariness of scale and location make it impossible to just compare numbers between different groups, but under certain circumstances we can establish that there’s a linear relationship between two distinct scales, and use items or students (or bills and politicians) which turn up in both groups in order to “link” [that’s the actual technical term] the two.
To be clear, when I talk about “correctness” above, that’s only meaningful in a testing situation: if you’re measuring depression or attitude to global warming, there’s no right answer, but certain answers will (in a good test) strongly correlate with the respondent’s underlying trait being at one end of the scale or the other, or past or before a certain point on the scale.
Finally, one of the things you’re getting at is that maybe left-right isn’t the only spectrum. It is indeed a choice here to declare that there is only one underlying trait being measured, and such a choice may be valid (if these are all topology questions) or invalid (if there are also questions about cooking). You can actually experiment with allowing multiple underlying traits, and compute the likelihood of your results under 1-trait, 2-trait, etc. assumptions. Unfortunately the greatest likelihood will occur when number-of-traits ≈ number-of-questions (or bills, or whatever), so you end up having to apply Ockham’s razor or some similar rule to cut things off at a reasonable point.
So you’re almost certainly correct that reality is better described by allowing for left-rightness *and* gun-friendliness or whatever, *but* you may find (as it would appear DW-Nominate has found) that you gain very little from the additional trait, and your model of the world is accurate-enough with only one trait which turns out to correlate highly with our left-right vocabulary. The circularity of social construction of political action in context of left-right vocabulary then becomes problematic in the same way as descriptions of single malt whisky.
“those who considered themselves ‘liberals’ or ‘leftists’ or ‘Democrats’ or ‘Big Endians’ were spraining their wrists patting themselves on the back for not being ‘herd-minded’ and ‘sheeple’ but having the courage and intelligence and forward-thinkingness not to cling on to old discredited out-moded traditions; those who liked to think of themselves as ‘right’ or ‘conservative’ or clear blue water types did likewise about not being selfish, self-absorbed individualists but rather concerned with the common good.”
It’s strange to hear the groups describe themselves like this, because currently in the US, the right is tarred as selfish and self-absorbed (while they describe themselves as individualistic) and the left is criticized as collectivist (while they call themselves unselfish) – though it was somewhat different during the Bush administration. My suspicion is that the out-of-power group is typically more individualist and the ins are more collectivist, regardless of whether left or right is in power.
Most traits can be framed positively and negatively, so ‘we’ care about the common good, but think for ourselves, while ‘they’ look after only their own interests, but are easily led.
Yes. Read the Scott Aaronson post linked on top. He talks about the ability of certain algorithms, like Google PageRank, to take seemingly circular ideas and wring information out of them.
Individual gun ownership is a relative disempowerment of two groups–law enforcement (and their bosses) and criminals, or at least the “illegality-curious.” Who supports it will depend on where they see their support in these three groups, I suppose, unless they are unusually committed to principle (which, if we are talking about elected officails, they aren’t).
Oh, and deer. But they don’t vote.
I’m worried that the outcome of this experiment would be largely determined by your assumptions. (Which may be better; one particularly deep fear of mine is that Western society accurately implements its citizens’ actual morality.)
Note that Aaronson’s eigendemocracy is intended to be a better democracy. The equivalent of eigendemocracy for a theocratic autocracy (people are good to the degree they are good to people who … who please the ruler who is pleasing to the gods)… would probably not result in anything resembling modern democracy. Morally, eigendemocracy mostly seems to return the assumptions you encode in it.
Similarly, DW-nominate recovers the Democrat/Republican split, but does so based on a long history of coalition politics (as pointed out by BJ Terry, above). E.g. American Christians tend to vote against abortion and for laissez-faire economics; South-American Christians may also be against abortion, but they can be quite “socialist” (from Pope Francis to liberation theology); African-American church leaders fought/fight really hard for civil rights (most famously, Martin Luther King.) Which is to say, (your use of) DW-nominate’s data set encodes certain assumptions (roughly, that the major American coalitions represent a coherent school of thought rather than a mutually-benificial alliance of not-too-opposed interests.)
Similarly, if you want to recover morality clusters – rather than, say, gender or social class or personality type – I’m worried that you’d have to be rather particular about which decisions you want to include. And if you exclude, e.g., prayer…
I’m honestly wondering why you think there are only three possible hypotheses and only two or a handful of possible factors.
“Morality” is a predicate so goddamn complicated and unreliable that my personal hunch is it’s a label we’re putting on something that is, underneath, much more variable than a simple two-class one-factor classifier.
Today’s smbc strikes me as strangely relevant
http://www.smbc-comics.com/?id=3397#comic
Came to post this. Beaten by anonymous once more!
BAH *shakes fist*
Yeah, morality is at least causality-hard (how do you determine responsibility otherwise), and I think is actually far far far far far more complicated than that. I was disappointed in the other Scott’s responses.
They would make good legislators for the upper house of parliament but not for the lower house, which should just reflect the opinions of the people.
In other words, even though you didn’t present a grand plan to fix everything, you came pretty close to formulating a thing equivalent to what I personally actually have as a grand plan for fixing everything.
In short, we should re-invent aristocracy as “those folks who have been measured to be the most virtuous”, have an upper house of parliament composed of such folks, and find out how much power we need to transfer to such an aristocracy in order to turn democracy into a system that actually works (I’ve written a bit more detail elsewhere).
I don’t think this will work Consider what would happen if it turned out that Nazis were more moral in these other ways than non-Nazis. Would you accept that as an argument for Naziism? No? Then why should it be acceptable as an argument against Naziism if the statistics go the other way?
I can think of plenty of scenarios where evil acts are associated with other good acts. For instance, consider a Nazi who is motivated to kill Jews because it gets him a promotion and that benefits his family. A selfish bastard who doesn’t care about his family would not be a Nazi for this reason. He would still want the personal benefits that the promotion brings him, but personal + family benefits are a greater motivation than personal benefits only.
It is entirely plausible–not guaranteed, maybe not even 50% likely, but certainly plausible–that the “I want to be a Nazi to benefit my family” effect associates Naziism with good things more than “I don’t mind being a Nazi because murder doesn’t bother me” associates it with bad things.
There is also the case of group-cohesion. This is a problem for the sincere Nazi who joined because he thought fascism was the way to restore Germany, yet who would be horrified by the concentration camps if he knew about them, and who then discovers the existence of concentration camps.
What he cannot do here is just automatically defect from Nazism because he finds they’re doing something morally wrong. Being part of a group means that you can’t always have your way. The question them becomes whether it would be worth sacrificing a restored German Empire to save these people.
Here, there is a smooth gradient in how much divergence from your own views you are willing to tolerate among people you consider to be your in-group. Set this to zero and you have social dysfunction on the level of some parts of the French Revolution. Set it very high and you have something like an idealized fascism that requires no secret police because everyone goes along with the program sincerely.
That’s not idealized fascism, that’s idealized moldbuggery…
Possibly relevant: more conscientious and agreeable people are more likely to “fail” the Milgram, which seems like evidence that Nazism is Empirically Good in the above-defined sense. The usual caveats about social psychology experiments apply but the result seems intuitively plausible enough.