In some ways I am sympathetic, but wouldn’t this critique apply straightforwardly to any discipline that employs significant idealizations? Wouldn’t it show that idealized models, in, say, economics ought to be discarded? Or that the discipline ought to shift significantly because they’re in effect playing chmess? Or is the difference that political philosophy is a normative discipline in a way that economics is not?
I feel like you may be overestimating the power of keeping models close to reality. There is a class of propositions in economics called "irrelevance propositions," which use wildly unrealistic assumptions to prove kind of insane theorems. The no-trade theorem is one example, and Ricardian equivalence is another. The point of these theorems is to show that under a range of very serious idealizations, some equivalence holds. But given that the equivalence doesn't actually hold, we can then isolate and test the assumptions to see what gives rise to the divergence.
I'm relatively confident that we can construe Rawls in this way, and the three unrealistic assumptions you point out are different reasons why the actual world diverges from the just society. It's a very useful roadmap to have. The hope is that we might be able to isolate how the violation of, e.g., sufficient reasonableness, a closed society, etc, contribute to the actual state of the world.
So my feeling is that these kinds of models are extremely useful, maybe even precisely *because* they diverge from the real world in such drastic ways. This could just be economist "cope", but I sort of doubt it.
"I'm relatively confident that we can construe Rawls in this way"
Based on what? You need to give reasons here, it has to be more than just an assertion of faith.
"It's a very useful roadmap to have."
Why? I've tried to give some reasons for thinking it is very much not a useful roadmap. Indeed not a roadmap at all, but a map of imaginary roads that don't exist. How useful is that?
"The hope is that we might be able to isolate how the violation of..."
Well, it's good to have hope, unless that hope turns out to be unfounded and leading you off in unhelpful directions
I am afraid that what you have said here looks like nothing other than "cope", and the main fact that you "sort of doubt it", based on your "feeling", may suggest that you are arguing this because you *want* it to be true, not because there is good reason to believe that it is true.
In truth, you actually offered no arguments that such idealizations aren't useful, you just asserted that it wasn't plausible! Here:
"Rawls insists that without his ideal theory (or something like it) “there is no rational basis for continually adjusting the social process so as to preserve background justice, nor for eliminating existing injustice”. But this is not plausible, and not only for the reason that it is a bare assertion, one we should no more accept than if somebody told us that until we have an ideal theory of the best ship then we have no rational basis for repairing our leaky boat. More than this, it is as if the ideal theory we were being told to consult turns out, upon closer inspection, not to be about water-borne vessels at all, but flying carpets. Or if you like, about chmess rather than chess."
You've basically just said: "this isn't plausible because Rawls is changing the subject." But whether idealizations are just a change of subject is exactly what's at issue. Asserting that they are is begging the question.
You've isolated three dimensions in which Rawls idealizes, and there is tons of empirical work to be done examining how far away we are from these ideals in the actual world, and this empirical work can generate policy proposals. Consider, for example, the work aiming to make public discourse more reasonable—nonprofits like Sway that try to change public discourse in order to get people to recognize reasonable disagreement and change their practices. If we don't have reasonableness as an idealization this work isn't motivated. But of course, that's only one dimension.
p.s. there isn't "tons of empirical work to be done" on these matters, the empirical work has been in for ages, and the answer is clear: we are light years away from his ideals being anything like the world we either live in, or have any reasonable expectation of ever living in.
These are not "idealizations" of actual political phenomena, they are imaginings of phenomena that could no longer be accurately described as political because they have departed too far from the necessary building blocks of what counts as politics
I guess the point of my comment about irrelevance propositions was to show that many things that people take to be "insane" or to have "departed too far from the necessary building blocks of [subject X]" actually turn out to be very useful.
In the piece you offered an extended analogy, the validity of which many people on this thread are calling into question. Their main point is that your analogy overgenerates, and would eliminate lots of very useful idealizations along with the ones that are not useful. But you offered very few arguments that Rawls' views fall into the latter category. Instead, as the quote below shows, you just relied on your analogy again.
Took the words right out of my mouth. I don't think Rawlsian ideal theory is actually defensible via the same arguments used for defending frictionless planes and expected utility maximizers--Appiah has a very nice discussion to this effect in his book on idealization--but I think the argument here is overbroad, in that it doesn't really let us distinguish good idealizations from bad ones.
As the author says near the end: 'Ideal theory of the kind practiced by Rawls and others does not involve simplifying assumptions; it involves counterfactual assumptions that mean it is theorising something altogether different.' As someone said 'All models are wrong. Some models are useful.' Judging whether a model contains usefully and acceptably simplifying assumptions or invalidating counterfactual assumptions is a critical aspect of assessing its utility.
But you could use that exact argument against all kinds of perfectly reasonable simplifying assumptions! Assume this spring were made of continuous matter (Hooke’s Law). Assume people were all self-interested (much of evolutionary game theory). Etc.
The missing idea in this article, IMO, is that we might be making simplifying assumptions in order to illustrate fundamental properties of a system that we all know is, in reality, more complicated. The point is to assume away some real features in order to isolate certain others. This isn’t always a good move, but you can’t always dismiss it as chmess.
An easier target than Rawls would be those political philosophers whose *policy proposals* only make sense in an ideal dream world—a world that ignores, tradeoffs, transaction costs, or what have you. Rawls is at least committed to caring about incentives etc. since these constrain the set of possible basic structures for society.
EDIT: but let me add that I loved the connection to Dennett on chmess. I’d forgotten about that passage!
i'm not 'always' dismissing 'it' as chess. Just ideal theories of politics, specifically.
Rawls thinks that real world political decisions (the realm of "nonideal theory") should be regulated by and made in reference to (his) ideal theory. So I'm afraid you can't get them off the hook that easy.
And yes, he cares a bit about incentives, so he's doing better than e.g. GA Cohen. But he doesn't care about them enough, because he doesn't have a realistic enough moral psychology.
I grant you that it would be chmess if Rawls were basing policy recommendations *entirely* on his theory. But isn't he pretty circumspect about doing so?
[Edit: Removed a false sentence -- Paul S is right. Rawls is not "circumspect" about welfare-state capitalism in sec. 42!]
To put it another way, I don't yet see your objection to the following kind of approach: we use a partly idealized theory to get our ideals and then use grubby empirical facts to figure out how to translate those ideals into policies. (Maybe your point is that idealizing can go too far even when trying to find ideals for a society?)
TJ says welfare state capitalism is ruled out, and we probably need to go for "property owning democracy", which is effectively a form of socialism, although he is very unspoken on the economics, which leaves him vulnerable on that point.
I don't have an objection to the kind of approach that you describe (so long as the idealisation does not lose touch with the basic building blocks of what politics necessarily has to be for creatures like us). My point is: Rawls (and such) are nowhere close to accurately being described as doing that.
What you describe is a lot more like what Dave Schmidtz advocate for in the first three chapters of his book Living Together. I like what Dave suggests.
Came here to glibly say, "now do econ", lol. But yeah, I think morally there is in fact a massive difference given that pol phil is (presumptively) normative and econ is (presumptively) descriptive. And that's not to say one should be able to get away with dubious assumptions more than the other. In general, houses of cards do not practically great mathematical frameworks make. And I don't think it's necessarily "better" that one can or should be able to quantify how off the idealization is from (what is for all practical purposes) reality. Yet IDK, maybe it's just my being a geek, but somehow it makes me less uncomfortable to say, "yeah this mathematical model might be garbage IRL but it's still kinda cute" than to say that moral intuitions about how to govern and live were fleshed out in a way that turned out to be...spurious. Ya feel me?
This reminds me of a joke: a farmer, whose chickens have been dying, employs the services of a biologist, a chemist, and a physicist. After dispatching with the theories of the biologist and the chemist (mites and toxic chemicals, respectively) the farmer asks the physicist what she thinks. The physicist, who has been busily scribbling on a note pad while the various investigations were taking place, says, “I believe I have the answer—but it only works for a spherical chicken in a vacuum.”
I'm a physicist and came here exactly to use the same joke, but with the opposite meaning.
I also studied *theoretical* physics and always found the joke utterly dumb (meaning also this whole post is actually wrong).
Of course in reality friction exists, so Newton's first principle (actually Galileo's) that an object mantain its status of linear motion if there is no force, seems unwarrented.
You know who started his physical thinking by what actually seem to happen? Aristotle. He got almost everything wrong. It took Galileo and Newton's "idealization" to build a theory that works.
So no, I'm sorry, your argument is unfounded and wrong.
In short, it's the claim that you shouldn't simplify a situation because that's unrealistic.
But in practice you should, at least as a start, because otherwise the problem is untractable.
Thank you!! My name is Nathan Woodard and I approve of your message!! I feel like the golden age of physics in the early 20th century created a lot of confusion in philosophy and social sciences--it became fashionable for theorists in other disciplines to steer towards the methods of the hard sciences (not hard as in difficult--hard in the sense of "hard determinism"). It seems to me that a lot of disciplines (like economics?) are still suffering a certain amount of brain damage that accrued during that period.
We physicists are expert wielders of idealized assumptions — a priori false but useful tools of thought. In social studies and many nascent sciences, however, such assumptions often ossify into dogma and become ideological cudgels that can persist for generations. Even in physics, whole subfields have at times gone astray — ether theory, caloric, steady-state cosmology, possibly even string theory today — but our field has generally managed these detours within one or two generations, thanks to the corrective pull of experiment and prediction.
The weirdest thing is, I know some people doing political philosophy who admit its like chmess and they dont have a problem with it. Neither do they defend chmess assumptions or try to make it relevant. They just treat it like a system of puzzles that you just solve. Their job is to take a set of assumptions and solve for results, without being concerned with applicability, rinse and repeat. It boggles my kind to call it "political" philosophy but here we are.
This is among the best essays I have encountered in Substack. Thanks! It seems like you are entering into territory that clicks very neatly into some of the biggest debates of the centuries:
1) Social constructivism vs biological determinism.
2) Tragic vision vs, constrained vision. (Sowell)
3) Positive vs. Negative Liberty (Isaiah Berlin)
4) Reason vs. Tradition (Burke vs. Rationalist reformers)
5) Empiricism vs. Rationalism in the human sciences
I’m willing to bet my prepper stash (dried beans, seaweed, an FM radio, and a giardia filter) that you had all five of these debates—and more—at the back of your mind as you wrote this. (Kidding about the stash, and apologies if I missed something you did cover.) If I’m wrong, I’d be thrilled for you to unpack that. If I’m right, did you consciously steer clear of these topics to keep things morally and politically neutral? Either way, I’d love to see follow-up essays that frame these debates within the scaffolding you’ve built here. Those essays would be legendary—a real gift to posterity.
Well, that stuff is in the background for me, because I've spent about 20 years immersed in moral and political theory - and crucially, their histories - and that is just the water I swim in. I didn't make any of it explicit, because I don't think Dan would've allowed me to run a 3,000,000 word guest post!
As Max Stirner would say, you can argue about rights all you want but at the end of the day who owns the means of production is whoever seizes them.
One factor that certainly contributes to Chmess is the incentive to regularly publish novel results. Chmess is much safer, after all you're just doing speculative theory so if you're wrong, there are no brutal regimes' blood on your hands. The h-index can march up risk-free.
Which is fairly endemic as far as the Symbolic Capital Bubble goes.
I don't think the chmess analogy works very well. The problem with chmess is that it's completely arbitrary - there's no rhyme or reason to the differences it has from chess, and it's impossible to imagine these differences being somehow normatively correct. The entire thought experiment rests on the fact that it just seems obviously absurd that learning about chmess could better help you play chess than learning actual chess theory, but this is the very claim that's at issue when it comes to ideal theories in politics. Maybe learning about the ideal is unhelpful, but you can't prove this by making an analogy about the unhelpfulness of an obviously non-ideal game like chmess.
Here are some plausible ways that ideal theories might be helpful for learning about politics even if they're not strictly true, and which are not captured in the chmess analogy:
1. They're simpler than the real world. The real world is just far too complicated to model in its entirety. You have to simplify it no matter what, and then add perturbations to you model later to make it more like the real world. Proponents of ideal theories do this openly, but people who claim to base their politics on the real world and not idealizations are still doing it. And all fields do this, not just political theory. Physicists do this all the time, to the point that idealizations made in physics problems have become a common joke, and yet no one doubts the utility of physics.
In this way, the analogue to an ideal theory would not be chmess, but something more like, "chess, but without castling or en pessant," or other simplifications of the rules. And it's a lot more obvious why learning about this simplified version of chess would help you to eventually understand the full game. In fact, it's how everybody actually learns chess: First you learn how all the pieces move under normal circumstances, and then you learn about the exceptions.
At least some of the assumptions that Rawls made, like the society being closed, seem to be simplifying assumptions of these sort. It's not clear that the differences between his ideal case and the real world invalidate his theory, and even if they do, it's not clear that they invalidate it in such a way that it should be scrapped entirely, rather than just adjusted to fit the real-world circumstances (There's a physics analogy for this process of adjustment, too: perturbation theory).
2. The ideal case is a normative ideal that we should strive towards. Assumptions like, "Everyone is perfectly rational," are obviously meant to capture this facet of the theory. When Rawls assumes perfect rationality and then derives conclusions from it, he's not trying to say, "This is what the world is actually like." He's saying, "Here's what rational people would do. We should be more like them." Objections on the basis of people not actually being rational miss the point.
The chmess analogy tries to counter this by suggesting that chmess is somehow the normative version of chess. The problem with this is twofold: First, it's just obviously not, to the point I have a hard time even imagining myself being convinced that it is. All of my intuitions about the utility of chmess are based, even if implicitly, on the obvious fact that the rules are completely arbitrary, not normative. Even when asked to imagine that they somehow are normative, I don't think I'm fully able to do this and overcome the intuition of them being arbitrary. That means the intuitions generated by the thought experiment aren't reliable. I can't trust intuitions about something I can't even conceive of properly, and I certainly can't when I'm trying to imagine something being normatively correct while my entire mind is shouting back, "No it isn't, even in a counterfactual scenario!"
Second, the chmess analogy relies on the rules of chess being impossible to change, so that even if you were convinced chmess was somehow better, you couldn't do anything about it. But that's not true for politics. The rules of politics can change to be brought closer to the ideal, and this has happened many times throughout history. That's the entire reason people study political theory. Furthermore, some ways of bringing the world closer to the ideal don't even require changing the rules of politics, but just changing object-level policies. For example, you can develop policies designed to give everyone what they would agree with if they were perfectly rational without actually having to make them perfectly rational.
The game of chmess seems to be a minor variant of chess, in such a way that studying it would offer very little insight into the game of chess. It would be roughly equivalent to a Rubik’s Cube with one square on each side with the wrong color. Let’s call this RubNik’s Cube. Yes, RubNik’s Cube would be challenging (for humans who use their color perception to solve it), but it’s not mathematically interesting. To a computer that begins with a definition of the cube that includes the original position of each square, it would be mathematically identical to a simple Rubik’s Cube. Chmess is hardly more interesting; in fact, it is far less interesting than Chess 960 (Fisher Random), a variant (to be exact, a group of 960 variants) that top players in the world do in fact play.
Now consider Rubik’s cubes of 4, 5, 6,….n layers. To study these objects, the original Rubik’s cube with 3 layers would be a valuable first step. It’s a simple case, a good place to start for the study of n-layer cubes. Of course, the simple application of the rules and techniques of solving a 3-layer cube would be inadequate for solving, say, a 10-layer cube, but it's difficult to imagine solving a 10-layer cube without solving the 3-layer cube first.
It seems that the assumption, in much political and economic theory, that humans are rational actors is such a first step. Such theories assume that human behavior can be modelled as an optimization function. The first approximation of that theory—that humans rationally pursue rational self-interest, is an extreme oversimplification of human behavior, but it’s not necessarily a bad place to start.
An optimization function that could actually model (and therefore explain) human political and economic behavior would have to possess many dimensions, starting with many of the factors that Dan Williams talks about all the time: confirmation bias, sense of self-worth, social approval, in-group loyalty vs out-group hostility, etc.
Even for one (such as I am) who is completely removed from the halls of academe, and particularly their philosophy departments, I found this essay a witty, enjoyable read. Also, the links toward the end to books examining more “chess like” approaches look fascinating, such as Robert Jubb’s “Unjust Authority” and Judith Shklar’s “The Faces of Injustice.”
I am also reminded, in reading this essay, of A.E. Housman’s “Introductory Lecture of 1892” (from which I took the quote I use for my Substack profile: “The house of delusions is cheap to build but drafty to live in, and ready at any instant to fall”).
I know nothing of Rawls, but, as described, his theorizing seems to show an advanced case of deductive reasoning disease of the type Housman dissects in his lecture. As Housman writes at the outset, “This method, conclusion first, reasons afterwards, has always been in high favour with the human race: you write down at the outset the answer to the sum; then you proceed to fabricate, not for use but for exhibition to the public, the ciphering by which you can pretend to have arrived at it. The method has one obvious advantage,—that you are thus quite sure of reaching the conclusion you want to reach.”
I look forward to following on (as best I can) where this exploration of political philosophy and its discontents may go.
I like the term ‘sober political realism’ and look forward to learning more about it. As a layperson I have occasionally wondered if philosophy in general should try harder to constrain itself to language, concepts and thought experiments that are realistic and can be widely understood. I have heard of ordinary language philosophy but I’m not sure if that captures what I’m saying here.
In physics it is common to simplify a system to be able to make a closed mathematical model and gain understanding. ‘spherical cows’. This is quite different to making an arbitrary change (chess->chmess is not a simplification it would have the same complexity). A lot of economic theory also has simplifying assumptions (e.g. rational consumers)- you can then understand the simpler system and then try and re-add complexity step by step. It sounds like the Rawls political model is more in these camps -which means this article completely misses the point as it is based on the erroneous chmess analogy.
What is your take on Amartya Sen's rather direct response to Rawls, "The Idea of Justice," which addresses "the plurality of impartial reasons" and his capabilities approach, which is much more material and pragmatic?
I'm essentially sympathetic to Sen, but unfortunately he doesn't really pull off what he needs to. The critique by Schmidtz is good on this: https://www.jstor.org/stable/10.1086/660816
Thank you for the paper. Unfortunately, it seems to me to be either a misapprehension or a mischaracterization of Sen's work, not out of neglect or malice, but as if trying to make intellectual points about what an ideal nonidealist theory should look like.
It both starts and ends with straw men. Sen's point about Mount Everest was concrete:
"The possibility of having an identifiably perfect alternative does not indicate that it is necessary, or indeed useful, to refer to it in judging the relative merits of two other alternatives; for example, we may indeed be willing to accept, with great certainty, that Mount Everest is the tallest mountain in the world, completely unbeatable in terms of stature by any other peak, but that understanding is neither needed, nor particularly helpful, in comparing the peak heights of, say, Mount Kilimanjaro and Mount McKinley. There would be something deeply odd in a general belief that a comparison of any two alternatives cannot be sensibly made without prior identification of a supreme alternative. There is no analytical connection there at all."
Schmidtz then permits the abstract leap onto some moral landscape:
"Sen says we need not know how high Everest is if our goal
is to compare lesser mountains (101). Simmons responds, “We can hardly claim to know whether we are on the path to the ideal of justice until we can specify in what that ideal consists.”
"The metaphor is Sen’s. If it misleads as astute a critic as Simmons, Sen has only himself to blame. Sen scarcely gestures at an argument here, or even at a clear thesis. However, I think what Sen needs to say is that the terrain’s outstanding landmarks are injustices: pits in an otherwise featureless plane."*
* then the footnote to that statement:
"There are passages suggesting otherwise, but I would not try to account for everything Sen says. I extrapolate from remarks such as..."
This is bad faith. Not just permitting the projective leap, but joining in it while also declaring what Sen "needs to say" about this landscape DESPITE evidence that he would NOT say this, while Schmidtz excuses himself from accounting for what the author says in a Review essay about that author's book. Not only did Sen do more than "scarcely gesture at an argument, or even a clear thesis," he contained and gave context to an explicit argument and wrote a book supporting a clear thesis which reinforces that very context.
The middle of the paper proves Schmidtz is familiar with transaction and opportunity costs, and yet it seems to escape him the social and economic landscape is not categorically equal to choosing which mountain to climb. Economic growth and development affords changing "the mountain you are currently climbing" rather than needing to descend and reascend some other mountain in a zero sum landscape. Schmidtz' further assertion, that Rawlsian, or other forms of idealism, as producers of "hypotheses of principles that are testable in practice," survive Sen's critique in any substantive way. The opportunity costs of which and whose "hypotheses of principles" to test in practice simply returns us to the original conundrum of whose ideals to implement in practice.
The worst offense is at the end:
"The West eliminated famine. The question is how. Sen seems to
think the secret has to do with the West securing a right to be fed. Feeding people undoubtedly has saved many lives in dire emergencies, but when it comes to explaining why so many countries no longer suffer from chronic, large-scale famines, it is hard not to notice that the real secret has more to do with securing the right to produce
than the right to be fed."
I struggle not to take vicarious offense at this. Not only did Sen not overlook "the right to produce" in favor of "the right to be fed," he developed the research question, proved BOTH of these as vulnerabilities, for which the West provided protection, WITHOUT directly idealizing and placing into the care of enforcement institutions... and he did it to the tune of a Nobel Prize for those very things, which Schmidtz knows.
Not only is this an ineffective critique of Sen's capabilities approach, or any element thereof as far as I can see, but Sen's book, being reviewed, and your very post here, are effective critiques of where Schmidtz went wildly wrong: he was playing 5D Chmess.
I experienced some cognitive dissonance while reading this essay. I big fan of Williams' due to his work on misinformation and his critique of naive political realism. But, at the same time, I have become increasingly partial to Rawls. When I was getting my phd in history, I barely knew who Rawls was. I guess you would say that we were studying chess instead of chmess.
But I started reading him about 8 years ago when I began to the increasing left authoritarianism surrounding me in academia. His ideal of pluralism in Political Liberalism seemed to offer an alternative to being forced into intolerant tribalism of the left or the right. Yeah, it's abstract and dense, but his basic idea seems pretty simple to me: "Hey we should be able to agree to some rules about how to live together fairly even if we don't agree about who God is or what He wants us to do with our lives." To me, that seems similar to the conclusions that I draw from Williams' critique of naive political realism: "Like, maybe we should chill out before we censor and punish those who disagree with us because maybe we're wrong and maybe they're not as evil as we think they are."
Maybe Rawls ideal game never did and never can exist, but I think that we were closer to playing it in the 20th century US than we are now. And I'd rather play open-minded tolerant liberalism than what we are playing at now even if the latter is more consistent with human nature.
This sounds to me like yet another of the constant demands from the "left-wing" and the "right-wing" (which are really siblings, not opposites) that we should give up on liberal democracy and just have a civil war. It also sounds like the constant left-wing attacks on economics by using the term "homo economicus" to make fun of all economic theory and claim that we should just throw out models and rely on dialectic (which never works, for anything, because it's just very sloppy medieval scholasticism).
"Not only is it manifestly the case that politics is exemplified by agents not exhibiting “strict compliance” regarding justice, but more fundamentally, the very reason that we have politics is precisely because human beings don’t exhibit strict compliance regarding things like justice." -- This is also left-wing/right-wing rhetoric. It dismisses the opposition as evil haters of justice. Whereas in reality, the reason we have politics is more because humans DO exhibit strict compliance with e.g. justice, but have widely varying beliefs about what "justice" is, AND insist on calling everyone who disagrees with them "evil" instead of admitting that they, too, want justice, but have different ideas about what is just.
I know I'm late to the party, but this was an incredibly interesting post, and it has given me a substantial amount of reading to do!
However, it strikes me that all three of Rawls' assumptions are simplifying. Two assumptions place restrictions on what behaviors we can expect, making behavior more predictable, and one assumption restricts the number of interactions we need to consider. While I don't believe these are the reasons Rawls' adopted these assumptions, they still seem like they could be useful.
That's not to endorse Rawls' philosophical project, or to say that it isn't political chmess. But since you acknowledge simplifying assumptions as potentially legitimate, and these assumptions seem to be simplifying to my taste, I think a Rawls' defender need not be compelled by your argument.
I had not heard of "chmess" before. I am reminded, very strongly, of academic economics (in which I have a Master's Degree). "Neoclassical economics is the celestial mechanics of an imaginary universe" (Kenneth Boulding).
Yes, economists are of course well aware of this criticism, but I never found their responses to it very persuasive. Milton Friedman argued that economic models should be judged by their accuracy, rather than their assumptions, but their accuracy is literally terrible, so ¯\_(ツ)_/¯
In some ways I am sympathetic, but wouldn’t this critique apply straightforwardly to any discipline that employs significant idealizations? Wouldn’t it show that idealized models, in, say, economics ought to be discarded? Or that the discipline ought to shift significantly because they’re in effect playing chmess? Or is the difference that political philosophy is a normative discipline in a way that economics is not?
Yes, one might even think that the displacement of political economy by economics was a big wrong turn...
I feel like you may be overestimating the power of keeping models close to reality. There is a class of propositions in economics called "irrelevance propositions," which use wildly unrealistic assumptions to prove kind of insane theorems. The no-trade theorem is one example, and Ricardian equivalence is another. The point of these theorems is to show that under a range of very serious idealizations, some equivalence holds. But given that the equivalence doesn't actually hold, we can then isolate and test the assumptions to see what gives rise to the divergence.
I'm relatively confident that we can construe Rawls in this way, and the three unrealistic assumptions you point out are different reasons why the actual world diverges from the just society. It's a very useful roadmap to have. The hope is that we might be able to isolate how the violation of, e.g., sufficient reasonableness, a closed society, etc, contribute to the actual state of the world.
So my feeling is that these kinds of models are extremely useful, maybe even precisely *because* they diverge from the real world in such drastic ways. This could just be economist "cope", but I sort of doubt it.
"I'm relatively confident that we can construe Rawls in this way"
Based on what? You need to give reasons here, it has to be more than just an assertion of faith.
"It's a very useful roadmap to have."
Why? I've tried to give some reasons for thinking it is very much not a useful roadmap. Indeed not a roadmap at all, but a map of imaginary roads that don't exist. How useful is that?
"The hope is that we might be able to isolate how the violation of..."
Well, it's good to have hope, unless that hope turns out to be unfounded and leading you off in unhelpful directions
I am afraid that what you have said here looks like nothing other than "cope", and the main fact that you "sort of doubt it", based on your "feeling", may suggest that you are arguing this because you *want* it to be true, not because there is good reason to believe that it is true.
In truth, you actually offered no arguments that such idealizations aren't useful, you just asserted that it wasn't plausible! Here:
"Rawls insists that without his ideal theory (or something like it) “there is no rational basis for continually adjusting the social process so as to preserve background justice, nor for eliminating existing injustice”. But this is not plausible, and not only for the reason that it is a bare assertion, one we should no more accept than if somebody told us that until we have an ideal theory of the best ship then we have no rational basis for repairing our leaky boat. More than this, it is as if the ideal theory we were being told to consult turns out, upon closer inspection, not to be about water-borne vessels at all, but flying carpets. Or if you like, about chmess rather than chess."
You've basically just said: "this isn't plausible because Rawls is changing the subject." But whether idealizations are just a change of subject is exactly what's at issue. Asserting that they are is begging the question.
You've isolated three dimensions in which Rawls idealizes, and there is tons of empirical work to be done examining how far away we are from these ideals in the actual world, and this empirical work can generate policy proposals. Consider, for example, the work aiming to make public discourse more reasonable—nonprofits like Sway that try to change public discourse in order to get people to recognize reasonable disagreement and change their practices. If we don't have reasonableness as an idealization this work isn't motivated. But of course, that's only one dimension.
p.s. there isn't "tons of empirical work to be done" on these matters, the empirical work has been in for ages, and the answer is clear: we are light years away from his ideals being anything like the world we either live in, or have any reasonable expectation of ever living in.
These are not "idealizations" of actual political phenomena, they are imaginings of phenomena that could no longer be accurately described as political because they have departed too far from the necessary building blocks of what counts as politics
I guess the point of my comment about irrelevance propositions was to show that many things that people take to be "insane" or to have "departed too far from the necessary building blocks of [subject X]" actually turn out to be very useful.
In the piece you offered an extended analogy, the validity of which many people on this thread are calling into question. Their main point is that your analogy overgenerates, and would eliminate lots of very useful idealizations along with the ones that are not useful. But you offered very few arguments that Rawls' views fall into the latter category. Instead, as the quote below shows, you just relied on your analogy again.
Agreed
Took the words right out of my mouth. I don't think Rawlsian ideal theory is actually defensible via the same arguments used for defending frictionless planes and expected utility maximizers--Appiah has a very nice discussion to this effect in his book on idealization--but I think the argument here is overbroad, in that it doesn't really let us distinguish good idealizations from bad ones.
As the author says near the end: 'Ideal theory of the kind practiced by Rawls and others does not involve simplifying assumptions; it involves counterfactual assumptions that mean it is theorising something altogether different.' As someone said 'All models are wrong. Some models are useful.' Judging whether a model contains usefully and acceptably simplifying assumptions or invalidating counterfactual assumptions is a critical aspect of assessing its utility.
What Alaistar said. More generally, see that Dave Schmidtz discussion about theories as maps linked to in the original
But you could use that exact argument against all kinds of perfectly reasonable simplifying assumptions! Assume this spring were made of continuous matter (Hooke’s Law). Assume people were all self-interested (much of evolutionary game theory). Etc.
The missing idea in this article, IMO, is that we might be making simplifying assumptions in order to illustrate fundamental properties of a system that we all know is, in reality, more complicated. The point is to assume away some real features in order to isolate certain others. This isn’t always a good move, but you can’t always dismiss it as chmess.
An easier target than Rawls would be those political philosophers whose *policy proposals* only make sense in an ideal dream world—a world that ignores, tradeoffs, transaction costs, or what have you. Rawls is at least committed to caring about incentives etc. since these constrain the set of possible basic structures for society.
EDIT: but let me add that I loved the connection to Dennett on chmess. I’d forgotten about that passage!
i'm not 'always' dismissing 'it' as chess. Just ideal theories of politics, specifically.
Rawls thinks that real world political decisions (the realm of "nonideal theory") should be regulated by and made in reference to (his) ideal theory. So I'm afraid you can't get them off the hook that easy.
And yes, he cares a bit about incentives, so he's doing better than e.g. GA Cohen. But he doesn't care about them enough, because he doesn't have a realistic enough moral psychology.
Thus Chmess it remains!
I grant you that it would be chmess if Rawls were basing policy recommendations *entirely* on his theory. But isn't he pretty circumspect about doing so?
[Edit: Removed a false sentence -- Paul S is right. Rawls is not "circumspect" about welfare-state capitalism in sec. 42!]
To put it another way, I don't yet see your objection to the following kind of approach: we use a partly idealized theory to get our ideals and then use grubby empirical facts to figure out how to translate those ideals into policies. (Maybe your point is that idealizing can go too far even when trying to find ideals for a society?)
TJ says welfare state capitalism is ruled out, and we probably need to go for "property owning democracy", which is effectively a form of socialism, although he is very unspoken on the economics, which leaves him vulnerable on that point.
I don't have an objection to the kind of approach that you describe (so long as the idealisation does not lose touch with the basic building blocks of what politics necessarily has to be for creatures like us). My point is: Rawls (and such) are nowhere close to accurately being described as doing that.
What you describe is a lot more like what Dave Schmidtz advocate for in the first three chapters of his book Living Together. I like what Dave suggests.
Great point -- thanks for the correction, Paul! (I've edited my earlier comment.)
Came here to glibly say, "now do econ", lol. But yeah, I think morally there is in fact a massive difference given that pol phil is (presumptively) normative and econ is (presumptively) descriptive. And that's not to say one should be able to get away with dubious assumptions more than the other. In general, houses of cards do not practically great mathematical frameworks make. And I don't think it's necessarily "better" that one can or should be able to quantify how off the idealization is from (what is for all practical purposes) reality. Yet IDK, maybe it's just my being a geek, but somehow it makes me less uncomfortable to say, "yeah this mathematical model might be garbage IRL but it's still kinda cute" than to say that moral intuitions about how to govern and live were fleshed out in a way that turned out to be...spurious. Ya feel me?
I feel ya!
This reminds me of a joke: a farmer, whose chickens have been dying, employs the services of a biologist, a chemist, and a physicist. After dispatching with the theories of the biologist and the chemist (mites and toxic chemicals, respectively) the farmer asks the physicist what she thinks. The physicist, who has been busily scribbling on a note pad while the various investigations were taking place, says, “I believe I have the answer—but it only works for a spherical chicken in a vacuum.”
I'm a physicist and came here exactly to use the same joke, but with the opposite meaning.
I also studied *theoretical* physics and always found the joke utterly dumb (meaning also this whole post is actually wrong).
Of course in reality friction exists, so Newton's first principle (actually Galileo's) that an object mantain its status of linear motion if there is no force, seems unwarrented.
You know who started his physical thinking by what actually seem to happen? Aristotle. He got almost everything wrong. It took Galileo and Newton's "idealization" to build a theory that works.
So no, I'm sorry, your argument is unfounded and wrong.
In short, it's the claim that you shouldn't simplify a situation because that's unrealistic.
But in practice you should, at least as a start, because otherwise the problem is untractable.
Nice joke though
All I would say here is: this is just yet more proof that political philosophy is fundamentally a different thing to physics :)
(in fact, I think it is true generally that philosophy is fundamentally a different thing to science… but that’s for another day)
Thank you!! My name is Nathan Woodard and I approve of your message!! I feel like the golden age of physics in the early 20th century created a lot of confusion in philosophy and social sciences--it became fashionable for theorists in other disciplines to steer towards the methods of the hard sciences (not hard as in difficult--hard in the sense of "hard determinism"). It seems to me that a lot of disciplines (like economics?) are still suffering a certain amount of brain damage that accrued during that period.
I'm also a physicist. As such, I am very envious of your name. :) :) :) :)
We physicists are expert wielders of idealized assumptions — a priori false but useful tools of thought. In social studies and many nascent sciences, however, such assumptions often ossify into dogma and become ideological cudgels that can persist for generations. Even in physics, whole subfields have at times gone astray — ether theory, caloric, steady-state cosmology, possibly even string theory today — but our field has generally managed these detours within one or two generations, thanks to the corrective pull of experiment and prediction.
Physics envy -- scourge of many a comparatively insecure discipline.
The weirdest thing is, I know some people doing political philosophy who admit its like chmess and they dont have a problem with it. Neither do they defend chmess assumptions or try to make it relevant. They just treat it like a system of puzzles that you just solve. Their job is to take a set of assumptions and solve for results, without being concerned with applicability, rinse and repeat. It boggles my kind to call it "political" philosophy but here we are.
This is among the best essays I have encountered in Substack. Thanks! It seems like you are entering into territory that clicks very neatly into some of the biggest debates of the centuries:
1) Social constructivism vs biological determinism.
2) Tragic vision vs, constrained vision. (Sowell)
3) Positive vs. Negative Liberty (Isaiah Berlin)
4) Reason vs. Tradition (Burke vs. Rationalist reformers)
5) Empiricism vs. Rationalism in the human sciences
I’m willing to bet my prepper stash (dried beans, seaweed, an FM radio, and a giardia filter) that you had all five of these debates—and more—at the back of your mind as you wrote this. (Kidding about the stash, and apologies if I missed something you did cover.) If I’m wrong, I’d be thrilled for you to unpack that. If I’m right, did you consciously steer clear of these topics to keep things morally and politically neutral? Either way, I’d love to see follow-up essays that frame these debates within the scaffolding you’ve built here. Those essays would be legendary—a real gift to posterity.
Well, that stuff is in the background for me, because I've spent about 20 years immersed in moral and political theory - and crucially, their histories - and that is just the water I swim in. I didn't make any of it explicit, because I don't think Dan would've allowed me to run a 3,000,000 word guest post!
That makes perfect sense. :) Thanks again for the brilliant essay! I just ordered all your books. :)
i’m afraid they are a lot more boring! Although I suppose they do retain a certain willingness to pick fights…
As Max Stirner would say, you can argue about rights all you want but at the end of the day who owns the means of production is whoever seizes them.
One factor that certainly contributes to Chmess is the incentive to regularly publish novel results. Chmess is much safer, after all you're just doing speculative theory so if you're wrong, there are no brutal regimes' blood on your hands. The h-index can march up risk-free.
Which is fairly endemic as far as the Symbolic Capital Bubble goes.
Oooh yes i like that. Seems like it could be a very good supplement to Joseph Heath's case regarding why all the Marxists became Rawlsians...
I don't think the chmess analogy works very well. The problem with chmess is that it's completely arbitrary - there's no rhyme or reason to the differences it has from chess, and it's impossible to imagine these differences being somehow normatively correct. The entire thought experiment rests on the fact that it just seems obviously absurd that learning about chmess could better help you play chess than learning actual chess theory, but this is the very claim that's at issue when it comes to ideal theories in politics. Maybe learning about the ideal is unhelpful, but you can't prove this by making an analogy about the unhelpfulness of an obviously non-ideal game like chmess.
Here are some plausible ways that ideal theories might be helpful for learning about politics even if they're not strictly true, and which are not captured in the chmess analogy:
1. They're simpler than the real world. The real world is just far too complicated to model in its entirety. You have to simplify it no matter what, and then add perturbations to you model later to make it more like the real world. Proponents of ideal theories do this openly, but people who claim to base their politics on the real world and not idealizations are still doing it. And all fields do this, not just political theory. Physicists do this all the time, to the point that idealizations made in physics problems have become a common joke, and yet no one doubts the utility of physics.
In this way, the analogue to an ideal theory would not be chmess, but something more like, "chess, but without castling or en pessant," or other simplifications of the rules. And it's a lot more obvious why learning about this simplified version of chess would help you to eventually understand the full game. In fact, it's how everybody actually learns chess: First you learn how all the pieces move under normal circumstances, and then you learn about the exceptions.
At least some of the assumptions that Rawls made, like the society being closed, seem to be simplifying assumptions of these sort. It's not clear that the differences between his ideal case and the real world invalidate his theory, and even if they do, it's not clear that they invalidate it in such a way that it should be scrapped entirely, rather than just adjusted to fit the real-world circumstances (There's a physics analogy for this process of adjustment, too: perturbation theory).
2. The ideal case is a normative ideal that we should strive towards. Assumptions like, "Everyone is perfectly rational," are obviously meant to capture this facet of the theory. When Rawls assumes perfect rationality and then derives conclusions from it, he's not trying to say, "This is what the world is actually like." He's saying, "Here's what rational people would do. We should be more like them." Objections on the basis of people not actually being rational miss the point.
The chmess analogy tries to counter this by suggesting that chmess is somehow the normative version of chess. The problem with this is twofold: First, it's just obviously not, to the point I have a hard time even imagining myself being convinced that it is. All of my intuitions about the utility of chmess are based, even if implicitly, on the obvious fact that the rules are completely arbitrary, not normative. Even when asked to imagine that they somehow are normative, I don't think I'm fully able to do this and overcome the intuition of them being arbitrary. That means the intuitions generated by the thought experiment aren't reliable. I can't trust intuitions about something I can't even conceive of properly, and I certainly can't when I'm trying to imagine something being normatively correct while my entire mind is shouting back, "No it isn't, even in a counterfactual scenario!"
Second, the chmess analogy relies on the rules of chess being impossible to change, so that even if you were convinced chmess was somehow better, you couldn't do anything about it. But that's not true for politics. The rules of politics can change to be brought closer to the ideal, and this has happened many times throughout history. That's the entire reason people study political theory. Furthermore, some ways of bringing the world closer to the ideal don't even require changing the rules of politics, but just changing object-level policies. For example, you can develop policies designed to give everyone what they would agree with if they were perfectly rational without actually having to make them perfectly rational.
The game of chmess seems to be a minor variant of chess, in such a way that studying it would offer very little insight into the game of chess. It would be roughly equivalent to a Rubik’s Cube with one square on each side with the wrong color. Let’s call this RubNik’s Cube. Yes, RubNik’s Cube would be challenging (for humans who use their color perception to solve it), but it’s not mathematically interesting. To a computer that begins with a definition of the cube that includes the original position of each square, it would be mathematically identical to a simple Rubik’s Cube. Chmess is hardly more interesting; in fact, it is far less interesting than Chess 960 (Fisher Random), a variant (to be exact, a group of 960 variants) that top players in the world do in fact play.
Now consider Rubik’s cubes of 4, 5, 6,….n layers. To study these objects, the original Rubik’s cube with 3 layers would be a valuable first step. It’s a simple case, a good place to start for the study of n-layer cubes. Of course, the simple application of the rules and techniques of solving a 3-layer cube would be inadequate for solving, say, a 10-layer cube, but it's difficult to imagine solving a 10-layer cube without solving the 3-layer cube first.
It seems that the assumption, in much political and economic theory, that humans are rational actors is such a first step. Such theories assume that human behavior can be modelled as an optimization function. The first approximation of that theory—that humans rationally pursue rational self-interest, is an extreme oversimplification of human behavior, but it’s not necessarily a bad place to start.
An optimization function that could actually model (and therefore explain) human political and economic behavior would have to possess many dimensions, starting with many of the factors that Dan Williams talks about all the time: confirmation bias, sense of self-worth, social approval, in-group loyalty vs out-group hostility, etc.
Even for one (such as I am) who is completely removed from the halls of academe, and particularly their philosophy departments, I found this essay a witty, enjoyable read. Also, the links toward the end to books examining more “chess like” approaches look fascinating, such as Robert Jubb’s “Unjust Authority” and Judith Shklar’s “The Faces of Injustice.”
I am also reminded, in reading this essay, of A.E. Housman’s “Introductory Lecture of 1892” (from which I took the quote I use for my Substack profile: “The house of delusions is cheap to build but drafty to live in, and ready at any instant to fall”).
I know nothing of Rawls, but, as described, his theorizing seems to show an advanced case of deductive reasoning disease of the type Housman dissects in his lecture. As Housman writes at the outset, “This method, conclusion first, reasons afterwards, has always been in high favour with the human race: you write down at the outset the answer to the sum; then you proceed to fabricate, not for use but for exhibition to the public, the ciphering by which you can pretend to have arrived at it. The method has one obvious advantage,—that you are thus quite sure of reaching the conclusion you want to reach.”
I look forward to following on (as best I can) where this exploration of political philosophy and its discontents may go.
I like the term ‘sober political realism’ and look forward to learning more about it. As a layperson I have occasionally wondered if philosophy in general should try harder to constrain itself to language, concepts and thought experiments that are realistic and can be widely understood. I have heard of ordinary language philosophy but I’m not sure if that captures what I’m saying here.
In physics it is common to simplify a system to be able to make a closed mathematical model and gain understanding. ‘spherical cows’. This is quite different to making an arbitrary change (chess->chmess is not a simplification it would have the same complexity). A lot of economic theory also has simplifying assumptions (e.g. rational consumers)- you can then understand the simpler system and then try and re-add complexity step by step. It sounds like the Rawls political model is more in these camps -which means this article completely misses the point as it is based on the erroneous chmess analogy.
What is your take on Amartya Sen's rather direct response to Rawls, "The Idea of Justice," which addresses "the plurality of impartial reasons" and his capabilities approach, which is much more material and pragmatic?
I'm essentially sympathetic to Sen, but unfortunately he doesn't really pull off what he needs to. The critique by Schmidtz is good on this: https://www.jstor.org/stable/10.1086/660816
Thank you for the paper. Unfortunately, it seems to me to be either a misapprehension or a mischaracterization of Sen's work, not out of neglect or malice, but as if trying to make intellectual points about what an ideal nonidealist theory should look like.
It both starts and ends with straw men. Sen's point about Mount Everest was concrete:
"The possibility of having an identifiably perfect alternative does not indicate that it is necessary, or indeed useful, to refer to it in judging the relative merits of two other alternatives; for example, we may indeed be willing to accept, with great certainty, that Mount Everest is the tallest mountain in the world, completely unbeatable in terms of stature by any other peak, but that understanding is neither needed, nor particularly helpful, in comparing the peak heights of, say, Mount Kilimanjaro and Mount McKinley. There would be something deeply odd in a general belief that a comparison of any two alternatives cannot be sensibly made without prior identification of a supreme alternative. There is no analytical connection there at all."
Schmidtz then permits the abstract leap onto some moral landscape:
"Sen says we need not know how high Everest is if our goal
is to compare lesser mountains (101). Simmons responds, “We can hardly claim to know whether we are on the path to the ideal of justice until we can specify in what that ideal consists.”
"The metaphor is Sen’s. If it misleads as astute a critic as Simmons, Sen has only himself to blame. Sen scarcely gestures at an argument here, or even at a clear thesis. However, I think what Sen needs to say is that the terrain’s outstanding landmarks are injustices: pits in an otherwise featureless plane."*
* then the footnote to that statement:
"There are passages suggesting otherwise, but I would not try to account for everything Sen says. I extrapolate from remarks such as..."
This is bad faith. Not just permitting the projective leap, but joining in it while also declaring what Sen "needs to say" about this landscape DESPITE evidence that he would NOT say this, while Schmidtz excuses himself from accounting for what the author says in a Review essay about that author's book. Not only did Sen do more than "scarcely gesture at an argument, or even a clear thesis," he contained and gave context to an explicit argument and wrote a book supporting a clear thesis which reinforces that very context.
The middle of the paper proves Schmidtz is familiar with transaction and opportunity costs, and yet it seems to escape him the social and economic landscape is not categorically equal to choosing which mountain to climb. Economic growth and development affords changing "the mountain you are currently climbing" rather than needing to descend and reascend some other mountain in a zero sum landscape. Schmidtz' further assertion, that Rawlsian, or other forms of idealism, as producers of "hypotheses of principles that are testable in practice," survive Sen's critique in any substantive way. The opportunity costs of which and whose "hypotheses of principles" to test in practice simply returns us to the original conundrum of whose ideals to implement in practice.
The worst offense is at the end:
"The West eliminated famine. The question is how. Sen seems to
think the secret has to do with the West securing a right to be fed. Feeding people undoubtedly has saved many lives in dire emergencies, but when it comes to explaining why so many countries no longer suffer from chronic, large-scale famines, it is hard not to notice that the real secret has more to do with securing the right to produce
than the right to be fed."
I struggle not to take vicarious offense at this. Not only did Sen not overlook "the right to produce" in favor of "the right to be fed," he developed the research question, proved BOTH of these as vulnerabilities, for which the West provided protection, WITHOUT directly idealizing and placing into the care of enforcement institutions... and he did it to the tune of a Nobel Prize for those very things, which Schmidtz knows.
Not only is this an ineffective critique of Sen's capabilities approach, or any element thereof as far as I can see, but Sen's book, being reviewed, and your very post here, are effective critiques of where Schmidtz went wildly wrong: he was playing 5D Chmess.
I experienced some cognitive dissonance while reading this essay. I big fan of Williams' due to his work on misinformation and his critique of naive political realism. But, at the same time, I have become increasingly partial to Rawls. When I was getting my phd in history, I barely knew who Rawls was. I guess you would say that we were studying chess instead of chmess.
But I started reading him about 8 years ago when I began to the increasing left authoritarianism surrounding me in academia. His ideal of pluralism in Political Liberalism seemed to offer an alternative to being forced into intolerant tribalism of the left or the right. Yeah, it's abstract and dense, but his basic idea seems pretty simple to me: "Hey we should be able to agree to some rules about how to live together fairly even if we don't agree about who God is or what He wants us to do with our lives." To me, that seems similar to the conclusions that I draw from Williams' critique of naive political realism: "Like, maybe we should chill out before we censor and punish those who disagree with us because maybe we're wrong and maybe they're not as evil as we think they are."
Maybe Rawls ideal game never did and never can exist, but I think that we were closer to playing it in the 20th century US than we are now. And I'd rather play open-minded tolerant liberalism than what we are playing at now even if the latter is more consistent with human nature.
This sounds to me like yet another of the constant demands from the "left-wing" and the "right-wing" (which are really siblings, not opposites) that we should give up on liberal democracy and just have a civil war. It also sounds like the constant left-wing attacks on economics by using the term "homo economicus" to make fun of all economic theory and claim that we should just throw out models and rely on dialectic (which never works, for anything, because it's just very sloppy medieval scholasticism).
"Not only is it manifestly the case that politics is exemplified by agents not exhibiting “strict compliance” regarding justice, but more fundamentally, the very reason that we have politics is precisely because human beings don’t exhibit strict compliance regarding things like justice." -- This is also left-wing/right-wing rhetoric. It dismisses the opposition as evil haters of justice. Whereas in reality, the reason we have politics is more because humans DO exhibit strict compliance with e.g. justice, but have widely varying beliefs about what "justice" is, AND insist on calling everyone who disagrees with them "evil" instead of admitting that they, too, want justice, but have different ideas about what is just.
I know I'm late to the party, but this was an incredibly interesting post, and it has given me a substantial amount of reading to do!
However, it strikes me that all three of Rawls' assumptions are simplifying. Two assumptions place restrictions on what behaviors we can expect, making behavior more predictable, and one assumption restricts the number of interactions we need to consider. While I don't believe these are the reasons Rawls' adopted these assumptions, they still seem like they could be useful.
That's not to endorse Rawls' philosophical project, or to say that it isn't political chmess. But since you acknowledge simplifying assumptions as potentially legitimate, and these assumptions seem to be simplifying to my taste, I think a Rawls' defender need not be compelled by your argument.
I had not heard of "chmess" before. I am reminded, very strongly, of academic economics (in which I have a Master's Degree). "Neoclassical economics is the celestial mechanics of an imaginary universe" (Kenneth Boulding).
Yes, economists are of course well aware of this criticism, but I never found their responses to it very persuasive. Milton Friedman argued that economic models should be judged by their accuracy, rather than their assumptions, but their accuracy is literally terrible, so ¯\_(ツ)_/¯
Thanks for the article! I enjoyed it. For me it’s exactly like martial arts and real life fighting.
that's a nice analogy. one looks cool, but the other is a lot more dangerous