The Structure of Evolutionary Theory

by Stephen Jay Gould

  But Darwin, ever so sly in his Victorian propriety, enshrouded this terminal line in such a flourish of benign prose that most readers, for more than a cen­tury, have construed his famous closing sentence as a poetic metaphor, in­tended only to ornament a revolution with a coda of ecumenical kindness. In fact, I am convinced that Darwin conceived this finale primarily as a mordent critique of the haughtiness and narrowness of physical scientists in debasing natural history, and as a defense of the greater interest and relevance of his own chosen profession. (I need only recall Darwin's extreme discomfort at Lord Kelvin's arrogant dismissal both of natural selection in particular, and Lyellian geology in general — see Chapter 6 for details. This famous incident should remind readers that Darwin may well have harbored angry feelings about the pretensions of mathematical physics and celestial mechanics to su­perior status over natural history among the sciences.)

  Note how Darwin contrasts the dull repetitiveness of planetary cycling (despite the elegance and simplicity of its quantitative expression) with the gutsy glory of rich diversity on life's ever rising and expanding tree. Darwin even gives his metaphor a geometric flavor, as he contrasts the horizontal so­lar system, its planets cycling around a central sun to nowhere, with the verti­cal tree of life, starting in utmost simplicity at the bottom, and rising right through the horizontality of this repetitive physical setting towards the heav­enly heights of magnificent and ever expanding diversity, into a contingent and unpredictable future of still greater possibility: “There is grandeur in this view of life . . . [and] whilst this planet has gone cycling on according to the fixed law of gravity, from so simple a beginning endless forms most beautiful and most wonderful have been, and are being, evolved” (1859, p. 490).

  Throughout the Origin of Species, Darwin stresses the beauty, and espe­cially the simplifying power, of historical explanation in evolutionary science as a cardinal feature of his view of life (as opposed to other versions of evolu­tion). In one of the most striking examples of “less is more” in the history of science (eloquently and elegantly more in this case), Darwin continually em­phasizes that the age-old perception of a “natural system” among organisms had always presumed a basis of order that must be complex, arcane and ab­stract; intricately numerical and geometrically lawlike; or divinely ordained, [Page 1335] and therefore of literally highest and deepest significance. Louis Agassiz, Dar­win's near contemporary and the last truly sophisticated scientific creationist in biological theory, had even argued (see Chapter 3 for an exegesis of his view) that since each species represents a single divine idea incarnated on earth, the “natural system” of taxonomic order among species must literally record the character of God's mind, for taxonomy discovers the principles of higher structuring among God's own unitary items of thought.

  But Darwin's profound, and wonderfully simple, alternative cuts through centuries of assumptions about the unresolvable depth and complexity of natural order with a breathtakingly direct and concrete resolution: the “natu­ral system,” or taxonomic order among species, just records the history of an unbroken genealogical sequence of historical descent, the arborescent topol­ogy of the tree of life. The height of arcane mystery becomes a record of sim­ple history: “As all the organic beings, extinct and recent, which have ever lived on this earth have to be classed together, and as all have been connected by the finest gradations, the best, or indeed, if our collections were nearly per­fect, the only possible arrangement, would be genealogical. Descent being on my view the hidden bond of connection which naturalists has been seeking under the term of the natural system” (1859, pp. 448-449). Moreover, this conclusion has important operational consequences, not just philosophical implications. If, for example, life's order records the connected pathways of a contingent and “messy” history, then a variety of formerly popular numerological schemes (like the “quinarian system” based on organizing taxa into rigid and invariable groups of five for each higher level) cease to make scientific sense.

  Over and over again, throughout the Origin, Darwin stresses that, for a large class of problems about species and interacting groups, answers must be sought in the particular and contingent prior histories of individual lineages, and not in general laws of nature that must affect all taxa in a coordinated and identical way (1859, p. 314):

  I believe in no fixed law of development, causing all the inhabitants of a country to change abruptly, or simultaneously, or to an equal degree . . . The variability of each species is quite independent of that of all others. Whether such variability be taken advantage of by natural selection, and whether the variations be accumulated to a greater or lesser amount, thus causing a greater or lesser amount of modification in the varying species, depends on many complex contingencies — on the variability be­ing of a beneficial nature, on the power of intercrossing, on the rate of breeding, on the slowly changing physical conditions of the country, and more especially on the nature of the other inhabitants with which the varying species comes into competition.

  Interestingly, one of the strongest modern critics of historicism in evolu­tionary science (Kauffman, 1993, as extensively discussed in Chapter 11), has explicitly identified the contingent status of the branching tree of life as his [Page 1336] primary source of discomfort with Darwin's system (Kauffman, 1993, pp. 5-6 in a section entitled “evolutionism, branching phylogenies, and Darwin”). Kauffman, of course, does not deny that the icon of branching correctly ex­presses the topology of life's history (at least for eukaryotic organisms). But he does argue, in the tradition of his intellectual mentor D'Arcy Thompson (see Chapter 11, pp. 1182–1208), that our Darwinian tradition places too much emphasis upon the particular history of a lineage to explain various evolutionary features that should, in his judgment, be encompassed under timeless and general laws as expressions of universal physics, and not explained as contingencies of unpredictable and individual pathways. Thus, although Darwin's own commitment to contingency has been underemphasized, or even unrecognized, by his later followers (largely in their own at­tempt to win more prestige for evolution under the misconception that sci­ence, in its “highest” form, explains by general laws and not by particular narrations), I am scarcely alone in identifying this central (and, in my judg­ment, entirely laudable) aspect of Darwin's view of life.

  Kauffman, on the other hand, makes the same identification as a sharp critic. His single page of discussion, devoted to doubts about particularism rooted in the tree of life, cites a form of the word “branching” no less than twenty times, a sure mark of Kauffman's discomfort with this model, and his good insight about an appropriate target for criticism. Kauffman writes (1993, p. 5), for example:

  The onset of evolutionism brought with it the concept of branching phylogenies. The branching image, so clear and succinct, has come to underlie all our thinking about organisms and evolution ... With the on­set of fullblown evolutionism and Darwin's outlook based on branching phylogenies, the very notion that biology might harbor ahistorical uni­versal laws other than “chance and necessity” has become simple non­sense. Darwin's ascension marks a transition to a view of organisms as ultimately accidental and historically contingent. Our purposes have be­come analysis of branching evolutionary paths and their causes on one hand, and reductionistic unraveling of the details of organismic machin­ery accumulated on the long evolutionary march on the other.

  It is important to recognize, and I'm sure that Kauffman and other critics would concur, that this debate between immanent vs. narrative styles of explanation contrasts different modes of factual knowing, and that both al­ternatives stand firmly opposed to trendy and nonsensical claims about the relativity of empirical “truth” in the light of social embeddedness for any transiently privileged intellectual procedure. When a champion of contin­gency (for the large chunk of nature properly falling under the aegis of narra­tive explanation) argues that he can explain with rigor after the fact what he could not have predicted in principle before the fact, h
e presents his best judg­ment about the empirical structure of historical complexity. Moreover, he does not confess thereby either any limitation imposed by an inferior form of science, or any irreducible subjectivity engendered by the admittedly ineluctable [Page 1337] interaction of human perception and mentality with external “reality” in all efforts to understand nature's ways.

  I would rather, and in the opposite direction, contend that our increasing willingness to take narrative explanations seriously has sparked a great po­tential gain, through admitting a pluralism of relevant and appropriate styles of explanation, in our accurate understanding of nature's wondrous amalgam of rulebound generalities and fascinating particulars. If I may return once more to Hatcher et al. (1907) on the extinction of ceratopsian dinosaurs (see p. 1331), the authors's inarticulated assumption that explanation must flow from general principles of evolutionary biology and uniformitarian geology allowed no intellectual space beyond the most conventional proposals about vectors of organic progress generated by the extrapolation of natural selec­tion in microevolutionary time, and on climatic change wrought by (at most) some intensification of ordinary geological processes. These presuppositions, in our current judgment, led Hatcher and his colleagues to factually incorrect conclusions based on false premises about inherent dinosaurian inferiority. In this case, I would argue that the introduction of narrative perspectives — particularly the idea that the K-T event should be explained as a singularity trig­gered by a bolide impact, and imposing its major effects fortuitously and exaptively upon particular features evolved in other contexts and “for” dif­ferent reasons — has enlarged our armamentarium of potential explanations, and has surely led to a gain in factual understanding through an increased range of permissible scientific approaches.

  As a first, and overly simplified, conclusion, one might then say that more adequate explanation in the evolutionary sciences demands that we titrate these two essential metacomponents of general theory and narrative particu­lars, or invariant predictability and contingent singularity, to achieve any sat­isfactory understanding of our primary subject matter — broad phenomena that embody sufficient regularity to exemplify the basic principles of theory, but that also engage, in their explicit reference to particular times, places and taxa, enough of the fascinating detail of historical events to ensure that even the most committed generalist will learn to appreciate, perhaps even to cher­ish, the antecedent details that ultimately fashion the empirical objects and events through which those basic principles become manifest.

  I would not argue that all conceivable evolutionary questions must invoke enough historical particulars to require a large contingent component in their full explanation. After all, a paleontologist could claim that he only cares about mass extinction in general, and remains entirely indifferent to the question of why trilobites died in the Permian and ammonites in the Creta­ceous. But what a heartless, gutless and uncurious soul he would then be­come. Indeed, James Hutton came pretty close to such total unconcern with the particular histories of geological sections in his “Theory of the Earth” — see Gould, 1987b. But then, Hutton's imperviousness to the fascination of history struck his friends and contemporaries as downright peculiar and mys­terious; and the longstanding impression about his opacity and unreadability stems as much from this peculiarly desiccated focus, as from any supposed inadequacy [Page 1338] in his prose style. Even in his own time, Hutton's friends felt that he could never prevail by his own wits, and that they would have to write “po­nies” to make his ideas accessible. The most famous of these guides (Playfair, 1802), one of the great works in the history of geology in itself, succeeded largely by applying Hutton's theoretical ideas to explain puzzling particulars that historically minded scholars had long found anomalous.

  In any case, and as a purely factual observation about the likes and habits of practicing scientists, hardly a natural historian, dead or alive, has ever failed to locate his chief delight in the lovely puzzles, the enchanting beauty, and the excruciating complexity and intractability of actual organisms in real places. We become natural historians because we loved those dinosaurs in museums, scrambled after those beetles in our backyard, or smelled the flow­ers of a hundred particular delights. Thus, we yearn to know, and cannot be satisfied until we do, both the general principles of how mass extinction helps to craft the patterns of life's history, and the particular reason why Pete the Protoceratops perished that day in the sands of the Gobi.

  This perspective on mixing immanent and historical styles of scientific explanation in the evolutionary sciences, places me, in concluding this book, into an oddly paradoxical situation, exemplifiable in four statements. First, I have championed the cause, and equal claim, of contingency (particularly in Gould, 1989c and 1996a) to the point of my ready identification as a propo­nent of this position (and with no complaint on my part, and no feeling that my critics have been unfair in any oversimplification). Second, the standard strategy for invoking contingency in natural history employs a device of argu­ment legitimately deemed restrictive in its negative criterion, and surely slated for abandonment as students of contingency develop their armamentarium of positive methods and preferential means of identification — but now accepted faute de mieux and in acknowledgment of current practice. That is, we tend to begin with a preference for explanation by predictability and subsumption under spatiotemporally invariant laws of nature, and to move towards con­tingency only when we fail. Contingency therefore becomes a residual do­main for details left unexplained by general laws.

  (Even so sophisticated a historian as McPherson (1988), studying so richly documented an episode as the American Civil War, grants the crucial North­ern victory at Gettysburg to contingency largely because all classically pro­posed general reasons, either for the Union's triumph in the entire war, or for success in this key battle in particular, have conspicuously failed. This being said, the host of fascinating details then evinced to explain Northern suc­cess at Gettysburg — each apparently trivial, each unpredictable, and each eminently changeable before its occurrence by the tiniest of different circum­stances — seems particularly impressive and conclusive as an example of con­tingent explanation, even for the most important events in history. Nonethe­less — for this key point remains especially troubling, and should serve as a sharp spur to both thought and action — however satisfactory the final interpretation, we might never have gotten to contingency at all unless the alterna­tive mode of explanation, so strongly privileged a priori, had failed. And I [Page 1339] need hardly remind evolutionary biologists that such approaches, based upon prejudicially ordered preferences, remain dangerous because the strengths of our (frequently unconscious) assumptions, and the “flexibilities” of nature in seeming to bow to our biases (because we push too hard, and often un­awares), may preclude any access to alternatives at all, as in our failure to consider fruitful and operational hypotheses that do not ascribe organismal traits to adaptation (Gould and Lewontin, 1979).)

  Third, Darwin himself followed this strategy in the Origin, opening up an admittedly considerable space for contingency when he could not devise a testable generality, or when he felt that he had reached a level of uniqueness in detail that required a similar uniqueness in antecedent generating condi­tions. Fourth, and finally, I therefore find myself in what most of my friends and colleagues — but not my own assessment of my deeper interests and con­cerns — might construe as the anomalous position of trying to “win back” for general theory a substantial realm of macroevolutionary phenomenology that, in its failure to emerge predictably from microevolutionary principles of strict Darwinism, would be granted (under point two) to the very realm of contingency that I have tried so strenuously to promote and enlarge.

  But I embrace this apparent paradox with delight. I have championed contingency, and will continue to do so, because its large realm and legitimate claims have been so poorly attended by evolutionary scientists who cannot discern the beat of
this different drummer while their brains and ears remain tuned only to the sounds of general theory. But this book — entitled The Struc­ture of Evolutionary Theory — does not address the realm of contingency as a central subject, and does fire my very best shot in the service of my lifelong fascination for the fierce beauty and sheer intellectual satisfaction of timeless and general theory. I am a child of the streets of New York City; and although I reveled in a million details of molding on the spandrel panels of Manhattan skyscrapers, and while I marveled at the inch of difference between a forgot­ten foul ball and an immortal home run, I guess I always thrilled more to the power of coordination than to the delight of a strange moment — or I would not have devoted 20 years and the longest project of my life to macroevolutionary theory rather than paleontological pageant.

  So yes, guilty as charged, and immensely proud of it! The most adequate one sentence description of my intent in writing this volume flows best as a refutation to the claim of paradox just above: This book attempts to expand and alter the premises of Darwinism, in order to build an enlarged and dis­tinctive evolutionary theory that, while remaining within the tradition, and under the logic, of Darwinian argument, can also explain a wide range of macroevolutionary phenomena lying outside the explanatory power of ex­trapolated modes and mechanisms of microevolution, and that would there­fore be assigned to contingent explanation if these microevolutionary princi­ples necessarily build the complete corpus of general theory in principle. To restate just the two most obvious examples at the higher tiers of time exemplified in this chapter: (1) punctuated equilibrium establishes, at the sec­ond tier, a general speciational theory of cladal trending, capable of explaining a [Page 1340] cardinal macroevolutionary phenomenon that has remained stubbornly resistant to conventional resolution in terms of adaptive advantages to organ­isms, generated by natural selection and extrapolated through geological time; (2) catastrophic mass extinction at the third tier suggests a general the­ory of faunal coordination far in excess (see Raup's quantitative argument on p. 1326) of what Darwinian microevolutionary assumptions about the inde­pendent history of lineages under competitive models of natural selection could possibly generate.