Self-help for a hedgehog stuck on a molehill.

struggle to inform the public about Darwinian evolution

by Stephen Jay Gould: Review of Climbing Mount Improbable by Richard Dawkins (W.W.Norton, New York and Viking Penguin, London. 1996). In Evolution (Vol. 51 June 1997 No. 3).

The Value of Metaphor

In a famously ambiguous, yet widely cited epigram, the 7th-century B.C. Greek poet Archilochus contrasted two disparate mammals in metaphor: "The fox knows many things, but the hedgehog knows one great thing." This epigram has caught the fancy of intellectuals in part because the human mind tends (however lamentably) to parse complexity into oversimplified dichotomies but also, and more specifically, because we sense that the favored working style of scientists and other scholars tends either toward the fox's mode of delighting in plurality and a wide range of explanatory tactics, or the hedgehog's strategy of trying to render all the overt variety of nature by one great and overarching truth.

Our allegiances to an insectivoran or carnivoran tactic depend strongly on scale and intent. The truth of evolution and the power of Darwinian explanation have been so liberating and transforming in the history of Western thought - while popular drive to oppose, or simple failure to understand, remain so entrenched (given the threat and uncongeniality of evolution to so much held dear both by pysche and society) - that writings for the general public must lean toward the hedgehog's task of documenting and defending the veracity and explanatory range of evolution itself. This genre, extensively pursued both by Dawkins and by me, enjoys a long and honorable legacy extending back to two Darwins (Erasmus and Charles) and on to nearly all natural historians with writing skills (including two Huxleys, also grandfather and grandson, Haeckel, Haldane, Medawar, Eiseley, Thomas, Simpson, Dobzhansky, Wilson, and others). The title of one of my own popular books, An Urchin in the Storm, invokes this image of an unfazed evolutionist fending off the slings and arrows of outrageous opposition from creationists, new-agers, demagogues, and other motley philistines and irrationalists from the persistent tradition of American anti-intellectualism.

In this important uphill battle for informing a hesitant (if not outrightly hostile) public about the claims of Darwinian evolution, and for explaining both the beauty and power of this revolutionary view of life, I feel collegially entwined with Richard Dawkins in a common enterprise. His metaphor of the "blind watchmaker" provides a brilliant epitome of Darwin's central principle, and no one now writing has done nearly so well in explaining the paradox (to the public mind, at least) of how a process without intentionality, and working only by a "selfish" principle of reproductive success, can yield organisms of such intricate, adaptive design.

Thus, although I applaud the extension of this shared concern in Climbing Mount Improbable (for example, Dawkins's passages, page 77 et seq., regarding the vital but limited role of chance in Darwinian theory, are particularly lucid), I must focus this review, written for professional colleagues in our leading trade journal, on how Dawkins's particular view of evolutionary theory affects his own concept of life's history and impacts his effectiveness in public presentation.

Our disagreements on these issues are well known, and I shall not abuse this forum to air a twice-told tale in general terms. I shall, instead, concentrate on the "plot line" and logic of Climbing Mount Improbable - a volume that I find admirable on the whole (for a generally lucid account of adaptation, presented with striking examples in Dawkins's attractive prose style) but seriously flawed by an ill-chosen, threefold metaphorical apparatus developed to carry the book's central logic. I find this apparatus more misleading than helpful in two crucial ways: first, in promoting an erroneous central image (with the first metaphor); and second, in a logical inconsistency that precludes union between this first metaphor and the other two leading images (and therefore introduces incoherence into the flow of the book's main argument).

I do not criticize the explicit use of metaphor to carry a central theme, for I regard such devices as indispensable in human reasoning. No one in the entire history of science ever made better use of metaphor than Charles Darwin himself. Several of Darwin's striking, and entirely helpful, images continue to play a central role in our teaching and understanding. Who can forget Darwin's wonderful image of the tree of life, placed at the climactic moment of his entire text (the summary of the argument for natural selection at the end of Chapter 4), and used both as a memorable literary trope and a stunning device for uniting the two crucial themes of evolution as genealogy (connectivity and branching of the tree) and natural selection as a primary mechanism of change (competition among growing branches, with extinction as a dropping of branches): ". . . so by generation I believe it has been with the great Tree of Life, which fills with its dead and broken branches the crust of the earth [a lovely image for fossils], and covers the surface with its ever-branching and beautiful ramifications." Or another great metaphor that captures the essence of Darwin's "tough" mechanism contrasted with our superficial hopes about the intrinsic goodness and overarching order of nature - the image of an overt "face" covering such a different operation within: "We behold the face of nature bright with gladness. We do not see, or we forget, that the birds which are idly singing round us mostly live on insects or seeds, and are thus constantly destroying life. . . ."

Metaphors are indispensable in human thinking, and particularly useful in explaining complex and unfamiliar material to nonprofessional audiences. The human mind is a wondrously obtuse and circuitous instrument. We think of ourselves as logical, as able to move in deductive order through a set of arguments from beginning to ineluctable conclusion. But we almost never work in such an idealized way. We always encounter obstacles and chasms that require a creative leap - a mode of imaginative bridging that America's greatest philosopher, C. S. Pierce, called "abduction," or "leading from," in contrast with the recognized and tamer methods of deduction, or logical sequencing, and induction, or generalizing from empirical cases. We need metaphors (literally "carriers from one place to another" in the original Greek; a moving truck, in modern Greece, is a "metaphora") to make such imaginative leaps.

But metaphors also represent twin dangers, literary and intellectual. On the literary side, some metaphors are so silly or obtuse that we end up laughing or scratching our heads rather than gaining enlightenment. (Among famous examples, I would submit the opening verse of Psalm 42: "As the hart panteth after the water brooks, so panteth my soul after thee, O God." This image plunged me into special confusion because I first learned Psalm 42 through oral recitation, pictured "heart" instead of "deer," and then couldn't, for the life of me, figure out why my innards should yearn for a river - not to mention an additional puzzlement about why my etherial essence should pant like a dog in the first place). On the intellectual side, many images (including some crucial metaphors used by all evolutionary biologists) are flat wrong - the ladder of life, the parsing of ants into queens and workers, or (in a newcomer that both Dawkins and I have singled out as especially misleading), Gaia conceived as the largest evolutionary individual in a hierarchy from nucleotides to the whole earth - a nonsensical image for many reasons but invalid prima facie if only on the obvious basis that Darwinian individuals must participate in a genealogical system, and planets don't have kids.

The distinction of metaphors - between useful and ineffective, brilliantly insightful and dismally misleading - therefore becomes an important consideration in making or evaluating any complex intellectual argument. Dawkins constructs the logic of Climbing Mount Improbable around three metaphors - Mount Improbable itself as a flagship, the Museum of Possible Forms, and his old hobbyhorse of organic bodies as Lumbering Robots housing the "real" units of evolutionary causality: the resident, replicating genes.

Impermissible Mount Improbable

As the preeminent hedgehog among contemporary Darwinians, Dawkins has staked his career on a belief that just about everything interesting in the evolutionary history of life either is, reduces to, or flows from the production of "one great thing" - OAC, or the "organized adaptive complexity" of biological forms - by the unitary principle of natural selection acting causally at the lowest possible level of selfish genes struggling to maximize their replicative success. In stating such a first principle as a comprehensive article of faith, Dawkins might take refuge in a scriptural source - a famous line from the preface to the Origin of Species, where Darwin states that the "classical" data of embryology, geology, and geographic distribution surely suffice to prove the fact of evolution, but that "such a conclusion, even if well founded, would be unsatisfactory, until it could be shown how the innumerable species inhabiting this world have been modified, so as to acquire that perfection of structure and coadaptation which most justly excites our admiration." But Darwin also knew (see chapter 4 of the Origin on mechanisms of speciation and diversification and his longest two-volume work of 1868, The Variation of Animals and Plants under Domestication, for sources and causes of channeling by variational constraint) that the production of adaptive design by natural selection could only serve as a necessary, but by no means sufficient, principle for explaining the full history of life for nature, above all, is a fox.

I share Dawkins's conviction that a popular presentation of evolutionary theory must feature a convincing explanation of what, to nonprofessionals (and especially to people with a skeptical or hostile attitude toward evolution), must inspire the greatest feelings of doubt and implausibility: how can a process without intentionality, and working only with raw material based on "random" variation in a system of causality that lacks any overarching directional principle, ever produce anything as complex (and as beautifully functional) as an eye? I also share Dawkins's pedagogical insight that such paradoxical concepts cannot be understood without a guiding metaphor to break through conventional beliefs that render the aim of instruction so utterly implausible at first hearing. (Consider for example, how geologists have always used a metaphor - the "cosmic mile" or "cosmic calendar" to cite the two classics - to inculcate the uncomfortable truth that earthly time comes in billions but that human history occupies only the last sliver of such cosmic immensity). But proper metaphors must accurately express the principles of the paradoxical theory thereby illustrated - and Mount Improbable, the conceptual foundation of Dawkins's book, is deeply invalid at its core, not just permissibly simplified for the appointed task.

Hill-climbing metaphors have a long history in evolutionary biology; they also have a locus classicus of conventional misuse (as beautifully documented by Will Provine in his biography of Sewall Wright). In the initial formulations (still most useful and most valid), adaptive "peaks" are genetic combinations (in their environmental expressions) and are therefore "of" the organism in relationship to surroundings and explicitly not a map of external features in the outside world. In the classic misuse, however, adaptive peaks are metaphorical places in an abstract "design space," biomechanically defined. These peaks exist prior to, and entirely apart from, the organisms that strive toward the summits. In short, the mountain range is a map of engineering solutions, with the highest peak as a global optimum and lower peaks as workable places where organisms may become stuck because natural selection can only push them upslope, and any passage to a higher peak can therefore only be accomplished if some other process (Wrightian genetic drift in the original and classical formulation) can move a population downslope into a valley between this suboptimal peak and the global summit of Mount Improbable.

Dawkins devises the metaphor of Mount Improbable for the worthy purpose of convincing skeptics that evolution can fashion complex designs one step at a time, with continuous adaptation maintained throughout. In Dawkins's image, skeptics first look at Mount Improbable from the wrong side and only see a sheer (and truly insurmountable) wall before them: "The towering, vertical cliffs of Mount Improbable can never, it seems, be climbed. . . . [Mountaineers] shake their tiny, baffled heads and declare the brooding summit forever unscalable" (page 73). But if skeptics then peek around the other side, they will find a gradual and continuous way: "There they would find not vertical cliffs and echoing canyons but gently inclined grassy meadows, graded steadily and easily towards the distant uplands. . . . The sheer height of the peak doesn't matter, so long as you don't try to scale it in a single bound. Locate the mildly sloping path and, if you have unlimited time, the ascent is only as formidable as the next step" (page 73). In successive chapters, Dawkins then applies this metaphor to three structures of mounting complexity: spider webs, wings of several phyla, and eyes of even more phyla. Of course, biology as engineering applies particularly well to such relatively isolatable, unambiguously optimizable organs with an evident primary function; but this brand of main-line adaptationism fares more poorly for most of the rest of biology.

All professional evolutionists understand the fallacy in trying to understand the differential occupation of organic morprospace as an adaptive mapping to a preexisting external landscape. We can all articulate the central error of such a perspective: since organisms help to create their own environments, adaptive peaks are built by interaction and undergo complex shifts as populations move in morphospace; organisms cannot climb stable mountains of an engineer's fancy. Dick Lewontin, my "partner in crime" for such issues, refers in our joint course on evolution to organisms moving on an environmental trampoline (and thus fashioning, to a large extent, their own landscapes). Populations propelled by natural selection are "myopic mountain climbers in a nightmare world," not modern hiking technocrats facing a stable and surmountable challenge with all the fanciest equipment purchased from Patagonia.

Since we all know this perfectly well, our only excuse for metaphors like Mount Improbable must lie in claims that such reductionist imagery constitutes a permissible simplification for legitimate instruction or pedagogical effect. But since evolution is the most contingent and pluralistic, the most irreducibly historic, of all major processes known to science, such imagery based on separation, linear causal chains, and predictable outcomes seems especially misleading in this field above all others. Shall we be satisfied with a primary image of populations as passive lumps, pushed by an external force called natural selection up peaks of a fixed landscape built by a firm of biomechanical engineers? I find such Dawkinsian imagery both more misleading than helpful and also unimaginatively conventional as a putative basis for the discipline that we all love and find so rich and exciting in its manageable subtlety: "Natural selection," Dawkins writes, "is the pressure that drives evolution up the slopes of Mount Improbable. Pressure really is rather a good metaphor. We speak of 'selection pressure,' and you can almost feel it pushing a species to evolve, shoving it up the gradients of the mountain." Surely, we can do better.

The Admirable but Unintegratable Museum of Possible Forms

And Dawkins does do better. After devoting two-thirds of his book to the metaphor of Mount Improbable, Dawkins shifts radically to a new metaphor of almost opposite intent - a most welcome reversal of the hardline view that he has promulgated throughout his career and a proper acknowledgment that organisms also shape their own history by positive forms of constraint "pushing out," as it were, from the inside. The first of two chapters, "the museum of all shells," uses Raup's famous analysis to explore the decidedly "clumpy" occupation of morphospace by coiled shells that grow as logarithmic spirals. Adaptation (not at all inappropriately) remains the central theme - for Dawkins emphasizes the domains of attainable form that are effectively unpopulated, presumably because shells of such shapes generally fail as adaptive designs (too weak to avoid predation, too clunky for effective motion, or too "expensive" to build, for example). But at least this metaphor departs sharply from Mount Improbable's passive mapping of form upon a static, external engineer's landscape to explain the patchy occupation of organic morphospace - and moves instead to a more adequate image of external design working in concert with internal rules of construction.

The subsequent chapter, entitled "kaleidoscopic embryos," then moves Dawkins so far from his former certainties that he virtually falls into the lap of those (including yours truly) who have long criticized his externalist adaptationism and have urged the importance of constraint as a positive evolutionary force for channeling (not only as a negative factor of limitation) and of natural selection as a hierarchical process working simultaneously at several levels of Darwinian individuality (from genes, to organisms, to demes, to species, to clades) and not only by Dawkins's ultimate (and logically false) reductionism to the selfish gene. This chapter explores how rules of embryonic construction (symmetries and serial repetitions, for example) can act as helpful promotors of evolutionary novelty - in other words, as constraints in the positive sense. This insight then leads Dawkins to embrace (however gingerly) the two implied and conjoined heresies that had been anathema to him: (1) a potential role for phenotypic saltationism based on underlying rules of continuationist style (I particularly applaud his willingness to consider the possible origin of shovelnosed lobsters by a homeotic mutation that would have to be ranked as a "hopeful monster"); and (2) an acknowledgment that higher-level selection may play an important role in macroevolution (through a principle that he calls, quite fairly in my view, "the evolution of evolvability").

Dawkins certainly recognizes the apostasy (with respect to his old hardline adaptationism) of these new explorations. He refers to his proposal "about some embryologies being 'better at evolving' than others" as "that rather odd suggestion" (page 225) and then explicitly (and correctly) acknowledges this mode as an endorsement of hierarchical selection theory. He writes on pages 224-225: "Moreover, a form of higher-level selection - what I have previously dubbed 'the evolution of evolvability' - may lead to the world becoming populated by types of creatures whose embryologies made them good at evolving. Coming from a dyed-in-the-wool Darwinist like me, this might sound like the rankest heresy. Natural selection is not supposed, in nice neo-Darwinian circles, to choose among large groupings." Dawkins then follows this assertion by recognizing its key implication for labeling developmental constraint as a potentially positive force in evolution: "But we shall see in this chapter that the symmetry constraint can turn out to be an enrichment: the very opposite of a restriction" (page 226).

The last paragraph of this chapter reiterates these themes in a manner almost suggestive of a personal epiphany. As a former anathamee, I can only cheer from the sidelines and say "bravo and welcome."

The central message of this chapter is that kaleidoscopic embryologies, whether working through segments and clusters of segments arranged in a line from front to rear as in an insect, or through "mirrors" of symmetry as in a jellyfish, are paradoxically both restrictions and enhancements. . . . Kaleidoscopic embryos have what it takes to inherit the earth. Whenever a major shift in kaleidoscopic mode or "mirror" has spawned a successful evolutionary radiation, that new mirror or mode will be inherited by all the lineages in that radiation. This is not ordinary Darwinian selection but it is a kind of high-level analogy of Darwinian selection. It is not too fanciful to suggest as its consequence that there has been an evolution of improved evolvability (page 255).

Although I applaud this section, and hope that Dawkins will extend his true (in both senses) apostasy to book length in future endeavors, the juxtaposition of this second metaphor with the earlier and far longer material on Mount Improbable produces a volume that cannot cohere in its overall logic of sequential presentation. And when Dawkins then appends to these two chapters (with their welcome and expansive metaphor justifying the hierarchical theory of selection) a straight line (and, in my view, retrograde) defense of his worn-out and illogical "selfish gene" theory (chapter 9, entitled "the robot repeater," and featuring his jejune metaphor of organisms as lumbering robots built by and for their selfish genes), then the general incoherence only increases - and a nonprofessional reader, lacking a historical scorecard to map these disparate arguments upon the constraints of Dawkins's personal intellectual genealogy, must really end up confused.

Because nature, as a fox, knows many things, we must be vulpine as well to map our evolutionary understanding upon her complexities Machiavelli, who knew a thing or two about practicalities, stated that the prince "must imitate the fox . . . to recognize traps." Mount Improbable is a trap. The plurality of external adaptation with internal (and largely positive) constraint, and of numerous and simultaneously effective levels of selection, provides a liberation. If carnivores can live happily with herbivores (wolves with lambs, leopards with kids) on the noblest of all improbable mountains (see Isaiah, chapter 11), then the hedgehog's hardline can surely yield to the fox's wise multiplicity.

...please excuse the interruption. Read this excerpt from Mark Ridley's mainstream textbook Evolution, to understand why the gene is the ultimate unit of selection, even when selection occurs at many levels in the hierarchy.


John Catalano
site map
what's new?
more pages
Behe's box
C is for Creation
the Gould Files
book of month
the green room
best & useful
"new?" central
science news
evo & creation
artificial life
other science
art, music, +