The idea that there are normal and abnormal ways for people to be and to behave is a very familiar one. So also is the idea that abnormality–or deviance–is something regrettable, deplorable, and even, in some cases, punishable. Here, however, my main concern will be to consider what might be meant by the claim that a person is, in some respect, normal. There is, no doubt, an unsophisticated usage according to which what is normal is what is familiar, and the unfamiliar is feared or condemned as abnormal. But since we are all sophisticated this need not detain us. Sophisticated philosophers, though, have often proposed conceptions of human nature, conceptions that presumably have implications for human normality. Currently, I am more concerned with scientific claims to provide universal accounts of human nature. Recently, I have been especially concerned with the claims of neoclassical economists and sociobiologists (or, as they now prefer to be called, evolutionary psychologists).
Both neoclassical economics and evolutionary psychology make broad claims about how we should think about human behavior, or at least very significant parts of human behavior. Science is on the whole concerned with generalizations, so these sciences are concerned with generalizations about human behavior. This immediately raises a problem: human behavior is, on the face of it, enormously diverse. Why should we expect there to be any very interesting generalizations about it at all. Economics–and a cluster of related scientific and philosophical approaches–offers an interesting answer to this question: its generalizations about people are purely formal. They say what people do given the things they want and what they believe. But they refuse to take any position on what people in fact want. This is most striking in the hostility shown by many economists toward the attempt to distinguish between mere desires and needs. Some people, perhaps many people, attach great importance to food, shelter, or medical care, others to jewellery, fast cars, and fine wine. For scientists or politicians to show a preference for one desire over another is insufferable paternalism. In its extreme version, perhaps only quite seriously entertained by economists of the Chicago school, the diversity of human behavior is fully explained by ultimately unexplained differences in preference: some choose the lives of playboys or film stars, others prefer to sleep under railway bridges and live in cardboard boxes. I shall not say a lot more about this problem later, though for the most part I shall be concerned with sciences such as evolutionary psychology that have something more substantial to say about what people, in general, are like. But first I want to discuss some more abstract philosophical issues about normality. This will also relate the question to the other official topic of this volume, taxonomy.
Philosophical Issues Concerning Normality
Judgements of normality are related in fundamental ways to the classification of things. Something cannot be normal without being a normal something or other. As J. L. Austin would have said, the word normal is substantive-hungry (Austin, 1962, pp. 68-70). I shall therefore approach the question of normality by way of some quite general remarks about the classification of things into kinds. My general thesis, finally, will be that the application of the term normal to people presupposes that people form a kind in a sense in which, I shall argue, they in fact do not. Since, moreover, the term normal almost invariably carries normative implications, its application to humans is not only a philosophical mistake, but potentially a dangerous one.
Two radically different ways of mapping the diversity of a domain of objects should be distinguished. First, there is the process generally associated with the term taxonomy, the process whereby the objects are assigned to a number of distinct kinds. If the properties of the objects tend to be grouped together in well-correlated clusters, the best way of carrying out this classificatory process will seem to be determined by the natures of the objects themselves, and we will be tempted to speak of the discovery of natural kinds. If, on the other hand, the respects of variation among the objects are little correlated, then it may be more natural to think of classification as the imposition of an arbitrary grid rather than the delineation of preexisting kinds. Extreme examples of such cases might be the classification of chemical elements according to the periodic table, and the classification of areas of the world by latitude and longitude. Although the difference between these visions of taxonomy, or the spectrum of kinds of diversity they delineate, is of great importance, and I shall return to it later, for the moment I want rather to contrast the whole range of such taxonomies with a quite different approach to the mapping of diversity. In this approach, the attempt is rather to define a standard or paradigmatic object, and characterize the remainder of the domain in terms of its deviations from this paradigm. In very rough analogy with the two familiar versions of coordinate geometry, we might refer to the first of these as Cartesian classification and the second as polar classification. We should note at once that these two modes of classification are incompatible with one another. We cannot provide a polar mapping of a domain that is acknowledged to comprise two or more Cartesian kinds, for this would presumably require two distinct paradigms. Thus it is natural to think of these mapping processes as sequential. A Cartesian classification provides us with a set of kinds, and a polar mapping provides us with a way to describe the variation within the kind.
One aim of this paper will be to address the problem of which of these approaches is the most appropriate to the mapping of human diversity. This is equivalent to the question of whether, or perhaps better, to what extent, humans form a single kind. The question of whether it even makes sense to talk about normal people arises because it is only to the extent that the variation within a domain of objects is appropriate for polar classification–that is, the extent to which we are concerned with variation within a natural kind–that the concept of normality has any grip. Normality is a concept that relates individuals to a paradigm for the kind–the normal member of the kind. A general way of understanding the arguments of this essay, on the other hand, is that most studies of human variation are better seen in terms of a Cartesian model of classification. If this is right, then the concept of normality has little or no relevant applications to humans.
I mentioned above that the word normal was substantive-hungry. It makes no sense to ask whether something is normal or abnormal without specifying what kind of thing of which it is supposed to be a normal or an abnormal instance. This echoes the point just noted, that we cannot superimpose a polar classification on a Cartesian one. For a Cartesian classification divides a domain into different kinds; and in applying the term normal we must know exactly which kind is in question. So, for instance, it is normal for men to grow facial hair but (perhaps) abnormal for women. It makes no sense to say this is either normal or abnormal for people generally. Similarly for cultural normalities and abnormalities. It is, perhaps, normal for schoolchildren in the United States to swear allegiance to the American flag each morning; this would be decidedly abnormal behavior almost anywhere else. Taxonomy must always precede judgments of normality and deviance. Conversely, to understand a judgment of normality correctly we must appreciate what taxonomic category is being applied to the subject of the judgement.
Another very general point involves something close to an ambiguity in the concept of normality. On the one hand, the normal may be thought of simply as the typical or common. On the other hand, it is sometimes intended normatively or evaluatively. The interplay between these usages is nicely illustrated by the dictionary. The following are the first six definitions in one standard dictionary (Costello, 1991):
1. conforming to the standard or the common type; usual; regular; natural;
2. serving to fix a standard;
3. of natural occurrence;
4. approximately average in any psychological trait, as intelligence, personality, or emotional adjustment;
5. free from any mental disorder; sane;
6. free from disease or malformation.
The ambiguity, I think, is introduced in definition 1, in the phrase “the standard or the common type.” A “standard” type, I suppose, is the kind “fixed” by a normal item in the sense of definition 2. Definition 2 says that a normal item can serve to stipulate a norm; definition 1 that an item can be normal by conforming to such a stipulation. On the other hand, an item can also be normal by conforming to a merely “common” type, presumably a statistical notion. I am not quite sure what is intended by definition 3, which surely contradicts at least the statistical reading of definition 1. Animals with two heads, snow storms in mid summer, or idiots savants all occur naturally; I’m not sure that they are normal. (Possibly “natural” is here, circularly, supposed to imply some kind of statistical normality.)
As a matter of fact, I suspect that in its application to people, the concept of normality is almost invariably normative. This is closely related to the fact that our interest in normality is most often driven by a concern to identify its absence, abnormality. This is perhaps illustrated by the fact that, to my ear anyhow, it is more natural to say that someone is of normal weight than that they are of normal height. In the latter case we would generally say “average height” to emphasize that the judgement was a purely statistical one. Nowadays, on the other hand, it is widely supposed that a particular, “normal” weight is highly desirable, whereas no such assumption is made about height. No doubt there is also a connection with the fact that weight, unlike height, is often considered to be a feature of people for which they are to some extent responsible, so that abnormality in this regard may even be thought culpable.
At any rate, of greatest interest and relevance to my present concerns are definitions 4 to 6.(1) All of these, though superficially matters of fact, are arguably normative. Psychological traits, sanity, and health, that is to say, are all plausibly normative notions. The question whether such concepts can be defined in objective, value-neutral ways is central to the question of whether there is any value-free concept of human normality.
I have already noted that judgments of normality only become possible after an object has been assigned to some kind or other. But if a classification into kinds is no more than a fairly arbitrary grid imposed on a continuously varying domain, it is hard to see what point there could be in distinguishing normal members of the kind. That an object is atypical of the kind would then reflect only the fact that it lay close to an arbitrary boundary that had been drawn between kinds. Attributions of normality seem to imply kinds with nonarbitrary paradigms. The concept of the normal is, thus, readily grounded in relation to natural kinds. On one view, membership of natural kinds is determined by possession of a common essence. We cannot, of course, identify normality with possession merely of the essential properties of the kind to which something belongs, since that is a condition on being a member of the kind at all, normal or abnormal. However, one might think of nonessential properties as being more or less normal by virtue of being more or less closely connected with the essence, presumably in some causal sense. Thus if the essence of a human were conceived as being some sufficiently general description of the human genome, one might argue that having four limbs was normal, on the grounds that the human genome has a very pervasive tendency to contribute causally to the development of four-limbed organisms.(2) On the other hand, being a plumber is neither normal nor abnormal, since the genome has little causal efficacy in determining whether an organism adopts a life of plumbing.
Nowadays, philosophers attracted to the idea that there are real natural kinds to be discovered generally conceive of the real essences of these kinds as properties that are to be discovered empirically. These properties are taken to be the properties that make a thing the kind of thing it is, but any necessity that a thing of that kind have that property is conceived as being causal rather than analytic. In the case with which I am concerned, humankind, it is generally assumed that the real essence that has been found causally necessary for belonging to humankind is the human genome, and a full account of this essence is currently being elaborated, it is held, in the Human Genome Project. Even if this is right, how widely this discovery of essence will license a concept of normality will depend on how many of the further properties of humans are determined by the genome. Hence it is no surprise that those theorists currently engaged in trying to elaborate an account of normal human behavior insist in seeing behavior as deriving centrally from genes. This is the position assumed within the currently thriving project of evolutionary psychology. But before turning to human behavior, however, I would like to consider some issues concerning human physiology.
Normal Human Physiology
Once again it will be useful to begin by recalling substantive hunger. Some taxonomic divisions of humans are based on physiological distinctions, and so some judgments of abnormality, to whatever extent they make sense at all, will apply only to narrower classes of humans. I have already mentioned an example of sex-specific physiological characteristics. The other obvious taxonomic divide is that of race. It seems natural to say, for example, that a dark skin color is normal for people of African origin (or better, recent African origin) or that well-developed epicanthic folds are normal to people from Asia. However on further consideration such judgments become more problematic.
Racial categories, as is now well known, are of little biological significance. From a genetic perspective, variation within a single race greatly exceeds any systematic variation between races. This in turn is reflected in the fact that physiological features may be characteristic of people with particular racial origins, but none, or almost none, are universal. Thus it is certainly possible to find, for example, a person of European origin with darker skin than a person of African origin. To say that dark skin is normal among people of African origin, then, is to say no more than that it is common or typical. What makes racial categories problematic is that, although they are of little biological significance, it would be a serious mistake, reflecting only an excessively all-encompassing view of science, to conclude that they are of little significance. In many cultures, racial categories are of great cultural significance, and in a perfectly legitimate sense this constitutes them as significant kinds (see Root, 1993). The criterion I have in mind here for a significant kind is that being a member of such a kind should provide the ground of characteristic causal powers. In most Western countries, being black is causally efficacious in, for example, attracting a low income. If there were really a value-neutral, purely statistical sense of “normality” as applied to people, one might, therefore, claim that it was normal for black people in Britain or the United States to have low incomes. This would generally be taken to express the view held by those such as the authors of the notorious book, The Bell Curve (Herrnstein and Murray, 1994), who seem to think it appropriate, given their view of racial differences in intelligence, that black people should have lower incomes.
The evaluative aspect of judgments of normality becomes clearer when we move from mere variation, as with height, to physiological features more obviously connected to functionality. It is natural to think that normal people have a standard range of capacities such as particular modes of locomotion, perception, cognition, and so on. To lack these capacities would be, in that respect, to be abnormal. The lack of such capacities is what is referred to by terms such as disability and handicap. Now it would be absurd to deny that there is a sense in which it is normal for people to have working eyes and ears, two legs, and so on. This is not just a statistical sense, but it is in an obvious way normative: our ideals of good health include possession of these features, and that is why we go to great lengths to intervene medically when some condition threatens a person with the loss of eyesight, the ability to walk, or other common physiological capacities. Nevertheless, it is important not to overstate the implications of such judgments of normality.
To analyze more fully the issues concerning normal and abnormal abilities, it is essential to distinguish between intrinsic capacities and relational capacities.(3) My capacity to move my fingers in more or less controlled and accurate ways is an intrinsic capacity. Only neurological or physiological damage to my body will remove this ability. My ability to send an e-mail, on the other hand, is a relational capacity. Only in so far as I have access to the instruments of a complex technology of a kind with which I am sufficiently familiar do I have this latter capacity. A necessary condition for this capacity is that I have certain intrinsic capacities: given the particular technology to which I have access, these include the capacity to make sufficiently controlled movements with my fingers. If I were quadriplegic, or had lost my hands in an accident, I could not possess the capacity to send e-mails on the basis of the intrinsic capacities that I currently use for this purpose. But of course it is quite possible that I could use some different technology, for example voice recognition software, to achieve the same purpose, and hence to realize the same relational capacity.
The question then is the following. Should we more fundamentally think of normal people as being normal in respect to a set of intrinsic capacities or a set of relational capacities? The way to see that the latter, possibly counterintuitive view is nevertheless correct is to notice that in fact most of the significant capacities of even fully able people are contextually determined. Consider, for instance, mobility. Perhaps if we lived in caves in rugged, mountainous terrain, our ability to move around the landscape would depend heavily on intrinsic capacities. But in fact the mobility that people in at least the developed world enjoy is a function of access to roads, footpaths, automobiles, trains, planes, and so on. With small but significant exceptions, these forms of mobility are accessible equally to people with fully functioning limbs and people confined to wheelchairs. Perhaps the most obvious exception is provided by the prevalence of stairs. Stairs are a contrivance that allow able-bodied people to move vertically through buildings, but do not provide this advantage for people in wheelchairs. But of course the able-bodied would be unable to achieve this mobility but for the presence of stairs, and the wheelchair-confined are only unable to achieve it through the absence of ramps or lifts. In complex, modern societies, very many capacities are open only to restricted classes of people with particular skills and access to particular kinds of technology (think again of the miscellaneous class of people able to send e-mail messages). There is nothing exceptional about the class of people unable to achieve vertical mobility in some buildings. Indeed as far as mobility goes, it is surely the case that the financial resources necessary for access to cars, train tickets, or plane fares are far more determinative than are control over one’s limbs. Similar remarks apply to sensory limitations of sight or hearing. In summary, though it is no doubt uncommon to lack the use of one’s legs or the evidence of one’s eyes, and though such deficits are undoubtedly liable to be hugely inconvenient to those who suffer from them, such deficits should not be seen as more than some among a range of characteristics that determine the vastly diverse range of capacities of different people in modern societies. As with the vast majority of such capacities, their presence or absence is determined by a range of intrinsic and environmental factors.
The moral of all this, I think, is a serious ontological one. We certainly can and do think of people as biological individuals, for example when we are considering them as medical patients. From this point of view intrinsic functional deficits are abnormalities. But the more fundamental view of a person is as socially embedded in a culturally and technologically complex context. We see that this is more fundamental when we realize that almost everything people currently are able to do depends essentially on the appropriate cultural and technological context. It is now something of a commonplace–if often ignored by rational choice theorists, economists, and others–that we are social animals. What is less often noticed is that we are uniquely social.(4) The dimension in which humans are uniquely social is that the things they can do as individuals depend to a unique degree on the context that is socially provided, so that it is of only secondary significance to talk about capacities abstracted from this context.
Let me quickly illustrate the foregoing points with the rather less contentious issue of height. Even at my own moderate elevation I have occasionally bumped my head on a door lintel or a low beam. I imagine that this kind of hazard must be a constant danger and concern to a person over seven feet tall. One might say that such a person was significantly handicapped in many contemporary built environments. On the other hand, it is easy enough to imagine that, in a world dominated by high doors and ceilings, and where, perhaps, light switches were commonly placed eight feet above the ground, short people would be equally disadvantaged. (Sociobiologists tell us that tall men are more attractive to women, so that in an evolutionary sense the short are drastically disadvantaged [Buss, 1994, pp. 38-40].) The desirability of typical or atypical physiology, then, is a function of the kind of environment in which a person finds herself. Normative assessments of atypicality, therefore, make no sense unless some particular state of the environment is taken for granted as somehow natural or unavoidable. Since normative implications of the concept of normality seem unavoidable, at least in application to humans, that concept is best avoided except in relation to the absence of the most extreme aberrations of physiology.
Normal Human Psychology
I have allowed that there is a thin and qualified sense, predominantly a medical one, in which it may be useful to talk about physiologically normal and abnormal humans. With psychology things get more complicated. The superficial contrast between these domains is easy enough to see. The notion that there are aspects of normal human physiology–vision, bipedal locomotion, audible speech production, and so on–the absence of any which is a significant disadvantage to many people, is plausible enough. Are there comparable aspects of normal human psychology? It is at any rate less obvious that this is so.
It may appear that a quick and affirmative answer can be reached simply by reflecting on cases of extreme mental derangement. Most people are not seriously psychopathic, and to be so is a considerable problem for the psychopath and others. The challenge, however, is to distinguish putative abnormalities from mere idiosyncrasies. Most people are not fanatical pinochle players or collectors of newts. But these eccentricities are hardly abnormalities. Why shouldn’t psychopathic behavior be a highly regrettable eccentricity? Certainly it is likely to be a good deal more regrettable than are the examples I suggested as eccentricities. But on the other hand, selling government influence to large corporations or defective parts to nuclear power stations are practices likely to be much more regrettable than the behavior of most psychopaths. Such activities are of course immoral, and sometimes criminal, but hardly within the purview of abnormal psychology.
One possible solution takes us back to kinds. One might be tempted to suggest that psychopaths, anorexics, schizophrenics, and so on are kinds of people in some sense that pinochle players or newt fanciers are not. The psychopath is expressing his essence in his antisocial behavior, perhaps, whereas someone just happens to be a pinochle player or a peddler of government influence. A salient and well-studied example with which to evaluate this suggestion is that of homosexuality. The first problem illustrated by this example is that it is now widely held that homosexuals, as a kind, were invented by the medical profession in the second half of the nineteenth century (Foucault, 1977). At that time what had been a form of behavior more or less frequently engaged in by a wide variety of people began to be seen as the definitive criterion for being a distinct and abnormal kind of person. The most interesting part of this story, perhaps, is the way that a group of people, distinguished as an object of study and treatment by medical professionals, eventually succeeded in taking control of their own identities and group definition. I do not mean to imply that by virtue of having been socially constructed such human kinds are not perfectly real kinds. Like racial kinds, they are real at least in the sense that being a member of such a kind appears to confer distinct causal properties. But a central goal of this seizing of control by the subjects of medical classification is to resist the definition of the characteristic behavior of homosexuals as abnormal. Recalling the substantive hunger of the word normal, we can say that certain sexual practices are normal for homosexuals or bisexuals, though not normal for heterosexuals. To ask whether they are normal for people is to make something akin to a category mistake.
Ian Hacking, in his recent book on multiple personality (1995), considers briefly the question whether multiples (as persons with multiple personalities prefer to refer to themselves), as a kind, may be at an early stage in the process of self-definition that homosexuals have already experienced. He remains inclined to doubt this, though with explicit hesitation (Hacking, 1995, p. 38). I take it that the reason Hacking considers it unlikely that multiples will succeed in defining themselves as a natural kind with distinctive behavior natural to members of that kind is that they are in some sense too grossly dysfunctional. Here we should recall the discussion of physical dysfunction. Is multiple personality an intrinsic source of incapacity or a relational one? Our existing society is one that depends in many circumstances on the ability to correlate a human body fairly reliably with a set of behavioral dispositions. This would not, I suppose, be possible in a society of multiples, or a society in which multiple personality was among the normally encountered range of human possibilities. It might be difficult to accommodate this possibility in our current societies, and it might not be worth the trouble; but surely it would not be unimaginable.(5) The picture that I am suggesting so far is that social organization and norms permit certain limits on behavior and place certain behaviors outside those limits. For behavior, or the people exhibiting it, to be normal is for it to fall within this boundary. Some of these behavioral limits are policed by the medical profession, and some by the judicial professions. The limits may be more or less justifiable and more or less subject to dispute by those placed beyond the pale. At any rate, what we have is a concept of normality that is normative through and through.
Might we still try to argue that the difference between homosexuals and multiples consisted in their constituting different kinds of kind? Perhaps normal and abnormal, or even natural and unnatural kinds? It is not easy to make sense of this suggestion. The normality of an individual member is determined by its relation to the kind to which it belongs, either by an appropriate degree of similarity to a paradigmatic member of the kind, or by partaking in the essence of the kind. So presumably the normality of a kind would have to be seen in relation to some higher-level kind of which it was held to be an abnormal subkind. But it makes no sense to think of a kind (as opposed to its members) having an essence that determined the normality or otherwise of its subkinds. And it is difficult to see what principled and nonarbitrary way there could be of determining that a particular subkind was a paradigm subkind for the higher level kind. It seems, therefore, that the abnormality of a subkind could only be derivative from the abnormality of its members. And this is what, I have been arguing, is typically a purely normative matter for the case of human kinds.
Against this emphasis on normativity, it is interesting that many homosexuals now seem quite enthusiastic about the project of reconstructing their condition as one that is biologically determined, and hence claiming that homosexual behavior is, in a purely factual sense, the normal behavior for members of a kind to which they belong. It is easy enough to understand the motives for this. Those opposed to the toleration of homosexuality typically refer to it as an unacceptable lifestyle choice. The claim of determinism denies that it is a choice at all, and therefore that it is an appropriate subject for moral condemnation or legal prohibition or constraint. The difficulty with this as a political strategy is that it may easily be seen as legitimating attempts to treat homosexuality as a medical problem, a possibility that is particularly salient in view of the idea that genetic tests for the predisposition to homosexuality might enable parents to elect the selective abortion of supposedly homosexually inclined fetuses. For medical sanctions, as opposed to moral or penal ones, are what become appropriate if we are dealing with a biologically determined condition. My earlier remarks about physical disability do not, of course, argue against the obvious preference for medical correction of the lack of some intrinsic capacity where this is possible, and the same might be held for the case of homosexuality. On the other hand, if a level of tolerance of homosexuality has been reached such that medical intervention would not seem reasonable, it is difficult to see what need there is for insisting, against the residual elements of social hostility, that this is at no level a freely made choice. So it seems difficult to see how the insistence on biological determinism can be a good strategy. Of course, homosexuality is not generally experienced as a free choice; but this is no reason to assume biological determination. I don’t experience a liking for oysters or chocolate as something I could freely reject, but I do not assume that this shows that these tastes are genetically determined.
Whether or not the appeal to biological determinism is good political strategy, or rendered compelling by experience, we should now turn to the question whether it is likely to be true. And, more generally, we must consider whether a concept of psychological normality could be grounded in an account of the normality of its physical basis. And here, surely, we encounter the most serious possible basis for a purely objective account of psychological normality. The contemporary scientific project with the strongest, or anyhow loudest, claims to provide such an account is evolutionary psychology,(6) and it is this that I must now consider, though very briefly.
A concise manifesto for evolutionary psychology can be found on the cover of an influential recent anthology:
With the advent of the cognitive revolution, human nature
can finally be defined precisely as the set of universal,
species-specific, information-processing systems that operate
beneath the surface of human cultural variability; this
collection of cognitive programs evolved in the Pleistocene
to solve the adaptive problems regularly faced by our
hunter-gatherer ancestors such as mate-selection, language
acquisition, cooperation, and sexual infidelity (Barkow,
Cosmides, and Tooby, 1992).
Beneath the business suits, saris, or loincloths of modern humans going about the diverse business of modern life, we thus learn, are really cavepeople seeking out mates or edible roots or sniffing out one another’s sexual infidelities. A further important feature of the program is that these information-processing systems are held to be domain-specific: the system that deals with mate-selection, for instance, is highly specialized for this task, and quite distinct, both functionally and structurally, from the one that distinguishes cooperators from cheats in trading situations. These specific systems are generally referred to as modules. Two questions we need to consider are first, whether there are likely to be such modules, and secondly, whether the current efforts of evolutionary psychologists can plausibly be taken to have revealed any. The second question can be answered rather confidently in the negative. Evidential standards in this part of science are ludicrously low. Although in contrast with the earlier days of sociobiology empirical research has begun to replace cliche and gossip, the empirical research is, to say the least, unimpressive.
To take one example, there has been a great deal of research on the question, a crucial one for evolutionary psychology, of the causes of sexual attractiveness. Sometimes this involves the investigation of hypotheses of such blinding bizarreness as to make even the most ardent Popperian cringe, for example the much researched theory, occasioning many laborious hours toiling over Playboy centerfolds with calipers, that men were most attracted to women with a waist-to-hip ratio of 0.7 (Singh, 1993). More typically, speculation on the exigencies of life in the Pleistocene lead to guesses about what might have been reproductively profitable decision-making criteria, and the hypothesis that these criteria apply to contemporary humans is investigated by sending out questionnaires asking people what features of other people they find attractive. What is epistemologically so distressing is that the predictably variable results to these surveys lead either to emphatic confirmation of the evolutionary hypothesis or the inference that interfering factors have somehow prevented the criterion from revealing itself. As many of the criteria investigated are wholly banal–people are found to prefer partners who are kind and understanding, or reasonably intelligent, for example–the fact that some of the results are somewhat confirmed shows nothing at all.(7)
The explanation for this shoddy methodology is fairly clear. Though recognizing at some level the need for empirical investigation, what carries much of the weight of the argument is a priori speculation on Pleistocene life. It is assumed that evolutionary change during the few tens of thousands of years of civilized human life will be negligible compared to the adaptation of the human genome that must have taken place over the several million years of relatively static Stone Age existence. It is also assumed, rather more dubiously, that we have a good idea what life was like for our ancestors during these millions of years. In view of these assumptions, the problem is not, contrary to appearances, really one of testing hypotheses about human nature at all, but a kind of quasi-Kuhnian normal science in which puzzles–failures to fit known evolutionary laws to the empirical facts–are resolved with various auxiliary, or ad hoc, hypotheses. But to someone who doubts the validity of the whole program, or has yet to see a convincing paradigm result from the program, the empirical research is not of the slightest use. Should we nevertheless be prepared to rest some weight on the wholly a priori adaptationist arguments? Clearly not. Even if we accept the standard story about the Pleistocene, we cannot legitimately argue from what would have been useful there to what must have evolved. Here the point can be made clear enough with a well-known little anecdote. Two anglers, so the story goes, are fishing for salmon in Alaska. Suddenly one of them sees a polar bear rapidly approaching. He hurriedly removes his waders and begins putting on his trainers. “You fool,” says his companion, observing this activity and its cause. “You can’t outrun a bear whatever you’re wearing.” “You fool,” says the first, “I don’t have to outrun the bear, I only have to outrun you.” The point of this little episode of natural selection hardly needs elaboration. Survival of the fittest doesn’t mean survival of the fittest possible, just survival of the fittest actual. Thus optimality analysis of hypothetical ancestral organisms has little significance unless we know a good deal about the range of behavior actually exhibited by actual ancestors. And even if the standard B-movie picture of Stone Age life is basically correct, we certainly do not know anything about the genetically encoded behavioral repertoire of Stone Age people.
It is also essential to add that much of evolutionary psychology is not only epistemically defective, but also potentially pernicious. Despite continuing attempts to brandish the fact/value distinction as a denial of any practical consequences of this theorizing, it is extremely difficult to see how such psychological theories can fail to have normative implications. Although evolutionary theorists almost universally deny any such goal, theories in evolutionary psychology cannot avoid offering at least partial justifications of behavior. If evolution has in fact shaped our behavior, it can only have done so by selecting physical structures, presumably in the brain, that cause the production of such behavior. To say that a certain behavior, which some find morally objectionable, is caused by a physical structure in my brain, is in effect to remove at least part of my responsibility for it. If my arm jerks convulsively and knocks your Ming vase to the floor you will feel rather differently about my behavior if you learn that I am subject to unpredictable brain spasms that cause my limbs to jerk wildly, than you would if you believed this to have been a freely chosen act of destruction. Similarly, if you are inclined to feel some moral disapproval at hearing that I have abandoned my wife and children and run off with a woman twenty-five years younger, your censure will surely be mitigated if you are convinced that there is a part of my brain the express function of which is to cause me to do just this.(8) Perhaps it is always possible to overcome the pull of our genes, but the very fact that there is such a pull to overcome must provide some mitigation of responsibility. So by presenting behavior as in an important sense biologically determined, its ethical dimension is removed or at least attenuated. There can be little doubt, moreover, that the increasing acceptance of such biological explanations of behavior would eventually lead to changes in the legal status of the behavior allegedly explained.
We still need to ask whether, however inadequate our current methods may be for disclosing the evolved structure of our minds, there is nevertheless such a structure, and a structure that constitutes normal human psychology. It may seem that there must be. I do not want to deny that we evolved. In the process of evolution only natural selection appears capable of producing complex functional structures. But to produce a structure of the complexity of the human brain would take many millions of years. It may seem just to follow that we must have brains that reflect millions of years of natural selection and that, therefore, our brains must be adapted to the conditions of the distant past.
Central to this argument is the idea that only genes can carry information across the generations, and that the selection of adapted genomes is an extremely slow process. In a recent edition of the London Times Higher Education Supplement it was asserted, among other hyperbolic remarks about the Human Genome Project, that this project will provide us with “the recipe book for a human, the key to understanding the basis of life” (Hodges, 1997). Such remarks are omnipresent in the popular and semipopular media, and something like this is, I suppose, widely held to be a contemporary truism, and a truism that underlies the assumption that the structure of our brains is more or less determined by features of our chromosomes. Yet far from being a truism, this statement is wildly misleading. The human genome is only a small fraction of the recipe for a human, if indeed there is any such recipe. A full set of human chromosomes has no chance of producing a human body unless it is situated in a human cell replete with all the proper transcription, and energy-processing cellular machinery; the cell must be properly situated in the right part of a human female body; and the human female must be situated in an adequately sustaining social and ecological environment, and so on. It is sometimes supposed that there is a fundamental asymmetry between the genes on the one hand and all these contextual prerequisites on the other. Probably this is most commonly expressed in the idea that the genes carry information for which these various levels of environmental context provide merely a channel. But this is an untenable position (see Oyama, 1985). Simply put, there is no sense of information in which the genes carry information that can be read off by the environment in which it cannot equally well be said that the environment carries information which is read off by the genes. More simply still, there are many sources of information that are required in building a human body–genetic, cellular, physiological, and, especially at later stages, cultural. All are necessary, none is sufficient. The atavism inherent in evolutionary psychology is, as I have noted, ultimately grounded in the idea that the transmission of human behavior from generation to generation is essentially based in the transmission of the physical structures of brains through physical structures of genes. And physical structures of genes are taken to be more or less constant over historical time. But there is no reason to assume this simple picture of the transmission of behavior, and there are excellent reasons for believing that the reality is far more complex. In the absence of this simplistic genocentric picture, finally, there is no reason to assume that our behavior is adapted to the conditions of our distant past.
Let me also comment briefly on the modularity assumption central to the program of evolutionary psychology. I’m not, in fact, at all clear what is at stake in claims about the modularity of mind. No one, I assume, thinks the brain is a wholly amorphous structureless mass, and no one supposes that it is exhaustively divided into wholly independent systems entirely dedicated to particular areas of behavior. The bottom line is that claims about mental modules are claims that fairly specific aspects of behavior are generated by genetically coded structures in the brain. Although there are now few strict genetic determinists–the causal chain from gene to behavior is too long and attenuated for such a view to be remotely plausible–specialized evolved mental modules are nevertheless the contemporary vehicle through which somewhat qualified versions of genetic determinism are currently expressed. With arguments for the existence of such mental modules almost exclusively a prioristic, and evidence for the behavior they are supposed to generate equivocal at the very best, we should treat this school of genetic determinism with all the respect that thinking people have come to accord to its predecessors.
If we reject evolutionary psychology it is hard to see where any substance can be found for the concept of normal human behavior. This may seem to leave a difficulty about what we should say about abnormal behavior. One way to approach this question is by asking once more what is the difference between a human kind like homosexuals, which we now consider as a kind within which a range of behavior uncommon among other humans is normal, and a kind such as Hacking’s multiples. Though otherwise unusual behavior may be common within a kind such as the latter, we are reluctant to allow such forms of behavior status as a normal part of the range of human behavior. I have noted that it is no help to claim that some human kinds are normal kinds whereas others are abnormal–even unnatural–kinds. But what I take to be the right way to understand all this should by now be clear enough. The decision to allow certain human kinds but not others status as domains of normal human behavior is a normative one, determined by whatever social processes determine the range of behavior acceptable within a given society.(9) Again, my paradigm is homosexuality. Whether the characteristic behavior of homosexuals was to count in the range of normal human behavior has been a fundamental ethical debate since homosexuals were first defined as a kind. Only recently, and by no means universally, has this ethical battle been increasingly won by homosexuals and their supporters. As I mentioned, Hacking considers the possibility that the same kind of battle might be fought and won by multiples. That he doubts that this will happen is a consequence of the judgment that having multiple personalities is fundamentally and inescapably dysfunctional. It may be, nonetheless, recalling the discussion of physical disability, that what is ultimately at stake is whether multiples manage to persuade the rest of us that we should change the way society functions in ways that will remove the disadvantage that currently accrues to people who lack a consistent and unitary personality. It is surely possible to imagine a society in which this would not be a disadvantage, though perhaps it is hard to imagine that we would choose to create such a society. Either way, normality in behavior depends mainly on what we, as a society, decide to accept as normal.
So far I have been concerned a good deal with kinds of people that might well seem in some respect abnormal, and arguing that the grounds of such a judgment are at any rate problematic. But of course a great deal of human variation does not seem, except to the very conservative or bigoted, either normal or abnormal. Now I am thinking of, in a very broad sense, cultural variation: the variations in behavior between people in different parts of the world, in different social strata within the same part of the world, between urban customs and rural customs, between gay people and straight people, and so on. Cultural variation is, prima facie, an absolutely fundamental fact about humans. Cultural variation is, moreover, itself quite variable in its nature. In relatively traditional cultures, or perhaps relatively traditional parts of less traditional cultures, it may be that we encounter good cultural natural kinds. To that extent it might be possible to speak of behavior normal for a Yanomamo Indian or for a rural British aristocrat in a purely descriptive sense. I suspect that in the more diverse, dynamic, and multicultural environments that increasingly many of us now inhabit, variation can only be measured with a fairly arbitrary grid.
However this variation may best be described, it is an explicit aim of evolutionary psychology to minimize its significance. To quote two leading exponents of the discipline, “cultural differences are vastly overstated, because beneath existing surface regularities all humans share the same set of preference-generating and decision-making devices.”(10) I have already said as much as I have time for about why I take the premise on which this conclusion is based to be quite unsupported. Let me conclude, however, by making a somewhat more positive suggestion as to how we might think of the origins of cultural variation in a way that would make the existence of real cultural difference unremarkable.
With some reservations, I think it is helpful to think here of the process of cultural evolution,(11) Without worrying too much about the causes of cultural change, we may, to begin with, think of cultural evolution as no more than a term for referring to the uncontroversial fact that over time cultures change. I would like to defend an analogy between biological and cultural evolution at least this far: biological evolution, as the title of its most famous statement makes clear, is in part about the origin of new kinds of organism; and similarly cultural evolution has led to the development of new and distinctive patterns of human behavior, what we might call cultural species. Such a process surely applies to the development of a great variety of early human cultures (allopatric speciation), and perhaps can be applied equally to more recent phenomena such as the emergence of a distinctive gay culture (sympatric speciation). A reasonable condition on using the term species here is that there be some mechanism, analogous to genetic inheritance, whereby the behavioral patterns characteristic of the culture are transmitted to novice members of the culture. But of course this is clearly the case: cultures are transmitted by a variety of processes, most obviously training and imitation. It is also plausible that some role is played by processes loosely analogous to natural selection: cultural practices are sometimes spread and perpetuated because of their perceived or experienced benefits, or because of features that promote their spread, intended or otherwise, from one person to another.
Despite reservations about how far the concept of cultural evolution should be pressed,12 I do maintain that the analogy between cultural and biological evolution goes far enough to license a belief in real cultural diversity that, in turn, should generate further deep skepticism about the possibility of a descriptive notion of normal human behavior. Behavior, that is to say, can be seen as normal ony relative to the particular culture in which it occurs. But if, as I have also suggested, modern societies, rather than dividing people into fairly distinct cultural kinds, display the kind of complex and overlapping cultural diversity only amenable to taxonomy with a more or less arbitrary grid, attempts at purely descriptive accounts of normal behavior can aspire only to the statistical. And this reinforces the idea that assertions of normality are of little significance unless they are intended or interpreted normatively. The general, extracultural concept of human normality extends only to the physiological, and even there the social and cultural embeddedness of human capacities makes the concept highly problematic. The recognition that judgments of the normality of people are almost always normative judgments, but often masquerade as merely descriptive, should motivate extreme caution in deploying the concept of a normal person.
(1) In fact definition 4, approximately average in any psychological trait, strikes me as wrong. Normal intelligence, I think, has to do not with an average but with a minimum threshold. The expression “abnormally intelligent” would be applied only to the very bright, never to the very dim.
(2) On this view, it might be added, it is abnormal to be born with any number of limbs other than four, but subsequently losing some, since this is generally beyond the control of the genome, should not count as an abnormality.
(3) The following discussion is greatly indebted to the work of John Perry and colleagues in the Archimedes Project at the Center for the Study of Language and Informations, Stanford University. See Perry, Scott, Macken, and McKinley (1996); and for philosophical background, Israel, Perry and Tutiya (1993); see also Amundson (1992).
(4) In one sense, no doubt, we are not as thoroughly social as at least the social insects. But it can be argued that ants or bees are not really individuals at all: they don’t, as individuals, do anything, but only contribute to things done by the colony. To say a bee collects pollen, on this view, is like saying my hand collects nuts–a synecdoche.
(5) I do not take a position on whether it would be desirable to make this accommodation. Certainly I do not mean to imply that social acceptance of a form of behavior is always desirable. It would almost certainly not be desirable for the case of pedophiles, for example, although it appears that pedophiles have recently made attempts at affirmative self-definition.
(6) Evolutionary psychology used to be referred to as human sociobiology, but the latter acquired a rather dubious reputation, and contemporary practitioners insist on the new name. They sometimes claim that the project has been radically transformed by the insight that evolution works to produce structures in the brain rather than, directly, forms of behavior (see Cosmides and Tooby, 1987). Whether early sociobiologists really doubted this, imagining perhaps that genetic control somehow bypassed the brain, seems to me questionable.
(7) For more detailed elaboration of this criticism, see Dupre (1998).
(8) A more distressing example is the attempt to portray rape as an evolved reproductive strategy for otherwise unsuccessful men. See, for instance, Thornhill and Thornhill (1983). For devastating criticism of this argument see Fausto-Sterking (1985).
(9) Here, as I might have at several points in this paper, I am happy to acknowledge the parallels between my position and that of Michel Foucault. Foucault has told in detail the stories of how kinds of people–criminals, mental patients, and of course homosexuals–have been defined by professional bodies for purposes that often have most to do with social control, and he has emphasized the use of the concept of normality to enforce preferred kinds of behavior on the subjects of the disciplines engaged in this process of definition.
(10) Cosmides and Tooby, 1994, pp. 327-32. This paper advocates the synthesis of evolutionary psychology with neoclassical economics, the former providing an account of the tastes about which economists have been notoriously unwilling to dispute. I discuss the prospects for this unholy alliance elsewhere (Dupre, 1998).
(11) This idea has received some serious attention. A pathbreaking discussion is Boyd and Richerson (1985). I consider the relevance of the process of cultural evolution to evolutionary psychology in Dupre (1987).
(12) I am skeptical, for instance, about the value of Richard Dawkins’s (1976) concept of a meme, the cultural analogue of a gene. I doubt whether it is helpful to atomize culture in the way that this tends to imply.
Amundson, Ron, “Disability, Handicap, and the Environment,” Journal of Social Philosophy 23 (1992): 105-119.
Austin, J. L., Sense and Sensibilia (Oxford: Oxford University Press, 1962).
Barkow, Jerome H., Cosmides, Leda, and Tooby, John, The Adapted Mind: Evolutionary Psychology and the Generation of Culture (New York: Oxford University Press, 1992).
Boyd, Robert, and Richerson, Peter J., Culture and the Evolutionary Process (Chicago: University of Chicago Press, 1985).
Buss, David, The Evolution of Desire: Strategies of Human Mating (New York: Basic Books, 1994).
Cosmides, Leda, and Tooby, John, “From Evolution to Behavior: Evolutionary Psychology as the Missing Link,” in J. Dupre, ed., The Latest on The Best: Essays on Evolution and Optimality (Cambridge, Mass.: MIT Press, 1987).
Cosmides, Leda, and Tooby, John, “Better than Rational: Evolutionary Psychology and the Invisible Hand,” American Economic Review 84 (1994): 327-32.
Costello, R. B., ed., Random House Webster’s College Dictionary (New York: Random House, 1991).
Dawkins, Richard, The Selfish Gene (Oxford: Oxford University Press, 1976).
Dupre, John, “Human Kinds” in J. Dupre, ed., The Latest on The Best: Essays on Evolution and Optimality (Cambridge, Mass.: MIT Press, 1987).
Dupre, John, “Against Reductive Explanations of Human Behaviour,” Proceedings of the Aristotelian Society, Supplementary Volume (1998).
Fausto-Sterling, Anne, Myths of Gender (New York: Basic Books, 1985).
Foucault, Michel, The History of Sexuality, Volume 1 (New York: Pantheon, 1977).
Hacking, Ian, Rewriting the Soul: Multiple Personality and the Sciences of Memory (Princeton: Princeton University Press, 1995).
Herrnstein, Richard J., and Murray, Charles, The Bell Curve: Intelligence and Class Structure in American Life (New York: Free Press, 1994).
Hodges, Lucy, “The Big Bang in Biology,” THES, December 5, 1997, pp. 22-23.
Israel, David, Perry, John, and Tutiya, Syun, “Executions, Motivations, and Accomplishments,” Philosophical Review 102 (1993): 515-39.
Oyama, Susan, The Ontogeny of Information (Cambridge: Cambridge University Press, 1985).
Perry, John, Macken, Elizabeth, Scott, Neil, and McKinley, Jan, “Disability, Inability and Cyberspace,” in Batya Friedman, ed., Designing Computers for People–Human Values and the Design of Computer Technology (Stanford: CSLI Publications, 1996).
Root, Michael, Philosophy of Social Science (Oxford: Blackwell, 1993).
Singh, Devendra, “Adaptive Significance of Waist-to-Hip Ratio and Female Physical Attractiveness,” Journal of Personality and Social Psychology 65 (1993): 293-307.
Thornhill, Randy, and Thornhill, Nancy, “Human Rape: An Evolutionary Analysis,” Ethology and Sociobiology 4 (1983): 137-73.
JOHN DUPRE, Versions of this paper were presented at the Universities of Oxford and of Southampton, and as an inaugural lecture at Birkbeck College, University of London. This version has benefited from the comments of various members of those audiences, to whom I am most grateful.
COPYRIGHT 1998 New School for Social Research
COPYRIGHT 2000 Gale Group