The Bactra Review: Occasional and eclectic book reviews by Cosma Shalizi   55

Explaining Culture

A Naturalistic Approach

by Dan Sperber

Oxford: Blackwell, 1996

How to Catch Insanity from Your Kids (Among Others); or, Histoire naturelle de l'infame

I don't know how many of my learned and refined readers have ever had occasion to pass through Roswell, New Mexico, located as it is in an especially bleak part of what is, on balance, a rather ugly state. While it perhaps does not deserve an actual visit, Roswell is certainly worthy of contemplation: like Lourdes, it has become famous for something which never happened. I refer, of course, to the infamous myth that a flying saucer crashed there in 1947, making the national news, only to be covered up by the US military. A shrine to this myth exists in the form of the grandiloquently named "International UFO Museum and Research Center"; but more diffuse homage is rendered to it in nearly every shop-window, down to wedding outfitters. The inside of the museum displays limitless credulity in its exhibits ("What I have seen refuted three times is true"), with tasteless and incompetent art for decoration, the latter all the more painful for its self-evident sincerity. To the skeptical but casual eye, the museum is a mis-guided monument to a particularly absurd fanaticism, to the strength of human gullibility, to a race of barely-thinking forked radishes' capacity for self-delusion. Having taken it thus in with a glance, that eye dismisses Roswell and its Museum, and moves on (probably on Route 285) to more pleasant prospects.

That eye moves on; but the truly philosophical eye will linger. The child of the Enlightenment, trained to uncover the hidden springs of conduct, penetrates the inner workings of the museum with ease. It is housed in an abandoned movie theater, obtained from the city at nominal cost. The list of founding sponsors of the museum, displayed as a discreet plaque in the lobby, contains no names distinguished among the believers in UFOs, but instead various members of the local business elite (who, as experience teaches, are the real rulers of any American small town), and their banks. This plaque is flanked by equally discreet awards from the local chamber of commerce. In sum: the museum was created as a tourist trap by the local rulers, to their direct benefit. Its existence and character are admirably explained by that blend of precepts which Marx and Engels took from the English and French historians of society and distilled into "historical materialism." This brew now meets with almost universal approval among the learned and sensible, as we have earlier remarked.

However satisfactory this potion may be in explaining deliberate products of human effort, like the Roswell Museum, it does not seem to work so well with other cultural phenomena, like the story of the Roswell UFO crash itself. For how are people to know what, in view of their stations in society, it would be advantageous for them to believe, and believe it sincerely? Have we all little Marxometers, pointing us in the direction of our socially preferred beliefs, of whose existence we are unaware, even when the connection to our social condition is so recondite that its elucidation requires hundreds of abstruse pages under the imprimatur of Verso or the Presses Universitaires de France? Do Gramsci's "organic intellectuals" act as external Marxometers? None of these conjectures is satisfactory. Nor is the obvious alternative: Agent Mulder, in his more (or, perhaps, less) paranoid moments, may suspect the UFO phenomenon in its entirety of being fabricated by the "military-industrial-entertainment-research complex," but this is neither a general explanation for wide-spread beliefs, nor one which a loyal member of the complex aforesaid, such as your humble reviewer, will adopt unless no other alternative presents itself. What is needed is a way of establish consensus, of getting the same belief into the minds of the multitude, without Hidden Persuaders actually implanting it there like so many buttock-dwelling microchips.

Now, it is a fact of observation that belief is catching. Here is a reasonably old recognition of this: in the 1764, the public prosecutor of Geneva ordered Voltaire's Philosophical Dictionary to be banned, condemning it as "contagious poison." I do not pretend this rather unimaginative civil servant had obtained a new insight into the nature of belief, or even had come upon a novel turn of phrase; I merely mean to show that this insight is an old one, and being no historian, nor even diligent, I merely pulled old books off my shelf until I found something to the purpose. The heart of the matter is that people have long realized that ideas can spread as though they were agents of disease and infection; in a word, like viruses, using that term in its original sense.

There is today a great and rapidly growing body of knowledge concerning the agents of physical infection, and the temptation to exapt that knowledge to the sphere of culture, to declare beliefs, traditions, practices and the rest of their kith and kin "viruses of the mind" is more than a biologist may be reasonably expected to resist. There are different vectors; there are opportunistic infections; there are parasites altering the behavior of their hosts; there are all manner of subtle and bizarre adaptations employed to get the viruses from mind to mind, from speaking in tongues to multi-level marketing to the axiomatic method. Best of all, we have the theoretical apparatus of Darwinian evolutionary theory, which fairly cries out to be put to use explaining culture as the result of the differential reproductive success of randomly varying replicators.

Again: this is by no means a recent idea. The first published attempt to explain social and cultural change in a selectionist, that is to say a genuinely Darwinian, manner was made in an 1880 essay by William James. (This essay is also notable for its discussion of the equivalents of founder effects and genetic drift in small populations, and of the dismal state of American politics in the gilded age.) The idea has been re-discovered at intervals ever since then: late in the 19th century, for instance, by the French sociologist Gabriel de Tarde (to whom Sperber, in the book under review, attributes its first formulation), and by the German psychologist Richard Semon, both of whom, alas, burdened it with unfortunate pieces of metaphysical baggage, and were on the losing side of academic quarrels, which combined to put an end to the notion among humanists and social scientists. Since then it has mostly been the province of biologists, such as the great demographer and pioneer of mathematical biology Alfred Lotka and the molecular biologist Jacques Monod, with some honorable exceptions among the other inhabitants of the commonwealth of letters: the philosopher Karl Popper, the geographer André Siegfried, the historian and philosopher of science Stephen Toulmin. The great breakthrough, however, was in the middle of the 1970s, when one of the biologists, Richard Dawkins, had the genius to give the idea one of its most compelling presentations; and the further genius to give the replicators a simple but neologistic name, "memes," which sticks in the memory, forces explanations (that is, opportunities for infection) and so helps propagation. Since then, even literary critics have grasped the notion --- at least, those literary critics who can actually understand popular science books.

This is all very fine stuff; it makes a great mind-toy and is a wonderful stick to beat over the heads of the infamous. The problem with it --- and I say this with real regret --- is that it's very far from clear that most of the elements of culture are stable replicators.

Here at last we come to the contribution of Sperber. Having completed the initiatory field-work proper to a cultural anthropologist, he returned to France to effect a synthesis of anthropological knowledge and cognitive science. One product of this effort was an excellent book he wrote with the linguist Deirdre Wilson, Relevance. This, depending on how it is approached, is variously an assault of the self-proclaimed science of semiotics, a new understanding of how we understand natural languages, a hypothesis on hypothesis formation, or an outline of a solution to the frame problem of artificial intelligence (this last has been aired in these pages before). Another product was the reluctant conclusion that most of what passes for explanation in anthropology is no more explanatory than the scribblings of a newspaper editorialist; like them, it is really either merely description, or interpretation, or sometimes vacuous. Genuine explanations of culture would show how its phenomena arise as the compounded effects of the myriad acts of individual people. Any methodological individualist would have agreed to such a formula; Sperber seems to have been the first to rise to the challenge it presents.

To explain how it comes about that every sapient creature in the United States knows the Roswell story, and similar facts, Sperber saw that we need an "epidemiology of beliefs," or even of "representations" in general. So far he has the agreement of the whole line of cultural Darwinists, from James to Dawkins to Aaron Lynch. Now, however, he reaps the fruit of his anthropological and cognitive studies. Unlike the genetic apparatus, the mind never leaves its contents alone; to think, even to remember, is to change. Nor are these changes, as it were, equally in all directions. Rather the mind works upon its representations in the direction of what Sperber and Wilson call "maximum relevance" --- roughly, extracting the most new information from them for the least processing. (This concept is made clear and precise in their book.) The correctness of their relevance theory is not essential to Sperber's present argument; it is enough that representations change in a highly non-random fashion. The changes undergone by stories in oral transmission, for example, have been studied by folklorists and psychologists at least since Bartlett's classical work in the 1920s, and the tendencies are both well-established and pronounced. The stories become more dramatic; more reasonable (according to the lights of the tellers); more stereotyped (according to the stereotypes of the tellers); etc. (These effects have appeared in these pages before --- in connection with the Roswell story, no less.) But these are all changes that make the story easier to remember, easier to retell, easier to accept and easier to relate to the rest of one's life, i.e., they would increase the fitness of the story. But the decoupling of variation from selection has long been recognized as essential to selectionist theories; as James put it in his essay, it is necessary that they belong to "irrelevant cycles" of causes. Relevance thus constitutes one objection to the selectionist account of culture.

The same transformations block selectionist theories at another point as well, undermining the need for selection by providing an another mechanism capable of inducing similar beliefs in a great many heads at once. Sperber imagines our representations (as instantiated in actual brains, not as abstracta) as elements of the space of a dynamical system, with attractors and basins of attraction. He further imagines that these attractors are the representations of maximum relevance, and that the attractors and basins of attraction of two persons with broadly similar mental contents will also be very similar. But then, from very vaguely related starting-points, those two persons will arrive at very similar ideas about various subjects:

[C]onsider your views on President Clinton. They are likely to be very similar to the views of many, and to have been influenced by the views of some. However, it is unlikely that you formed your own views simply by copying, or by averaging other people's views. Rather, you used your own background knowledge and preferences to put into perspective information you were given about Clinton, and to arrive by a mixture of affective reactions and inferences at your present view. The fact that your views are similar to many other people's may be explained not at all by a copying process, and only partly by an influence process; it may crucially involve the convergence of your affective and cognitive processes with those of many people towards some psychologically attractive type of views in the vast range of possible views on Clinton. [p. 106]
(The same mechanisms allow us, for instance, to correct typos, or even to read what we know should have been written in their place, without noticing that it isn't.)

It will not have escaped the attention of certain professionally-deformed readers that this verbal picture cries out for agent-based modeling. This would let us get at questions like when selection is more important than attraction, and the effects of introducing reliable, non-transformative storage --- writing and the kindred arts. There are fascinating issues in evolutionary game-theory as well, since the representational dynamics will to some extent depend on what representations are already held; though not so much as one might suppose. The reason for the limitation is connected to Sperber's account of religion.

Somewhere in the above-mentioned Philosophical Dictionary --- the passage eludes me just now --- Voltaire has fun with the effects of trying to believe an impossible or contradictory religious dogma. Sperber explains our ability to do so as follows. We can hold (says Sperber) "reflective" beliefs, "believed in virtue of second-order beliefs about them." There is no reason why these reflective beliefs must be comprehended or even comprehensible by those who hold them.

Now, consider the following case. Young Bobby has in his belief box the two representations:
What Mom says is true.
Mom says that God is everywhere.
Bobby does not fully understand how somebody, be it God or anyone else, can be everywhere. However, his mother saying so gives him sufficient ground to exhibit all the behaviors symptomatic of belief: he will readily state that God is everywhere, will assent when the same statement is made by others, and may even refrain from sinning in places where (apparently) nobody can see him. That God is everywhere is for Bobby a reflective belief.... Here is a belief which, like most religious beliefs, does not lend itself to a final, clear interpretation, and which therefore will never become an intuitive belief. Part of the interest of religious beliefs for those who hold them comes precisely from the fact that you are never through interpreting them. While the cognitive usefulness of religious and other mysterious beliefs may be limited... it is not too difficult to see how their very mysteriousness makes them "addictive." [p. 90]

Reflective beliefs are not limited to such statements as "God is everywhere," but include all of our knowledge which is not either direct perception, or arrived at from direct if not unconscious inference from perception: history, empirical science, most of mathematics. (But not all, apparently: see Stainslas Dehane, The Number Sense.) Such "well-understood reflective beliefs ... include an explicit account of rational grounds to hold them. Their mutual consistency and their consistency with intuitive beliefs can be ascertained, and plays an important, though quite complex, role in their acceptance or rejection." By contrast, mysterious reflective beliefs are rationally held (when they are rational) solely upon the authority of those whom we learned them from. (Bobby may not be faulted for believing both his mother and his science teachers; but the one authority is valid, and the other a mere phantasm.) Thus, even if Sperber is correct, and the memeticists quite wrong, we do not lose a handy weapon with which to belabor the fanatical and the superstitious.

Sperber's book carries on from the "epidemiology of representations" in two directions. One is an explication of social institutions and social facts. Most if not all of these --- money, say, or marriage, or the existence of a research institute on Hyde Park Road in Santa Fe --- require, or even consist of, shared mental representations. (Money is whatever we all agree is money, be it stone circles, cowrie shells, bales of tobacco, appropriately worked bit of soft metals, or the numbers stored in certain bit-fields of certain computers.) This is only going to work if the representations we hold internally are all sufficiently similar that no serious conflict as to our expectations arises: precisely what his conjectured mechanisms are designed to accomplish. This opens the way to fulfilling the demand made long ago by Friedrich Hayek, to "analyze social relations in terms of correspondence and noncorrespondence, or compatibility and noncompatibility, of individual aims and desires" and expectations.

The other direction in which Sperber pursues the implications of his basic notion (that "culture is the precipitate of cognition and communication in a human population") is an explanation, grounded in evolutionary psychology and its view of the mind as an agglomeration of specialized modules, of how our representational attractors and basins of attraction form, and of how cultural diversity grows in this uniform mental soil. While this is, again, highly convincing, it defies adequate summary in less than the space (just above thirty pages) in which Sperber expounds it, and accordingly will be left to the reader.

Sperber concludes by reflecting on "What Is at Stake" in a naturalistic account of human culture; these are the dangers of success. (The inevitable charges of reductionism are brushed aside thus: "Let them [the critics] show that the so-called reductionist approaches are ill-conceived, or else let them articulate the moral reasons for their censorship" [p. 152].) The gravest of these is that it might distract the social sciences from their "fundamental role" of "enlightening citizens"; and another, more subtle one.

As in a mirror, we look for our image in the social sciences. When we do not recognize ourselves in the reflected image, we are disturbed. Cognitive psychology does not reflect an immediately recognizable image of ourselves; nor would an epidemiology of representations. Worse, what we think of as essential and primary --- that is, our existence as conscious persons --- comes out, at best, as a changing pattern, socially projected on to a biological structure, itself precarious. [p. 155]
But Sperber views both of these possibilities as remote dangers, to be overcome by theoretical pluralism and the stubbornness which refuses to see the physical world as the crawling molecular chaos of the physicists.

Sperber's arguments carry conviction, at least at the hand-waving level at which the social sciences are accustomed to conducting themselves. Whether they can be carried further, to precision and even accuracy, is another question, technical and empirical. (It may even be possible to recover a Darwinian-Jamesian-Dawkinsian theory, complete with memes, by treating as equivalent all representations within the same basin of attraction, thus looking at the "slow modes," the infectious transmission that happens after the inner cognitive transformations have come to a halt.) Like the memeticists, Sperber raises the prospect of bringing to life the "specter of a natural science of the social": the question now is whether those dry bones can live.


ix + 175pp., endnotes, bibliography, index, 3 diagrams
Anthropology / Cognitive Science / Cultural Evolution / Sociology, Social Theory, etc.
Currently in print as a paperback, ISBN 0-631-20045-2, US$37.75 [Buy from Powell's]
26 August -- 2 September 1998