1. Introduction.

Some philosophers, notably Michael Friedman (1974) and Philip Kitcher (1981), have argued that the unification of ever more phenomena under one theoretical scheme lies at the heart of scientific explanation, suggesting, if not quite entailing, that the horizontal expansion of theories is an unqualified scientific good. On the other hand it is sometimes suggested that broad generalizations are of little use in understanding human culture. For example Michael Oakeshott has written: "[K]nowledge of [a tradition of behavior] is unavoidably knowledge of its detail: to know only the gist is to know nothing. What has to be learned is not an abstract idea, or a set of tricks, not even a ritual, but a concrete coherent manner of living in all its intricateness" (1962, 129; Quoted in Shapin 1994, xix.) This perspective, I take it, is broadly shared by historians and cultural anthropologists; the theoretical conflicts between the latter and their more evolutionarily inclined colleagues are perhaps the clearest contemporary schism over issues of local vs. global approaches to human behavior. Though I would not go so far as to say that to know the gist is to know nothing, I do believe that the local specificity of human culture is such as to limit greatly the possible illumination that can be gleaned from theories claiming to apply across the whole range of human cultural diversity. And this, if correct, reveals serious limitations to the general goal of horizontal scientific unification.
 

2. Imperialist Theories in Action.

One should surely note at the outset that this starting point introduces, to put it mildly, some analogies. I suppose that a trip to the local singles bar has something in common with a trip to the supermarket: one may go to either place with the hope of bringing back something that one wants, and the most active singles bars are even occasionally referred to as "meat markets." But the analogy is very limited. The "goods" on display at a singles bar do not come with price tags, and those customers who leave at 2:00 a.m without the desired product generally do not do so because they have surveyed all the price tags (or even shadow price tags) and found that the prices are all too steep. And needless to say, many "risky sexual trades" arise in situations even less market-like than this. Of course the economist will reply that I have simply ignored the relevant definition of a trade just cited, that of any activity perceived as mutually beneficial by two consenting parties. To this one might respond that the picture of careful rational deliberation of costs and benefits (suitably discounted by estimated probabilities) seems more inappropriate to the contract for risky sex hastily drawn up at a singles bar even than to more traditional routes to sexual intimacy. The more fundamental question, to which I shall return, is whether such an abstract account of human interaction is of much use outside its core application, in this case the buying and selling of commercial goods. It is this skeptical question that casts general doubt on the desirability of this imperialistic approach to human behavior.

Philipson and Posner arrive at some surprising conclusions from their economic perspective. Most striking is the following. Policy responses to the HIV epidemic have generally advocated and subsidized widely available voluntary testing. The assumption, presumably, is that someone aware that they are HIV infected is less likely to endanger others by engaging in behavior liable to spread the disease. Philipson and Posner argue that this may well be counterproductive and increase the spread of the infection. A trader contemplating risky sex will estimate the benefits of the activity and subtract the costs of contracting a very unpleasant and ultimately fatal disease, discounted by the probability of contracting the disease. The latter will depend on the prevalence of infected persons in some relevant class and the likelihood of contracting the disease even from an infected person, both of which are quite low in most circumstances. But a person who has been able to determine, by means of a reliable test, that he or she is already infected will have no chance at all of contracting the disease (just as the only people whom it is absolutely impossible to kill are the already dead). Thus for such people the HIV-related costs of a sexual transaction are zero, and they will be much more likely to engage in risky sex, thereby spreading the infection. Philipson and Posner do acknowledge that there may be some altruists in the population, people for whom infecting a sexual partner with a fatal disease would involve some disutility. Since the proportion of altruists is unknown, and due to further complexities, the actual effects of widely available voluntary testing remains an empirical issue. Nonetheless, it is clear throughout the book that Philipson and Posner are strongly inclined to believe that the effect will be deleterious. (Even altruism can have perverse effects. Since altruists, who would prefer not to kill their sexual partners, will generally be less active in the risky sex market, the market will have a higher proportion of egoists and therefore be all the more dangerous.) Unsurprisingly, this fits into a pattern of skepticism about government interventions of various kinds, a pattern characteristic of devotees of the generally beneficial consequences of unfettered markets.

I do not know of any comparably detailed studies from the point of view of evolutionary biology except for those that focus on the evolution of the virus itself. It is plausible, however, that speculative evolutionary biology could be deployed here in support of the economistic perspective. To begin with, evolutionary biology will certainly treat sex as a major behavioral imperative, while notoriously viewing altruism as a problematic and limited hypothetical motive. Thus biology could readily be deployed in support of Philipson and Posner's evaluation of the relative significance of these motivational factors. Much more speculatively, it is widely held that a function of the often surprising mate selection criteria encountered in nature is the selection of mates with relatively low parasite loads (Hamilton and Zuk 1982). But if there is evolutionary pressure to find ways of selecting relatively parasite-free partners, there will be equally strong selective pressure to conceal the parasites that one has. Thus if, as evolutionary imperialists frequently appear to do, we see human deliberation as a Darwinian enabling mechanism, the behavior of Philipson's and Posner's egoistic HIV spreaders represents the implementation of a deep biological imperative. (It should not be objected to any of this that the sexual behavior in question is not primarily reproductive, since contemporary biology generally does not connect behavioral dispositions with explicitly held final goals. On the other hand if, as some sociobiologists seem to think, homosexuality is an evolved behavior with a unique and distinctive function, then the preceding suggestions may be irrelevant to this important case.)

Although economism and evolutionary speculation often support one another naturally, evolution explaining particular behavioral preferences, and economics exploring the consequences of choices in accordance with those preferences, relations are not always so smooth. My second example is of a phenomenon that is surely a great embarrassment to human sociobiologists, the spectacular decline in fertility in recent decades in the world's wealthiest countries. Although heroic attempts might be made to argue that somehow or other, contrary to all appearances, people are continuing to maximize reproductive success, this would have all the plausibility of the thesis that we live in the best of all possible worlds. Evolutionary approaches to this question must at least back away from direct explanation of behavior towards more causally distant psychological mechanisms. Of course, the claim that people have some serious interest in sex, an interest with an evolutionary basis in the connection between sex and reproduction is hardly controversial. But it is evident that modern Western humans have managed to decouple the interest in sex from its consequences in reproduction, something facilitated by widely available and effective methods of contraception. Evidently decisions about reproduction are now typically made in some much broader context of individual or even social goals, and evolutionary biology appears to have ceased to have any significant implications for human fertility.

Economics may therefore seem much better placed to address this phenomenon. The locus classicus for such an attempt is Gary Becker's Treatise on the Family (1981; revised ed., 1991). Becker's narrative begins with an account of the marriage market within which men and women attempt to acquire partners with whom they can create utility-maximizing families. Such partnerships having been formed, the members must the determine how to allocate their resources to achieve the highest accessible level of utility. Among the major items that they wish to produce and--in a chillingly Swiftian phrase--to consume, are children. Already we see, in contradistinction to a naive evolutionary perspective, that raising children is just one among many possible tastes, so that we should have no a priori expectations about natural fertility rates. Becker then makes a further distinction between two possible sources of family utility, quantity of children and quality of children. These interact in obvious ways: high quality children are costly, involving expenditures on such things as health and education, and thus a demand for superior children reduces the demand for numbers of children. Various explanations are offered for the changes in preference between quality and quantity of children with economic development. Whereas subsistence agriculture creates a demand for lots of children as a source of cheap labor, developed economies provide a range of economic opportunities that increase the returns on investment in the production of high-grade offspring. And, remarkably enough, there is a correlation between the level of education of mothers and of their children, though Becker cautions (1991,153) that this may not be a direct causal connection, but rather a consequence of the lower demand for quantity of children among more educated women, and the inverse relation between the demand for quantity and demand for quality.

I do not want to deny that the economic, and even the evolutionary, speculations just reviewed may have some grain of truth in them, and might provide a legitimate basis for empirical investigations of various kinds. And certainly economists and biologists engaged in the kinds of speculation I have just described will generally be ready to admit that various other factors are also relevant to the phenomena under consideration. So it may be wondered what my objection is. I now turn to some more critical reflections on this kind of scientific work.
 

3. Problems with Imperialist Science

Economics, notoriously, conceives of human behavior as the exercise of "rational" choice. A person faced with a decision is conceived as estimating which action will generate the largest expected excess of benefits over costs to the agent. Thus in the case of the person contemplating risky sex, the benefits are the expected pleasure to be derived from the sexual contact, and the costs are the estimated risks of acquiring a fatal illness. Against the objection that human motivation is a bit more complex than this, it may be replied that this economistic perspective is at worst benignly vacuous. If it is true that people are motivated by some concern for the well-being of others, or some wish to behave in ways customary among their social group, or according to their conception of morality, duty, etc., then behavior satisfying these various criteria will simply be seen as providing some contribution to the utility of the agent. "Utility" here need not be taken as referring to any real measurable quantity, but only as a fictional device for describing a consistent set of preferences. (The normative evisceration of utility, through marginalism, ordinalism, and finally revealed preference theory, is a dominant theme of the last hundred years of economic thought (see Gagnier 1993).) Both the economic works I have been discussing exemplify such a strategy. Thus Becker suggests that parents have a utility function in which utility from their own consumption is traded off against the utility of their children. Posner and Philipson conceive of altruists as people who derive disutility from communicating a fatal disease to another person. And so on.

I want to suggest that this may in the end be vacuous, but it is hardly benign. Here I touch on a wide range of philosophical work that has criticized the rational choice foundations of economic thinking, and I can only summarize (see, e.g. Sen 1979; Anderson 1993). The most obvious point is that to treat altruism, morality, or accepted social norms simply as tastes that some people happen to have--I like candy and fast cars, you like morality and oysters--is grossly to misplace the importance of norms of behavior in peoples' lives. A second point is that whereas in principle the rational choice perspective may be capable of encompassing any mix of self-interested and non-self-interested concerns, in practice it almost invariably reflects the assumption that the former are far more significant. The most prominent contemporary defender of the project of analyzing ethics in terms of economically conceived rational choice, David Gauthier, for example, writes: "It is neither unrealistic nor pessimistic to suppose that beyond the ties of blood and friendship ... human beings exhibit little positive fellow feeling" (1986, p.101). A third point is perhaps the most important for my present concerns. The economistic perspective on human behavior is in an important sense scientistic. By this I mean here that it conceives itself as an objective and (ironically) disinterested reflection on the facts of human behavior. But human behavior is not an immutable set of phenomena awaiting the correct scientific analysis, but is rather subject to constant historical evolution. And theorizing about human behavior is always to some extent an intervention in this evolution. Striking empirical confirmation of this claim in the present context is provided by recent research by the economist Robert Frank and his associates (Frank, Gilovich and Regan 1993). Frank administered tests of the disposition to cooperate to students before and after completing introductory courses in microeconomics and in astronomy, and found that the economics students were strikingly less cooperative at the end of the course. (The astronomy students were marginally more cooperative.) In confirmation of this result he also discovered that economic professors made fewer contributions to charity than professors in other fields. I share the conviction of Adam Smith (1776, 15), confirmed by Frank's first result, that on the whole people do not select their occupations because of their different dispositions, but acquire different capacities and dispositions as a result of their occupations. An unusually highly refined awareness of self-interest may be harmless or even appropriate for analyzing the futures market in pork-bellies, but it does not seem to me an asset for analyzing familial relations or the spread of AIDS.

My second more traditionally methodological point is also something I take to be a central aspect of scientism. "Serious" economic approaches to behavioral questions do not merely suggest ways in which economic factors might impinge on matters of human concern: it is a professional responsibility to dress up these suggestions in some quasi-mathematical guise. Becker's Treatise on the Family is an outstanding example of this problem. (Since Becker is a Nobel laureate it is of course no surprise that he adheres to the professional demands of his discipline.) The book addresses matters of obvious importance--the selection of marriage partners, decisions about child-bearing and rearing, and so on--and on occasion has illuminating insights and even empirical data that bear on these topics. On the other hand, the work as a whole is rendered largely unreadable by a continuous and obfuscating veneer of mathematics. I say "obfuscating" because the mathematical modelling at almost every point requires a level of abstraction that removes the discussion from any serious connection with the phenomena. Thus, for example, the first chapter on marriage markets, which addresses the economics of monogamy versus polygamy, assumes for purposes of model construction, that all men and all women are identical. The object of the exercise is to determine how many wives will maximize a husband's income, which is total family income minus the income of his wives. Becker is at pains to emphasize (e.g. 1991, 96) that men are not assumed to value wives for their own sakes, but only for their contributions to family productivity. While this does help the analogy between marriage and the market in, say, cars (or better, oil, since for analytic simplicity the number of spouses is allowed to vary continuously), it also makes it most unlikely that any conclusions will have much relevance to human mating patterns. As the chapter progresses, it is true, the men are allowed to differ in quality (superior men have characteristics that positively effect the marginal productivity of their identical wives), and in the following chapter, on assortative mating, high-quality and low-quality women are also introduced. Again though, differences in quality are only variations in some set of properties that contributes to efficient family production.

This, I am inclined to claim, is all fairly self-evident nonsense on its face--these abstractions would seem harmlessly ridiculous in a prose discussion of the nature of the family. The problem, however, is that its face is not easily seen. As I mentioned, the book is largely unreadable. It is, however, skimmable--informal discussion leads me to believe that even economists usually skim this kind of work--and conclusions can be seen to emerge. There is some danger that such conclusions will be given some weight as appearing from the serious scientific work (see all that mathematics) of a world-renowned scientist. One might even be tempted just to read the quite clear and intelligible summaries and conclusions at the end of each chapter. If one were to do so, one might read, for example, that "one of the more surprising conclusions of our analysis is that progressive taxes and expenditures may well widen the inequality in the long-run equilibrium disposition of disposable income" (1991, 231) Although many will be happy just to welcome this conclusion, those less enthusiastic about it might go back and see where it came from. In the section on Government Redistribution of Income (1991, 218) we may read, before diving into the alphabet soup, Becker's model for a progressive taxation system. This amounts to a flat rate tax on income, combined with a lump sum redistribution for everyone. Although this does provide an asymptotic approach towards the flat tax rate with increasing income, and thus might strictly be counted as progressive, it is hardly the kind of thing serious proponents of progressive taxation have in mind. This strikes me as rather weakly supporting the surprising conclusion in the summary. Although this is a particularly egregious example, it serves to me to illustrate clearly enough the rhetorical dangers of the formalized abstractions that fill the pages of Becker's book.

I have not tried to give detailed critiques of the examples discussed above, although I am currently engaged in a study of economics that includes critical study of several areas of economic thinking. Indeed, in some cases I am inclined to wonder whether any more than an intelligible exposition of the ideas in question, unobscured by jargon or irrelevant formalism, is required to display the implausibility of many of these account. With more space I would also like to compare the above examples with some examples of evolutionary imperialism, which I have also criticized elsewhere. One point I would like to mention here is that, in contrast with the potential vacuity of economic theory--whatever people do must reflect what they most want--evolutionary biology is hampered by a central concept with rather more content. Although evolutionary biologists have notoriously been accused of making up any old story that will link some observed behavior to reproductive success, the need for such a story is a non-trivial requirement, and stories can be investigated for plausibility. One cannot simply say that these particular animals prefer service to their kind or a glorious death to reproductive success. Thus, as indicated in the case of fertility, reproductive optimality must be sought in many cases some causal distance behind overt behavioral performance.
 

4. Conclusions

Advocates of the semantic view of theories have emphasized that much of science is more realistically described as consisting in sets of models than in universal laws. It seems to me that this is indeed a more realistic account of the way science is practiced in biology and economics. One important point about this general conception is that it rests uneasily with a commitment to expansionist science. Whereas it is natural to think of the power and interest of a law of nature being a function of the variety of cases to which it applies, it is much less clear that breadth of application is a virtue for models. Models are, after all, constructed to reflect as accurately as is practicable, particular kinds of situations. The further they are transported from their paradigm applications the less realistic they will become, and the more they will need to be modified to account for differences between their original home and their new areas of application. As new environments introduce different causal factors, alien models from distant domains will be increasingly partial in their relevance. This seems to me exactly the way to see the partiality and even irrelevance of the applications of economic and evolutionary models discussed in earlier parts of this paper.

The home turf of economics is the market. And while there may perhaps, in a suitably extended or analogical senses be said to be markets in children, risky sex, marriage partners, etc., paradigm markets are in oil, wheat, pork belly futures, or houses: that is, relatively interchangeable commodities that people acquire by paying money. As we move into these merely analogical or even metaphorical markets, the attempt to apply market models involves more and more abstraction from the diverse range of factors that actually affect behavior in these areas. To mention only one striking point pertinent to marriage "markets", the conception of romantic love which, for all its ideological problems, surely plays some role in matrimonial decisions in Western societies, is precisely an extreme commitment to the non-substitutability of one item for another.

I believe that the inadequacies of applications of such models so far from their natural home would seem obvious in most cases if there were not some strong motive inclining scientists to be sympathetic with the projects they represent. And indeed the standard response to such skepticism is to admit that the treatments are as yet too simplified to be altogether realistic, but that they constitute the first step towards some very important goal. What goal? The goal of applying the preferred imperialist strategy of explanation to the phenomena in question. Since I am suggesting that this is not a goal we should have any interest in pursuing, I urge that these simplistic and unconvincing explanatory endeavors should be taken at face value. Moreover, as is very clearly illustrated by incursions of economics into various domains, alien intellectual strategies may import inappropriate and even dangerous assumptions into the colonized domains. Certainly we may sensibly want to acquire a range of modelling strategies that are ready to hand when the phenomena call for them. But Procrustean fitting of phenomena into an antecedently preferred type of model is another matter.

I shall conclude with one final remark about the relevance of these suggestions to biology. I have concentrated on explanations of human behavior, in part because the intuitions here about the inadequacy of certain proferred explanations are sharpest. But I believe that concerns about scientific imperialism have much wider relevance to the practice of biology. It is true that evolution by natural selection is a process of sufficient generality and significance to have some bearing on a very wide range of biological phenomena. But what bearing, and the detailed nature of the process, seem to me to be as variable as the phenomena of human behavior. Thus the widely held belief that some basic mathematical models from population genetics provide us with the fundamental mechanisms of natural selection, the attempts by philosophers to provide universal, axiomatized accounts of the theory of evolution by natural selection, and even the attempt to provide a homogeneous and uniform account of biological classification across the whole biological domain, seem to me as misguidedly imperialistic as the projects I have been criticizing in the human sciences.
 

References

Becker, G. S. (1981; enlarged ed. 1991), A Treatise on the Family. Cambridge, Mass.: Harvard University Press.

Dupré, J. (1993), The Disorder of Things: Metaphysical Foundations of the Disunity of Science. Cambridge, Mass.: Harvard University Press.

Frank, R. H., Gilovich, T. and Regan, D. T. (1993), "Does Studying Economics Inhibit Cooperation?", Journal of Economic Perspectives 7: 159-171.

Friedman, M. (1974), "Explanation and Scientific Understanding", Journal of Philosophy 71: 5-19.

Gagnier, R. (1993), "On the Insatiability of Human Wants: Economic and Aesthetic Man", Victorian Studies 36: 125-154.

Gauthier, D. (1986), Morals by Agreement. Oxford: Oxford University Press.

Hamilton, W.D. and Zuk, M. (1982), "Heritable True Fitness and Bright Birds: A Role for Parasites?", Science 218: 384-387.

Kitcher, P. (1981), "Explanatory Unification", Philosophy of Science 48: 507-531.

Oakeshott, M. (1962), "Political Education", in Rationalism in Politics and Other Essays. New York: Basic Books, pp. 111-136.

Philipson, T. J. and Posner R. A. (1993), Private Choices and Public Health: The AIDS Epidemic in Economic Perspective. Cambridge, Mass.: Harvard University Press.

Sen, A. K. (1979), "Rational Fools: A Critique of the Behavioural Foundations of Economic Theory", in F. Hahn and M. Hollis, (eds.), Philosophy and Economic Theory. Oxford: Oxford University Press, pp. 86-109.

Shapin, S. (1994), A Social History of Truth. Chicago: Chicago University Press.

Smith, A. (1776), The Wealth of Nations. Edited by E. Cannan, New York: The Modern Library (1937).