The Structure of Evolutionary Theory

by Stephen Jay Gould

  [Page 17]

  1-3. The famous frontispiece from Scilla's treatise of 1670 defending the organic nature of fossils. The solid young man, representing the truth of sensory experi­ence, shows a fossil sea urchin in his right hand to a wraithlike figure represent­ing the former style of speculative thinking. With his left hand, the solid figure points to other fossils found in Sicily. The text proclaims: “Vain speculation un­deceived by the senses.”

  [Page 18]

  this particular specimen just happen to occupy the exact places that I needed a priori to make my central point about lower choppings that destroy theo­ries, middle choppings that change theories in a Falconerian way (major al­terations in structure upon a preserved foundation), and upper choppings that change theories in the lesser manner of Darwin's Milanese metaphor (smaller excisions that leave the framework intact as well).

  The central trunk (the theory of natural selection) cannot be severed, or the creature (the theory) dies. (The roots, if you will, represent sources of evi­dence; any one may be excised, if recognized as incorrect by later study, so long as enough remain to anchor the structure). This central trunk then di­vides into a limited number of major branches. These basic struts — the three

  1-4. Agostino Scilla was also a celebrated painter as well as a scien­tist. The plates of his 1670 trea­tise are therefore particularly well done. This figure, representing a fossil coral that Scilla found near Messina, fortuitously (and without any alteration whatsoever), pre­sents a detailed picture of the basic logic of Darwinian theory as recog­nized in this book. See text for details.

  [Page 19]

  branches of the Darwinian essence in this particular picture — are also so es­sential that any severing of a complete branch either kills, or so seriously compromises, the entire theory that a new name and basic structure becomes essential.

  We now reach the interesting point where excisions and regraftings pre­serve the essential nature of an intellectual structure, but with two quite dif­ferent levels of change and revision, as characterized by Falconer's and Dar­win's competing metaphors for the Duomo of Milan. I would argue that a severing low on any one of the three major branches corresponds with a revi­sion profound enough to validate the more interesting Falconerian version of major revision upon a conserved foundation. (The Falconerian model is, in this sense, a Goldilockean solution itself, between the “too much” of full de­struction and the “too little” of minor cosmetic revision.) On the other hand, the severing of a subbranch of one of the three branches symbolizes a less portentous change, closer to Darwinian models for the Milanese Duomo — an alteration of important visual elements, but without change in the basic framework.

  My fascination with the current state of evolutionary theory, at least as I read current developments in both logic and empirics, lies in its close con­formity to the Falconerian model — with enough continuity to make the past history of the field so informative (and so persistently, even emotionally, compelling), but with enough deep difference and intellectual fascination to stim­ulate anyone with a thirst for the intriguing mode of novelty that jars previ­ous certainty, but does not throw a field into the total anarchy of complete rebuilding (not a bad thing either, but far from the actual circumstance in this case).

  To summarize my views on the utility of such a model for the essence of Darwinian logic, I will designate three levels of potential cuts or excisions to this organic (and logical) coral of the structure of evolutionary theory, as originally formulated by Darwin in the Origin of Species, and as revised in a Falconerian way in recent decades. The most inclusive and most fundamental K-cuts (killing cuts) sever at least one of the three central principles of Dar­winian logic and thereby destroy the theory tout court. The second level of R-cuts (revision cuts) removes enough of the original form on one of the three central branches to ensure that the new (and stronger or more arborescent) branch, in regrowing from the cut, will build a theory with an intact Darwin­ian foundation, but with a general form sufficiently expanded, revised or re­constructed to present an interestingly different structure of general explana­tion — the Falconerian model for the Duomo of Milan. Finally, the third level of S-cuts (subsidiary cuts) affects only a subbranch of one of the three major branches, and therefore reformulates the general theory in interesting ways, while leaving the basic structure of explanation intact — the Darwinian model for the Duomo of Milan.

  I wrote this book because I believe that all three pillars, branches, or tri­pod legs, representing the three fundamental principles of Darwinian central logic, have been subjected to fascinating R-cuts that have given us at least the [Page 20] firm outlines — for the revised structure of evolutionary explanation remains a work vigorously in progress, as only befits the nature of its subject, after all! — of a far richer and fascinatingly different theory with a retained Dar­winian core rooted in the principles of natural selection. In short, we live in the midst of a Falconerian remodeling of our growing and multiform, yet co­herently grounded, intellectual mansion.

  I will not, in this chapter, detail the nature of the K-cuts that failed (thus preserving the central logic of Darwinism), the R-cuts that have succeeded in changing the structure of evolutionary theory in such interesting ways, and the S-cuts that have refurbished major rooms in particular wings of the edi­fice — for these specifications set the subject matter of all following chapters. But to provide a better opening sense of this book's argument — and to clarify the nature of the three central claims of Darwinian logic — I shall at least dis­tinguish, for each branch, the K-cuts that never prevailed (and therefore did not fell the structure) from the R-cuts that have affected each branch, and have therefore provoked our current process of building an enriched struc­ture for evolutionary theory.

  Returning to Scilla's coral (Fig. 1-4), consider the central branch as the first leg of the tripod (agency, or the claim for organismal selection as the causal locus of the basic mechanism), the left branch as the second leg (efficacy, or the claim that selection acts as the primary creative force in building evo­lutionary novelties), and the right branch as the third leg (scope, or the claim that these microevolutionary modes and processes can, by extrapola­tion through the vastness of geological time, explain the full panoply of life's changes in form and diversity).

  The cut labeled K1 on Figure 1-4 would have severed the entire coral by disproving natural selection as an evolutionary force at all. The cut labeled K2 would have fully severed the second branch, leaving natural selection as a legitimate cause, but denying it any creative role, and thereby dethroning Darwinism as a major principle in explaining life's history. (We shall see, in Chapters 3–6, that such a denial of creativity underlay the most common anti-Darwinian argument in the first generations of debate.) The cut labeled К3 would have fully severed the third branch, allowing that natural selection might craft some minor changes legitimately called “creative” in a local sense, but denying that Darwin's mechanism could then be extended to explain the panoply of macroevolutionary processes, or the actual pageant of life's his­tory. The success of any one of these K-cuts would have destroyed Darwinian theory, plain and simple. None of them succeeded, and the foundation of Darwinian central logic remains intact and strong.

  In striking, and most positive, contrast, I believe that higher R-cuts — leav­ing the base of each major branch intact, but requiring a substantial regrowth and regrafting of an enlarged structure upon the retained foundation — have been successfully wielded against all three branches of Darwinian logic, as the structure of evolutionary theory developed in the last third of the 20th cen­tury (following too rigid a calcification of the original structure, a good ad­umbration of the coral metaphor!, in the hardening of the Modern Synthesis [Page 21] that culminated in the Darwinian centennial celebrations of 1959). On the first branch of agency, the cut labeled R1 (see Fig. 1-4) expanded Darwin's unilevel theory of organismal selection into a hierarchical model of selection acting
simultaneously on several legitimate levels of Darwinian individuality (genes, cell-lineages, organisms, demes, species, and clades). I shall show in Chapters 3, 8, and 9 how the logic of this pronounced expansion builds a the­ory fascinatingly different from, and not just a smooth extension of, Darwin's single level mechanism of agency — my reason for portraying the hierarchical model as a deeply interesting R-cut rather than a more superficial S-cut.

  On the second branch of efficacy, the cut labeled R2 accepts the validity of Darwin's argument for creativity (by leaving the base of the branch intact), but introduces a sufficient weight of formalist thinking — via renewed appre­ciation for the enormous importance of structural, historical, and develop­mental constraint in channeling the pathways of evolution, often in highly positive ways — that the pure functionalism of a strictly Darwinian (and ex­ternalist) approach to adaptation no longer suffices to explain the channeling of phyletic directions, and the clumping and inhomogeneous population of organic morphospace. The strict Darwinian form of explanation has thereby been greatly changed and enriched, but in no way defeated. I shall discuss the historical aspect of this branch in Chapters 4 and 5, and modern reformula­tions of this R2 cut in Chapters 10 and 11.

  On the final branch of scope, the cut labeled R3 accepts the Darwinian contention that microevolutionary modes and principles can build grand pat­terns by cumulation through geological immensity, but rejects the argument that such extrapolations can render the entire panoply of phenomena in life's history without adding explicitly macroevolutionary modes for distinctive expression of these processes at higher tiers of time — as in the explanation of cladal trends by species sorting under punctuated equilibrium, rather than by extended adaptive anagenesis of purely organismal selection, and in the ne­cessity of titrating adaptive microevolutionary accumulation with occasional resetting of rules and patterns by catastrophically triggered mass extinctions at time's highest tier. Chapters 6 and 12 discuss historical and modern cri­tiques of Darwinian extrapolationism.

  For now, I will say little about the even higher and more superficial S-cuts of subbranches, but I will at least indicate how I construe this category by stating a hypothetical example for each branch: an S1 cut, for example, might accept the selective basis of evolutionary change in a purely mechanical sense, but then deny full force to Darwin's deliciously radical philosophical claim that all apparent “higher level” harmony arises consequentially, through the invisible hand of lower levels acting for personal reproductive success. One might, in principle, propose such a revision by arguing that a higher force, operating by an overarching principle of order, “employs” natural selection as its mechanical agent. (I speak only hypothetically here, for no such defend­able scientific hypothesis now exists, although the concept certainly remains intelligible. Explicitly theological versions don't count as science, whatever their kind or form of potential validity.) An S2 cut might be assayed by a [Page 22] developmental saltationist who accepted the selectionist basis of adaptive change but felt that, at a sufficient relative frequency to be counted as impor­tant, the initial steps of such changes may be larger than the pure continuationism of Darwinian selection can admit. And an S3 cut might accept the full validity of microevolutionary extrapolationism, but deny the subsidiary de­fense of progress that Darwin grafted onto this apparatus (see Chapter 6) with ecological arguments about plenitude and the priority of biotic over abiotic competition.

  As a paleontologist and part-time historian of science by profession, my reading of these important R-cuts arose from a macroevolutionary perspec­tive framed largely in terms of longstanding difficulties faced by Darwinism in extending its successes for explaining small changes in palpable time into equally adequate causal accounts for broader patterns and processes in geo­logical history. I have, in this effort, also benefited from my personal study of Darwin's life and times, and especially the late 19th century debates on mech­anisms of evolution (as promulgated largely by professionals who could nei­ther fully understand nor accept the radical philosophical commitments un­derlying Darwin's view). This historical study allowed me to grasp the continuity in basic themes from Darwin's own formulation, through these foundational debates, right down to the major theoretical struggles of our own time. An appreciation of this continuity allowed me to discern and de­fine the distinctively Darwinian view of life.

  But I recognize only too well that every strength comes paired with weaknesses. In my case, a paleontological focus leads me into relative ignorance for an equally important locus of reform in the structure of Darwinism — in­creasing knowledge of the nature of genomes and the mechanics of develop­ment. (I try to cover the outlines of important theoretical critiques from this “opposite” realm of the smallest, but the relative weightings in my text reflect my own varying competencies far more than the merits of the cases. For ex­ample, although I do discuss, and perhaps even adequately outline, the im­portance of Kimura and King's neutralist theory in questioning previous as­sumptions of adaptationist hegemony, I surely do not give an appropriate volume of attention to this enormously important subject.)

  Nonetheless, I hope that I have managed to present an adequate account of the coordinating themes that grant such interest and coherence to modern reformulations of the structure of evolutionary theory. Such thematic consis­tency in revision becomes possible largely because Darwin himself, in his characteristically brilliant way, tied the diverse threads of his initiating argu­ment into an overall view with a similarly tight structure — thus granting clear definition to his own commitments, and also permitting their revision in the form of an equally coherent “package.” I would argue, moreover, and with­out wishing to become extravagantly hagiographical (for I wrote this book, after all, primarily to discuss a critique and revision of strict Darwinism), that our modern sense of limitations in the canonical version arises from decisions that Darwin made for tough and correct reasons in the context of his initiat­ing times — reasons that made his account the first operational theory of [Page 23] evolution in modern science. In particular, as Chapter 2 will discuss in detail, Darwin converted evolution from untestable speculation to doable science by breaking through the old paradox (as embedded most prominently in Lamarck's system) of contrasting a palpable force of small-scale change that could do little in extension, with a basically nonoperational (and orthogonal) mechanism of large-scale change putatively responsible for all the interesting patterns of life's history, but imperceptible and untestable from the uniformitarian study of modern organisms.

  By claiming that the small-scale mechanics of modern change could, by extension, explain all of evolution, Darwin opened the entire field to empirical study. And yet, as Hegel and so many other students of change have noted, progress in human (and other) affairs tends to spiral upwards in cycles of pro­posal (thesis), then countered by opposition (antithesis), and finally leading to a new formulation combining the best aspects of both competitors (synthe­sis). Darwin's thesis established evolution as a science, but his essential com­mitments, as expressed in the three legs of his necessary logical tripod (or the three branches of his conceptual tree or coral, as in the alternate metaphor of Fig. 1-4), eventually proved too narrow and confining, thus requiring an an­tithesis of extension and reformulation on each branch, and leading — or so this book maintains as a central thesis of its own — to a still newer and richer synthesis expressing our best current understanding of the structure of evolu­tionary theory.

  In fact, and to repeat my summary in this different form, one might encapsulate the long argument of this book in such a Hegelian format. Pre-Darwinian concepts of evolution remained speculative and essentially nonoperational, largely because (see Chapter 3) they fell into the disabling para­dox of contrasting an effectively unknowable large-scale force of cosmic progress against an orthogonal, palpable and testable small-scale force that could generate local adaptation and diversity, but that couldn't, in principle, explain the macroevolutionary
pattern of life. Then Darwin, in his thesis (also an antithesis to these earlier sterile constructions), brilliantly argued that the putative large-scale force did not exist, and that all evolution could be explained by upward extrapolation from the small-scale force, now prop­erly understood as natural selection. In a first stage of debate during the late 19th and early 20th centuries (Chapters 3–6), most critiques of Darwinism — one might designate them as a first round of ultimately destructive antithe­ses — simply denied sufficient agency, efficacy and range to natural selection, and reasserted the old claim of duality, with selection relegated to triviality, and some truly contrary force sought as the explanation for major features of evolution. Strict Darwinism eventually fended off these destructive critiques, reasserted itself in the triumphant, and initially (and generously) pluralistic form of the Modern Synthesis, but eventually calcified into a “hardened” ver­sion (Chapter 7).

  Then, in a strikingly different, and ultimately fruitful, second round of antitheses, a renewed debate about central theoretical issues arose during the last three decades of the 20th century, and reshaped the field by recognizing [Page 24] that selection needed to be amplified, reformulated and invigorated by other, noncontrary (and, at most, orthogonal) causes, not rejected as wrong, or scorned as trivial (Chapters 8–12). The one long argument of this book holds that a synthesis (still much in progress) has now sufficiently coagulated from this debate to designate our best current understanding of the structure of evolutionary theory as something rich and new, with a firmly retained basis in Darwinian logic — in other words, and following the organizing and opening metaphor of this chapter, as a validation of Falconer's, rather than Darwin's, concept of the historical growth and change of Milan's cathedral.