The Structure of Evolutionary Theory

by Stephen Jay Gould

  When we move from simple tubes and sheets (aspects of universal geome­try, even though evolved modes of growth in particular lineages must elicit [Page 1261] the forms) to more complex developmental architectures that record the contingencies of particular lineages (rather than the general geometry of Euclid­ian space), both the range and the number of potential spandrels, and their capacity for future exaptive utility, must broaden enormously. I have already mentioned Darwin's (and Owen's) example of non-fusion in neonatal mam­malian skull sutures (p. 326).

  In complex sexual animals, a particularly interesting class of spandrels originates as consequences of a developmental constraint leading both sexes along an initially shared embryological pathway that later branches to differ­entiate a set of homologous structures into the two major facies of a species's sexual dimorphism. In the most widely discussed (literally since Aristotle), and still very much unresolved, case of so-called “male mimicking” genitalia in female spotted hyenas, the peniform clitoris, and the labia majora that fold and fuse in the midline to resemble a scrotal sac, may have originated as span­drels of high testosterone levels that, in female development, correlate with the attainment of a larger adult size than males, and with behavioral domina­tion over males. (We should all remember a lesson from our introductory Bi­ology 1 course: the penis and clitoris, and the scrotal sac and labia majora, are homologous pairs of organs, specialized along divergent paths by disparate hormonal titers in the development of the two sexes.)

  This fascinating case remains controversial (see Frank, 1997), and some re­cent information indicates that high levels of androgens may not induce “male-like” features in female development (R. Wrangham, personal com­munication). But the explicit formulation of such a nonadaptationist hypoth­esis, based upon spandrels, after more than 2000 years of unchallenged adaptationist speculation, has certainly sparked the debate and inspired a vast outpouring of research that will eventually resolve the issue. At least we may

  11-10. Exaptation of raised area at the shoulders as a hump (with secondary ad­aptation of coloring), presumably significant in sexual selection in Megaloceros. We only were able to learn about this unfossilizable feature of this extinct deer because our Cro-Magnon ancestors painted these animals on cave walls, as here from Cougnac.

  [Page 1262]

  be confident that no future researcher of such high status and such empirical expertise will again commit the bald version of the Nietzschian fallacy of con­fusing current utility with historical origin — as did Kruuk (1972), the most eminent student of spotted hyenas, when he argued that he had resolved the adaptive origin of “malemimicking” genitalia by documenting the un­doubted utility of these features in facilitating recognition of previously soli­tary clan members in the “meeting ceremony.”

  I wanted to call my popular article (Gould, 1987a) on my favorite human example “tits and clits,” but desisted because readers would have assumed a sexist bias when I really meant to designate male tits, not to denigrate women. (I eventually settled upon the rhythmically catchy, but not quite so pithy, “male nipples and clitoral ripples.”) The male question may be largely facetious, although the issue has evoked substantial and explicit discussion ever since Buff on: why do men grow apparently non-functional nipples? The puzzlement of so many people, including several accomplished scientists, flowed from adaptationist biases that demanded an explanation in utilitarian terms: perhaps males can suckle babies in certain circumstances; or perhaps they once did, and male nipples persist as a vestige? But the probable resolu­tion, based on a quite different (albeit simple) perspective, requires the codi­fication of a concept of nonadaptive spandrels for recognition and under­standing: males probably grow nipples because females need them for an evident purpose, and many aspects of development follow a single pathway. So females grow nipples as adaptations for suckling, and males grow smaller and unused nipples as a spandrel based upon the value of single developmen­tal channels.

  The female counterpart, however, has evoked much argument, and im­posed substantial grief upon millions of women during the 20th century: why do most female orgasms emanate from a clitoral, rather than a vaginal (or some other), site? The male biologists who fretted over this question sim­ply assumed that a deeply vaginal site, nearer the region of fertilization, would offer greater selective benefit to the Darwinian summum bonum of enhanced reproductive success — hence the supposed puzzle. In the tragedy heaped upon this error, Sigmund Freud defined the almost anatomically im­possible transition from clitoral to vaginal orgasm as the defining criterion of sexual maturity in women. He even regarded, and so labelled, the persis­tence of perfectly satisfactory and exciting clitoral orgasm as a form of frigid­ity. Under his influence, millions of women with normal sexual responses struggled to meet this impossible criterion of “true” realization — with conse­quences ranging from the sad to the tragic.

  But a more probable resolution — if any mystery remains once we shed our adaptationist biases — might identify the clitoral site as a spandrel by the same argument applied to male nipples. The female clitoris is the developmental homolog of the male penis, and the adaptive value of male orgasm seems no more problematical than the biological function of the female breast. The clitoral site of female orgasm need not hold any special adaptive value per se, but may arise as a developmental consequence of selection upon the same or­gan in males. [Page 1263]

  This argument about the nonadaptive nature of the clitoral site (as ad­vanced by Symons, 1979; and Gould, 1987a) has been widely misunder­stood as a denial of either the adaptive value of female orgasm in general, or even as a claim that female orgasms lack significance in some broader sense. Only a grossly overextended commitment to exclusivity for the simplest form of adaptationist argument could ever have led to such foolish misunderstand­ing. I cannot speak from direct experience of course, but I accept the clear testimony that clitoral orgasm plays a pleasurable and central role in female sexuality and its joys. All these favorable attributes, however, emerge just as clearly and just as easily, whether the clitoral site of orgasm arose as a spandrel or an adaptation. (As a spandrel, the clitoral site would represent the different expression of a male adaptation, just as male nipples may be the spandrels of a female adaptation.) After all, we defined the concept of span­drels largely in terms of their rich potential for exaptive utility (Gould and Vrba, 1982). To state Nietzsche's principle in another way: origin as a spandrel implies no diminution of potential for crucial and joyous exaptive use later on.

  As to the important question of a potentially adaptive nature for female or­gasm in general, I have no firm opinion, and certainly feel no hostility to­wards functional hypotheses in conventional Darwinian terms of enhanced reproductive success. (I have only questioned the adaptive interpretation of the specifically clitoral site, since basic developmental anatomy would seem to dictate such a placement for other, and prior, reasons. Alfred Kinsey (1953), a very fine evolutionary entomologist who achieved far greater fame in his “second career” as a sociologist of sexual behavior, upbraided Freud, who had begun his own professional life as a comparative anatomist (with a doctoral thesis on the neurology of amphioxus), for failing to draw the obvious inference from a developmental homology that he knew perfectly well, and from easily available information about the rich innervation of the clito­ris, and the virtual anesthesia of the vaginal canal.

  Some common hypotheses for the adaptive nature of female orgasm have not fared well, including the standard argument that muscular contractions during orgasm help to draw sperm down the vaginal canal towards the site of fertilization (see Buss et al., 1998). Other arguments of a more psychological nature, especially the claim that orgasm may have positive value in stimulat­ing pair bonding, make more sense in a too obvious way that always manages to incite my skepticism. In any case, the establishment of an important and potentially exaptive value of female orgasm does not challenge the hypothesis that the clitoral site originated a
s a nonadaptive spandrel (in Darwinian terms for enhancing reproductive success, not in human terms of female pleasure) of selection upon the homologous organ in males, the penis.

  RELATIONSHIP TO COMPLEXITY. As a primary correlation regulating the distribution and importance of spandrels vs. primary adaptations, increasing complexity of an organ must imply a rising relative frequency of nonadaptive side consequences with potential future utility. With greater complexity in number and form of components, cooptable side consequences must rise to [Page 1264] exceed, or even to overwhelm, primary adaptations. The chief example in bi­ology may be a unique feature of only one species, but we obviously (and properly) care for legitimate reasons of parochial concern. The human brain may have reached its current size by ordinary adaptive processes keyed to specific benefits of more complex mentalities for our hunter-gatherer ances­tors on African savannahs. But the implicit spandrels in an organ of such complexity must exceed the overt functional reasons for its origin. (Just con­sider the obvious analogy to much less powerful computers. I may buy my home computer only for word processing and keeping the family spread sheet, but the machine, by virtue of its inherent internal complexity, can also perform computational tasks exceeding by orders of magnitude the items of my original intentions — the primary adaptations, if you will — in purchasing the device.)

  A failure to appreciate the central role of spandrels, and the general impor­tance of nonadaptation in the origin of evolutionary novelties, has often op­erated as the principal impediment in efforts to construct a proper evolution­ary theory for the biological basis of universal traits in Homo sapiens — or what our vernacular calls “human nature.”

  I welcome the acknowledgment of self-proclaimed “evolutionary psycholo­gists” (compared with the greater stress placed by the “sociobiology” of the 1970's on a search for current adaptive value) that many universal traits of human behavior and cognition need not be viewed as current adaptations, but may rather be judged as misfits, or even maladaptations, to the current complexities of human culture. But most evolutionary psychologists have coupled this acknowledgment with a belief that the origins of such features must be sought in their adaptive value to our hunter-gatherer African ances­tors. (Much of the daily practice of current “evolutionary psychology” focuses upon efforts to identify and characterize the EEA (their term), or “environ­ment of evolutionary adaptation,” for the origin of cognitive universals as direct adaptations in the common ancestral population of all modern humans.)

  I applaud this use and recognition of the Nietzsche-Darwin principle of discordance between reasons for historical origin and bases of current utility (or disutility). But I also believe that “evolutionary psychology” will remain lim­ited and stymied in its worthy and vital goal — to understand the human mind in evolutionary terms — so long as its practitioners place such unwarranted and effectively exclusive weight upon conventional adaptationist explana­tions for the origin of universal cognitive traits, and fail to recognize the cen­tral role (I would say dominant, but the issue obviously remains open) of con­straints and nonadaptations in the initial construction of the cognitive and emotional modules and attributes that we collectively designate as “human nature.”

  A central principle about constraint from each of my two chapters (10 and 11) on the subject would broaden the range of hypotheses and lead to a richer and ultimately more accurate “evolutionary” psychology, both in immediate empirical terms of understanding the human mind, and in conformity with the true depth and range of modern evolutionary theory, rather than invoking [Page 1265] an almost caricatured version of adaptationism as the only ground of evolu­tionary explanation for the origin of traits.

  1. At a sufficient depth and distance, original adaptations now act primar­ily as historical constraints, and must be so characterized and analyzed (the central theme of Chapter 10). When we recognize a cognitive universal of hu­man mentality as ill-fit to the complexities of modern social life, we do not then achieve an explanation of its human origin in adaptationist terms simply because we can state a good case for its initial phyletic appearance as an ad­aptation. We need to specify the evolutionary distance and the environmental context of initial appearance before we can render any judgment. In general, I would accept the statement that if we can locate the feature's adaptational or­igin in the last common ancestor of Homo sapiens, or even as far back as the common ancestor of the hominid line (after splitting from the lineage of great apes), then we may legitimately argue that this initial adaptive context establishes the “evolutionary meaning” of the feature in our quest to understand its appearance in human phyletic history.

  But suppose that the feature had a far more ancient, but still fully adapta­tional, origin in a distant ancestor of very different form and neurological function, and also living in a very different environment — say, in the basal gnathostome fish of early Paleozoic times. Suppose also that this mental at­tribute has persisted ever since as a plesiomorphic aspect of the basic opera­tion of the vertebrate brain. When we then try to explain the evolutionary sig­nificance of this mental mode in contemporary human life — especially when we try to identify its role in quirky and clearly suboptimal characteristics of human reasoning in the modern world — would we wish to claim that an ad­aptational analysis (in recognition of the feature's Darwinian origin in such a distant ancestor) will provide our best understanding? Clearly, we do not so proceed in most evolutionary analyses — and for good reasons discussed at length in Chapter 10 on the evolution of development. Rather, we treat such features predominantly as historical constraints because, as invariant and plesiomorphic traits of our entire clade (not only of all hominids and pri­mates, but also of all mammals and tetrapods), they operate as unchanging constraints upon any subsequent evolution of mental modes, despite their ad­aptational origin in such a distant ancestor of such different form and envi­ronment.

  I suspect that many puzzling features of human mentality would be better resolved if we conceptualized them as historical constraints derived from dis­tant adaptational origins. To cite a hypothetical example (that would attract my substantial and favorable wager were I a betting man): I agree with a ma­jor theme of structuralist philosophy and research, as developed most co­gently in our times by Claude Levi-Strauss and his followers, that identifies our tendencies to parse natural variety into pairs of opposed and dichotomous categories as an inherent property of human mental functioning — with male and female, night and day, and culture vs. nature as primary examples. I think that most people would also identify this strong preference as a con­straint with highly unfortunate consequences for human life — not only because [Page 1266] we so often construct invalid dichotomous taxonomies in our real world of complex continua, but primarily because we so often impose an­other conceptual module for moral judgment upon our pairings (the Manichean good vs. bad), and then proceed to identify one side of the dichotomy (including ourselves and our preferences) as righteous, and the other side (in­cluding “foreigners” and competitors) as worthy of anathematization or even ripe for burning. (I need hardly add that yet another aspect of human mental­ity, our capacity to devise grisly means of death and torture, and our techno­logical ability to apply such means to large numbers of people in short peri­ods of time, makes our innate preferences for dichotomization particularly dangerous.)

  Now I am perfectly willing to believe that our brain's preference for dichotomization arose as a highly adaptive attribute in a very distant and ancient small-brained ancestor that, to enhance its prospects for survival, needed to make limited, quick, and twofold decisions that exhausted the maximal capacity of its judgment in any case: mate or wait, eat or sleep, fight or flee. But, whatever the adaptational basis of origin, dichotomization then persisted throughout the subsequent phylogeny of vertebrates as a historical constraint that became more and more quirky, and more and more limiting, as the brain enlarged into the much more sophisticated instrument of a lin�
�eage that eventually generated our exalted, but curiously freighted, selves.

  2. At the level of immediate reasons for persistence and flourishing of the hunter-gatherer common ancestor of Homo sapiens in Africa, many distinc­tive mental attributes of our species, including major features of “human na­ture” that define our evolutionary success, must have arisen as nonadaptive spandrels (later exapted, in several cases, as vital bases of our current domi­nation), and not as primary adaptations (the central theme of Chapter 11). This conclusion necessarily follows from the previous argument that, at the level of maximal natural complexity represented by the human brain, conse­quential spandrels must, at least in number, overwhelm the primary adapta­tions that generate them. Therefore, in terms of exaptive potential for evolu­tionary futures, the brain includes more cooptable spandrels than primary adaptations. Any “evolutionary psychology” that neglects the nonadaptational origin of many features now useful (or at least used, however dubiously), and that limits the domain of evolutionary inquiry to arguments (of­ten speculative) about initial adaptive causes and benefits, will become more misleading than enlightening in restricting investigation to such a narrow scope of inquiry. We must abandon the largely unconscious bias of an overly strict Darwinian approach that equates all “evolutionary” explanation with adaptationist analysis.