The social sciences are useless. So why do we study them? Here’s a good reason:

Back when I taught at Berkeley, you could always get a reaction from the students by invoking Stanford. The funny thing is, though, if you’re at Stanford and mention Berkeley, nobody cares. You have to bring up Harvard to get a reaction.

Similarly, MIT students have a chip on their shoulder about Harvard, but Harvard students don’t really think much about MIT. There’s an asymmetry. Harvard and Yale, that’s more symmetric.

This came up the other day in conversation with a biologist who was saying that, when they consider the scientific method, they think of physics as the gold standard. But when we do social science, we often look up to biology.

The thing about social science is that it hasn’t produced much. We social scientists don’t have an inferiority complex; we really are inferior. Physics has produced locomotives, semiconductors, and the atomic bomb. Chemistry has produced amazing new materials. Biology has produced the coronavirus vaccine, and lots more. Social science has produced . . . what, exactly? A method of evaluating redistricting plans? Better polling? The Big Five? The Implicit Association Test? A better auction rule? Some cool marketing tricks? The past two hundred years of social science have given us nothing as useful and important as what gets produced every day in biology, chemistry, and physics.

But then the question arises: What’s the point of social science? Why do we do it at all? That’s a good question for me, given that I teach in the political science department and write papers on districting, voting power, social penumbras, gaydar, and all the rest.

Here’s my answer. We study the natural sciences because they help us understand the natural world and they also solve problems, from vaccines to the building of bridges to more efficient food production. We study the social sciences because they help us understand the social world and because, whatever we do, people will engage in social-science reasoning.

The baseball analyst Bill James once said that the alternative to good statistics is not no statistics, it’s bad statistics. Similarly, the alternative to good social science is not no social science, it’s bad social science.

The reason we do social science is because bad social science is being promulgated 24/7, all year long, all over the world. And bad social science can do damage.

In summary: the utilitarian motivation for the natural sciences is that can make us healthier, happier, and more comfortable. The utilitarian motivation for the social sciences is they can protect us from bad social-science reasoning. It’s a lesser thing, but that’s what we’ve got, and it’s not nothing.

166 thoughts on “The social sciences are useless. So why do we study them? Here’s a good reason:

  1. This is ignoring pretty much all basic science research. Applied science yields those tangible benefits, but basic science exists just to explore a concept. Maybe down the road there’s a tangible benefit to society, and you can be sure that possibility is included in the grant application. But the real reason the basic science is being done is “this is extremely cool, and I would like to know more about it please”.

    • Anja:

      I agree. I was not intending to suggest that all, or even most, natural science research is useful in any way. My point was that the entire edifice of natural science, going up and down the ladder between the most theoretical and the most applied, has produced and continues to produce useful developments in the way that social science does not.

      As I said in my above post, I still think social science is important, as it offers an alternative to the sorts of “piss-poor monocausal” reasoning that would be done in its absence. That’s one reason I get annoyed at bad social science: if the whole point (in my mind) of social science is to serve as a replacement for bad forms of reasoning, it’s kinda horrible when the bad reasoning is done inside the house. Indeed, bad social science can be worse than bad informal social reasoning, as it comes with this unearned air of certainty. Remember Bertrand Russell’s line about “views which are so absurd that only very learned men could possibly adopt them.”

      • I think that giving institutional support to what’s basically back justified tip down social control is more harmful than just letting shamans do it. Both for science, which loses credibility every time a psychologist opens their mouth or picks up a pencil and for society which is constantly being reorganized to a new totalitarian ideal.

        • Soren:

          When I do social science, I don’t think I’m “giving institutional support to what’s basically back justified top down social control.” Indeed, much of my social science research is explicitly pushing against that idea.

  2. Andrew, I agree with your assessment about the social sciences. And it’s probably moot to point out once more that compared to physics, which tries to carve nature at its true joints, we’re much too far downstream from the level of observation where you can actually still observe causal chains first-hand. We are at the level of mass effects, where too many factors shape the visible outcomes. You can consdier yourself lucky if you do find an island of replicable predictability somehwere in there. But at the back of my mind there’s still Asimov with his idea, articulated more than 60 years ago in the Foundation books, that perhaps one day we will have the tools (mathematical?) to predict the behavior of humans at a grand scale and eventually steer the course of political processes and history instead of only interpreting it or making empirical sense of it post-hoc. Let’s face it: if we in the social sciences had really understood what leads to a Hitler, a Stalin, or the other butchers of recent history, we would have prevented their successors ever since. We haven’t, and that’s embarrassing and painful for us scientists and certainly for those people who feel the wrath of their mad leaders directly. Just in the last couple of years, the world had another 1933 moment, and it’s been neither predicted nor prevented by social scientists. So yes, preventing bad social science is a worthy goal. But an even better one is trying to reach higher. Asimov painted a scientific utopia. Perhaps we can get there eventually.

      • Social science helps us make better collective choices. Living in social groups is a human adaptation. While other animal species also have social adaptations, we rely quite extensively on reasoning to cooperate with each other in mutually advantageous ways.

    • You say that if we “really understood what leads to a Hitler…. we would have prevented their successors ever since.” You might have said “their successors would have been more successful.” Understanding cuts both ways. “Progress” in understanding – whether in the physical or social sciences – can be used for good or evil (and/or anywhere in between). Also, Asimov’s “scientific utopia” sounds ominous to me. What role is left for humans when we can predict behavior perfectly? I suppose we get a lifetime of enjoyment and leisure, but it isn’t obvious there is much else for us to do.

      • Dale, I don’t think that this really cuts both ways. Almost all of the recent authoritarian strongmen of world politics tried to ignore science — it’s too complex for them and doesn’t yield starightforward claims of the sort “…..!” (the exclamation mark at the end is the really important part). Typically it’s democratic governments whose memmbers are used to dealing with tiresome complexity who listen to science. Not always, not pefectly, but certainly more than the dictators. But yes — an intelligent dictator who understands science and knows how to use it to his (could it ever be her?) advantage — that would indeed be a frightening vision. But it would also be an exremely rare phenomenon.

        And I realize that I may have misrepresented Asimov: his vision was more of a new mass psychology that allows you to predict the broad strokes of historic developments but not individuals and certainly not perfectly. If I remember correctly, in his trilogy the protagonist used the knowledge not to change history itself but to make sure that certain achievements of civilization would make it safely through the next dark ages and be available for further development after humankind has emerged from its most recent dip into chaos.

        • I accept your revised Asimov description in the 2nd paragraph, though I still see the scientific utopia (recognizing that it is not Asimov’s utopia, but it is articulated by some others) as ominous – and all too possible.

          As for authoritarian leaders and science – it isn’t a matter of what they publicly say. I’ve heard the anti-science sentiments, but on another level they use science to gain and stay in power. Considerable science has been used to harvest votes and manipulate voters. The more we can predict human behavior, the greater the temptation will be to further control people to stay in power. So, I do think it cuts both ways – better understanding of human behavior need not prevent authoritarians – I’m not even convinced it makes them less likely.

        • I think Hitler is an exceptionally poor example, the 1930s were a rather unusual time frame and Germany in a fairly unique position among the nations of the time.

          ‘Strongman’ leaders in general are a far more common phenomenon (I’d argue that at least since the dawn of written history, they’ve been more the norm than the exception; most premodern monarchies are tied up with the military in origin).

          But I think both claims (authoritarian leaders are poorly placed to use social science vs. it provides tools easy to misuse) are correct to a degree. The authoritarian personality type may not be good at using it, and may not produce institutions that are good at using it – but modern complex society IMO has far more direct insight into what is going on in the population, due to vastly greater information flow and vastly better tools to understand it, and things like ubiquitous ID requirements – so much more direct control of a large population is possible than it ever was to, say, Roman emperors or George III.

          I think in the “Western world” we are now in a position where takeovers are significantly harder but the powers of a government that did go definitively bad would be greater.

  3. Some specific applications of economics with very direct utilitarian benefits:

    – Kidney exchange programs ( save lives much in the same way medical progress does. They rely greatly on the tools of market design. Related applications are in school choice and all the other types of matching programs.
    – Central banks are much better at conducting monetary policy than they would be without econ (see e.g. the bad decisions they made in the 70ies — thought that was admittedly partly the fault of bad economics).
    – How you structure patenting laws has great effects for what type of R&D firms invest in. There’s currently big issues with that in cancer research. Without a good understanding of the effects of these incentives you can’t make the laws perform well.
    – People save millions of dollars because academic finance research has pushed the market from high-fee, actively managed funds to low-fee index funds. And yes, it really was academia driving this; it’s not at all in the interest of the financial industry (which keeps pushing back)

    Overall, you’re probably still right that the social sciences are inferior to physics, biology, and the engineering disciplines. But there are some very tangible, direct benefits of social science research.

    • Sandro:

      1. Sure, but think about it. Biology developed kidney transplants. Social science developed . . . an improved method of assigning these transplants. I agree that this is not nothing, but the contribution of the social sciences here is much much much less than the contribution of the physical sciences. It’s like the difference between the people who developed airplanes and the people who developed . . . ummm, I dunno, a system for pricing airline tickets.

      2. Your second example, though, that one’s better. If we think of accounting and banking and finance as a social science—and I guess it is!—then, yes, this has had a huge impact on the world.

      • To reframe a l little: Things that are important range from the spectacular to the very mundane. We need the whole range. A brilliant idea is of no practical good unless we can do the range of less brilliant things that are needed to implement it well.

      • I know this is late (sorry!) but two thing I think it’s important to ask about: where are distinctions drawn between various elements of science and how do we allocate “usefulness” across very disparate sources of innovation?

        Taking the kidney example–who ought get “credit” for different parts of the iterative process of improvement? Should we give any credit to the long forgotten attempts at skin grafting and limb transplantation conducted by early “surgeons?” If so, then we aren’t crediting scientists per se, but doctors: for they largely didn’t follow the scientific method, which ought exclude them from the realm of a “natural scientist” as defined today. (Tagliacozzi, for example, never attempted to suggest any underlying causal mechanism of graft rejection beyond his assertion that it was the “strength of the patients’ own will.”) How much should we credit Kocher, who substantially advanced the field of organ removal, though did little (if anything) related to organ replacement?

        More directly related to the overall topic, let’s consider the issue of paired kidney exchange, one of the developments that has led to significant life saving in the United States. It was first proposed in the early 1970s by doctors, however nothing came of it due to the perceived legal ban on exchanging organs (since it was illegal to exchange for “anything of value,” which was interpreted by many to mean “another kidney”). And while the first paired donations in the US were conducted in 2000 and 2001, the economists who were the driving force behind regional and national registries were right there with the surgeons since the beginning (with the groundbreaking paper by Roth, Sonmez and Unver being first circulated in 2003). Likely without their work, US legislation wouldn’t have been clarified in 2007, and exchanges wouldn’t have proliferated nearly as far nor as fast. Could their work have happened without the surgeons? Of course not. But It’s not really clear that the economists were secondary themselves.

    • I agree: if you consider economics to be social science (do you?), then there’s been some benefit.

      What do you call the field that produced the thinking that led to the American Revolution and Constitution? Political science? Moral philosophy? Maybe it’s not social science.

      • Darin:

        The thinking that led to the American Revolution and Constitution was pretty cool. It’s also 250 years old! In my above post I was thinking about more recent work. In contrast, physics, chemistry, biology, etc., have led to lots of new things in the past 250 years.

  4. Well, let me just give a little plug for economics. The astonishing increase in living standards over the last 400 years is not independent of the economic system which fostered it. Economists didn’t create the steam engine, or electricity, or markets, or free trade policies, but they provided (through Smith, Mill, and Ricardo and continuing to the present day) the intellectual underpinnings which were needed for a system which has always been under threat despite its unprecedented historical successes. These threats come from the fact that these successes also generate concomitant problems; creating a framework to celebrate the successes, understand their origins and a framework to address the shortcomings is what economics has done when it is doing its best. It’s a derivative kind of success relative to biology and physics and engineering — more explanation and justification than generation, but the framework underlying those gains is fragile, capable of being extinguished in every generation, and the best of economic thought stops that from happening. (And you’re right… the worst of economic thought is what it’s aimed at.)

    • +1

      Yes, strongly agree, IMO this is far and away the greatest success of social science: the understanding of how free markets contribute to bettering the human condition.

        • I’d consider Smith, Mill, and Ricardo more classical political economy than social science as it is today. If we’re including them as important economists we should also include Aristotle and the tradition of Liberal philosophers as important political scientists, since Democratic governance and society are at least as important as and prerequisites to free markets.

          But again, I would honestly call all of them more philosophers than scientists–their work is more utopian theorizing than empirical and they were as much creators of these systems as students of their behavior.

        • I think Aristotle and Plato *were* important political scientists, and Smith, Ricardo and Mill were not just theorizers, they used a lot of data. As to the importance of democratic governance vis-a-vis markets, I suspect we disagree, but they are not completely independent.

          And sure, modern economics journals don’t look much like The Wealth of Nations, and the focus of the great bulk of articles is much narrower than Smith’s vision. But that doesn’t mean they aren’t ultimately aiming at the same thing: characterizing the strengths and weaknesses of markets and plumbing the proper role of gov’t policy to enhance human welfare.

          And the same goes for political science vis-a-vis Aristotle as well, I guess. I just don’t know the literature at all.

    • Don’t forget that medicine also had a framework (or many …) but until recent times it did untold amounts of harm because the framework was not aligned with empirical knowledge).

      I suggest that the same has been the case in other areas, including economics. The problem is how to know that you know enough to be leaving the “untold amounts of harm” arena and to be entering the “public” good one.

      • Modern medicine is currently injecting huge numbers of people with literal poison (chemotherapy). There has still never been a single clinical trial that accounts for caloric restriction due to the effects on appetite and GI tract.

    • I agree, but in fairness it could be broadened to include some other disciplines by referring to what you mention about economic policy to the role some social science disciplines have had in fostering better government in general.

      For psychology, I think psychology could be credited as a way to study nonmedical interventions to improve mental health of both individuals and communities in some cases. This may not be flashy but it’s a legitimate improvement.

    • But Smith didn’t invent neither created the capitalism.

      Interesting comment but I do not think it resist an historical research (funny I mention a social science as an argument). In wealth of nation, Smith describe the change is see around him, he also use older philosophy in an attempts of normative justification. He create insight and *justification* to somethings that is going on in the society around him.

      But if we look at what “produce” science, we have a problem. Religion did produce a lot of things, incredible books and cathedral for example.

      I partially agree of a mutual influence, but I have a hard time to believe that without Smith (or others) they will be no modern capitalism or consumerism society today. Smith work is a comment.

      For me, social science primary goal is justification. Good social science is an introspective for the society while bad social social science is an internal justification toward current academic theory (an academic construction for the existing academic construction). Often internal coherence is solid, but very weak as soon as we go outside the given field.

  5. Couldn’t you say that a better understanding of how humans behave in social environments, which is part of the contribution of the social sciences, has contributed to a lower frequency of wars over time, which created a supportive environment for things developed in the natural sciences to be created and enjoyed?

    • Possible… but there are confounders. I would argue that the lack of world wars during the last 75 years is primarily the result of nuclear weapons making the risk too high (and secondarily, and more speculatively, an economy more based on innovation than land & raw resources means that conquest is no longer so economically favorable for a nation).

      And smaller-scale wars have not exactly been absent.

  6. “The utilitarian motivation for the social sciences is they can protect us from bad social-science reasoning. It’s a lesser thing, but that’s what we’ve got, and it’s not nothing”.

    Well, yes, that – plus when someone in a bar asks what you do for a living, you can say “well, stuff like understanding why some groups have more political clout than others” – which gets you a free drink – instead of “I’m investigating the electric dipole moment of the neutron”, which won’t.

  7. Science also produced nuclear weapons, biological weapons, global warming, etc. and has failed to stop genocide (although it has wiped out smallpox), domestic abuse, etc. My point is that you can’t cherry pick scientific achievements. The difference between those that improve life and those that don’t (the jury is still out on many, such as social media and AI, in my opinion) has much to do with social science. As Jonathan says, our economic system has much to do with the successes – as well as the failures (I’m going beyond what he says here, but I’ll own that). I think the distinction between physical and social science is a false one. The superior progress in physical science certainly exceeds that in social science, for a number of (mostly) legitimate reasons. But the realized superior progress in physical science cannot be separated from social science.

    • Dale:

      Yes, I mentioned the atomic bomb in my above post! “Useful” is not always the same thing as “good.” Whether the natural sciences have been a net good for humanity, that’s an open question. There’s still debate as to whether development of agriculture is a net plus or minus, for example.

      • Or as Zhou Enlai apocryphally said about the French Revolution: “Too soon to tell.” For the answer to agriculture question to be a net minus, you’re going to have to show me a helluva counterfactual.

    • “Science also produced nuclear weapons, biological weapons, global warming, etc”

      And yet both life expectancy and the standard of living continue to rise, well into the nuclear and global warming ages.

      • Sure, jim, but there’s what economists call a peso problem. (And I’ll continue the mutual backscratching with Dale Lehman by agreeing with him.) If your innovation has a 99% chance of a 10% improvement in living standards and a 1% chance of destroying life on earth, its expected value is really, really poor.

        • “If your innovation has a 99% chance of a 10% improvement in living standards and a 1% chance of destroying life on earth, its expected value is really, really poor.”

          Is that a verifiable probability estimate? :) Of course there’s no rational way to calculate the chance that either will destroy life on earth, humanity, or even have a negative impact on civilization. I’d argue that neither has had or ever will have a negative impact on balance. The idea is – to use a non-verifiable probability estimate – 99% sentiment. :)

        • I think actual destruction of life on earth or extinction of humanity is extremely improbable, but a nuclear war could certainly collapse technological civilization*.

          On the optimistic side, I tend to think that geopolitical relationships will change radically (and likely be more friendly) once the Cold War generation ages out of politics. So we may be well more than halfway through the ‘danger zone’.

          Of course, the technology won’t vanish. But it’s the ‘one hour war’* aspect of thousands of ICBMs ready to launch that really makes it an existential threat to civilization — not so much the existence of nuclear weapons *per se*. Isolated use wouldn’t necessarily wreck civilization (as it didn’t in 1945).

          *It is quite possible to destroy cities totally without nuclear weapons; even the Romans did it. But historically one side was obviously losing before total destruction and could surrender.

        • “but a nuclear war could certainly collapse technological civilization*.”

          Yes, I agree this is possible. I rate it as highly unlikely, but not vanishingly small.

        • Yes. I think it’s fairly unlikely too, but given the huge potential risk, well worth mitigating. Which in the short term would mostly mean changing procedures keeping an accidental or irrational launch from happening (the President receiving an incorrect report of inbound ICBMs leading to a response, or something like that).

          Long term, the answer is IMO going to be ending mutually-assured destruction”, either due to social/geopolitical changes that reduce suspicion between the major powers (IMO economics push in this direction, which is why I think we might see major change when the Cold War generation in US, Russia, and China age out of power) or due to technology development that makes strategic (sufficient development of lasers could do this* – with good enough laser weapons, missile defense would be trivial).

          *Nikola Tesla thought development of energy-beam weapons would end war – large fixed installations would ensure the defender always won, vaporizing ships and airplanes for no more than the cost of electricity, and so war would never be attempted. He might turn out to have been right.

          (Although more realistically, I think proxy wars would continue in areas too poor to afford the defenses. But at least it would remove war as an existential threat to civilization.)

        • Confused said
          “On the optimistic side, I tend to think that geopolitical relationships will change radically (and likely be more friendly) once the Cold War generation ages out of politics”


          “It is quite possible to destroy cities totally without nuclear weapons; even the Romans did it. But historically one side was obviously losing before total destruction and could surrender.”

          Both good points (’cause it ain’t simple)

  8. Do you count psychology as a social science? (Not a rhetroical question, I’ve worked at two different psychology departments, once it was listed under social sciences and once it was under natural sciences.) If so, then I would say that the advent of psychotherapy and other modern non-pharmaceutical psychiatric care has certainly gone a very long way to making people healthier, happier, and more comfortable!

    • Joakim:

      I consider psychology to be a mix of natural and social science. Consider, for example, developmental psychology, which is about what happens in the brain (which I’d consider natural science, in the same way that biology is a natural science) but also things like language which are inherently social.

      Good point about psychotherapy. That indeed is a social science which has had a large effect. Not as large an effect as penicillin, airplanes, and TNT, but it’s certainly affected the world.

    • My personal issue with the idea of psychotherapy as an useful technology coming from psychology is that psychotherapy has developed almost orthogonal to scientific psychological theory. Sure, some interventions are heavely based on some form of scientific psychology (e.g., desensitization coming from basic elements from behaviorism) but many psychotherapeutical interventions are loosely based on some scientific theories (some CBTs are not as scientifically based as thay make us think; EMDR is another example that really wants to be scientific, has evidence in its favor, but the overall theory is really sketchy) and some downright deny their own scientificity (like some branches of psychoanalysis). Nonetheless, empirical evidence from outcome research has a hard time sistematically separating approaches in term of net benefit (the so-called ‘Dodo bird verdict’ in the literature), even though the average effect seems to be positive for most approaches and conditions.
      Bottom line: I have to agree with Andrew – scientific psychology hasn’t really given us better psychotherapies; at most, scientific research about the issue has shown we know very little about the subject, certainly much less than psychotherapy theories makes us think.

    • Joakim said, “Do you count psychology as a social science? (Not a rhetorical question, I’ve worked at two different psychology departments, once it was listed under social sciences and once it was under natural sciences.) ”

      When I was an undergraduate at the University of Michigan, there was just one psychology department, but psychology classes were classified individually as either “social science” or “natural science”.

      A the University of Texas at Austin (where I have worked for many years), there are two psychology departments:

      1. “Department of Psychology” in the College of Liberal Arts. Its home page describes its mission as
      “To advance knowledge through exceptional science and education, fostering open dialogue and collaborative engagement among a diverse group of scientists, educators, clinicians, and community partners.”


      2. “Department of Educational Psychology” in the College of Education. Its home page says,
      “Our department is dedicated to advancing knowledge of human behavior as it relates to education. We train our students to work as teachers, counselors, evaluators, and researchers in educational environments. Students develop skills in the areas of human development and behavior, psychology of education, learning theory, and research and evaluation methods. This year, U.S. News and World Report ranked our department 6th among public institutions and 8th overall.”

  9. I view government policy as the final outcome of social science. Public policy is grounded, in theory at least, by research in the social sciences.

    And while they are often knocked (and are definitely impermanent), modern western governments have moved forward a tremendous distance in the past 100 years or so. I think you can attribute a lot of that to advances in the social sciences.

    • Andrew:

      Fair enough. Hobbes, Locke, Marx, and other social science thinkers have affected our systems of government in a big way. Indeed, I guess we could say that political science all the way up to the 1950s has had big effects on policy. “Mutual assured destruction” and other security policies have been informed by game theory, and then there was Nixon with his madman theory of international relations, which could be thought as just another form of the age-old tactic of brinksmanship but had a modern flavor. And, even now, social science work on gerrymandering, opinion polling, etc., has effects on policy. I alluded to some of this in my above post. I still think that it’s only been some very occasional bits of social science that have had the impact of lots of routine work in the natural sciences. But, sure, the political science ideas of Hobbes, Locke, Marx, etc., along with the economics ideas of Smith, Ricardo, Mill, Keynes, Friedman, etc., have had huge effects on the world.

      I guess the difference is that in the natural sciences, modern ideas such as gene therapy, solid state physics, the construction of ceramics, etc., are clear advances on what came before, whereas in social science, I don’t see recent advances as having anything like the impact of what came before.

      • I can agree with that, although I’m not sure the rate of advancements is slowing.

        I think the difference is that it is so much harder to identify a true advancement in the social sciences than it is for the hard sciences, because the error rate is so high (as evidenced by the discussions on this blog) and the evidence tends to be obscure.

        The high error rate is doesn’t mean that there aren’t real advancements being made, though. It’s just harder to see them among all of the mistakes and dead ends.

        Beyond that, social science achievements are more attenuated. It’s a lot easier to say “physicists theorized a new particle, and then discovered it, as evidenced by this extensive data from a detector under the ice in Antarctica” than it is to conceptualize that “we came to a consensus view that homosexuality is not a choice and, as a result, judges determined that there is no rational basis on which to distinguish between opposite-sex and same-sex marriage, and now we as a matter of public policy allow same-sex couples to marry and to enjoy the rights afforded to others.” But the latter is still an advancement grounded in the social sciences.

  10. Andrew –

    > The thing about social science is that it hasn’t produced much.

    I think you might be selling social science a little bit short. Consider as just one example that children were considered as small adults, and there was no need to consider the developmental aspects of childhood, etc.

    I also think that there’s a lot of overlap between physical science and social science that’s relevant – for example, the fact discrimination is no longer widely considered justified due to biological inferiority is an extremely important improvement that is rooted in both physical and social science.

    But yeah, the main value in social science is that as flawed as it is, it’s better (in balance) than the alternatives of superstition, prejudice, and other poor forms of causal reasoning.

  11. Another factor to consider: In the past 300 years or so, how many individuals have devoted their careers to science, engineering, or medical research? How many have done social science? Even today, the conference of the Allied Social Sciences Association (the biggest econ conference) draws in on the order of 10k participants. Even specialized medical conferences are larger by an order of magnitude.

    The causal direction is not clear, of course (is social science smaller because it’s less useful, or is it less useful because it’s smaller). But it might help compare like with like.

    • Sandro:

      Good point. If my argument that the main societal purpose of natural science is to develop new things, while the main societal purpose of social science is to point people away from confused modes of thinking, then it makes sense that lots more sustained effort will go to natural sciences than social sciences. It takes more work to develop new things than to point people away from confused modes of thinking.

    • That said, ironically at universities I have always felt that social sciences have a disproportionately high influence and representation compared to the relative size metric we are discussing above.

      Take committees etc. and there’s more social science representation maybe not in absolute numbers but relative to achievement, funding and the other metrics discussed up thread.

    • “Even specialized medical conferences are larger by an order of magnitude.”

      Yes but medicine isn’t science. The number of people at any natural sciences conference (physics, chemistry, geology, ecology etc) would also be dwarfed by a specialized medical conference. The number of researchers at most university hospitals probably outnumbers all of natural and social sciences put together.

      • Excellent point. In fact doctors HAVE to attend those conferences with some regularity in order to maintain their credentials (only so many CME credits can be gained by reading articles and answering questions). And any specialist has to attend their speciality’s national conference at least every few years to maintain good standing. To a non-trivial extent this is because running a professional society costs money, and requiring attendance at the conference (which is an order of magnitude more expensive than the national conferences I’ve attended for several biological science societies) helps raise that money.

        Only a very small number of attendees at a medical conference are there presenting research or actively involved in something other than just Attending… where as I think a majority of people at the conferences I’ve been to are presenting (or are the advisors of students who are talking), judging, sitting on a panel, organizing a symposium, holding sub committee meetings etc.)

    • Even today, the conference of the Allied Social Sciences Association (the biggest econ conference) draws in on the order of 10k participants. Even specialized medical conferences are larger by an order of magnitude.

      I was once at a medical research conference with ~40k participants. I thought that was huge.

      Years later I was talking to someone else near the same convention center and they said there was a doorknob conference of ~400k people there. I was told that was typical.

  12. Incidentally, I have been drawn to the social sciences later in life precisely because of the inherent difficulty of wrangling with observational data. While it may be true that much (most?) of what people do in the field will be useless, mostly because uselessness still gets published and is infinitely easier to do than trying to understand things, there is a flipside to that coin. For those trying to do science proper there is no more exciting place to be than the social sciences because of the innumerable number of confounding issues to contend with.

  13. To add onto Jonathon’s point above, one can argue that the massive poverty reduction (and all the health, well-being, and livelihood benefits that come with that) in China can be at least partially attributed to economic growth theory. It’s hard to say exactly how much that scholarship influenced policy but even if it had a small influence – which I think most would agree with – then that’s a huge impact, on par with some of the best medical advancements. To a lesser extent, we see impacts from a lot of social welfare programs.

    I also think I disagree with the notion that creating a vaccine, for instance, is more impactful than the economic / market conditions that enabled the vaccine to be created and distributed. I don’t think you can have the one without the other.

    In short, I don’t think social sciences are definitely inferior to the natural sciences – it’s just that the influences are more indirect and less obvious.

    Also, side note but I always thought economics’ inferiority complex was with respect to physics, not biology. At least that was my inferiority complex when I was studying it!

    • “I also think I disagree with the notion that creating a vaccine, for instance, is more impactful than the economic / market conditions that enabled the vaccine to be created and distributed. ”

      Excellent. Exactly.

      • I’d consider Adam Smith to be a social scientist, if that’s the sort of thing you’re referring to. But if we include, say, Thomas Jefferson and Mao Zedong as ‘social scientists’ then I’m afraid the term is too broad to be meaningful.

        • I don’t consider politicians to be social scientists but I do think they’re influenced by social science research, either directly or via their advisors. That’s really the biggest way social science can have any sort of impact since it’s just ideas, so to be useful someone needs to use / implement them.

        • “I’d consider Adam Smith to be a social scientist, if that’s the sort of thing you’re referring to.”

          More to the economic theory and concepts that he and others uncovered than to Smith himself; but no, not to Mao or Xi or any political leadership.

  14. Not to pat economics on the back too much, but we can see pretty clearly the results of a shifting academic consensus regarding government spending during a recession, no? Twelve years ago American enacted what we later found out was a much too small stimulus, and this made America’s recovery more painful than it needed to be (although still better than austerity, which was much more prevalent in Europe). Now we are trying something that is quite literally a trillion dollars bigger. This certainly fits in with your framing of social science discovering what not to do rather than making something concrete like an airplane, but I do think we can look around and see the impact of social science research every year inflation remains stable.

    As another example, over the last 200 years recessions in the US have gotten less frequent and less severe, with a particular drop following social science’s better understanding of monetary policy. I don’t know how to put a price tag on this, but it might be in the trillions?

  15. Andrea – You raised an interesting issue. Two follow up comments:

    1. We recently had a discussion, among Israeli statisticians, on the role of statisticians in the COVID19 response by government and policy makers. It looked odd to us that the key advisors on such issues where physicists and computer scientists, some epidemiologists, less so, statisticians. After all organizing data and developing models, is our profession. Why did physicists and computer scientists enter our playground? In a subsequent meeting, focused on surveys of population response and attitude to COVID19 I was surprised to hear a similar complaint, but from social scientists. They emotionally made the case that social scientists should lead the COVID19 response effort, since the pandemic is, mostly, a social phenomena. I guess they must think that statisticians could not really contribute here…

    2. I am involved in a large project focused on developing communities qualified as “benevolent”. Most participants are social scientists. I did some litt search on methods developed in social science in this area and found very little. What I found was mostly qualitative (story telling ad descriptive) which I would not qualify as methodology.

    So, I tend to say, like you, that I am not sure what social science has provided that can help build up communities.

    What you should consider, however, is ask social scientists what they think of statistic’s contribution. Some I met, do not think it i that much.

  16. I’m not sure there is a single scientific development (either progress or danger, whichever you prefer) that has taken place without simultaneous involvement of the social sciences. Our economic, political, and social systems are intimately involved with the ability of science to progress. So, in that sense, I think comparing the rates of progress is meaningless.

    Now, the post may be about the rate of academic progress within universities (and other research settings). It is an empirical matter, and I am tempted to agree that social scientists appear to work on more trivial problems (on average) than scientists. But I’m not so sure. I think there is plenty of irrelevant research by scientists. I also am leery of falling into the base rate fallacy trap, since the numbers of people that work on the two sets of problems probably differs substantially. Also, I can think of plenty of silly research done by social scientists who make scientific pronouncements (e.g. economists willing to recommend decisions about whether 2nd doses of vaccines are needed) as well as plenty of silly research done by scientists willing to wander into the social sciences (such as the physicist Andrew recently blogged about who felt he had reinvented all of economic theory).

    What I tend to believe is that a lot more silly research is done on social science subject matter, because humans are just a lot sillier than atoms or cells.

  17. First, Joan Robinson: “The purpose of studying economics is not to acquire a set of ready-made answers to economic questions, but to learn how to avoid being deceived by economists.”

    Second, I want to push back against some of the discussion about economics in the comments. For better and worse, major changes in economic organization, especially those that occured before WWII, were largely made in advance of economic analysis of them. Wrt to market-reliant organization in general, at least in England, I recommend P. S. Atiyah’s The Rise and Fall of Freedom of Contract. There was mutual interaction between law, philosophy and developments on the ground, and economics came in somewhat later and actually resisted the inevitable and necessary rebalancing that took place in the latter half of the nineteenth century. The book is over 40 years old now, but I’m not aware of any changes in historical understanding, except perhaps for the role of colonialism and slavery, that would alter its conclusions.

    But since the mid-twentieth century economics has played a more concurrent role. This is obscured somewhat by massive selection pressures from the political system (or more precisely the political economy) that enshrine some theories and proposals and ignore others, but when we are finally able to undo this it turns out there were prescient thinkers a few decades back who can guide us today. These lags and selection effects make it difficult to identify causal processes, but I think they can be teased out.

    • The argument made in other comments is not that economics caused changes in the economic structure, but that economics justified those changes when they were threatened. Some developments in finance are different, as one person noted in another comment academia did push for changes, e.g. index funds.

  18. Your time will come if you wait look long enough. Look at the poor linguists, who spent centuries studying talking, which everyone knew how to do anyway. But now suddenly machine translation is a very big thing.

      • I don’t know to what extent the recent advances in machine translation would be considered driven by the social science of linguistics, vs. just computer science. The best translators today were made by completely ignoring most of the theoretical results of linguistics in favor of just reducing sentences to word tokens and treating translation as a statistical problem of sequence-to-sequence prediction to be solved with big data machine learning techniques.

  19. These discussions easily and often wander into sloppy essentializing about some fundamental difference in the truth value of natural and social science.

    The sociologist Randall Collins has a wonderful (and self-incriminating) paper on, “Why the Social Sciences Won’t Become High-Consensus, Rapid-Discovery Science.”
    (JSTOR link:

    He argues that there is not some core epistemological distinction but instead natural sciences are communities organized around tinkering with measurement tools while social sciences are communities organized around just ideas. All the inventions we attribute to the natural sciences have nothing to do with epistemology and everything to do with the accidents and learning that come out of communities actively tinkering with lineages of machines.

    One recent development Collins did not predict is that computational models in the social sciences seem to have started generating tinker communities. If he is right, that should result in more of the high-consensus discoveries we associate with natural science.

    • I actually just posted a comment (apparently it hit moderation?) saying that possibly the big difference is that observational tools have improved radically in the physical sciences.

      I don’t know whether that is likely to change in the future so much… I think part of the difference is that pre-scientific understanding of the physical world was much more thoroughly wrong than understanding of human behavior, so there was more ground to make up.

    • “High consensus” is a good point too. If we point at social changes that are to some degree driven by social science, many of them might be seen as advancements by many people, but as bad things by others (look at how politicized economics can be, which is arguably the most “hard numbers” of the social sciences). The fact that antibiotics or airplanes or computers work is a lot less arguable.

  20. Excellent post. As a new social scientist, I have no criticisms, but I want to offer my perspective, which may not be too far off from what you talk about here.

    Social science produces cultural technologies. Perspectives, ways of thinking, anecdotes, and metaphors to make sense of the social world and guide behavior. Intersectionality, Myers-Briggs, 6-degrees of separation, the marshmallow test, the invisible hand of the market, Milgram’s shock experiment, and so many more find their way into everyday communication and motivate real-world policies and actions (regardless of whether their effects are good or bad). Even at a personal level, ideas like the 10,000 hour rule (as imperfect as it is) contribute to how people live their lives.

    So if the goal of a natural scientist is to design a new drug or develop a novel material, then the social scientist aims to inject a new idea into the cultural conciousness. As you say, someone will come up with these ideas, and they should probably come from social scientists than the alternatives.

  21. I think one issue is that the tools of observation have not improved as they have in the physical sciences – and doubly so if we separate out the more biological/neurological aspects of psychology as closer to physical science.

    Aristotelian ‘physics’ was actually a rather good model for what was observable pre-telescope. With the instruments of the day, there was no clear reason to prefer a heliocentric model of the Solar System to a geocentric one (in fact, even post-Copernicus there was serious confusion due to the failure to observe stellar parallax*). On Earth, with air resistance, heavy things *do* fall faster, and no one back then had ever seen anything falling in vacuum.

    But throughout history people have been able to observe social interactions, and think about them, while it wasn’t necessarily formalized as a social science. (Although the Greek philosophers did formalize politics, and there was plenty of religious writing on the topic of human behavior, etc.)

    We have statistics now and a tradition of scientific thought, and that helps, but the observations are not fundamentally different in the way physical ones are.

    • “But throughout history people have been able to observe social interactions, and think about them, while it wasn’t necessarily formalized as a social science.”

      Baddabing. Humans have been social scientists for some hundreds-odd millennia, but natural scientists for only three millennia. The social science revolution is already ten thousand of behind us. That’s why it’s hard to, for example, improve on teaching methods – they’ve been being improved and refined for as long as humanity has existed.

  22. There could a bias: researchers may tend to dismiss/undervalue/trivialize the type of research that they have been familiar with. A biologist who by chance does not invent a new vaccine in their career might find political sciences fascinating as well.

  23. Am I correct that this is your conclusion?

    Doing bad social science is easier (at least moneywise) than doing bad natural science so that there is a lot of it out there.
    And people (in and outside of the sciences) can easily be made to believe bad science.
    Further, there are some groups that (e.g., CDC in US) that try to guide people towards the useful natural science, but in the social sciences often groups don’t direct people to the more useful stuff, but often to useless (or worse) stuff. This suggests that as well as doing better social science we should be highlighting useful (or good) and criticizing useless (or bad) research.

  24. It may be reasonable to expect that the partial unification of social science with technology, both as a subject matter (social networks) and as a tool (big data) will accelerate its development. Then again, those advances may be subsumed by computer science, information theory, etc.

    Along those same lines, each time an area of social science becomes more rigorous and borrows tools from other fields, it tends to get gobbled up by those other fields. Drug treatments for mental health problems were originally studied by social scientists. The same for the study of human perception, neurology, and even computation. Andrew’s responses to some comments demonstrate that this gobbling up is happening in real time: developmental psych is only partly social science because it involves the brain? Doesn’t every social science, inherently? I suspect Andrew means that medical tools like EKGs and CAT scans are used to confirm and advance dev psych theories, whereas heuristics are studied using tools that do not come from medicine. For now–in the end, all social science is built on a foundation of biology.

    Put more succinctly, we keep redrawing the line around what counts as social or soft science, based on the tools that are used rather than the subject matter, so that any program of research with the potential for real impact is almost by definition not social science.

    • I think this is a pretty strong (albeit incomplete) argument! Sometimes the question of progress gets asked about philosophy, and one of the arguments is that when philosophy progresses, it soon becomes another field, and so no credit is given to it. It’s not immediately clear to me that this is the case for the social sciences, but if there’s sufficient evidence for it, I’d be willing to buy it.

  25. For the pecking order of sciences, see
    Where we can model social science problems and study them rigorously, I think the result has often been less than earth-shaking, because the results are upper bounds which instincts and institutions evolved over generations are not too far from, rather than a breakthrough new way of working: Arrow’s Theorem revolutionise democracy partly because it said that no voting system can achieve everything we might want of it.

    • >>which instincts and institutions evolved over generations are not too far from

      Yeah, exactly. I think pre-scientific social understanding was a lot farther ahead than pre-scientific understanding of the physical universe, since we don’t have fundamentally better observational tools than pre-science thinkers did. We have a culture of science and statistical tools, but the observations themselves aren’t really different. There’s no social science equivalent of the telescope or microscope or DNA analysis or particle accelerators…

  26. You wrote:
    “Chemistry has produced amazing new materials.”

    No. It’s Materials Science that has produced all those kewl materials*.

    Chemistry produces things like LSD**…

    *: Really. Chemists don’t get materials. It’s a different field, different ways of thinking. They may think they do, but they don’t.
    **: There used to be a T-shirt with “Better Living through Chemistry” on the front and an LSD molecule on the back. Of course, LSD is seriously kewl in that it’s the most potent drug ever discovered by at least 2 orders of magnitude. But if you want a bridge, you don’t want a chemist.

  27. > The baseball analyst Bill James once said that the alternative to good statistics is not no statistics, it’s bad statistics. Similarly, the alternative to good social science is not no social science, it’s bad social science.

    This is a great quote (both the Bill James part, and applying it to “why study social science). I’ve seen you mention the Bill James quote a few times, but I can’t find a source for it, only other blog posts that mention it. Do you happen to remember where you saw the original?

  28. I would say that standardized testing and psychometrics is a useful product of the social sciences. It’s not perfect, but it makes some helpful and useful predictions. For example, it’s useful to know that unintelligent people are likely to consistently make certain kinds of mistakes and that they aren’t just uneducated or somehow morally inferior. To my knowledge, most education systems and countries use some kind of standardized testing, and I think they benefit from using it.

    One recurring problem with the social sciences is that a lot of the research is about fighting for a team rather than a search for objective truth. For example, any research on whether the Chinese government or the US government is “better” is bound to get entangled with which government the researchers want to win. So a lot of research will come out saying that the US gov is better and a lot will say the Chinese gov is better, and each piece of research might help one side to win the conflict, but the net effect of doing larger volumes of research basically cancels out.

    You have these conflicts with: libertarians vs. interventionists (economics), conservatives vs. liberals (sociology), centralists vs. democrats (political science).

  29. Social Science has been and can be insightful. The problem I have is understanding the jargon and quantification. The writing is so obtuse sometimes. My understanding from a couple of literary critics is that novelists and screenwriters are able to capture more accurate insights. Serge Lang and Robert Nozik also stated that a literary sensibility is most welcome. Perhaps why fictio and documentaries are so consulted.

  30. Having worked in and studied physics, math, computer science and cognitive science, worked in research-oriented psych departments, and now at a major tech company. I find this whole perspective and argument facile–enough so that I cannot see how to read it as sincere. First, it’s not clear (really not) how to distinguish social sciences from natural sciences. Which is computer science? It’s clearly part of the ‘science of the artificial’, the study of our constructs–and so social.

    This matters, because at my current job I see *every day* the huge, huge value created by methods designed by people who saw/see themselves as social and cognitive scientists, who you might decide to classify as mathematicians or computer scientists, and whose algorithms are creating vast wealth, are a huge part of the economy, and for better or worse, are having a huge practical impact on people’s lives. Latent Semantic Analysis? LDA and Topics Models? Deep networks? Much of the foundations of AI? These things are having a huge, practical, everyday impact, and come directly from social science. Heck–have you heard of this little think called STAN? Making statistics better in the ‘hard’ sciences all the time, and also creating demonstrable value for at least a few companies I know. Built with mostly social applications in mind. Never mind the huge impact that I/O psychology has had in making organizations more effective, which I think is slightly hard to trace but no less impactful. Social science products are an entrenched and hugely successful part of our economy, because they actually make lots of things hugely better.

    I get that social scientists are supposed to believe that the natural sciences are better. But…It just doesn’t seem to hold up at all. Hard sciences produce lots of valuable stuff. Social sciences produce lots of valuable stuff.

    • David:

      My post is indeed sincere! The comments are helpful too, in allowing us to explore these ideas further. You make a good point about the impact, for good and bad, of the information sciences. Search engines, self-driving cars, online shopping, drones, machine translation, . . . these really are changing the world. And we could go back to the 1930s and 1940s, when modern statistics allowed the U.S. and other governments to do efficient war production, and of course information science was used to crack the Enigma code.

      I’m not sure how to classify information sciences (including statistics, computer science, and some aspects of linguistics). You argue that they are social sciences, and I see what you’re saying . . . but as someone who’s both a statistician and a political science, I see things differently. When I’m writing algorithms for Stan, I think of this as statistics, or mathematics, or statistical computing, not as social science. I think of social science as being the study of society in some way, such as political science, economics, some aspects of psychology, etc.

      • Thanks for clarifying! I can see some of why this is complicated for you. I guess having come out of cognitive science, there’s a more direct connection for me between the social and information sciences: for instance, for me “Bayesianism” is both a family of perspectives on statistics, *and* a family of theories about how the mind makes inferences. Ditto for the Church-Turing thesis, etc. Landauer was always very clear about the connection between LSA and philosophical problems of interpersonal meaning–which sure seems social. Self-driving cars depend on a study of driving patterns, laws, and skills, all of which are part of society…it’s not *just* information-processing. It’s harder for me to cut any clean lines here.

        I guess what I suspect is that when we partition things into “Is it social science, or is it statistics or math math?” we are sneaking in a lot of prejudice, in a way that really makes the rest of the conversation moot. When I was working in solar physics, for sure I had days which were more focused on methodology, days that were focused on understanding solar structure, and days that were very math-heavy. I didn’t decide that only some of those were “really math”, while the others were “just physics”. If you’re spending a week doing math, so that you can solve a social science problem, seems to me that that’s part of the social science, and the social science deserves ‘credit’ for it, just as physics deserves credit for the development of many mathematical techniques that it motivated.

        Your Stan work, of course, may well be very different from that–but lots of work in information sciences seems to fit the paradigm of (a) motivated by problems that are squarely social, (b) applied to problems that are clearly social, and (c) characterized as ‘math’ or ‘statistics’ only because those fields are higher prestige than social science.

        • David:

          Hmmm . . . I’m not sure.

          Take Bayesian Data Analysis, or for that matter Bayesian data analysis, much of which has been motivated by social science problems. My colleagues and I came up with Mister P and developed Stan in order to better understand public opinion and voting, among other things. Or, to take something much bigger, the people at Google developed their search algorithm for the social goal of being able to better categorize and access information. And Turing etc. developed a lot of ideas in information theory cos they wanted to win the war.

          But, to say that something was motivated by social problems or even directly motivated by social science (as has been the case with many methods of statistical modeling and survey research) is not the same as saying that the method it itself social science.

          Here’s an example. Jet planes and the atomic bomb were motivated by social concerns of wanting to win a war. But these are fundamentally developments in physics (and I guess some chemistry with the planes because they had to develop the right fuel mix). Similarly, I feel like google search, deep learning, Bayesian statistics, etc., are part of the information sciences, not the physical or social sciences.

          In any case, this discussion was helpful, in that I wasn’t even thinking of the information sciences as a category when writing the above post.

        • I agree with everything you’re saying here, completely….and also believe you’re omitting key pieces of the story that reveal the deep social motivations for a lot of this work. For instance, Google Search wasn’t just *motivated* by social goals (which is too low a bar), the Page Rank algorithm was trying to model the problem of how *humans* attribute importance based on citation structures, in lots of social situations but academic journal citation was the most commonly discussed. It was directly inspired by, and in dialogue with work by, ‘pure’ social/cognitive scientists like Steve Sloman and Brad Love.

          It seems to me that this kind of deep entwining of thinking about and studying interpersonal social structures, and then trying to automate some version of them in a scalable way, is bread and butter work for a big portion of the information sciences. I don’t know the exact proportion…but plenty. It seems ad hoc at best to say “well, when you’re trying to mathematically model the interpersonal transfer of information, as it happens in societies, that’s really math, not social science”. I agree, though…these processes don’t seem to be that similar to what happens, in, say political science, anthropology, and other disciplines–so I do see why you’d want to pull them out as a relatively unusual social science.

        • Super late to this discussion, but thanks for spending all this time expounding on this point — this post rang false to me as well for similar reasons. It is very hard not to see cognitive science/psychology as incredibly intertwined with computer science research when findings from one field are implemented in the other so often, and your summary/articulation of it is great. It feels like one pretty decent rule of thumb for writing a computer science paper is “see how the brain seems to do something, and then see if is useful elsewhere”, and one decent rule of thumb for writing a cognitive science paper is “find something useful in computer science, and then see if the brain seems to do that too”. Obviously not every paper, but that seems like not a totally unfair description of a decent chunk of really productive/useful work in both fields.

        • Anon:

          The information sciences are important, sure. But I don’t think of “see how the brain seems to do something, and then see if is useful elsewhere” as necessarily being social science. The division is arbitrary, but for example two important areas of cognitive science are visual processing and reasoning (“judgment and decision making”). I consider these to be natural sciences (some combination of biology and mathematics), not social sciences. I do consider the study of language to be a social science. I see an overlap between cognitive and social science, but I don’t see cognitive as a subset of social science. I agree, though, that lots of useful developments have come from the study of language, and that was not captured in my original post.

    • “designed by people who saw/see themselves as social and cognitive scientists”

      I’m not familiar with what “cognitive science” is used in the tech industry to make $billions. Could you give an example?

      • Programming language and compiler design arise from formal languages, a field of study created by linguists. Of particular importance are Chomsky’s hierarchy of recursive grammars.

        • I believe compilers were running before Chomsky started publishing.

          PS. I know an amusing story about LR(k) grammars and Ohm’s Law but the most of the audience here probably would not enjoy it.

        • @Bob76: What makes you think that? Chomsky published on context-free grammars in 1956, and Backus-Naur form was directly based on them. BNF, of course, was part of the foundation of early dialects of ALGOL, Fortran, and other imperative languages. So…it’s a race, and there were lots of precursors to both compilers and Chomsky’s work coming out of philosophy…but no question that Chomsky’s generative grammars greatly influenced the development of compilers. And they were a study of how humans exchange information using language across societies. I’d’ve thought this all was pretty clearly part of social science, though in this conversation it appears that scientifically studying societies is not enough to get one labeled as a social scientist….

        • > I believe compilers were running before Chomsky started publishing.

          The first compiled language was Autocode, released in September of 1952, while Chomsky finished his Masters thesis in 1951. Though this is all irrelevant to whether or not his linguistic work influenced the trajectory of compiler design, which it definitely did. Hilariously, Donald Knuth took a copy of Chomsky’s “syntactic structures” on his honeymoon.

          > PS. I know an amusing story about LR(k) grammars and Ohm’s Law but the most of the audience here probably would not enjoy it.

          Try me.

        • Somebody said, “Try me.” which I read as a challenge to supply an amusing story about LR(K) grammars and Ohm’s Law.

          Here it is. I’m walking on Main Street in Cambridge, Mass.–near the intersection with Ames Street—with an MIT professor (course VI). He makes a comment along the lines of “Our field has changed. I was on an examination committee today for a doctoral candidate whose PhD topic is something called LR(k) grammars—whatever they are. Being unable to ask any relevant question about her field, I asked her ‘What is Ohm’s Law?’ She replied ‘I don’t know.'”

          I think you have to have a specific background to find this amusing.

          I agree that formal language theory had very major influence on the design of both computer languages and compilers. Compilers also have important elements that are outside formal language theory. I actually used to know something about this topic. But most everything I knew is either forgotten or obsolete.


      • @jim: Deep Learning (which was championed by, among others, Geoff Hinton and Peter Dayan, and is heavily based on techniques like backprop, developed by McClelland), Latent Dirichlet Analysis, Latent Semantic Analysis were examples I used in my first post, but yeah, compilers are also a good example. Any number of functional techniques. Any search algorithm that was influenced by, say, Simon and Newell’s work (that is, all of them). Heck, PCA was developed by Hoteling, who as I understand it was an economist, and certainly factor analysis was developed by psychologists, but is deeply connected to space reduction techniques. As far as I can tell, all this stuff gets used every day, by basically every major tech company.

        • Interesting discussion but I’m not convinced.

          For my money there is a real distinction between physical and social sciences. The study of a human organ – the brain, for example – is natural science in my world, while the study of how that organ and it’s organism interact with other members of its species is social science. The big problem with putting the study of the way the brain works into “social science” is the fact that there are hundreds of similar organisms (mammals) with highly similar brains which all work in a similar way to a greater or lesser degree which all fall into the category of biology.

          I’ve been listening to a book about the evolution of astronomy and physics (cant recall name or author, don’t want to run to car). The author points out again and again that physics is what it is because that’s actually the way the world works. Physics *is not* a social construct, in which our understanding of it is in any way cultural.

          IMO the same is true for language, the brain, networks and information. These things are the way they are because that’s how they occur in the natural world. Information is no different to a dog or a beetle than it is to a human. Each has different ways of perceiving information, but the information *means* the same thing. IOW, information, the networks that bear it, the languages that communicate it, and the brains that understand it are natural phenomena.

          I’m not sure how to wrap that up but those are my thoughts for the time being.

        • Jim:

          I agree, but certain topics are on the boundary. For example, linguistics is partly an information science because it involves informational constraints; but it’s also a natural science because it involves constraint of how many sounds our auditory systems and brains can tell apart, along with how language is arranged in our brain, and the physical constraints of what sorts of sounds we can make with our mouths; and it’s also a social science because language has evolved to suit our social needs. Similarly, the study of animal behavior is both a biological and a social science.

          Then there’s the role of information sciences. Traditionally we think of mathematics and computing as tools to be applied to problems in the natural sciences, social sciences, and engineering. But for something like google search, that seems to miss the point. Yes, search is an engineering problem involving social data, and self-driving cars are an engineering problem involving physical and social constraints, but as discussed elsewhere in this thread, it does not quite make sense to me to consider self-driving cars, say, as a triumph of the physical sciences in the same way that I would describe conventional cars.

        • jim –

          > Information is no different to a dog or a beetle than it is to a human. Each has different ways of perceiving information, but the information *means* the same thing. IOW, information, the networks that bear it, the languages that communicate it, and the brains that understand it are natural phenomena.
          You might find this intersting:


          Most of us believe we perceive reality as it generally is, with perhaps some slight distortions. But mathematical models based on natural selection suggest that our strongest intuitions may be false. On this episode of Conversations with Tom, Donald Hoffman and Tom Bilyeu discuss the fascinating possibilities that stem from the theory that evolution gave us the equivalent of a desktop interface that hides the real truth from us so that we can take effective action. They discuss the nature of consciousness, free will, space-time, causality and the self.

        • I am not sure the two ideas are really in conflict.

          Yes, the brain ‘filters’ raw sense data (a horrible oversimplification). But I don’t think that means that the physical data are different, just that perception of them is imperfect.

          Objective reality is what it is, and if our perception of it is imperfect, we can improve greatly on that (with better measurement tools, and procedural tools to deal with bias).

          (And I think people tend to really overstate the significance of that filtering! “the nature of consciousness, free will, space-time, causality and the self” is pretty sweeping…)

  31. First, let us recall the definition of science: “the intellectual and practical activity encompassing the systematic study of the structure and behavior of the physical and natural world through observation and experiment.” Basic is “systematic study”. And of course humans and their behaviors are part of the natural world.

    I have always thought that in fact the social sciences are the harder disciplines because the opportunities for experimentation are more restricted. Now of course “harder” has usually been interpreted as more mathematical, but that has changed, at least for economics.

    As for the impacts on how economics has affected our lives, I mention the influence of Alfred Cecil Pigou and his work The Economics of Welfare, 1920 Central to our understanding of the negative impacts of “free market” activities is his development of the concept of market externalities which underlies environmental control policies. This is the foundation of the proposed carbon tax and also is part of the rationale for cigarette taxation.

    His work generated an enormous subsequent literature exploring the implications of social cost and motivating the regulation of markets.

    Not to mention the development of macroeconomic policy tools starting with John Maynard Keynes. It has taken so long to refine these precisely because of the inability to experiment. Thus we have to wait for “natural experiments” such as the current pandemic.

    • Haynes:

      Yes, several commenters have pointed out the huge influence of economists from Smith to Marx and beyond, and others have pointed out the impact of political scientists such as Hobbes etc.

      So let me rephrase it. Some fundamental ideas of social science (the invisible hand, comparative advantage, Marx’s theory of capitalism, the idea that a governmental budget is not the same as a family budget, the theory of externalities, monetarism, game theory, and a few other easy-to-say but subtle ideas) have been very important in the world. But it’s not clear to me that cutting-edge social science is so impactful.

      To put it another way: in biomedicine we have core theories such as evolution, genetics, energy flow within the body, etc., and engineering-type work such as developing vaccines, intermediate work that enables the engineering, and also science to understand the natural world (discoveries of new species, etc.). In physics we have the laws of thermodynamics and electromagnetism, quantum mechanics, etc., and engineering such as more efficient batteries and solar cells, intermediate work such as the theory of solid state physics, and research on climate change and other things that are happening in the world (sorry to burst your bubble on that one, Freakonomics team!). In social science, I’m convinced by the commenters that certain core theories have been very impactful, but the engineering-type work seems much less important than the comparable engineering being done in the natural sciences. As I said in my post, I do think that work such as Red State Blue State is important, but more as an antidote to David-Brooks-style bad social science than making a positive contribution on its own. It’s necessary work, but it doesn’t compare to covid vaccines, more efficient batteries, etc.

  32. Two different perspectives, the first has been alluded to above.

    1) How much has business administration, economics, organizational psych etc. contributed to enhance individual and organizational performance? Let’s just say 10% (which probably is far too low). So, in a cumulative perspective, over decades, this means that social science is the most important science of them all? – although tbf, stupid social science has also created substantial obstacles.

    2) Social science is different. Clever people can generate relevant insights into human and organizational behavior without having to do lengthy, complicated and complex scientific studies. For example, challenges like groupthink and power dynamcis clever people have solved before the terms were even invented. In contrast, clever people don’t just come up with insight into quantum mechanics, DNA etc.

  33. This is a bit of an aside, but does anyone get the sense the grading system for social sciences at the college level gives way to bad social science? I just wrote a course evaluation the other day which basically asked the question: why if I’m in the 90th percentile (as far as scores on standardized testing) for math do I get consistently mediocre marks, but in the social science courses I take, which I’m in a lower percentile (as far as reading/writing scores on standardized test) I consistently get higher marks?

    I realize I’m writing this anecdotally, so, perhaps, my experience is not representative. It does seem, though, there are some obvious datasets for evaluation here, but I don’t know where to find them; if anyone does, please let me know.



    • I’ll give you an anecdotal observation from India:

      Departments compete to get more students to protect their case for funding and existance. This existential crisis is greater for the social science departments. And hence greater grade dilution.

      It’s the race to be wanting to be perceived as an easy major.

      • Rahul:

        I’m not sure. It’s my impression that the social sciences have been the most popular major for a long time. A few decades ago the most popular major was psychology; now it’s economics/business. I’d guess that math departments are in as much of a crunch as social science.

        A few years ago we discussed “why weren’t the instructors all giving all A’s already?” As a teacher, I can assure you that it’s less effort to give high grades than to give low grades. One thing I hadn’t thought about, that comes up implicitly in your comment, is that some majors are overwhelmed with students, and they give low grades as a way of driving students away, or as a way to screen students if only some fraction of them will be allowed to continue in the major. I don’t think this is the whole story, but I guess it’s part of it.

        • Hey Andrew,

          I majored in math at the University of British Columbia (grad 2014). When I was admitted to the major, it was not very competitive if competitive at all. From what I recall, I just had to take calculus 1 – 3, linear algebra, a introduction to computer science course and a proofs course. Assuming I passed those, I was in. If there was a bar for admission beyond the course requirements, I assume it was low: I barely passed my proofs course and I got Cs and Bs in the other courses. Here are the admission standards now: To your point, you’d expect the math department to have an upward grade inflation too, but this doesn’t seem to be the case, at least in my experience.

          The solution proposed by Val Johnson makes a lot of sense to me. The only aspect that doesn’t seem clear to me is controlling for year to year variation. Intra-year variation control seems obvious since you can compare a students answers from the same test to one another. Inter-year variation, however, seems tough to control since presumably you’d have to write different tests so that students would not be able to short-cut the learning process. Further, I would assume some years the student population is more competent than others, especially given how universities are prone to swings in age demographics, so normalizing the marks to an average grade of 75 every year would dilute marks as an accurate representation of student proficiency. Maybe Val Johnson assumes some level of average mark variation each year though, so my point is moot.

          To put my cards on the table, and get back to the original point, I think there is connective tissue between the ease at which you can get a high mark in social science course and the level of bad social science being promulgated. The low bar for high marks in social science, as opposed to natural sciences, is reflective of the low bar for bad social science to be published. Obviously, I’m making quite a few assumptions here, but my basic hypothesis is this: increase the rigor for grading and admission to the social sciences and the average quality of research will improve. Maybe I’m wrong about this – I realize my evidence wholly anecdotal – but I want to be clear about what I’m proposing, so as to not make the discourse too tangential rather than just putting out vaguely worded questions which are left up for interpretation.

      • Hey Rahul,

        It has not been my experience that social science departments are hurting for students, quite the opposite. When I graduated undergraduate a little over five years ago, the math majors walked with the psychology and sociology majors. I think the master of ceremony did the math majors in one breath.

        More generally, it looks like the most common majors tend to be humanities/business: “Of the 1,956,000 bachelor’s degrees conferred in 2016–17, the greatest numbers of degrees were conferred in the fields of business (381,000), health professions and related programs (238,000), social sciences and history (159,000), psychology (117,000), biological and biomedical sciences (117,000), engineering (116,000), communication, journalism, and related programs (94,000), and visual and performing arts (91,000).”

        That being said, I’m totally open to the idea the Indian system is both structurally and culturally different; only speaking for the American/Canadian system here.

    • Grayson –

      I wonder if you feel the content you were tested on really has “validity” as a measure
      (IOW actually measures what it is intended to measure) with respect to the content, and more importantly the learning context, for your social science courses?

      Also, I would imagine you might bring a different attitude towards the coursework compared to mastering skills in a more abstracted format as in testing – and that difference might help to explain the contrast in how your output is evaluated?

      • Hey Joahua,

        “Valid” seems a little tough to tie down: different people have different thresholds for what’s valid. In my opinion though, the grading system for math/stem tended to be more valid as in the course of evaluating results it would be pretty clear what the answer is/is not, whereas the grading system for humanities courses seemed a little more hand wavy. This is pretty cliché observation, so take it for what its worth.


        …. I think I’m trying to start to think in circles on the topic (argh!?). Maybe the best way to solve this problem is for me to ask your criteria for “valid”, and I can provide you further information for you to evaluate. So, what’s “valid”? Examples will be helpful for my interpretation.

        • I think your perception is correct, but it is an unfortunate reality. Math/Stem is often taught that way, with clear right and wrong answers. I believe that leads easily to thinking along the lines of NHST, p-value cutoffs, and dichotomous decision making. It need not be that way, but it is an easy path to follow. Social sciences tend to be more open-ended, particularly if they use less mathematics (e.g., compare econ and sociology). Again it need not be that way. But the reality is that grading is “easier” in social sciences (I don’t have an authoritative data source, but I’ve seen papers confirming this and I’ve looked at data sets myself that show a clear difference in grades across disciplines). And, I believe it is generally true that there is a higher proportion of shoddy work in the social sciences.

          I view all of these realities as rather unfortunate. As Andrew often says, “statistics is hard.” I personally think that much of that “hardness” is related to the fact that nothing has clearly right and wrong answers – at least no questions worth asking. But once you accept that all questions must be interpreted, and that all answers involve caveats and are subject to improvement, then you are left with subject judgements about quality. This applies to the quality of student work as well as the quality of published research. When the definition of “valid” becomes subjective, then it is hard to decide how to grade students, and hard to decide how to rank researchers. None of us like (at least I don’t) relying on pedigrees, citation counts, publication counts, TED talks, or other “objective” measures, but we don’t like the alternatives either. And, when we keep measuring students with such objective measures, then how can we refuse to apply them to ourselves?

        • Hey Grayson –

          Joahua gave me permission to respond in his stead…

          > Maybe the best way to solve this problem is for me to ask your criteria for “valid”, and I can provide you further information for you to evaluate. So, what’s “valid”? Examples will be helpful for my interpretation.

          It’s hard to know how to respond to that without knowing more about the subjects of the classes, but what I was going for is the general issue with “validity” in standardized testing. In other words, maybe the standardized testing you took did a good job of testing how well you do on a standardized test on that subject, but didn’t do a very good job of measuring the skills and ways of thinking that would lead to you doing well in the classroom.

          It would be interesting to know how scores in math testing line up on grades in math-related classes as compared to how well scores in social science testing line up with scores in social science classes – but if there testing in math-related subjects line up more closely overall, that might just tell you that the math testing was more “valid” – and not necessarily tell you anything much of meaning about the grading in the social science classes per se.

          > “Valid” seems a little tough to tie down: different people have different thresholds for what’s valid. In my opinion though, the grading system for math/stem tended to be more valid as in the course of evaluating results it would be pretty clear what the answer is/is not, whereas the grading system for humanities courses seemed a little more hand wavy. This is pretty cliché observation, so take it for what its worth.

          Well, the larger amount of hand-waving may just be a function of the comparison of the fields – Dale speaks to above. A domain being taught in a more clear right/wrong frame isn’t necessarily an improvement. Maybe math instruction would be better off if the rightness and wrongness of testing were less obvious. I have worked with many students who did quite well when it was easily clear to them how to follow the rules and do what they were told and and memorize algorithms and such, but who did less well when they were expected to figure out for themselves how to address complex problems where there wasn’t a clearly designated path forward. In some ways, that can turn out to be something of a disadvantage in the real world, or even life in the academics once one gets the graduate school level and students are expected to chart their own course more.

        • Hey,

          Thanks for you response. As couple for comments.

          “In other words, maybe the standardized testing you took did a good job of testing how well you do on a standardized test on that subject, but didn’t do a very good job of measuring the skills and ways of thinking that would lead to you doing well in the classroom.”

          Honestly fair point. I’m not inclined to think that though because there seems to be a general uniformity of people getting worse grades in math than social science (i.e. average math grade is lower than average social science grade).

          “Well, the larger amount of hand-waving may just be a function of the comparison of the fields – Dale speaks to above. A domain being taught in a more clear right/wrong frame isn’t necessarily an improvement. Maybe math instruction would be better off if the rightness and wrongness of testing were less obvious.”

          Also a fair point, and I have a tough time getting around the fact that social sciences just are inherently more subjective. Although, conversely, rather than making “math instruction rightness and wrongness less obvious” couldn’t you make rightness and wrongness of other disciplines more obvious? Curious on your thoughts which direction each subject should go an why?

        • > Honestly fair point. I’m not inclined to think that though because there seems to be a general uniformity of people getting worse grades in math than social science (i.e. average math grade is lower than average social science grade).

          I’m not sure what meaning you take from that. (I must admit, I don’t accept grading as a particularly useful measurement of much). But still, why would lower average math grade be mutually exclusive with the idea that standardized testing in math more closely approximates the criteria used to evaluate what goes on in math courses, as compared to standardized testing in humanities versus the criteria used to assess performance in humanities courses?

          > and I have a tough time getting around the fact that social sciences just are inherently more subjective.

          Maybe, but again I’m not sure what conclusion you draw from that if it’s true.

          > Although, conversely, rather than making “math instruction rightness and wrongness less obvious” couldn’t you make rightness and wrongness of other disciplines more obvious?

          I’m not sure how you’d do that. Seems to me that you’d wind up just making the assessment criteria more limited but also more likely in an arbitrary fashion – not necessarily in a useful fashion. I think that you’re accepting a limited, but perhaps sub-opitmally useful grading criteria for math, and just assuming that’s a better outcome because it’s easier to see a binary comparison of right vs. wrong.

        • > (I must admit, I don’t accept grading as a particularly useful measurement of much).

          I think this is the crux of it. You don’t believe grades matter much. To my mind, they do (or at least they should), otherwise I’m having a hard time seeing the point of grading at all. Maybe I’m missing something though.

  34. Andrew,

    We are roughly the same age and I’ve been thinking similar thoughts about reduced form applied microeconomics.

    I view applied economists to be detectives. We collect our data and estimate a historical correlation. In environmental economics, these can be funky correlations. I co-authored a paper documenting using billions of “Weibo” tweets in China that people are angrier their on hotter and more polluted days.

    Zheng S, Wang J, Sun C, Zhang X, Kahn ME. Air pollution lowers Chinese urbanites’ expressed happiness on social media. Nature Human Behaviour. 2019 Mar;3(3):237-43.

    As economists document these facts, we expand the imagination of the public about what will occur again in the future if we don’t adapt. We can adapt through market products or through government intervention or by changing our routines. We control our destiny once we understand “past cause and effect”. In this sense, the applied researcher helps people and firms to avoid hitting the “Titanic” because we can now better see it.

    If people already know all of the correlations that the NBER researchers are uncovering, then our value added is lower!

    I discuss these themes in this YouTube video.

    best, matt

    • So, the economist “discovers” that people are angrier on hotter and more polluted days and then the scientist invents a cooling fabric that prevents the buildup of anger on such days. As a result, we find the scientist’s research to be more impactful than the economists’. In this case, that might be fair enough, but it is possible that the scientist would never have made the invention without the economists’ “discovery” of the linkage. I find the linkages get in the way of meaningfully talking about the difference in usefulness of social vs natural science.

    • Matt:

      I agree with the scientist-as-detective model. Indeed, in political science we often talk about “puzzles.” I think the “detective” thing is something that the natural sciences and social sciences have in common, indeed one reason why it makes sense to consider both of these as sciences. Of course, detective work is not the only thing that’s necessary for something to be a science; after all, actual police detectives are generally not considered to be scientists, nor for that matter do we think of a doctor as a scientist when he or she is doing the detective work of trying to figure out what’s going on with a patient. Conversely, when a detective or a doctor is figuring things out, we can say that he or she is acting like a scientist. To pursue this argument a bit further, this suggests that what makes people “scientists” is not just that they act as scientists. Maybe for something to be a “science” as opposed to just “detective work,” it needs to be involved in uncovering some general truths? . . .

      OK, I know that lots has been written on the question, “What is science?”, and we can’t resolve this in a comment thread. Nonetheless I think this sort of discussion has value. At least, it helps me in thinking these things through.

  35. Hi – I work in a social science focused psychology department in the UK. Much of the social science in the faculty is telling others how the world “should” be but done in a “this is research”, self-righteous, style of narrative. I think it just needs culling and the “researchers” need to join a moral/ethical club or society. Yes Mr Feynman

  36. Fisheries and quota based management has been a strong success to deal with open access resource depletion, which was developed by economists.

  37. Hi – apologies but I don’t really get that post of yours and I think “social sciences” are ultimately more important than natural sciences. Sure, I’m happy about the invention of electronic equipment, synthetic fertilizers, antibiotics and so on. But “social sciences” allow us to investigate how societies can differ and how the environment shapes society. Sure, nowadays there are very few hunters and gatherers but their societies are very different in many aspects or very similar in others.

    If natural sciences can give us technological innovation, then “social sciences” allow us reflection. It doesn’t really make sense to mourn over the lack of quantitative results. I don’t know why people care so much about anecdotes but then fail to appreciate systematic qualitative research.

    And when we talk about “social sciences”, which field are we actually talking about? Historical anthropology? Ethnology? Linguistics, philology or philosophy? Or are we talking about fields that try themselves in quantitative studies like sociology, economics, psychology, politics and so on. Then I’d agree, they are overrated.

  38. This seems to underplay the downstream effects. User interface and experience design shape large parts of our lives, along with applied behavioural economics and public relations.

    What is an iPhone without a user interface? What is Netflix without its experience design? What is uber without its behavioural modelling? Google is built on advertising, Facebook is one big social experiment.

    How can social science be a poor cousin if it creates the most successful corporations on the planet?

  39. Social science isn’t mature yet. I am willing to bet that in the coming century psychology and neuroscience will be unified, and we will have a mathematical understanding. This will influence statistics, engineering, optimal control theory, game theory and other branches of math. This is similar to how physics and the development of calculus influenced each other. Psychology has made a large database of facts in the last 50 years (which were reproduced and now well established). I don’t see any other way to get artificial intelligence, the current paradigms are too primitive.

  40. For better or worse, modern capitalism is driven in part by economics, including studies of finance/business, consumers, organizational psychology, user-interface, etc.

  41. The question is much more complex than you think. I’d say you are asking the wrong questions. Saying that physical sciences are better because it produces more for some people somewhere. As technology increases so does world hunger – if we consider world hunger as an institution between one third to one half of humanity went to bed hungry now. In the stone ages it must have been much smaller we can see as science has increased so has the institution of starvation.

  42. “Making lives better” is relative. Does technology make lives better? For example, imagine if guns were never invented how many people would have been saved. But, because we have guns, we need medicine and a bunch of other scientific advances to develop ways of treating gun wounds and developing new surgical techniques. The CHASE never ends. Human beings are in a constant CHASE. My point is that AWARENESS is what is most important in life, it’s not how long you live, or what you die of. That sense of inner peace and awareness is likely not to come from the natural sciences, it’s most likely to be discovered in the social sciences and humanities, because at minimum, those disciplines often make you think HARD about important issues. E.g., “Why is my neighbor poor, and I’m rich?” “Why is my father angry at me all the time and drinks a 6-pack every night?” “Why do I hate people who aren’t like me?” “Why is someone content on killing a mocking bird?” Social science may not provide definitive answers to those questions, but it often ASKS them, which many would argue is more important than asking questions regarding technological or even medical advance. Asking questions initiates constructive thought. To the point: I’d rather die at 50 with awareness than at 70 without it because science got me to 70 (even as I dodged cell-phone distracted drivers, guns, bombs, pollution, chemical food additives, and dead or suffering deer on the highway as a result of our technological advances, etc.). Whatever gets you to awareness and inner peace is what is generally good to read and study. Those topics are usually found in the humanities and social sciences. I count pure mathematics in those areas as well, because I consider it a humanity at its highest (or deepest) reaches that promotes harmony, just like deep philosophical thinking does as well. The distinction between mathematics and humanities vanishes when you think about it deeply. The divorce rate is 50%. Psychological theory provides narratives for understanding your spouse that could save your marriage and not scar your children forever. WHAT IS MORE IMPORTANT THAN YOUR CHILDREN? You think your spouse is a collection of atoms? Maybe, from one perspective, until our current theory of physics is replaced by another (something you learn in history of science, a humanity). Just because psychological theory may not always provide a perfect SCIENTIFIC narrative, it doesn’t mean the narrative is not useful. The “utility” of a model of explanation is what is most important, not necessarily whether it accords perfectly with the religion of science. Humanity is always “chasing” a “better world” via its advances in science. Humanity is in a constant state of RESTLESSNESS. The humanities have long told us the “better world” need not be “chased” at all for it is right in front of us. If you’re unhappy before you got the COVID-19 vaccine, you’ll be just as unhappy after. Medical, physical, or technological science will never save humanity, though it may give your cancer-ridden child a shot at a GOOD life, but the medical is only providing the OPPORTUNITY for actualization, it isn’t in itself going to actualize the child. Only self-awareness will, which is available to any of us if we are prepared to do the work and ask good questions. Much of the social sciences and humanities try to help us along that path of self-discovery. It’s imperfect. Have they helped humanity? Very difficult to measure because the question is out of context. Just like asking whether nuclear bombs have helped or hindered humanity is a dissertation-long answer. And by the end of it, you still won’t have a definite answer. We don’t NEED the airplane. We’d do fine without it. You know why? Because human beings HABITUATE to new stimuli and ways of doing things. We’d be just as happy without airplanes or a new car, or this or that NOVELTY. If anything, chasing those things is simply chasing a DISTRACTION. We know this because social science tells us about habituation and provides some insight into how we behave. Yes, science has brought us the new big-screen TV, but social science has told us (or “suggested”) we don’t need it because we will habituate to it within a few days and not even appreciate the bigger screen. Wasted money. Chase continues. Social science tells us “STAY STILL DAMMIT AND DON’T LOOK FOR THE ANSWER OUTSIDE”. Take that idea to the extreme, and you’ll see that “progress” in humanity has nothing to do with science at all. It has to do with SPIRIT and SOUL. So the problem with “making lives better” depends entirely on what “better” means. In the end, making one’s life better usually means figuring yourself out and deepening to true, unconditional happiness so you may spread that love energy to everything you come into contact. Where did I get the IDEAS that led me to that state of heavenly deep and true happiness that I can now impart onto everything I come into contact? Social sciences and humanities.

  43. I don’t agree with your title, and you did not even scratch the surface of the importance of the social sciences. It’s not difficult to acknowledge the indispensability of social research in a world that’s rife with social problems.

    • Ada:

      I agree that there are many important social problems, and I agree that society is important. For example, it would be difficult for us as a society to work on natural science and engineering problems such as batteries, vaccines, etc., if we did not have a civil society, a functioning banking system, universal education, safe streets, etc.—and all of these are social problems, not natural science problems.

      So I’m not denying the importance of society, and social problems, and attempts to understand and solve social problems.

      What I’m denying is the direct relevance of social science. As I wrote in my above post, some examples of successful social science are a method of evaluating redistricting plans, better polling, the Big Five, the Implicit Association Test, better auction rules, and cool marketing tricks. All of these are worth studying—indeed, I’ve worked on some of them myself!—but I see their connection to social problems as pretty indirect. As I wrote above, I think the biggest value of social science is to make people less ready to believe naive social science reasoning. And, yes, I get that social scientists work on important problems ranging from effective teaching to prison reform, but it’s my impression that whatever progress happens in this area is more a product of political and economic processes than from social science findings.

      I guess we can come up with a few examples of social science research that really changed the world. I guess that the writings of Smith, Ricardo, etc., were important steps in bringing the world closer to free trade, and the writings of Freud and Marx had a lot of influence too, for better or worse, I guess mostly worse but who am I to say there. Margaret Mead etc. gave us a relativistic perspective that, imperfect as it was, is still in my view an improvement over naive racial determinism. So, yes, there are some examples, and I’m sure that you could come up with several more. Modern polling has changed the world for the better in many ways (I’m with George Gallup on that one), and that builds upon a lot of social science research. And, arguably much of modern biology builds upon the social science research of Galton and others. (But I’m not going to count work such as that of Piaget or Tversky and Kahneman that is really cool and is brilliant science but does not have direct application to the social world.) So, yeah, sure. But compare those to the massive changes in the world arising from natural science research, and it’s just much much less.

      So, yes, society is super important, it’s essential for everything else. Social science, though, that’s another thing entirely.

      Again, I’m not recommending that social science funding be set to zero, or anything like that. I think social science is important! Just mostly in this indirect way that by understanding the social world better we can avoid some big mistakes.

Leave a Reply to Robert Matthews Cancel reply

Your email address will not be published. Required fields are marked *