CHAPTER 10
The Proper Study of Mankind
Know then thyself, presume not God to scan,
The proper study of mankind is Man.
—Alexander Pope, “An Essay on Man”
When Alexander Pope published his “Essay on Man” in 1732 our understanding of the world was very different from what it is today. Written just decades after Isaac Newton’s masterpiece, Principia, had laid out the mathematical principles of planetary motion, Pope’s essay arrived when intellectuals were still wrapping their heads around a concept that must have been staggering at the time—that the laws governing the motion of everyday objects here on Earth were exactly the same laws as those governing the heavenly spheres. In fact, they were still grappling with the idea that physical “laws” of any kind could be written down in terms of mathematical equations, and that these equations could then be used to predict with uncanny precision the future behavior of everything ranging from tomorrow’s high tide to the return of distant comets. It must have been a magical time to be alive when the universe, so long an enigma, seemed suddenly to have been conquered by the mind of a single man. As Pope himself said,
Nature and Nature’s laws lay hid in night:
God said, Let Newton be! and all was light.1
For the next three centuries, the knowledge of mankind would swell inexorably, sweeping before it the mysteries of the world. The results have been impressive. We have theories of the universe that go all the way back to the big bang, and telescopes that peer across galaxies. We have sent space probes out of the solar system and put men on the moon. We have built bombs that can level an entire city, and missiles that can fly through a window. We have measured the earth to great precision and understood its inner workings. We have engineered immense buildings and bridges, and changed the shape of rivers, mountains, even coastlines. We have clocks that can measure time in billionths of a second, and computers that can search through every word ever written in less time than it takes to write just one. In science, it seems, we can make the angels dance on the head of a pin.
It’s pretty obvious that social science has not kept up, but it’s easy to infer the wrong lesson from this observation. I was reminded of this problem recently by a physicist colleague of mine who complained to me that he’d been reading a lot of sociology, and that in his opinion the problem with the discipline was that it hadn’t discovered any laws of human behavior that were anywhere near as general or accurate as those he was accustomed to in physics. Instead, it seemed to him, sociology was just an endless conglomeration of special cases, when someone did something for some reason one time, and someone else did something else for some other reason at another time. As a physicist, he found this lack of lawlike behavior particularly frustrating. After all, it is hard to imagine how any of the remarkable progress in physics could have occurred in the absence of laws like Newton’s that apply generally across time and space. So integral to the success of science have these laws been, in fact, that they have come to be associated with the very idea of science itself. Surely, he felt, the inability of sociologists to come up with anything remotely comparable meant that social science didn’t really deserve to be thought of as science at all.
PHYSICS ENVY
As it turns out, this tendency to judge sociology by the standards of physics is an old one, going all the way back to Auguste Comte, the nineteenth-century philosopher who is often credited as the founding father of sociology. Comte imagined that sociology, which he even called social physics, would take its place alongside mathematics, astronomy, physics, chemistry, and biology as one of the six fundamental sciences describing all of reality. Sociology in Comte’s view would be a “total theory” of all human experience, encompassing all the other sciences and extending them to account for cultures, institutions, economies, politics, everything—exactly the kind of general theory that my physicist friend was looking for. Comte never got around to articulating this theory in any detail, but his philosophy of positivism—the idea that social entities and forces can be described and analyzed in the same way as physical entities and forces—set the stage for all the grand theories that followed.
One of the first such theories was proposed shortly after Comte by the philosopher Herbert Spencer, a contemporary of Darwin’s. Spencer advanced the notion that societies could be understood as organisms, where individual humans could be thought of as cells, institutions played the role of organs, and development was driven by some loose analog of natural selection. It was Spencer, in fact, not Darwin, who coined the phrase “survival of the fittest.” Spencer’s specific ideas were quickly rejected as naïve, but his basic philosophical claim that societies are organized the way they are in order to serve some holistic function persisted alongside Comte’s positivism, and informed the thinking of sociologists such as Émile Durkheim, who is still considered one of the giant figures of the discipline.
The apotheosis of grand theorizing, however, didn’t arrive until the mid-twentieth century in the work of Harvard sociologist Talcott Parsons, who advanced a brand of theory that became known as structural functionalism. According to Parsons, social institutions were made up of networks of interlocking roles, which in turn were played by individuals who were motivated by rational ends. At the same time, however, individual action was constrained by social norms, laws, and other mechanisms of control that were encoded in the institutions of which the individuals were a part.2 By exhaustively classifying all the various functions that different sorts of behavior could satisfy, along with the different social and cultural structures in which they took place, Parsons attempted nothing less than to describe all of society. It was a grand edifice indeed, and Parsons’s name is generally listed among the great social theorists of the ages. But as with Spencer and Comte before him, the ink was scarcely dry on Parsons’s “general theory” before the critics tore it apart: it said little more than that “people do things because they want to,” it was not really a theory at all, but just a “set of concepts and definitions,” and it was so complicated that nobody could understand it.3
Looking back on the wreckage of Parsons’s theory some years later, Robert Merton—the sociologist whose work on the Matthew Effect I discussed in the previous chapter—concluded that social theorists had been too quick to try to emulate the theoretical successes of their physicist colleagues. It wasn’t that Merton didn’t sympathize with the envy that physicists could inspire in others. As he put it, “Many sociologists take the achievements of physics as the standard for self-appraisal. They want to compare biceps with their bigger brothers. They, too, want to count. And when it becomes evident that they neither have the rugged physique nor pack the murderous wallop of their big brothers, some sociologists despair. They begin to ask: is a science of society really possible unless we institute a total system of sociology?” Although sympathetic, however, Merton cautioned that “this perspective ignores the fact that between twentieth century physics and twentieth century sociology stand billions of hours of sustained, disciplined, and cumulative research.” In physics, it was only after Copernicus and Brahe and a host of others had conducted centuries’ worth of painstaking observations that astronomers like Kepler sought out mathematical regularities that could account for the data they had inherited. And only then was a singular genius like Newton in a position to reduce these regularities to bona fide laws. By contrast, the social theorists Merton was describing had gone about it the other way around, proposing whole systems of thought at the outset and only then worrying about what it was that they needed to measure.4 “Perhaps,” Merton lamented, “sociology is not yet ready for its Einstein because it has not yet found its Kepler—to say nothing of its Newton, Laplace, Gibbs, Maxwell, or Planck.”5
Rather than questing after grand theories or universal laws of human behavior, therefore, Merton instead advocated that sociologists should focus on developing “theories of the middle range,” meaning theories that are broad enough to account for
more than isolated phenomena but specific enough to say something concrete and useful. For example, the “theory of relative deprivation” states that people feel distressed by circumstances only inasmuch as their hardship exceeds that of the people around them. Thus if your house burns down in a freak fire, you’re devastated, but if your whole city is wiped out in an earthquake and hundreds of your neighbors die, you feel lucky to be alive. It’s not a completely general theory, claiming only to predict how people respond to adversity, but it also aims to apply to perceptions of adversity quite broadly. Likewise, the “theory of the role set” stresses that each individual plays not only multiple roles—a teacher at school, a father at home, a catcher on the weekend softball team—but also that each of these roles is itself a collection of relationships: between a teacher and his students, between him and his colleagues, and between him and his principal. Again, the theory is somewhat specific—saying nothing about markets or governments, or other important features of the social world—but also somewhat general, applying to people of all kinds.6
Merton’s call for middle-range theories is generally regarded as sensible, but it didn’t quell the ardor for theories of a grander nature. Barely a year after Merton published his critique, in fact, the economist John Harsanyi, who shared the 1994 Nobel memorial Prize in Economics for his work on game theory, proposed that rational choice theory—the theory of human decision making that I discussed in Chapter 2—was ready to provide precisely the kind of general theory that Merton has just concluded was wildly premature. And so another cycle began, with rational choice theorists drawing parallels between their efforts and Newtonian mechanics, while critics increasingly leveled the same complaints against it that the rational choice theorists themselves had made about the previous round of theories like Parsons’s.7 Nor has the growing realization that rational choice theory cannot provide a universal theory of human behavior any more than its predecessors could yet delivered social science from the green-eyed monster of physics envy.8 Quite to the contrary, if the complaint of my physicist colleague is anything to go by, even if sociologists have finally gotten tired of grand theories of everything, there is a whole generation of physicists waiting to step into the breach.9
When you think about the sheer complexity of human behavior, this approach to doing social science seems kind of implausible. As I discussed in Chapter 2, individual behavior is complicated by dozens of psychological biases, many of which operate outside of our conscious awareness and interact in as-yet-unknown ways. And as I discussed in Chapter 3, when individuals interact with one another, their collective behavior may simply not be derivable from their individual attributes and incentives, no matter how much you know about them. Given that the real complexity of the social world—involving not just people and groups, but also a bewildering array of markets, governments, firms, and other institutions that we have created for ourselves—is so much greater than I have even begun to describe here, why on earth would any one person even think that they could write down a set of rules that could explain it all?
My answer is that social theorists are people too, and so they make the same mistake that planners, politicians, marketers, and business strategists make, which is to dramatically underestimate the difficulty of what they are trying to do. And just like planners, politicians, and so on, no matter how many times such grand theories fail, there is always someone new who thinks that it can’t be that difficult—because, after all, “it’s not rocket science.” If much of what sociology has to offer seems like common sense, in other words, it is not just because everything about human behavior seems obvious once you know the answer. Part of the problem is also that social scientists, like everyone else, participate in social life and so feel as if they can understand why people do what they do simply by thinking about it. It is not surprising, therefore, that many social scientific explanations suffer from the same weaknesses—ex post facto assertions of rationality, representative individuals, special people, and correlation substituting for causation—that pervade our commonsense explanations as well.
MEASURING THE UNMEASURABLE
One response to this problem, as Lazarsfeld’s colleague Samuel Stouffer noted more than sixty years ago, is for sociologists to depend less on their common sense, not more, and instead try to cultivate uncommon sense.10 But getting away from commonsense reasoning in sociology is easier said than done. In large part the difficulty is that for most of the history of social science, it simply hasn’t been possible to measure the elements of social phenomena the way we measure the elements of physical and biological phenomena. Social phenomena, as I have already noted, consist of large populations of people interacting with and influencing one another as well as with the organizations and governments they create—none of which is easy to observe directly, let alone put in a lab.11
Recently, however, the world has begun to change in ways that may lift some of these historical limitations on social science. Communication technologies, like e-mail, cell phones, and instant messaging now implicitly trace out social networks among billions of individuals, along with the flow of information among them. Online communities such as Facebook, Twitter, Wikipedia, and World of Warcraft facilitate interactions among people in ways that both promote new kinds of social activity and also record it. Crowdsourcing sites like Amazon’s Mechanical Turk are increasingly being used as “virtual labs” in which researchers can run psychological and behavioral experiments.12 And Web search, online media, and electronic commerce are generating ever-increasing insight into the intentions and actions of people everywhere. The capability to observe the actions and interactions of potentially billions of people presents some serious issues about the rights and privacy of individuals, and so we must proceed with caution.13 Nevertheless, these technologies also exhibit enormous scientific potential, allowing us for the first time in history to observe, in high fidelity, the real-time behavior of large groups, and even societies as a whole.
For example, the Music Lab experiments that I discussed in Chapter 3, which showed the importance of social influence in determining success, involved nearly thirty thousand participants. One could have dreamed up this experiment fifty years ago—back when social psychologists were first pioneering experimental studies of influence and group decision making—but it would have been impossible to conduct it until recently for the simple reason that you can’t fit that many people in a physical lab. Likewise, the “influencers” study on Twitter that I discussed in Chapter 4 was designed to answer a question—about whether or not special individuals cause information to spread—that has been around for decades. Answering it, however, required us to track the diffusion of more than 70 million URLs over the entire Twitter network for a two-month period. Prior to social networking services like Twitter and Facebook, which, remember, are just a few years old, that level of scale and resolution would have been impossible.14
Other experiments that I have described, like the Small World experiment from Chapter 4, were certainly possible in the pre-Internet era, but not on the scale at which they can now be conducted. Milgram’s original experiment, for example, used physical letters and relied on just three hundred individuals attempting to reach a single person in Boston. The e-mail–based experiment that my colleagues and I conducted back in 2002 involved more than sixty thousand people directing messages to one of eighteen targets, who in turn were located in thirteen countries. In the course of being delivered, the message chains passed through more than 160 different countries; thus for all its limitations, the experiment was at least a crude test of the small-world hypothesis on a truly global scale. Likewise, David Reiley and Randall Lewis’s field experiment on ad effectiveness, described in Chapter 8, was similar in design to experiments that had been conducted in the past, but with 1.6 million participants, it was many times larger. The sheer scale of the exercise is impressive simply on the grounds that it can be done at all, but it’s also important scientifically—because it’s possible t
hat the effects, while real, can be small, in which case one needs very large numbers to tease them out of the noise.15
BIRDS OF A FEATHER
Another kind of study that would have been impossible to conduct until recently concerns one of the most widely observed patterns in social life, known in sociology as the homophily principle—the idea that “birds of a feather flock together.” For several decades now, wherever sociologists have looked they have found that friends, spouses, coworkers, and social acquaintances are more similar than strangers with respect to a whole range of attributes—like race, age, gender, income, education—and also attitudes. But where does all this similarity come from? At first, the answer seems obvious: People are likely to form ties with others who are similar because, rightly or wrongly, that’s whom they’d prefer to spend their time with. But what this commonsense explanation overlooks is that people can only choose their friends from among the people they actually meet, which is determined to a large extent by the people they work with, or who belong to the same organizations, or to whom they are introduced by mutual acquaintances. And as sociologists have also shown, many of these social environments tend to be highly homogeneous in terms of race, gender, age, and education. As a result, it is entirely possible that the similarity we see around us has less to do with our own psychological preferences than the restricted opportunities that the world presents to us.16
Everything Is Obvious Page 24