The Structure of Evolutionary Theory

by Stephen Jay Gould

  Punctuated equilibrium does not merely assert the existence of a phenomenon, but ventures a stronger claim for a dominant role as a macroevolutionary pattern in geological time. But how can this vernacular notion of “dominant” be translated into a quantitative prediction for testing? At this point in the argument, we encounter the difficult (and pervasive) method­ological issue of assessing relative frequency in sciences of natural history. If species were like identical beans in the beanbag of classical thought experi­ments in probability, then we could devise a sampling scheme based on enumerative induction. Enough randomly selected cases could establish a pattern at a desired level of statistical resolution. But species are irreducibly unique, and the set of all species does not exhibit a distribution consistent with requirement of standard statistical procedures. It matters crucially whether we study a clam or a mammal, a Cambrian or a Tertiary taxon, a species in the stable tropics, or at volatile high latitudes. Moreover — and es­pecially — the “ideal case study” method has often failed, and led to paro­chialisms and false generalities, precisely because we tend to select unusual cases and ignore, often quite unconsciously, a dominant pattern. Indeed, proclamations for the supposed “truth” of gradualism — asserted against ev­ery working paleontologist's knowledge of its rarity — emerged largely from such a restriction of attention to exceedingly rare cases under the false belief that they alone provided a record of evolution at all! The falsification of most “textbook classics” upon restudy only accentuates the fallacy of the “case study” method, and its root in prior expectation rather than objective reading of the fossil record.

  Punctuated equilibrium must therefore be tested by relative frequencies among all taxa (or in a truly randomized subset) in a particular fauna, a par­ticular clade, a particular place and time, etc. If we can say, as Ager did (see p. 753) that all but one Mesozoic brachiopod species displays stasis, or as Imbrie did (see p. 760) that all but one Devonian species from the Michigan Basin shows no change, then we have specified a dominant pattern, at least within a particular, well-defined and evolutionarily meaningful package. I cannot give a firm percentage for what constitutes a “dominant” relative fre­quency — for, again, we encounter a theory-bound claim, where “dominant” specifies a weight, beyond which the morphological history of a clade must be explicated primarily by the differential success of species treated as stable en­tities, or Darwinian individuals in macroevolution — and not by anagenetic change within species. More research must be done, largely in the testing of mathematical models under realistic circumstances, to learn the relative frequencies [Page 774] and rates required to impart such dominance to species-individuals in the course of macroevolution. For now, and for empirically minded pale­ontologists, the study of relative frequencies in entire faunas, rather than the extraction of apparently idealized cases, should be pursued as a primary strategy of research.

  Critics have sometimes stated that punctuated equilibrium rests upon declaration rather than documentation. (Maynard Smith once compared the the­ory to “Aunt Jobisca's” maxim about ancient verities “known” by folk wis­dom a priori.) We do indeed assert that working paleontologists know the fact of dominant stasis in their bones — but this claim represents a fair con­sensus about the history of a field, and does underscore a paradox of non-concordance between deep practical knowledge and imposed theoretical ex­pectation. We have never tried to argue that such a “professional feeling” constitutes documentation for punctuated equilibrium. As with all scientific theories, punctuated equilibrium will live or die by concrete and quantifiable evidence. As with any good hypothesis, punctuated equilibrium becomes op­erational when workable definitions can be provided for key claims and ex­pectations — in this case, for stasis, punctuation, and relative frequency. Con­trary to the impression of some critics who have not followed the primary literature of paleobiology during the last 25 years, punctuated equilibrium has proven its fruitfulness and operational worth by being tested — and usu­ally confirmed, but sometimes confuted — in a voluminous literature of richly documented cases (see pp. 822–874).

  Microevolutionary links

  Eldredge and I coined the term-punctuated equilibrium in a paper first presented (Gould and Eldredge, 1971) at a symposium entitled “Models in Paleobiology” at the 1971 Annual Meeting of the Geological Society of America. T. J. M. Schopf, the organizer of the symposium, conceived the en­terprise as a tutorial in modern evolutionary theory for professional inverte­brate paleontologists. By accidents of history, invertebrate paleontologists generally receive their advanced academic degrees from geology departments, not from biology. Fossils became primary tools for stratigraphic correlation long before the development of evolutionary theory, and even before all scien­tists had accepted them as remains of ancient organisms! Given traditions of narrowness in postgraduate education — particularly in Europe, where stu­dents often attend no formal courses at all, and certainly no courses for credit, outside the department that will grant their degree — most paleontolo­gists, before the present generation, did not receive any explicit training in evolutionary biology, and could not articulate the basic concepts of popula­tion genetics or theories of speciation. In paleontological usage, “evolution” designated little more than the inferred pathway of phylogeny. This “little learning” often became the “dangerous thing” of Alexander Pope's classic couplet, as paleontologists derived their understanding of evolution from memories of old textbooks, or from shared impressions amounting to little more than the blind leading the blind. This situation has now changed dramatically — [Page 775] and Eldredge and I do take pride in the role played by punctuated equilibrium in encouraging this shift of interest — as a profession of paleobiology, supported by several new journals dedicated to the subject (Paleobiology, Historical Biology, Lethaea, Palaios, and Palaeogeography Palaeoclimatology Palaeoecology (or P-cubed to aficionados), for example), has arisen to accommodate burgeoning research in the application of evolution­ary theory to the fossil record, and in enlarging and revising the theory in the light of novel macroevolutionary data.

  In any case, Schopf's symposium featured a series of presentations; each suggesting how one aspect of paleontological work might be enlightened by modern microevolutionary theory, particularly as expressed in the applica­tion of models, preferably quantitative in nature. Eldredge and I drew the topic of species and speciation — and our original article on punctuated equi­librium (Eldredge and Gould, 1972) emerged as a result. (As I have often stated, the basic idea had been presented in Eldredge, 1971. We had been graduate students together at the American Museum of Natural History, un­der the tutelage of Norman D. Newell. We had discussed these issues often and intensely throughout our graduate years. We had been particularly frus­trated — for we had both struggled to master statistical and other quantitative methods — with the difficulty of locating gradualistic sequences for applying these techniques, and therefore for documenting “evolution” as paleonto­logical tradition then defined the term and activity. When I received Schopf's invitation to talk on models of speciation, I felt that Eldredge's 1971 publica­tion had presented the only new and interesting ideas on paleontological im­plications of the subject — so I asked Schopf if we could present the paper jointly. I wrote most of our 1972 paper, and I did coin the term punctuated equilibrium — but the basic structure of the theory belongs to Eldredge, with priority established in his 1971 paper.)

  I mention this background to clarify the original context and continuing focus of the theory of punctuated equilibrium — a notion rooted in the explicit goal that Eldredge and I set for ourselves: to apply microevolutionary ideas about speciation to the data of the fossil record and the scale of geological time. Before we proposed the theory of punctuated equilibrium, most paleon­tologists assumed that the bulk of evolutionary change proceeded in the anagenetic mode — that is, by continuous transformation of a unitary popula­tion through time (see Fig. 9-5). In this context,
most paleontological discus­sion about species centered itself upon a contentious issue that constantly cir­culated throughout our literature (see Imbrie, 1957; Weller, 1961; McAlester, 1962; Shaw, 1969) and even generated entire symposia dedicated to potential solutions (see Sylvester-Bradley, 1956): the so-called species problem in pale­ontology.

  This supposed problem — more philosophical and definitional than empiri­cal (once one accepts the underlying assumptions about anagenesis as a domi­nant factual reality) — arises because a true continuum cannot be unambigu­ously divided into segments with discrete names. If population A changes so extensively by anagenesis that we feel impelled to provide the resulting population

  [Page 776]

  9-5. Typical textbook illustration of evolution by continuous anagenetic transformation of an unbranched population through time. This textbook labels the figure explicitly and exclusively as its icon of “evolution” itself, not of gradual­ism or any other subcategory of evolutionary change. From the standard paleontological textbook of my student generation, Moore, Lalicker, and Fischer, 1953.

  with a new Linnaean name (as species B), then where should we place the breakpoint between A and B? Any boundary must be arbitrary — if only by the illogic of the unavoidable implication that the last parental generation of species A could not, in principle, breed with its own immediate offspring in species B. (We may abhor human incest for social reasons, but we can scarcely deny the biological possibility — hence the perceived societal need for a taboo.) This problem generated a large, tedious, and fruitless literature, pri­marily because the issue always remained available, unresolved and therefore ripe for yet another go-round whenever a paleontologist needed to deliver a general address and couldn't think of anything else to say.

  Punctuated equilibrium took a radically different approach by admitting unresolvability under the stated assumptions, but then denying the focal em­pirical premise that new species usually (or even often) arise by gradualistic anagenesis. Instead, Eldredge and I argued that the vast majority of species originate by splitting, and that the standard tempo of speciation, when ex­pressed in geological time, features origin in a geological moment followed by long persistence in stasis. Thus, the classic and endlessly fretted “species problem in paleontology” disappears because species act as well-defined Dar­winian individuals, not as arbitrary subdivisions of a continuum. Species then gain definability because they almost always arise by speciation (that is, by splitting, or geographic isolation of a daughter population followed by genetic differentiation from the parental population), not by anagenesis (or transformation of the entire mass of an ancestral species). To be sure, a new species must pass through a short period of ambiguity during its initial differentiation from an ancestral population, but, in the proper scaling of macroevolutionary time, this period passes so quickly (almost always in the [Page 777] unresolvable geological moment of a single bedding plane), that operational definability encounters no threat.

  Of course, gradualists did not deny that speciation often occurs by branch­ing. They just didn't grant this process of splitting any formative role in the accumulation of macroevolutionary change for three reasons. First, they con­ceived speciation only as an engine for generating diversity, not as an agent for changing average form within a clade (that is, for the key macroevolu­tionary phenomenon of trends — see quotes of Huxley and Ayala, and Mayr's response, on p. 563). Trends arose by anagenesis (see Fig. 9-6), and speciation only served the subsidiary (if essential) function of iterating a favorable fea­ture, initially evolved by anagenesis, into more than one taxon — thus provid­ing a hedge against extinction.

  Second, they granted little quantitative weight to the role of speciation (splitting as opposed to anagenesis) in the totality of evolutionary change. In a famous estimate that became canonical, Simpson (1944) stated that about 10 percent of evolutionary change occurred by speciation, and 90 percent by anagenesis.

  Third, when gradualists portrayed speciation at all (see Fig. 9-7), they depicted the process as two events of anagenesis proceeding at characteristi­cally slow rates. Thus, they identified nothing distinctively different about change by speciation. Some contingency of history, they argued, splits a pop­ulation into two separate units, and each proceeds along its ordinary anagenetic way. Punctuated equilibrium, on the other hand, proposes that the geological tempo of speciation differs radically from gradualistic anagenesis. (We also argue, of course, that such anagenesis rarely occurs at all!)

  The theory of punctuated equilibrium therefore began as a faithful re­sponse to Schopf's original charge to Eldredge and me: to show how standard

  9-6. A standard illustration from Simpson (1944), showing that all trends, and all stability for that matter, originate primarily in the anagenetic mode — that is, by change during the lifetime of individual species, with branching serving pri­marily to diversify and iterate the favorable designs originated by anagenesis, and thus to prevent extinction of the lineage.

  [Page 778]

  9-7. Another illustration from the standard student paleontological textbook of the 1950's, with speciation depicted merely as two events of gradualistic change, following a separation of lineages. From Moore, Lalicker, and Fischer, 1953.

  microevolutionary views about speciation, then unfamiliar to the great ma­jority of working paleontologists, might help our profession to interpret the history of life more adequately. (As a best testimony to this unfamiliarity, I note that most paleontologists didn't even recognize the conceptual and ter­minological distinction between “speciation” defined as a process of splitting, and the accumulation of enough change by anagenesis to provoke the coining of a new Linnaean name for an unbranched single population.)

  In this crucial sense, the theory of punctuated equilibrium adopts a very conservative position. The theory asserts no novel claim about modes or mechanisms of speciation; punctuated equilibrium merely takes a standard microevolutionary model and elucidates its expected expression when prop­erly scaled into geological time. This scaling, however, did provoke a radical reinterpretation of paleontological data — for we argued that the literal ap­pearance of the fossil record, though conventionally dismissed as an artifact of imperfect evidence, may actually be recording the workings of evolution as understood by neontologists.* This empowering switch enabled paleon­tologists to cherish their basic data as adequate and revealing, rather than pitifully fragmentary and inevitably obfuscating. Paleontology could emerge from the intellectual sloth of debarment from theoretical insight imposed by poor data — a self-generated torpor that had confined the field to a descriptive role in documenting the actual pathways of life's history. Paleontology could now take a deserved and active place among the evolutionary sciences.

  The major and persisting misunderstanding of punctuated equilibrium among neontologists — a great frustration for us, and one that we have tried [Page 779] to explicate and resolve again and again (Gould and Eldredge, 1977, 1993; Gould, 1982c, 1989e), though without conspicuous success — involves the false assumption that if we are really saying something radical, we must be staking a claim for a novel mechanism of speciation, or for a different (read non-Darwinian) style of genetic change. When our critics then join this false assumption to our terminology of “unresolvable geological moments” or “punctuations,” they begin to fear that the dreaded specter of saltationism must be lurking just around the corner, trying yet again to raise its ugly head after such a well-deserved burial. Vituperation then trumps logic, angry as­sumption precludes careful reading, and punctuated equilibrium becomes a loathed doctrine of ignorant and grandstanding paleontologists who ought to stay in their own limited bailiwick, and get on with the job of documenting large-scale patterns generated by mechanisms that can be recognized and comprehended only by neontologists.

  But punctuated equilibrium makes no iconoclastic claim about speciation at all. The radicalism of punctuated equilibrium lies in the extensive conse­quences of it
s key implication that conventional mechanisms of speciation scale into geological time as the observed punctuations and stasis of most spe­cies, and not as the elusive gradualism that a century of largely fruitless paleontological effort had sought as the only true expression of evolution in the fossil record. The central intellectual strategy of our original 1972 paper rests upon this premise. We took Mayr's allopatric theory (as expressed in his classic treatise of 1963, deemed “magisterial” by Huxley), and tried to eluci­date its implied expression when scaled into geological time. We did not select this theory to fit a paleontological pattern that we wished to validate. We choose Mayr's formulation because his allopatric theory represented the most orthodox and conventional view of speciation then available in neontological literature — and we had been given the task of applying standard evolutionary views to the fossil record. I recognize, with 30 years of hindsight, that our original assessment both of Mayr's theory and of professional consensus may have been both naive and overly dichotomous, but we could not have stated our intent more clearly — the reform of paleontological practice by the para­doxical route of applying a fully conventional apparatus of neontological the­ory. We wrote (1972, p. 94): “During the past thirty years, the allopatric the­ory has grown in popularity to become, for the vast majority of biologists, the theory of speciation. Its only serious challenger is the sympatric theory. Here we discuss only the implications of the allopatric theory for interpreting the fossil record of sexually reproducing metazoans. We do this simply because it is the allopatric, rather than the sympatric, theory that is preferred by biol­ogists.”