The Structure of Evolutionary Theory

by Stephen Jay Gould

3. In the late 18th century, the great poet (and naturalist) Goethe developed a fascinating (and, in the light of modern discoveries in evo-devo, more than partly correct) archetypal theory in the structuralist or formalist mode — and explicitly critical of functionalist, teleological and adaptationist alternatives [Page 65] — for the diversity of organs growing off the stems and roots of plants. He viewed cotyledons, and all the standard parts of flowers (sepals, petals, stamens and carpels), as modifications of a leaf archetype.

  4. The famous early 19th century argument, culminating in the public debate of 1830 between Georges Cuvier and Etienne Geoffroy St. Hilaire (and analyzed by Goethe in his final paper before his death), did not, as com­monly misinterpreted, pit evolutionary theories against creationist accounts (although Geoffroy favored a limited theory of evolution, while Cuvier re­mained strongly opposed), but rather represented the most striking and en­during incident in this older and persistent struggle between formalist (Geoffroy) and functionalist (Cuvier) explanations of morphology and taxonomic order. Geoffroy advocated the abstract vertebra as an archetype for all animals, beginning (largely successfully) with a common basis for anatomical differences between teleosts and tetrapods, moving to the putatively common design of insects and vertebrates (still with some success, partly confirmed by the Hoxology of modern evo-devo, but also including some “howlers” like the homology of arthropod limbs with vertebrate ribs), and crashing with the proposed homology of vertebrates and a cephalopod doubled back upon it­self (the comparison that sufficiently aroused Cuvier's growing ire into a call for public debate). Geoffroy's theory of dorsoventral inversion between insects and vertebrates was not a silly evolutionary conjecture about “the worm that turned” (as later caricatures often portray), and did not represent an evolutionary explanation at all, but rather expressed a formalist comparison based upon a common underlying structure, ecologically oriented one way in vertebrates (central nervous system up), and the other way in arthropods. The common impression of Cuvier's victory must be reassessed as a complex “draw,” with Geoffroy's position abetted by the fortuity of his longer life and his courting of prominent literary friends as supporters (including Balzac and Georges Sand).

  5. Adaptationist preferences have enjoyed a long anglophonic tradition, beginning with the treatises of Ray and Boyle, in Newton's founding genera­tion, on final causes; then extending, in creationist terms, through Paley and the Bridgewater Treatises; and finally culminating in the radically reversed evolutionary explanations (but still retaining the same functionalist and adaptationist commitments) of Darwin, extending forward to Fisher and the Modern Synthesis. By contrast, continental traditions have favored formalist and structuralist explanations of morphology, from the creationist accounts of Agassiz, through the transitional systems of Goethe and Geoffroy, to the fully evolutionary accounts of Goldschmidt and Schindewolf in the mid 20th century. Interestingly, the complex views of Richard Owen, so widely misun­derstood as an opponent of evolution (when he only rejected the predomi­nant functionalism of traditional British approaches to morphology), may best be grasped when we understand him as a rare anglophonic exponent of a predominantly formalist theory. Owen, following Geoffroy, tried to explain the entire vertebrate skeleton, including the skull and limbs, as a set of modi­fications upon a vertebral archetype.

  6. Darwin maintained a genuine interest in formalist constraints upon [Page 66] adaptationist optimality for individualized features of anatomy — a theme that he epitomized as “correlations of growth.” But he developed an explicit framework and rationale, most thoroughly discussed not in the Origin but in his longest 1868 book on The Variation of Animals and Plants Under Do­mestication, that relegated such formalist effects to a clearly subservient and secondary status, compared with natural selection and adaptation, in evolu­tionary causality.

  Chapter 5: Channels and saltations in post-Darwinian formalism

  1. Galton's Polyhedron, the metaphor and model devised by Darwin's bril­liant and eccentric cousin Francis Galton, and then fruitfully used by many evolutionary critics of Darwinism, including St. George Mivart, W. K. Brooks, Hugo de Vries, and Richard Goldschmidt, clearly expresses the two great, and both logically and historically conjoined, themes of formalist (or structuralist, or internalist, in other terminologies) challenges to functionalist (or adaptationist, or externalist) theories in the Darwinian tradition. This model of evolution by facet-flipping to limited possibilities of adjacent planes in inherited structure stresses the two themes — channels set by internal con­straint, and evolutionary transition by discontinuous saltation — that struc­turalist alternatives tend to embrace and that pure Darwinism must combat as challenges to basic components of its essential logic (for channels direct the pathways of evolutionary change from the inside, albeit in potentially posi­tive and adaptive ways, even though some external force, like natural selection, may be required as an initiating impulse; whereas saltational change violates the Darwinian requirement for selection's creativity by vesting the scope and direction of change in the nature and magnitude of internal jumps, and not in sequences of adaptive accumulations mediated by natural selection at each step).

  2. Orthogenesis, as a general term for evolutionary directionality along channels of internal constraint, rather than external pathways of natural se­lection, existed in several versions, ranging from helpful auxiliaries to Dar­winism, to outright alternatives that denied any creative potency to selection. Theodor Eimer, who coined the term orthogenesis, presented a middle ver­sion that tried to integrate internal channels of orthogenesis with external pathways of functionalist determination. But Eimer defended Lamarckian mechanics for his functionalism, thus leading him to oppose natural selection (he spoke of the Ohnmacht, or “without power,” of selection, contrasted with Weismann's Allmacht, or “all power”) despite his pluralistic linkage of formalist and functionalist explanations.

  3. The orthogenetic theory of the late 19th century American paleontolo­gist Alpheus Hyatt embodied maximal opposition to natural selection, and must be viewed as alternative, rather than auxiliary, to Darwinism. Hyatt conceived the pathway of ontogeny, modified only by heterochronic changes permitted under the biogenetic law, as the internal directing channel that nat­ural selection could tweak, but not derail. Illustrating the influence of theory over perception, Hyatt found several parallel lineages of snails, running along [Page 67] different segments of a common pathway, but all supposedly living in an identical environment — where others had reconstructed typical Darwinian monophyletic trees of phylogeny from the same stratigraphic section of fresh­water planorbids. Hyatt, who engaged in a long and ultimately frustrating correspondence with Darwin on this subject, believed that lineages followed a preordained “ontogeny” of phyletic youth, maturity and old age, thus at­tributing the different internal responses of lineages living in the same envi­ronment to their residence in different stages of an ontogenetically fixed and shared phyletic pathway (a preset internal channel with a vengeance).

  4. Charles Otis Whitman, a great early 20th century American naturalist, developed the most congenial auxiliary theory (to Darwinism) of ortho­genesis in his extensive work on the evolution of color patterns in Darwin's own favorite organism, the domestic pigeon. Whitman argued that domestic pigeons in particular, and dove-like birds in general, followed a strong chan­nel of internal predisposition leading in one direction from checkers to bars, and eventually to the obliteration of all color. (Darwin, by interesting con­trast, argued for a reverse tendency from bars to checkers, but also held, as his basic theory obviously implies, that selection largely determines any par­ticular event and that no internal predisposition can trump the dictates of im­mediate function.)

  5. In his 1894 book on Materials for the Study of Variation (where he coined the term homeosis), William Bateson presented an extensive catalog of cases in discontinuous variation among individuals in a population and be­tween populations of closely related organisms. He used
these examples to develop a formalist theory of saltational evolution, strongly opposed to the adaptationist assumptions of Darwinian accounts. (Bateson's acerbic criti­cisms of adaptationist scenario-building and story-telling in the speculative mode emphasize a common linkage between structuralist preferences for me­chanical explanation, and distaste for the adaptationist assumption that func­tional necessity leads and the evolution of form follows.) Although Bateson coined the term genetics, his personal commitment to a “vibratory” theory of heredity, based on physical laws of classical mechanics — an intuition that he could never “cash out” as a testable theory — prevented his allegiance to the growing influence of Mendelian principles.

  6. Hugo de Vries, the brilliant Dutch botanist who understood the logic of selectionism so thoroughly and acutely (but largely in contrast with the only other biologists, Weismann and Darwin himself, who also grasped all the richness and range of implications, but with favor), developed a saltational theory of evolution, but explicitly denied any predisposition of lineages to follow internal channels of constraint. (He thus showed the potential inde­pendence of the frequently linked formalist themes of channeling and salta­tion, a conjunction espoused by Bateson and Goldschmidt for example, but denied in the other direction by Whitman, who favored channeling but de­nied saltation by supporting a gradualist theory of orthogenetic change.) This fascinating scholar regarded Darwin as his intellectual hero and never forgot the kindness and encouragement conveyed by his mentor and guru during [Page 68] their one personal meeting early in de Vries's career. But de Vries, who devel­oped the theory of intracellular pangenesis (the ultimate source for the term “gene”) in the late 19th century, and then (quite fortuitously and long after he had reached saltational conclusions for other reasons) became one of Men­del's rediscoverers, based his truly saltational theory of immediate macromutational origin of species on his work with the evening primrose, Oenothera lamarckiana, where he mistook an odd chromosomal organization that gen­erates occasional saltations for a biological generality. De Vries, who under­stood the logic of selectionism so well, who knew that his macromutational theory refuted several essential components of Darwinian logic, but who could not bear (for largely psychological reasons) to forsake his intellectual and personal hero, insisted upon his larger fealty to Darwin, even though he had banned Darwinian mechanisms from the master's own realm of the ori­gin of species. So de Vries developed a hierarchical theory that, while denying selection for the origin of species, restored selectionist logic at the higher level of phyletic trends by explicitly proposing “species selection” (his term) as a mechanism for generating broader phylogenetic patterns.

  7. By proposing a comprehensive formalist theory in the heyday of developing Darwinian orthodoxy, Richard Goldschmidt became the whipping boy of the Modern Synthesis — and for entirely understandable reasons. Gold­schmidt showed his grasp, and his keen ability to utilize, microevolutionary theory by supporting this approach and philosophy in his work on variation and intraspecific evolution within the gypsy moth, Lymantria dispar. But he then expressed his apostasy by advocating discontinuity of causality, and pro­posing a largely nonselectionist and formalist account for macroevolution from the origin of species to higher levels of phyletic pattern. Goldschmidt in­tegrated both themes of saltation (in his concept of “systemic mutation” based on his increasingly lonely, and ultimately indefensible, battle to deny the corpuscular gene) and channeling (in his more famous, if ridiculed, idea of “hopeful monsters,” or macromutants channeled along viable lines set by internal pathways of ontogeny, sexual differences, etc.). The developmental theme of the “hopeful monster” (despite its inappropriate name, virtually guaranteed to inspire ridicule and opposition), based on the important con­cept of “rate genes,” came first in Goldschmidt's thought, and always occu­pied more of his attention and research. Unfortunately, he bound this inter­esting challenge from development, a partially valid concept that could have been incorporated into a Darwinian framework as an auxiliary hypothesis (and now has been accepted, to a large extent, if under different names), to his truly oppositional and ultimately incorrect theory of systemic mutation, therefore winning anathema for his entire system. Goldschmidt may have acted as the architect of his own undoing, but much of his work should evoke sympathetic attention today.

  Chapter 6: Pattern and progress on the geological stage

  1. Darwin based his argument for a broad and general vector of progress in life's history not on the “bare bones” operation of natural selection (where he [Page 69] had explicitly denied such an outcome as the most radical implication of his theory), but on subsidiary ecological claims for the predominance of biotic over abiotic competition, and for a geological history of plenitude in a persis­tently crowded ecological world, where one species must displace another to gain entry into ecosystems (the metaphor of the wedge). Darwin used these ecological sequelae, along with the gradualist and incrementalist logic of natural selection itself, as primary justifications for his third essential claim of selection's scope, or the uniformitarian extension of small-scale microevolution, in a smoothly continuationist manner, to explain all patterns of macro-evolution by accumulation of increments through the immensity of geological time.

  2. Such a claim requires that the geological stage operate in an appropri­ate, and “Goldilockean,” manner — not too much change to debar the opera­tion and domination of this slowly and smoothly accumulative biological mode, and not too little to provide insufficient impetus (within Darwin's externalist and functionalist theory) for attributing the amount of change ac­tually observed to natural selection.

  3. The primary claim of “too much” derived from the school of “catastrophism” in geology — a movement that has been unfairly stigmatized by later history, following Lyell's successful and largely rhetorical mischaracterization (he was a lawyer by profession), as an unscientific defense of super-naturalism to cram the observed results of geology into the strictures of bibli­cal chronology, but that actually took the opposite position of strict empirical literalism (whereas uniformitarians argued that the numerous literal appear­ances of rapidity in the geological record must be “interpreted” as misleading consequences of how gradual change must be expressed in a woefully imper­fect set of strata). The great catastrophist Cuvier, in particular, was an En­lightenment rationalist, not a theological apologist — and he based his defense of catastrophism upon his literalist reading of the paleontological and geolog­ical record.

  4. The primary claim of “too little” geology followed Lord Kelvin's increasingly diminished estimates for the age of the earth (incorrectly made — although Kelvin accurately described the necessary, but (as it turned out) empirically false, logic required to validate his views — by assuming that heat now flowing from the earth represented a continuing loss from an originally molten state). Darwin worried intensely over Kelvin's claims, even referring to him as an “odious spectre” in a letter to Wallace. Darwin feared that Kel­vin's low estimates would not permit enough time to generate the history of fife under his slowly acting theory of gradualistic and accumulative change. Although this story has been told often, and has become familiar to scientists, an important (and decisive) aspect of the tale has rarely been exposed: Darwin fought this battle alone, and his strong distress illustrates the maximal, and unique, extent of his gradualistic and continuationist commitments. His closest colleagues, Wallace and Huxley, did not find Kelvin's low estimates unacceptable, but argued that we had only been led to expect such slow change from our previous conception of the earth's age, and that faster rates [Page 70] of phyletic change, implied by Kelvin's dates, were entirely acceptable under their reading of evolution.

  Chapter 7: The modern synthesis as a limited consensus

  1. From the anarchic situation that prevailed at the Darwinian centennial celebrations of 1909 (confidence in the factuality of evolution, linked with agnosticism about theories and mechanics, as the first fruit
s of Mendelism seemed, initially, to refute the gradualism and incrementalism of natural se­lection), the Modern Synthesis eventually emerged in two stages (following the union of Darwinian and Mendelian perspectives in the work of Fisher and others): first, by a welcome restriction that eliminated Kellogg's three al­ternatives in oppositional modes that would have destroyed Darwinism (Lamarckism as a substitute functionalism, and saltationism and orthogene­sis as formalist alternatives), and reasserted, now in a context of Mendelian particulate inheritance, the adequacy of natural selection as a creative force; and second, by an increasingly dubious hardening, culminating in centennial celebrations for the Origin in 1959, that substituted an increasingly rigid adaptationism for an earlier pluralism that embraced all mechanisms (includ­ing genetic drift) consistent with known genetic principles, while favoring se­lection as a primary force.

  2. In his founding book of 1930, The Genetical Theory of Natural Selec­tion, R. A. Fisher showed how slow, gradualist evolution in large, panmictic populations (treated almost as an ahistorical system, analogous to effectively infinite populations of identical gas molecules free to move and diffuse by physical principles) could validate strict Darwinism under Mendelian partic­ulate inheritance (with Darwin's own acceptance of blending inheritance ex­posed as a more serious impediment than Darwin himself had realized), and disprove saltational alternatives by the inverse correlation of frequency and magnitude in variation. To these mathematical and general chapters, Fisher appended a long closing section devoted to his eugenical theory that Western society had begun to degenerate seriously as a consequence of the social pro­motion of infertility (the rise in class level of “good” genetic stock, largely by their correlated tendency to have fewer children, thereby husbanding their economic resources to potentiate their social elevation). Fisher conceived this eugenical “blight” as entirely Darwinian in character — invisible in its gradual expression generation by generation, but ultimately more deadly than the ex­plicit saltational degenerations stressed by most eugenicists.