Needed, an oxymoron: a general scientific theory of individual differences. To focus upon individuality is to celebrate particularity, whereas any general theory must, by definition, submerge the individual case in a wider sea of pattern. Each of us cherishes our own separate, individual personhood, making much of the “fact” that we are different from everyone else (while also insisting, of course, that we aren’t all that different). But attention to individual differences runs the risk of being unscientific, insofar as science aims at generalizing, raising our heads above the individual trees to recognize the forest. Yet the need is there. When Kierkegaard insisted that his tombstone say “That Individual,” he was identifying both an existential truth and a profound scientific dilemma.
One of the unspoken secrets in basic scientific research, from anthropology to zoology (with intervening stops at physiology, political science, psychology, psychiatry, and sociology) is that, nearly always, individuals turn out to be different from one another, and that—to an extent rarely admitted and virtually never pursued—scientific generalizations tend to hush up those differences. It can be argued that that is what generalizations are: statements that apply to a larger class of phenomena and must, by definition, do violence to individuality. But since science seeks to explain observed phenomena, it should also be able to explain the granular particularity of such phenomena. In fact, generalities lose potency if they occur at the cost of artificially leveling otherwise significant features of reality.
No geneticist would dispute that in every sexually reproducing species, individuals possess distinct genotypes (monozygotic twins excepted). In the best-studied species, Homo sapiens, we know that individuals differ apparently even in traits that don’t provide an adaptive advantage, like fingerprints, as well as in physical appearance and personality. Field biologists can often distinguish individuals among their study animals by distinct physical and/or behavioral traits. Almost certainly such individuality is at least as apparent to the animals themselves. Every pet owner knows, moreover, that individual dogs, cats, or horses are not interchangeable, yet theoretical constructs in biological science, in particular, often proceed as though they were. Thus biologists theorize about adaptive traits for “the” adult male or “a” juvenile female, knowing full well that there are only individual adult males and juvenile females—each of them distinct, albeit similar enough to be grouped together.
Medical science, by contrast, is unusual in that it has long acknowledged the importance of individuality among its subjects. Thus, in their training, physicians are repeatedly urged to treat the patient, not the disease. Although there are typical syndromes and basic commonalities among organ systems and ailments, good doctors know that individual Homo sapiens may, for example, develop tuberculosis without fever, or idiosyncratic unresponsiveness or hyper-responsiveness to certain drugs. That is why The New England Journal of Medicine and most medical-specialty journals devote considerable space to individual case reports, something rarely found in other sciences. One of the hottest current areas at the interface of medical genomics and pharmacology concerns the prospects of directing particular pharmaceuticals and their optimal dosages, not to Homo sapiens generally, but rather to the DNA profile of each individual patient.
One of the frustrating things about reading older classics of natural history is the extent to which observers like Ernest Thompson Seton and C. Hart Merriam reported on prey-catching behavior of “the” lynx, or vocalizations of “the” golden eagle, without specifying whether the subject in question was male or female, juvenile or adult. Modern biologists demand that information because we recognize that neither “the” lynx nor “the” golden eagle exists; rather, there is only this lynx and that golden eagle. In fact, one of the triumphs of modern biology has been precisely the overcoming of a tendency to think in something like Platonic ideals. Taxonomists no longer concern themselves with the “type species” of a genus, implicitly recognizing the crucial role of individual variation.
Yet even as we yearn for more-detailed identifying information about lynxes and golden eagles, we do so to combine them into yet another conceptual group, smaller than the species but larger than the individual, like “the” adult male lynx or “the” juvenile female golden eagle. Most textbooks in animal behavior, including my own, contain extensive material, both descriptive and theoretical, concerning the behavior of such larger categories, but not a single entry for behavioral individuality. Generalization thrives, but individuality suffers.
Whether they are comparative psychologists looking for laws of learning, ethologists seeking to identify species-typical behaviors, or evolutionary psychologists concerned with the adaptive value of behavior, researchers in behavioral biology typically view deviations from the statistical norm as aberrant, either genetically or experientially. And for good reason. Few of us would credit as “science” a lengthy rendition of seemingly disconnected anecdotal accounts of individual cases. Furthermore, such detailed descriptions quickly become downright boring to anyone not intimately concerned with the individuals in question.
Accordingly, scientists try to reveal underlying processes, to identify and enunciate principles, to go inductively from the specific to the general. Yet, especially when it comes to living things, each specific case really is distinct. That is why the biological (and social) sciences are so involved with statistics. Chemists can concern themselves with “the” sulfuric acid molecule, or physicists with “the” neutron, confident that having seen one, they have pretty much seen them all. But students of the life and social sciences are always confronted with diversity. As a result, we lean heavily on complex mathematical techniques that tell us whether it is safe to generalize and, if so, how far we can go, and with how much confidence. That is what statistics is all about, and no reputable report in biology or the social sciences will present empirical data without accompanying confidence limits, correlation coefficients, or similar attempts to aggregate findings in a meaningful way—which means, subject to valid generalizations.
Just as Galileo, in the course of his enforced recantation, is said to have muttered of the Earth, “Nonetheless, it moves,” many a researcher, considering the homogenization of disparate data so neatly massaged into a satisfying generalization, is likely to have muttered, of the various individuals thereby erased, “Nonetheless, they are different.”
Which leads us to ask: Why are they different?
For biologists, understanding ideally takes place at both the proximate (immediate) and ultimate (evolutionary) levels of causation. On the proximate level, several factors appear likely. Genetic differences among individuals are obvious sources of individual variability, such differences being produced by mutation as well as (in sexually reproducing species) meiosis and sexual recombination, the basic processes whereby novel DNA is produced and then rearranged as each new individual is formed.
Proximate causes of individual differences must also include the different environments experienced by each individual, with “environment” defined broadly to include all experiences, personal and social. Age-related effects would thus also be expected, as the passage of time provides an opportunity for both genetic and environmental influences—not to mention their interaction—to be more thoroughly expressed. A newborn deer hiding in the brush, for example, will become immobile in response to almost any intruder, whereas adult males’ response will depend on whether that intruder is another buck or a cougar.
Individual differences should probably be distinguished, however, from differences based on distinctive biological and social roles. Thus, among the hoary marmots, Marmota caligata, which I have studied extensively, adult males are typically either socially dominant within their colony, or clearly subordinate to a dominant individual. In a sense, the differences between dominant and subordinate males reflect important aspects of their behavioral individuality. When and if a satellite male assumes the role of dominant male, his behavior becomes that of such males generally. Reproductive females, for their part, are more aggressive than their nonreproductive counterparts, spending more time near their burrows; those roles switch when their reproductive roles reverse. Although social and biological roles are crucial to each specific behavior, it seems most useful to control for “role effects,” and to restrict the concept of behavioral individuality to distinctions among individuals that are socially and biologically as similar as possible in all other respects, notably age, sex, social status, physical health, residence situation, and reproductive state.
Even in cases of genetically identical individuals, idiosyncratic differences in personal experiences can nonetheless be expected to generate a gap in observable characteristics among individuals. Such experiences may begin quite early in life: Intrauterine positioning, for example, can influence phenotypic variation among rodents. A fetus surrounded by males on either side is liable to be androgenized compared with one surrounded by two females. Despite extensive and intensive studies of behavioral development, we still know remarkably little about how individual differentiation actually occurs.
At the ultimate, or evolutionary level, individual differences are even more problematic. One deceptively simple explanation is that the adaptive significance of individual differences is directly equivalent to the adaptive significance of sexual reproduction itself, a subject that has received substantial attention from evolutionary biologists, but that still remains oddly resistant to straightforward explanation. Here is the problem: Given the many costs of reproducing sexually compared with asexually, it isn’t clear why so many creatures opt for the former. A sexually reproducing individual projects only 50 percent of its genes into each offspring, while for asexual creatures, it is 100 percent. That would seem to convey a twofold benefit to any organism whose ancestors opted out of sexual reproduction.
It seems increasingly likely that sexual reproduction enhances the fitness of its practitioners by generating an array of offspring, at least some of which are likely to be adapted to an ever-changing environment, and/or by keeping ahead of parasites and other disease-causing organisms. In any event, a common thread amid diverse theories is that the adaptive significance of sex relates to the production of genetic diversity.
But in order for such genotypic diversity to convey a fitness benefit, it must be reflected in phenotypic diversity, which is to say, it must have some demonstrable effect on the way each individual looks, acts, or responds physiologically. In other words, there must be individual differences. It is therefore ironic that many biologists who are quite familiar with the theoretical dilemma associated with sexual reproduction, and with the received wisdom as to its adaptive significance, nonetheless tend to disregard the existence of substantial individual differences among their research subjects, or scratch their heads when asked to explain its prevalence.
Before biologists fully appreciated—and were confounded by—the genetics of sexual reproduction, Charles Darwin recognized individual differences as essential to the process of natural selection itself. Thus differential reproduction produces evolutionary change only if the more fit are different in ways that can be inherited from the less. Individuality therefore occupies a fundamental place in our understanding of basic evolutionary biology, italicizing the paradox that it has been so rarely investigated.
There are other possible ultimate explanations for the existence of individuality. In some cases, at least, it might be neutral or nonadaptive, a byproduct of selection that maintains genotypic differences for other reasons, such as the well-known case of sickle-cell anemia’s being maintained in the human population because of a benefit conferred upon individuals whose sickle-cell gene is obscured by its genetically dominant alternative. Or it might be the unavoidable result of genetic “noise” that simply has not been selected against. It may even be maladaptive, although it stretches credulity that so fundamental characteristics of living things should carry a pervasive evolutionary cost. On the other hand, individual differences may persist, at least in certain cases, because, even though a “best” behavior or anatomy or physiology might exist, the vagaries of genetics combined with experience necessarily cause random and idiosyncratic departures from that ideal. It would presumably be adaptive, for example, for everyone to be well coordinated, but some are more so than others. In all probability, that isn’t because of an evolutionary payoff to being a klutz, but because in the trajectory from one-celled zygote to adult human being, there are lots of opportunities for things to go at least somewhat awry.
It may also be that individual differences are directly selected if individuals benefit by being distinct from others. For example, when predators develop a “search image” of particular prey, individuals that differ from the most abundant form(s) of such images can experience an advantage. Selection of that sort could favor a continuously varying array of phenotypes, behavioral no less than physical. Individual differences could also be the result of sexual selection, if mate choice favored individuals who differ from the chooser, as a means of reducing inbreeding and its attendant disadvantages in fitness.
Several other factors could select for behavioral individuality, and (along with those described above) they are not mutually exclusive. Thus individual recognition between parent and offspring appears to be adaptive, predictable, and widespread. Especially when possible mix-ups could occur, selection should favor parents whose offspring are distinctive, hence not easily mistaken for a nonrelative. The inclusive fitness benefits of being able to identify kin beyond offspring/parents could also select for individual differences. The study of “kin recognition” has, in fact, become a cottage industry among students of animal behavior. When such recognition is demonstrated, attention is then typically directed to the mechanism whereby it is achieved, whether based on instinctive identification, imprinting from early experience, or simple physical proximity. Only rarely, however, does the chain of causation run the other way, to the possibility that individual differences may have been selected as a means of providing for the adaptive dispensing of nepotistic benefits.
It is also possible that the exigencies of reciprocal altruism have selected for individual differences among would-be reciprocators, as in the case of vampire bats, in which well-fed individuals donate blood meals to those less fortunate. In such cases, selection would probably be especially intense on the ability of initial donors—whether bat or human—to discriminate among their beneficiaries, all the better to insist upon subsequent recompense. By the same token, since the donor could subsequently be identified by the beneficiary, the donor is more likely to be paid back, and therefore more prone to benevolence.
Finally, environmental heterogeneity—defined broadly to include social as well as biological and physical environments—could result in proportionately more individual differences, both through simple adaptation to diverse experiences and through the selection of behavioral flexibility to exploit diverse environments. We might consider shared traits of a species or population as a kind of coarse adjustment in pursuit of fitness, and individual differences—however achieved—as the fine tuning.
Generalizations about behavioral individuality are, at this stage in our knowledge, difficult to support. It is tempting, for example, to suggest that “higher” animals with more-complex brains exhibit more individual variability than do their “lower” counterparts, which rely more on automatic, species-typical reactions to a narrow range of fixed stimuli. It would be surprising if jellyfish or barnacles turn out to demonstrate as much behavioral individuality as elephants or human beings do.
It may thus be significant that some of the most effective portrayals of behavioral individuality come from studies of large-brained animals like chimpanzees and gorillas. And yet, in their basic morphology, two oak trees are also likely to differ more from each other than two elephants are, at least in their physical structure, although we may assume that the inverse is true when it comes to behavior. Our inattention to such matters is emphasized by the fact that no common metric exists with which to make such comparisons.
When seeking to extrapolate from a sample to a larger population, life scientists typically present research results in terms of either mean or median, all the while knowing that there is no “average” individual. (Since there are roughly equal numbers of men and women, the average human being would have one ovary and one testicle.) The nature of statistical inference is such that results must be accompanied by measures of variation. It might seem that such measures effectively take notice of individual variability. But let’s face it: We are overwhelmingly more interested in measures of central tendency, whether it be the area of a panda’s paw, the alarm-calling frequency of a scrub jay, or the incarceration rates of young, unmarried men. We give at best only passing attention to statistical measures of dispersion, largely as unavoidable indices of irrelevant noise. Or—especially if one’s own data are at issue—such measures are considered with trepidation, since if too great, they threaten to keep the results from “reaching significance.” Very rarely are such indicators of individual differences seen as significant in themselves.
Like it or not, however, it is clear that individuality exists, and that it matters. The social structure of coyotes (Canis latrans), for example, is apparently influenced by interactive patterns among littermates; it has also been suggested that wolves (Canis lupus) are predisposed to social niches in their packs by individual traits that characterize them as pups. In yellow-bellied marmots (Marmota flaviventris), interactions among the young born in the same year are strongly influenced by their individual behavioral profiles, which in turn appear to be more consequential than patterns of genetic relatedness per se. Similarly, the greatest part of the variance in female reproductive success is explained by variance among individual females, rather than by the variance among different social groups or different genetic lines.
The impact of individual variability is probably especially great in highly social species (chimpanzees, elephants, human beings, hoary marmots) as opposed to relatively solitary ones (aye-ayes, rhinoceroses, woodchucks), which are less likely to encounter diversity. On the other hand, the fact that a species experiences a high level of social integration may suggest that individuals of the species are relatively unaffected by social vagaries, which may explain why they are capable of getting along in proximity. Species that are comparatively social might possibly develop greater behavioral individuality, since they are likely to occupy social roles that are more clearly defined. We simply do not know, for example, how eusocial versus solitary bees compare in their individuality. Nor, at this stage, can we even make cogent predictions.
I want to urge acknowledgment of the existence and importance of individual differences, and to disagree with Goethe’s maxim “Individuum est ineffabile” (Individuality cannot be explained). Individual differences, I am confident, will eventually be explained. First, however, they must be recognized.
In part, the resistance encountered by human sociobiology, Darwinian psychology, evolutionary psychology—call it what you will—may reflect that none of the “ultimate” interpretations thus far offered account for the enormous amount of (perceived or actual) individual variation that human beings identify among themselves. Perhaps there is something about the human psyche that believes a theory of individuality will do insufficient justice to our own deeply cherished individuality.
The animated movie Antz begins with a hilarious scene in which Z, a troubled ant, is speaking (with the voice of Woody Allen) to a therapist about his feelings of “insignificance.” To Z’s consternation, however, the therapist approves enthusiastically: “Being an ant is being able to say, ‘Hey-I’m meaningless, you’re meaningless.’ ... Remember—let’s be the best superorganism we can be!” The reality is that the best superorganism a person can be is a terrible letdown for a species that has a hard time reconciling social homogenization with its insistence on being “me.”
The current dearth of “individuality theory” may thus reflect the fact that, until recently, advances in applying evolutionary biology to human behavior have been almost entirely the work of biologists, who typically have given individuality short shrift. By contrast, psychologists—stimulated in part by the early work of Darwin’s cousin, Francis Galton—have generally been more receptive to individual differences, with anthropologists occupying a more or less intermediate position (although with no small amount of individual differences!). Perhaps the growing involvement of the latter disciplines in attempts to flesh out a truly evolutionary theory of human nature will result in fuller incorporation of behavioral individuality.
Western science since Aristotle has sought to identify and understand classes of phenomena, looking beyond the particular to organize knowledge into general categories. Accordingly, my request for greater attention to individual differences may seem strangely retrograde. Maybe the best way to justify so perverse a preoccupation is to substitute individual differences for the famed question about climbing mountains: Why study individual differences? Because they are there.