Put a kid in a room with a tempting cookie, and nothing else but that cookie, and see if he chooses to eat that cookie immediately or wait ten minutes, so that he is rewarded with another cookie. If he can’t wait, he prefers immediate gratification. If he waits, he prefers delayed gratification. According to Daniel Kahneman’s Thinking Fast and Slow, those kids who wait for the second cookie turn out to be more intelligent later in life, and can even be seen as more “rational.” Might, however, those kids who exhibit more self-control turn out to be more intelligent simply because a conducive social environment determines both? What contributes to “intelligence” in advanced industrial society, if not the self-control needed to sit through 16+ years of primary, secondary, and tertiary education? Most crucially, why would we say that these kids are more “rational,” given that the preference for delayed gratification varies enormously by culture and time period, and that such “rationality” happens to coincide with that which determines success in our own advanced industrial societies? None of these questions occur to Kahneman.
To be fair to Kahneman (and his discipline of cognitive psychology generally), he does not claim to uncover dynamics outside of the individual mind (and would no doubt demur from claiming an absolute causal connection between delayed gratification and intelligence). Nevertheless, these kinds of questions do point to certain lacuna in the discipline of psychology. Psychology aims to isolate causal phenomena to the individual mind, in good reductionist fashion, without consideration of those surrounding social factors which may have equal or greater causal weight. In doing so, it often blinds itself to the social or cultural contingency of its causal claims, and ignores social context to the extent that it defines as “normal” that which is merely normal in our specific cultures and time periods.
Take sex, for example. In contemporary Western societies, sex is taught in our schools as a healthy and pleasurable activity between consenting adults, the chief danger of which is disease. This is a direct result of psychological research in the 1950s and 1960s that both rationalized and contributed to the opening up of sexual mores. This may in fact be true (and very few of us growing up in such an environment could credibly claim not to believe this), yet it goes against the wisdom of perhaps 99% of the cultures that have existed (and perhaps 99.9% of the “culture-years” we could collect in a dataset). Any legislation of “normality” based on our own particular norms of “health” and “pleasure” rests on the crucial assumptions that these are among the goals that are most desirable in life (which may not be true, for example, of cultures more concerned with honor or sacrifice), that relations between the sexes are meaningfully equal to the extent that “consent” is not a stand-in for power, and that certain other dangers of sex (for example, castigation by a community more concerned with female chastity) pale in comparison with those related to “health.” There is, in other words, great cultural and historical contingency in the ways that contemporary psychologists evaluate “normal sexuality.” Not mindful of this contingency, they readily impose upon us their vision of “normality,” based upon the “authority of science,” in the face of which most of us willingly accept that we may be “abnormal” if they say so, when in fact “science” has become a cloak for cultural bias. None of this is to say that they might not be right, but that any research not taking into consideration such extreme and improbable contingency may rely on such narrow assumptions of human nature that it misunderstands who we are.
The methodological root of the problem is, of course, that psychological research is often based on college students from a single country (most often the U.S.). This cannot account for possible differences based on culture or age. Richard Nisbett, among others, has shown via cross-cultural psychological research that, for example, East Asians are more inclined to be overconfident in their causal generalizations, while Westerners are more inclined to limit their perceptions to focal objects rather than backgrounds. (In fact, there are entire research programs devoted to showing that the Chinese mind does not work exactly the same way that Western minds do.) Another dimension of variation that cannot be even in principle be accounted for experimentally is that of time. We know, for example, that ideas of sex, health, and happiness differed greatly in the West just one century ago. If, through the “authority of science” (the proof of which may be that the U.S. National Science Foundation funnels a great deal of money to experimental psychology), certain ideas of human nature become enshrined as “universal,” we then approach the kind of dangerous universalism that was responsible for the American occupation of Iraq. The analogy is a direct one. Psychologists claim to discover regularities in human nature based on a limited population, and generalize to other populations based on the (often) unproven assumption that all people are alike. Neoconservatives and other proponents of American universalism claim to see in the U.S. experience a democratic model that would work for other countries, based on the unproven (and highly unlikely) assumption that other nationalities are just like Americans. They both ignore the effects of culture, the latter as further mediated through institutions.
In addition to this implicit universalism, psychologists can also be seen as guilty of reductionism when they try to apply individual findings to group interactions over time. Of course, all variants of science including the natural and social sciences engage in some variant of reductionism when they try to explain wholes in terms of isolated parts, whether that is chemical reactions in terms of molecules or political outcomes in terms of social movements. In contrast to the social sciences, however, psychologists may be less aware of the myriad factors that determine any social outcome and less conscious of second-order social effects that are exogenous to their causal models. Take, for example, Kahneman’s recommendation that business meetings should convene only after the participants have written a memo stating their views. This is to avoid the observed, experimental result that meetings are adversely framed by the sequence of who speaks, often led by the most powerful and authoritative. His recommendation may work perfectly fine in an experimental setting with only one iteration, but in a real business situation, it would likely encounter many problems. For example, senior managers would likely resist writing the memos, if they think they lack the time or feel that seniority ought to give them the right to speak instead of write (especially those who speak well but write awfully, a condition all too common in the business world of which Kanheman likely has no inkling). This, in turn, may lead to decreased meeting frequency and overall less effective corporate communications, net of any increase in effectiveness brought on by the memo-meetings. In short, any regularly occurring group interaction (i.e. institution) such as a business, cannot be easily modeled by psychological factors alone, in the absence of their subsequent influence on social factors (i.e. concerns about legitimate hierarchy) and the influence of these first-order factors on second-order social factors (i.e. decreased meeting frequency). While Kahenman’s suggestion does have reductive value (and is a good starting point for thinking about how to make meetings more productive), the audacity with which psychologists claim that their findings really will work in real-life situations should give us pause, when all they have are highly simplified models, whose range of application somehow magically shifts from people to institutions. The world is not as simple as psychologists imagine it to be, and while social scientists also endeavor to simplify the world via causal explanation, they are arguably more aware of their limitations.
None of this is to say that social scientists have always been more humble than psychologists, (witness predictions of Soviet stability prior to 1989 or the havoc wrought by proponents of one of the founders of social science, Karl Marx), but that today’s social scientists, who struggle to explain more than 25% of the variance of any phenomenon via multiple regression (in contrast to the experimental sciences, whose explanatory yield is much higher), know that the social world is inherently complicated, and is only predictable (if at all) via the specification of dozens of causal factors. This recognition of the difficulty of social explanation ought to make academics of all kinds correspondingly humble in making public policy recommendations. The world can only be changed if we know how it works, and if even with the most advanced statistical and data collection methods we can only determine 25% of how it works, then activism ought to be tempered with humility. In other words, the difficulty of social science should make us all Burkean conservatives. Consider, for example, Cass Sunstein’s observation (based on psychological research) that group discussion tends to polarize the participants, making people feel more strongly about their original positions when they argue with differently-minded people. A radical would take such a finding and try to restructure the public sphere so as to discourage polarization and restore the kind of moderation that makes for stable democratic politics. An extreme radical would go so far as to suggest that legislative bodies ought to be restructured as well, perhaps making legislators write down their opinions before they speak, or do most of their debate via some kind of online forum and only convene to ratify their collective decisions. A Burkean conservative, on the other hand, would point to the continuing existence of Anglo-American legislatures for hundreds of years as prima facie evidence that they continue to serve a useful purpose and work for the greater polity in ways we cannot fully explain. On this view, any isolated academic finding like Sunstein’s or Kahneman’s is likely to lead to unintended consequences if divorced from the social fabric and institutional ecosystem (e.g. surrounding the making and legitimation of laws or the running of a business), and should only be taken into consideration as fodder for reform at the margins.