2 Two Evolutionary Principles

As a preliminary to our later discussions, I shall now enunciate two basic principles governing the evolutionary process. The principles are closely related and are commonplaces in evolutionary theory. Given their familiarity, I do not think it is necessary to defend these principles, but clearly articulating them can be useful when considering various putative evolutionary explanations. I shall call these principles the principle of ancestral preservation and the principle of incremental adaptation.

The first principle is well captured in a resonant and carefully honed phrase used by Darwin in the very first sentence of The Descent of Man: “modified descendant of some pre-existing form.” Darwin is about to argue that man is just such a modified descendant of a pre-existing form, namely an ape, and he does so by observing that man shares many of the characteristics of (other) apes. The hypothesis is that this is because the process of natural selection has preserved the earlier forms in the later forms. Natural selection has operated on an ancestral stock and modified that stock in various ways, but the characteristics of the earlier stock have persisted though time and now (partly) constitute the present stock. In other words, preexisting forms have been inherited, so that there are “remnants” of the earlier creatures in the later ones. These remnants may be adaptive or neutral or even slightly disadvantageous; but the reason they are there is simply that the earlier forms had them and passed them on genetically. Later forms modify earlier forms, retaining them over time.

This principle stands in contrast to the theory Darwin is opposing, namely that each species was created separately. Here there is no preservation of ancestral traits, simply because the present species did not evolve from an earlier species. There is no mechanism whereby the evolutionary process must use earlier forms as raw material on which to build later forms, since all species are created independently. Thus, according to the “separate creation” theory, it is just an accident that species share many essential characteristics (God could have created them as all completely unlike each other). On Darwin’s theory, by contrast, the present embeds the past, so that later species retain and contain features of earlier species. The long evolutionary history of a species is recorded within the present makeup of the species. The more recent the descent, the more will be preserved, but even remote ancestors can leave their imprint on subsequent creatures, marking them as having had a particular evolutionary history. This preservation of features will be most evident in anatomy, but in principle it also applies to behavior and psychology. Traces of the past will persist in the later creatures, shaping their bodies, minds, and behavior. Thus biologists often speak of “vestigial” features in human organisms that no longer serve any function—the appendix, the earlobes, body hair, toe nails perhaps. But even the most adaptive organ is an inheritance from ancient forms; it is not created anew when a species evolves. For example, the feet and wings of mammals and birds ultimately derive from the fins of fish, and accordingly share much of their anatomy with earlier fins. Apparently, the bones of the inner ear in mammals derive from the jawbones of ancestral reptiles.¹ Quite literally, earlier forms, often of animals now extinct, are alive in present species. Evolution works with the materials it has, and its mechanism is the modification of these materials over time. Every species now in existence derives from a preexisting species that has been modified over time; no species is created de novo, and no organ lacks its evolutionary lineage. The evolutionary process brings in the new by tinkering with the old, and the old is preserved—pickled, we might say—in the new.

This preservationist picture can be compared with the growth of a city over time. There is the most ancient part of the city, often crude and simple, but this has been worked over to create more recent structures—not demolished and replaced, but modified. A single building might thus incorporate architectural forms from widely separated eras. What was once a stable might now be converted to housing; a jail might become a hotel. Relics from the distant past are everywhere, mixed in with more recent structures. The old persists within the new, sometimes usefully, sometimes accidentally. The difference between a city and a species is that the city can be intentionally planned, but evolution works by trial and error. And evolution can’t demolish a species and then replace it with another from scratch (as God could if He didn’t like the first species He created). Evolutionary change is thus inherently conservative and haphazard. It is rather like geological change: one stratum of rock can be superimposed on another, but the original rock survives, sometimes hidden, forming “bedrock.” The deeper you dig, the further into the past you travel. This is why it is often noted that human DNA contains the remnants of DNA from ancestral species, which may no longer be active, but are preserved because they are not harmful to the organism. There is no reason why this principle should not be as true for the mind as for the body—relics of earlier minds may persist in descendant minds, even if no longer accompanied by the original environment in which they evolved. There is a kind of “law of inertia” at work in the evolutionary process—persistence unless actively thwarted.²

The principle of incremental adaptation states that evolution works gradually, not radically. There are no sudden leaps forward, no abrupt revolutions. In traditional evolutionary language, there are no “saltations”—there are only “incrementations” (to introduce a neologism). This principle too is implicitly contained in Darwin’s pithy formulation, “modified descendants of pre-existing forms”—evolutionary change consists of modifications, not outright inventions. The textbook example is the giraffe’s neck: it did not change from short to long in a single evolutionary event, but gradually lengthened over a considerable period of time, as small mutations were positively selected for, one after the other. The lengthening was cumulative, not abrupt and unprecedented. Much the same can be said about an elongation more germane to the subject of this book: the human thumb is a lengthened version of earlier primate thumbs, and the lengthening was gradual, not sudden. It is easy to miss this point if you tacitly presuppose a “homuncular” view of evolution, according to which some sort of intelligence directs the process toward some chosen ideal end. But the process is not teleological in this way, so the end point is achieved “accidently,” no matter how apposite it may appear. Each increment must confer an adaptive advantage in order to be passed on, and the sum of them confers a large adaptive advantage. It is logically possible that a single mutation might produce an unprecedented trait with a brand new adaptive advantage, but in practice the probability of this is vanishingly small. Thus a mouse might (as a matter of logical possibility) undergo some extraordinary genetic convulsion that results in forming a mouse brain as intelligent as the human brain, but such things simply never happen. Mutation is a very blunt and blind instrument and cannot be expected to work miracles: it offers up modest changes that natural selection either allows to persist or eliminates. We never get something from nothing in evolution. New species don’t spring into existence overnight by dint of some spectacular mutation, and the same is true of new organs or traits. Evolution always takes the form of a gradient, not an upsurge of abrupt steps.³

Biologists have a useful term to express this fundamental point: preadaptation. Using this term, I can state the incremental principle in the form of a law: for every adaptation there must be a preadaptation. A preadaptation gives the appearance of anticipation, as if the lengthening neck or thumb were aiming at some ideal end-state of optimal length. But it is not real anticipation; it is entirely accidental, mechanical. The entire process is guided by blind natural selection, one piece at a time. It is just that every new adaptation needs a platform in the prior structure of the animal’s ancestors—something to work with. If we want to explain how a given adaptive trait evolved, we have to identify the preadaptation from which it arose by small coherent steps—by incremental modification. It is never theoretically satisfactory to postulate a sudden leap in the dark that had a positive result; we need to identify the intermediate steps. Worms cannot evolve from bacteria in a single generation, and speakers cannot evolve from nonspeakers in a single generation. There has to be a bridge between the two states. This places a demanding constraint on evolutionary explanations: no saltations allowed! The evolutionary process works incrementally, so our explanations must respect this natural fact. That is basically why evolutionary change takes so long. It is thus a condition of adequacy on evolutionary explanations that they respect the principle of incremental adaptation. We need precursors, precedents, preconditions. Later we will see what this involves in particular cases, notably language; for now, I am just laying out some methodological desiderata. Very often it turns out that some prior trait that worked well enough in a given habitat now acquires a different adaptive function in a new habitat, for which it is then selected and perhaps modified. We shall see that this is conspicuously true of the hand, when the habitat of our ancestors changed quite dramatically.

The picture of evolution suggested by our two principles is then this: evolutionary change is tightly constrained by the past and proceeds by small adjustments going into the future. Earlier forms persist and are modified gradually over long periods of time. There are no sudden departures and revolutionary developments in real evolutionary time, and earlier forms set the parameters for later transformations (this is quite compatible with so-called punctuated equilibrium, by the way, once timescales are properly appreciated).

I can explain this picture by means of what I shall call adaptive space. Consider all logically possible adaptations to an environment consisting of sea, land, and air—an environment like that obtaining on Earth. This is probably an infinite set, or in any case an extremely large one. The set of actual adaptations on Earth is only a small subset of this enormous set. On the independent creation story, with God as creator, any of these logically possible adaptations could occur at any time, according to God’s choice. There is no constraint stemming from the past or necessity for incremental modification. But our two principles tell us that it matters tremendously how the process begins. Suppose life begins in the sea and then proceeds to land and finally ascends to the air—fish, reptiles, and birds (to simplify). Given that fish come first, they establish the constraints on later developments, so that their form will be preserved in subsequent species of reptiles, mammals, and birds. For example, the pectoral and pelvic fins of fish will act as the templates for reptilian and mammalian feet and for avian wings. Later species will accordingly be tetrapods, with each limb ending in extremities possessing a five-pronged ray structure. This will be the basic animal design, preserved across newly evolved species. But if we imagine a planet on which life began in the air, say, because of special conditions obtaining there, and then moved down to the land and finally to the sea, then we will find a very different evolutionary trajectory, in which birdlike features are preserved down through the generations. The fins of fish will resemble the wings of their early bird ancestors, not vice versa.

Adaptive space will take on a different shape on this planet from that obtaining on Earth. The preexisting forms will be quite different, and the modifications will preserve them to some marked degree. If we suppose that finally a species evolves that possesses language, reason, a rich social life, art, and science, this species nevertheless will not resemble us very closely (except in those respects)—since it will preserve different preexisting forms. The basic body plan might incorporate eight limbs with twelve fingers on each of four hands, with vestigial feathers and a high chirpy voice. Any given evolutionary setup will trace a particular path through total logically possible adaptive space, and this path will reflect the starting point. It will not proceed according to what would be ideally adaptive at any given time, as we might expect on the creationist model, but according to the available existing materials for mutation and natural selection. There will be fixed initial conditions and then evolutionary trajectories that preclude saltation. The fact that a particular species will go extinct at a certain point unless it acquires trait T has no tendency to make that species acquire T—however “nice” that would be. There is no benevolent supervisor of the evolutionary process; there is just small-scale random mutation and (non-random) natural selection working on mutational input. The trait T may be highly adaptive, even indispensable to survival, but it is not an “accessible trait” for that species at that time. At some later time, after much incremental variation, T might become accessible; but it is not accessible unless sufficient preadaptations are already in place to provide a platform for its development.

Thus, high intelligence might be a generally adaptive trait for many species in many environments, but it is not an accessible trait for the vast majority of species, simply because there is too much discontinuity between the present form of the species and that desirable trait. To explain how a trait arises (say, language), we need to show that it is accessible given the antecedent attributes of the organism—that is, we need to tell an incremental story.⁴ We cannot just announce that a very fortunate adaptation occurred in which the entire apparatus of human language was suddenly and spectacularly installed by freak mutation in the human brain. If we are trying to give an evolutionary explanation of the Transition, we need to respect the principle of incremental adaptation and recognize that antecedent forms constrain present potential. It must be a matter of small steps, fine adjustments, and steady progress, with previous traits constraining the entire process—smooth ascent, not sudden elevation. As I remarked at the beginning of this chapter, these points are quite orthodox in evolutionary thinking, but it is worth making them maximally explicit, especially when we come to consider the evolution of cognitive traits, such as language, that appear to have no obvious antecedents in the rest of nature.