Creating the Creators

If creation demands a visionary creator, then how does blind evolution manage to build such splendid new things as ourselves?

By Stephen Jay Gould|Tuesday, October 01, 1996
Perhaps Darwin was feeling a prick of conscience for having torn away the mainstay of human smugness with his documentation of evolution-- after all, we would never again be able to view ourselves as created rulers of a world made expressly for us. But for whatever reason, as he wrote the last paragraphs of his epochal Origin of Species, Darwin felt compelled to summarize the few bastions of traditional hope that evolution might buttress. Life’s long continuity could at least inspire some confidence in an extended future; and the pathway from squishy invertebrate to transcendent human must mean that evolution implies progress. Darwin wrote: Hence we may look with some confidence to a secure future of great length. And as natural selection works solely by and for the good of each being, all corporeal and mental endowments will tend to progress towards perfection.

This comforting view still defines our general cultural understanding of evolution and its implications. Two rarely questioned beliefs, however, stand at the center of this vernacular interpretation:

First, that even though evolution has produced an enormously complex tree of branching lineages, life as a whole has moved from a world inhabited only by bacteria to a modern biota now dominated by the paragon of neural advance, Homo sapiens. In this general sense, evolution is inherently and predictably progressive.

Second, that evolution, as Darwin taught us, works by a process called natural selection. This mechanism requires that survivors in the struggle for existence be better adapted to local environments. Thus, each step in an evolutionary sequence must feature a precise and intricate fit of organism to environment. Natural selection tracks environmental change, as organisms remain intricately adapted while gaining in general complexity.

A kind of wonderful irony, both instructive and amusing, permeates this common understanding of evolution as a progressive sequence of creatures, each exquisitely well adapted to local environments. We believe that such an account of evolution makes our own appearance both sensible and predictable. Indeed, our preference for viewing evolution as progressive and strictly adaptational arose largely to validate our own presence as an unsurprising consequence of nature’s intrinsic order. And now, the irony: to produce a creature with our structural and neurological complexity, evolution must be creative in the vernacular sense of this word--that is, evolution must be able to develop novel structures with previously unrealized functions. How else could a process that began with bacteria ever add the number of novelties required to evolve a human being (or any complex multicellular creature)? Yet, if evolution truly worked simply by fashioning exquisitely adapted creatures in an ascending series, humans could never have originated at all.

Precise adaptation, with each part finely honed to perform a definite function in an optimal way, can only lead to blind alleys, dead ends, and extinction. In our world of radically and unpredictably changing environments, an evolutionary potential for creative response requires that organisms possess an opposite set of attributes usually devalued in our culture: sloppiness, broad potential, quirkiness, unpredictability, and, above all, massive redundancy. The key is flexibility, not admirable precision. Ironically, then, to make us at all, evolution must work by processes contrary to the prejudicial hopes that we invest in Darwin’s legacy to validate our traditional status as lords of all by right of residence atop life’s pinnacle. Going even further, humans could arise only because evolution disproves what we have so long promulgated as our natural right and status. So choose your alternative: either evolution can work as a sop to our hopes, and we can’t ever arise; or evolution cancels our hopes and permits creatures like us to originate. What choice do we have since we do, after all, exist!

I would argue that three basic principles define and permit the creativity of evolution in the vernacular sense noted above--the capacity to originate novel structures and functions. All three share the common property of emphasizing flexibility and latent potential, rather than the admirably precise adaptation that serves as a paradigm for textbook illustrations of evolution--long giraffe necks to eat high leaves, showy peacock tails to win more female attention, complex mimicry to resemble another species, or a stick, or a leaf, or a piece of dung, all in order to fool predators.

Quirky shifts and latent potential. Consider the paradox of the peacock: the magnificent showy tail wins the most precious of immediate Darwinian advantages for individual males--more sexual access to more females, and more genes passed to future generations. But what else can you do with such an encumbrance? Change of circumstance and environment is the only constancy in evolution. If organisms are locked into complex structures with elaborate and inflexible functions, how can they evolve to meet these inevitable changes? And if they can’t evolve, they will die. Immediate success based on inflexible complexity therefore spells geologic doom.

Any elaborate and particular adaptation presents the same paradox. We contemplate the leaf-mimicking insect with amazement. How can natural selection devise such elaborate and detailed camouflage, down to the right color, the irregular external shape, and even each leaf vein mimicked by an insect wing vein? But what else could such an insect ever do but fool predators with this otherwise cumbersome device. What if the mimicked tree dies out locally? What if the predator moves away? Again, committed and inflexible intricacy for advantages of the moment implies limited potential for future change--a virtual guarantee of short existence in geologic time.

But how can species escape the paradox? Natural selection cannot anticipate the future and can adapt organisms only to challenges of the moment. If the necessary flexibility for future change cannot be evolved explicitly, then such lability must arise as a fortuitous side consequence of the ordinary operation of natural selection. Fortunately, the inherent architecture of genetic programs, developmental processes, and adult anatomies guarantees that any structure built by natural selection also maintains the latent potential for a wide range of other uses. The later exploitation of these latent potentials permits the evolution of novelty by quirky and unpredictable shifts in function--as fins become legs, forearms become wings, and big brains permit us to read and write.

This crucial principle of quirky shifts based on latent potentials exists in two versions, the first less radical and recognized by Darwin himself as a necessary theme to explain the evolution of novelty, and the second more unconventional because the latent potential exists in originally nonadaptive structures, thereby implying an important evolutionary role for features not directly built by natural selection--a proposition that strict Darwinians (I am not one) view with dismay.

The first may be characterized as the principle of Tires to sandals in the Nairobi recycling market. We who live in wealthy Western nations are not well equipped for appreciating this structural principle, vital both to human technology and to evolutionary change. We now throw out and buy again, rather than fix, a large range of objects from watches to radios. We also rarely rebuild materials for radically different uses. But poorer nations must recycle and reuse extensively, often finding a strikingly different purpose for material too worn out to perform in an original role.

I will never forget a fascinating visit to the recycling market of Nairobi, Kenya--where old telephone wire becomes jewelry, tin cans get sawed in half to be used as kerosene lamps, oil drum tops are beaten into large cooking pans, and treadless automobile tires become sturdy sandals. In fact, I own three pairs of sandals made from worn-out automobile tires-- one purchased in Nairobi; one in Quito, Ecuador; and one in India. Tires make very good sandals, but one would never argue that Goodrich (or whoever) built the tires to provide footwear in Third World nations. Durability for sandals is a latent potential of auto tires, and the production of such sandals defines a quirky functional shift.

Evolution works like the Nairobi market, not like the throwaway society of the wealthy West. You can evolve further only by using what you have in new and interesting ways. Organisms have no equivalent to currency for acquiring something truly new; they can reconstruct only from their own innards.

If organisms could not reuse old material in strikingly new ways, how could evolution ever produce anything novel? This classical dilemma has a fancy name dating from mid-nineteenth century debates following the publication of Darwin’s book: The problem of the incipient stages of useful structures. I prefer a catchier label based on a primary example: The 5-percent-of-a-wing problem. To cite the defining case: wings and feathers work wonderfully for flight; we can easily understand their adaptive function as fully developed organs. But how can a wing ever be constructed if evolution must pass through a long series of intermediary stages--for 5 percent of a wing confers no benefit whatsoever in flight. How can evolution ever build a bird’s wing from the forearm of small running dinosaurs if early stages in the putative transition cannot function for flight at all?

In a brilliant resolution of this conundrum, Darwin proposed that organs explicitly adapted by natural selection for one function also possess latent potential for working in other ways, if later environmental shifts encourage such an evolutionary response. (This latent potential arises as a fortuitous consequence of structural design, not as a direct and explicit result of natural selection. Evolution can’t anticipate an unknown future.) A row of feathers on a forearm (5 percent of a wing, so to speak) cannot aid flight, but feathers also work superbly as thermoregulatory devices for conserving heat. Thus, feathers may have evolved from reptilian scales for an initial function in thermoregulation-- and only later were they co-opted for flight when they became numerous and elaborate enough to provide aerodynamic advantages. (Experimental studies on insect wings--where the same evolutionary problem applies--show that tiny wings confer thermodynamic but no aerodynamic benefits. In a sequence of increasing wing size, advantages for flight kick in just when further growth stops providing any additional thermodynamic benefits.) Thus, structures evolved to retain heat have a latent potential for use in flight--an originally unexpected capacity that may become important as the organs get more elaborate or as environmental conditions change. Much of evolution’s novelty arises from the actualization of such latent potentials, not from slow and explicit improvement of an unchanged function by natural selection.

This principle of co-optation in the evolution of novel functions underlies much of evolution’s quirkiness and tendency to change course in unpredictable ways. If an intelligent extraterrestrial had visited the late Triassic Earth and watched a small running dinosaur sparsely clad in forearm feathers that worked only for thermodynamic effect, could the spaceman possibly have foreseen a future Earth with 8,000 species of flying birds? If an earlier visitor to the evolving Earthly zoo had seen a small lineage of fully aquatic fishes with lobe-shaped fins evolved only for scuttling along the bottoms of ponds, could he have foreseen an entire history of vertebrate evolution on land, and the eventual transformation of the forward pair into hands capable of sitting at a typewriter and composing this article?

The second version of quirky shifts based on latent potentials may be called The spandrels of San Marco, or Milton’s principle of ‘They also serve who only stand and wait.’ The principle of the Nairobi recycling market lies entirely within the larger and conventionally Darwinian theme of evolution by continuous adaptation--for this principle speaks only of a quirky shift in function from an original use to something quite different. But must every novel function be co-opted from some previous and different adaptation of the same organ? How about the possibility of co-opting a later use from the latent potential of features that didn’t have any adaptive value when they first arose? Such a principle, if common and important in the history of life, would add an interesting twist to conventional evolutionary theory--for the strictly Darwinian approach now generally favored views adaptation as ubiquitous, and the only important reason for evolution.

I use an architectural analogy as my name for this principle. Venice’s celebrated Basilica of San Marco contains several hemispherical domes, each mounted on four rounded arches. As a structural necessity, not an analogue of adaptation, such a geometric arrangement must yield four tapering triangular spaces--one at each corner below the dome where two arches meet at right angles. The basic decision to mount domes on four arches may be viewed as an analogue of adaptation--builders knew that such an arrangement was structurally sound and aesthetically pleasing. But once this primary decision has been made, the four tapering triangular spaces must come along for the ride as necessary architectural by-products, and not for any specific utility in themselves. (Such filled-in spaces between arches, domes, pillars, and so forth are called spandrels.)

The four triangular spandrels beneath each dome are by-products of a basic architectural decision, not adaptations in themselves. But since the spandrels must be present, and since they occupy a good deal of space, some later (and quite clever) use may be found for them. The walls and ceilings of San Marco--including the spandrels--are covered with beautiful mosaics. Four necessary spandrels suggest possibilities for Christian themes well suited to the preexisting space--and two of San Marco’s domes contain lovely representations of the four evangelists (including the basilica’s patron, Mark) in the spandrels.

But consider what a foolish error we would make if we noted the superb fit of the mosaic design to the space (a secondary adaptation to a preexisting geometry) and then said: Now I know why the spandrels exist; they were built to house the evangelists. We recognize such an argument as ridiculously backward: the four spandrels formed as a nonadaptive by- product of a larger architectural scheme and were later co-opted for a secondary use in representing a key theme of Christian faith.

Similarly, any biological adaptation also produces a host of structural by-products, initially irrelevant to the organism’s functioning but available for later co-optation in fashioning novel evolutionary directions. Much of evolution’s creative power lies in the flexibility provided by this storehouse of latent functional potential.

To cite a pair of simple, but intriguing, examples: as a snail builds its shell by winding around an imaginary axis of coiling ( just as Earth turns on an imaginary axis of rotation), a long and narrow cylindrical space, called the umbilicus, must form in the position of the axis. The umbilicus is a geometric necessity--a consequence of winding a tube around an axis--not an adaptation. But since this space must form as a spandrel, the snail can later co-opt the umbilicus for novel functions. In a remarkable example, one group of snails pushes fertilized eggs into the umbilicus, thus co-opting this space as a well-protected brooding chamber!

The extinct giant deer (popularly known as the Irish elk) grew the world’s largest antlers, up to 13 feet across and 75 pounds in bulk on a skull that weighed only 5 pounds. To hold up such a maximally heavy head, the Irish elk evolved powerful muscles and ligaments running from the neck to the vertebral column at the shoulder. To provide enough attachment area for these ligaments, the Irish elk evolved (as do many lineages of mammals with large and heavy heads) high projecting spines on the shoulder vertebrae. These spines necessarily produce a broadly raised area on the animal’s back in the shoulder region. Many large mammals grow such a raised area as a geometric by-product of the vertebral spines beneath, not as a direct adaptation for anything. In the Irish elk, this raised area later evolved into a large and distinctive hump, accented by dark patches of color and radiating lines. This unique hump is presumably a secondary elaboration (perhaps for sexual display or as a recognition device) of an originally nonadaptive structure--the raised area necessarily produced by the vertebral spines beneath. Interestingly, we only know about the Irish elk’s hump because our Cro-Magnon ancestors painted these animals, colors and all, on cave walls. Fatty humps, made entirely of soft tissue, do not fossilize.

As a striking example much closer to home, most of the distinctive mental features that form our human nature probably arose as co-opted spandrels, not as direct adaptations. I don’t doubt that our brains reached their unparalleled size and complexity by an ordinary process of natural selection, working for some set of functional advantages that higher mentality provided. But even if selection and adaptation produced the increased size, the human brain, as nature’s most elaborate neural device, also gained the capacity to do thousands of additional things as structural by-products of this increased complexity, and not as direct adaptations. For example, human brains obviously didn’t get large so that we might learn to read or write--for these functions arose tens of thousands of years after our brains reached their current size. Yet reading and writing (and a thousand other attributes of mind) have become crucial components of human life and nature. Thus, without the flexibility imparted by our spandrels (the latent potentials of our evolved mental complexity), we would not have become such a miraculous nuisance in the history of this planet.

Redundancy. Consider another paradox that will help us explain why exquisite adaptation cannot provide a primary source of evolution’s creativity (but will, instead, usually act as an impediment to substantial evolutionary novelty). Life began with bacteria, and bacteria possess relatively few genes compared with humans and other complex multicellular creatures. Now suppose that you are an optimally adapted bacterium in this ancient world that knows nothing more complex. You have been honed to immediate adaptive perfection by natural selection; you are therefore, to use an anachronism in your time, happy as a clam. You are the meanest and leanest of optimal machines. You contain no slop, no extras. Every one of your genes does its one (or few) things superbly well. You couldn’t be better--but how could you ever change, at least in any substantial way?

Oh, evolution might tinker a bit here and there: a reaction rate or a metabolic pathway may require a little fine-tuning if environments change. But nothing major can be altered because you need every gene you possess for something vital to your life. To make anything truly different, you would have to adapt one of your existing genes to a novel use. But how then could the old and still necessary function be performed? In other words, you are stuck--optimally adapted to be sure, but in a permanent rut, mired at a structural level that cannot be transcended.

To resolve this paradox, we must recognize that this ideal of adaptive optimality--the leanest and the meanest--works as badly in evolutionary reality as it does for human morality. To undergo creative evolutionary change, organisms need the flexibility provided by the opposite phenomena of slop, redundancy, and latent potential. But how does this glorious messiness arise if natural selection knows no future (or any conscious intent for that matter) and can work only to improve immediate adaptation? Fortunately, structural constraints and principles, independent of natural selection, preclude the development of lean and mean optimality and therefore permit evolution (in the long run) to overcome its own tendency to limiting specialization (in the short run).

All biological structures (at all scales from genes to organs) maintain a capacity for massive redundancy--that is, for building more stuff or information than minimally needed to maintain an adaptation. The extra material then becomes available for constructing evolutionary novelties because enough remains to perform the original, and still necessary, function.

In organs and body parts, the principle of redundancy finds primary expression in the concept of overdesign, or margin of safety. Two (or more) structures often perform the same basic function. This generosity may benefit an organism in the immediate present ( just as a spare tire saves many a driver), but extra capacity also permits creative evolution in novel directions--because the spare tire can turn into something marvelously different, and the car still runs.

Consider two crucial examples in the evolution of vertebrates: Since a fish’s air bladder is the same organ as a mammal’s lung, many people assume that air bladders evolved into lungs (because mammals are supposedly higher than fishes). In fact, evolution took the opposite path: the lungs possessed by all early fishes became air bladders in most modern fishes but remained as lungs in the ancestors of terrestrial vertebrates. Since more than half of all vertebrate species are fishes with air bladders, this creative evolutionary transition represents a key event in the history of vertebrates ( however demoted or ignored because we are so hung up on our own supposed superiority and don’t like to credit fishes for altering a primitive organ like our lung to a different and highly successful function).

But how can a lung become an air bladder? How could such a transition occur without suffocation of the intermediary forms? A famous song tells us that fish gotta swim, but they also gotta breathe. The principle of redundancy resolves this riddle. Early fishes breathed with two organs: gills and lungs (as do modern lungfishes, technically known as dipnoans--meaning two breathing). Thus, fish could continue to breathe with gills while lungs evolved to the novel function of air bladders.

The malleus and incus (hammer and anvil) bones of the mammalian middle ear evolved from precursors that articulated the jaws of our reptilian ancestors. But how could such a creative transition occur? A vertebrate cannot survive with an unhinged jaw. Creationists have used this argument to claim that evolution is impossible and that mammals must have been specially created, not evolved from reptiles. But the principle of redundancy resolves this problem as well--not only by a clever theoretical argument but as a demonstrated fact, because the intermediate forms have been found as fossils. The intermediates evolved a double jaw joint--one between the old reptilian bones that would later enter the mammalian ear, and the other between the two bones that now form the jaw joint of mammals. Thus, one joint could disappear as evolution moved and transformed the bones for a different primary function of hearing--while the other joint continued to function in the necessary task of articulation.

At the genetic level, the principle of redundancy has an even more general expression in the phenomenon of gene duplication. If, as in many bacteria, each gene exists as a single copy and codes for an essential enzyme or protein, how could substantial change ever occur--for any major shift in function would annihilate an original use that remains essential for life?

The solution to this most general statement of a central paradox resides in a property of genetic material in eukaryotic organisms (nonbacterial creatures with complex cells, including such unicellular forms as amoebas and paramecia, and all multicellular organisms). For a set of dimly understood and complex reasons, the genetic programs of eukaryotic organisms maintain a high level of redundancy, largely because many genes tend to duplicate themselves within the genetic program and therefore exist in multiple copies. As natural selection has no consciousness and cannot work for future benefits, this repeated dna does not originate in order to provide the requisite flexibility that creative evolutionary change requires. Rather, such creative flexibility emerges as an evolutionary legacy, a fortuitous and unintended side consequence of dna’s tendency to produce multiple copies within the genetic programs of eukaryotic organisms. When multiple copies exist, the essential function can be maintained by some copies while others become available for evolutionary modification in substantially new and creative directions. If repeated dna did not exist for its own immediate reasons, our world would probably be inhabited only by organisms of bacterial grade--a perfectly good alternative world to be sure, but one that could not include the writer and readers of this essay.

Selected flexibility. The first two principles are entirely general in evolution. They share the property of providing flexibility (by latent potential and redundancy) against the tendency of natural selection to produce an exquisite fit of form to environment, thereby dooming the organism in the long geologic run as environments inevitably change in major ways. But we should also ask whether, in certain cases, natural selection can work directly for flexibility. The answer is yes, though perhaps not often--but the yes applies to the case of greatest parochial interest for us: namely, to human evolution.

As a general statement, natural selection operates to produce a better fit of organism to prevailing local environment. In most cases this fit entails greater specialization and consequent loss of flexibility. ( Therefore, in the key argument of this essay, flexibility must arise as an unintended side consequence of natural selection--that is, from such structural principles as latent potential and redundancy.) But if better local adaptation can sometimes arise by increased flexibility, then natural selection might also operate directly toward such a result. The unique cognitive abilities imparted to us by our large brains, particularly our capacity for learning, may have placed us in an unusual situation favoring direct selection for flexibility. The basic argument has a long pedigree (an oldie but goodie in my book), dating at least (in a pre-evolutionary version) to the great seventeenth-century English philosopher John Locke.

Most mammalian babies grow up quickly--and need to do so, given the greater vulnerability of juveniles to all causes of mortality. Humans, however, have evolved an inordinately extended period of childhood dependency on parents and other adults. Moreover, nearly every aspect of our growth seems excessively slow and delayed relative to patterns in other mammals. We don’t even become sexually mature until well into our second decade of life. What possible advantage could such delay impart in a Darwinian world that measures success in terms of reproductive prowess?

Locke’s argument, in its later evolutionary version, holds that humans developed this profound delay because our unique cognition requires long periods of learning to achieve proper use. Most mammals reach adulthood rapidly and leave the company of their parents and other potential teachers. (In another relevant mammalian pattern, only juveniles indulge in play behavior and retain flexibility for learning; adults become set in their ways.) But humans need a long period of socialization and learning to develop their mental capacities, and this can best be achieved by an extended childhood, with retention of the flexibility so common in mammals during this period alone.

An evolutionary process called neoteny (literally, holding on to youth) works by selection for slowing down rates of development, leaving the adults of descendants with features that characterized juvenile stages of their ancestors. Many technical arguments beyond the scope of this essay indicate that neoteny has been a dominant theme of human evolution. In this sense, and speaking only partly metaphorically, human adults are childlike. We have evolved an extended childhood, presumably for the advantages imparted by prolonged flexibility for learning. And we retain some of this crucial flexibility into an adult stage that, in most mammals, entails rigidification of behavior.

In summary, then, Homo sapiens, this most peculiar, powerful, and dangerous of current species, originated because evolution’s sloppy flexibility permits complex creatures to arise--and not (as we might like to believe) because we were meant to appear as a natural result of inevitable improvement constructed by a process (natural selection) that continually makes successful creatures better and better. We are here because distant unicellular ancestors evolved multiple copies of many genes, thereby allowing some to change while others retained needed functions. We are here because the fins of ancestral fishes held latent potential for transformation to the different role of bearing weight on land; because reptilian ear bones could be co-opted to become mammalian hearing bones; and as a result of a thousand other quirky and unpredictable transitions based on the inherent potential of anatomic structures to work in ways that were not the selected function of their original design. And we are here because our odd mentality set an unusual context that placed an explicit selective value upon flexibility. Every complex species owes its unpredictable existence to the sloppy sources of evolution’s creativity. We are quirky, if glorious, accidents not to be repeated on this planet. May we, then, take some care for our own fragility.
Comment on this article