In chapter 1, I defined biological evolution as the naturalistic production of difference, and differentiated it from some of its homonyms. Stellar “evolution” for example, involves a transformation of state dictated by physical law. A yellow star will probably eventually become a red giant, because that is what stars do when they have run out of hydrogen atoms to fuse. A star’s transformation is highly predictable, and determined by a small number of variables. “Cultural evolution” may involve the direct and conscious production of solutions to environmental problems. And the “evolution” by which an embryo is eventually transformed into a geezer is determined neither by physical law nor by the need to survive, but by the enactment of a genetic program, itself the product of eons of evolution, but which encodes a life cycle.
None of these is what we mean today by “evolution,” though. We use the term specifically to refer to the manner by which descendants come to differ from their biological ancestors. Darwin called it “descent with modification.” There are, of course, many ways to theorize the relationship between descent and modification. They might be different processes, or different aspects of the same process. The modification might be brief in relation to the descent, or the descent might be brief in relation to the modification. There might be different roles for males and females, or there might be different kinds of responsiveness to the needs set by the environment for survival.
For a notable example, take the fit between an organism and its environment, which we call “adaptation.” Aristotle believed it was the result of species simply having been built that way. Darwin argued that it was rather the result of a long-term bias in survival and reproduction of organisms that differed slightly from the average, in the direction of a better fit. In other words, that adaptation is the result of history, rather than miracle.
But how does that bias operate in nature? The British polymath Herbert Spencer had been thinking along similar lines, and convinced Darwin that his phrase “survival of the fittest” was effectively synonymous with Darwin’s “natural selection.” In 1868, Darwin even said so, in The Variation of Animals and Plants under Domestication:
This preservation, during the battle for life, of varieties which possess any advantage in structure, constitution, or instinct, I have called Natural Selection; and Mr. Herbert Spencer has well expressed the same idea by the Survival of the Fittest.1
But there is a crucial difference between the two phrases. If only Spencer’s “fittest” survive, then the descendant populations can be expected to be very fine-tuned to the environment, since they were not merely fitter, but fittest. The pores of the sieve, so to speak, were very small. Natural selection, on the other hand, makes no claim as to the relative size of the pores. Under extraordinary circumstances, only the fittest may survive, but it is primarily simply the fitter that survive. That necessarily implies a bit more unfitness, or “slop,” between the organism and its environment than we might expect if only the fittest were surviving. And sure enough, that question—Just how precisely attuned to the environment are you?—is an unresolved tension in evolution that crops up in unexpected ways.
Biologists since Aristotle haven’t doubted the basic fit between what an animal does and where it lives. But that fit was explained by Aristotle by analogy to a human creation, a tool. To Aristotle, a problem is posed: How do I cut this wood? And one makes a saw to solve the problem. That is why the saw cuts wood—it was made to do so. It would be absurd to imagine making a saw for no reason, and then asking what you can do with it, and serendipitously discovering that it is good for cutting wood. Likewise, concluded Aristotle, body parts are made for particular functions, as solutions to environmental problems. The problem came first, and the body part was fashioned to solve it.2
Darwinian evolution reverses this relationship: the body part preceded the use, and was merely tweaked to fit. Organisms that could survive a bit longer and more prolifically with a slightly tweaked body part in a particular place became the progenitors of a disproportionate number of descendants, similarly tweaked. The hand, which was once held open to support a monkey’s body weight, became modified to suspend an ape’s body like a grappling hook, and to support the ape’s body on the ground while closed; and later, in people, to hold and manipulate a sharp stone or a pen or a baseball, and not to support the body’s weight at all. Aristotle had got it backward; the hand was always there in some form (at least since our ancestors were fish), and it changed over the eons in use and form.
Nevertheless, it is still clear that the fit between an organism and its environment exists. Polar bears are adapted to the arctic, and Gila monsters to the desert. If you study the ecology, behavior, or anatomy of animals, you can’t help but see it. If you study the human body in a comparative context, you can’t miss the way the human foot is similar to the ape foot, but more stable and rigid, just as its weight-bearing role in human locomotion necessitates.
It is, after all, bodies that adapt. They do so genetically, as in having the right genes turned on at the right time. They also do so developmentally (and irreversibly): the body grows in certain characteristic ways in response to hypoxia or oxygen stress, for example. And they adapt physiologically (and reversibly) as well, as in tanning or shivering or callousing under the stimulation of ultraviolet light or cold or abrasion.
On the other hand, if you study the human genome in a comparative context, all you see is how similar the human genome is to the ape genome. You don’t see the feet; for there are no feet in the genome. Nor tans nor shivers nor callouses. There are genes there, not bodies, and it has proven remarkably difficult to match up human genes to human adaptations in any but a small handful of cases. Indeed, it is hard to find adaptation at all reliably in the genome.3
The best-known cases of human genetic adaptations to environmental pressure are those to malaria, incorporating a range of blood diseases and other genetic variants, including sickle-cell anemia and thalassemia. But human populations commonly have their own non-adaptive idiosyncrasies—notably, elevated risks of other genetic diseases. These are accidental, not adaptive—for example, porphyria variegata (another blood disease) among white Dutch South Africans, the genetic legacy of a seventeenth-century settler.4
Along the lines of sickle-cell anemia, the prevalence of Tay-Sachs (a neurological disease) in the gene pool of Ashkenazi Jews has been suggested as a genetic adaptation affording protection to heterozygotes against either tuberculosis or stupidity.5 Carriers, in this framework, may be more resistant to tuberculosis, or may instead be a bit smarter than non-carriers. Nevertheless, it is unclear from the population genetics whether selection has operated at all, with over 80 percent of the Tay-Sachs alleles in Ashkenazi Jews being identical, suggestive of a strong “founder effect.”6 After all, the higher prevalence of the disease in French Canadians and Cajuns is interpreted in this way. Carrying the cystic fibrosis allele, more common in northern Europeans than in other populations, has been associated with resistance to many different diseases, all plausible, but none established.7 While the existence of many alleles causing cystic fibrosis is consistent with an inference of selection, the preponderance of a single one—ƊF508, comprising locally between 40 and 80 percent of the CF alleles in Europe—suggests the complex interplay of stochastic and deterministic forces.8
The point is that we ought to be able to distinguish between these alternative explanations, selection and drift. But usually, even with the finest-grained genetic data, we cannot. Usually the best we can do is show that some feature of the genome is more uniform and less diverse than we think it ought to be, and speculate about the reason that its patterns of difference might be so unexpected.
We have two facts about genetics at work here. First, bodies adapt, because they actually interact with environments; and genomes do not, at least not directly. Consequently, where an anatomist can look at the precision engineering of an eye or a hand, the geneticist looking at the genome sees more of a tinkerer at work than an engineer, in the famous metaphor of the French molecular biologist François Jacob.9 Second, the units of the genome do not map onto the units of the body. We have genes, units of hereditary instruction; and we have elbows, units of the arm—but we don’t have “elbow genes.” In fact, long after the completion of the Human Genome Project, we still know remarkably little about the production of a four-dimensional (space-filling and maturing) body from a one-dimensional set of instructions (the DNA sequence). We have known for a long time, though—this was known as the “unit-character problem” to an earlier generation10—that although the DNA (or genotype) somehow encodes the body (or phenotype), the genetic elements don’t correspond to the body parts in any simple way.
Consequently, where an anatomist can see adaptation, and inferentially the invisible hand of natural selection, a geneticist can see sloppiness and wiggle room, produced by a lot of randomness and historical accident.11 The patterns they see, the questions they ask, and the explanations they invoke differ correspondingly. The geneticist sees a genome in which most DNA changes are neither good nor bad, mutation is a constant but light pressure on the integrity of the system, and DNA sequences are consequently expected to change, indeed to degrade, with some degree of regularity. In fact, the regularity is so much of an expectation that the amount of detectable genetic difference between two species is generally taken as a chronological indication of how long ago their gene pools separated, not of how differently adapted they may or may not have become.12 When we compare humans and chimpanzees genetically, for example, we see far more readily how similar their genomes are, not how behaviorally, ecologically, demographically, and cognitively different they are. The DNA sequences of two animals that have recently become differently adapted are expected to be very similar, but for the constant pressure of mutation, and the very rare “really good” mutation that actually translates into a physical benefit. Consequently, when examining their genomes, we will expect to find differences, and we explain sequences that are too similar as being constrained by selection, because they are more functionally important than other sections of DNA, where differences are accumulating.
In some cases, DNA sequences that are too different can be identified, but the adaptive story behind them is often thin and insubstantial. The gene called FOXP2 impairs cognitive linguistic competence when mutated. Three coding-sequence mutations differentiate the human gene from the mouse gene, two of which occurred recently in human evolution because even the chimpanzee lacks them. It is certainly a gene involved in language, but is it a language gene? After all, rhesus monkeys and chimpanzees have the same coding sequence, but have quite different vocalizations and cognitive properties. The orangutan has a unique coding-sequence mutation, but no obvious special communicative faculties. And one of the unique human mutations arose in parallel in Carnivora. So one can make a strong case for this gene being nebulously “involved” in cognitive linguistic function, but a considerably weaker case for this gene to be a selectively driven master human language gene, as it is often represented.13 The problem is that selection occurs on phenotypes, and genotypic data are difficult to translate phenotypically; to think of FOXP2 as a master language gene is to fall into the trap of unit-characters.
The anatomist, on the other hand, focuses on the particular observable differences among species and explains them in terms of the adaptive differences between the species. The similarities require no explanation; one queries not the choice to remain on four legs, made by myriad primate species, but the change to two legs. One does not query the retention of body hair in all other primates, but its loss in one lineage. It is obviously good to be able to speak, but all the species that can’t speak seem to make do. We anticipate anatomical stability, which requires no explanation, and we interrogate change, which does require an explanation in terms of Darwinian selection. In contrast, the geneticist expects change, and interrogates stability.
It is hard to overstate the implications of these divergent ways of approaching evolutionary data. Geneticists can see animals that look pretty much the same, but whose genomes are scrambled—for example, gibbons and siamangs. Gibbons and siamangs are both known as “lesser apes,” and despite some anatomical distinctions, they are clearly similar kinds of animals, variants on an anatomical theme. Yet gibbon cells have twenty-two pairs of chromosomes, and siamang cells have twenty-five. But that overstates their similarities, for most of the siamang chromosomes cannot even be identified in their gibbon counterparts, because so many rearrangements have arisen between them. Homologous human and chimpanzee chromosomes, by contrast, can be readily matched up and identified almost perfectly. Yet a gibbon sperm with twenty-two chromosomes can fertilize a siamang egg with twenty-five chromosomes, and produce a living hybrid “siabon.”14 It is hard to avoid the conclusion that shuffling the genes around, while leaving them fairly intact, just doesn’t interrupt the production of gibbons from their DNA sequences very much. It is a system that cries of slop, not of precision.
Biochemical systems are often characterized by their redundancy, rather than by the efficiency that anatomical systems seem to show. Structurally different proteins can work in other species; form doesn’t necessarily follow function so precisely when one deals in biochemicals. Efficiency and adaptation are what you expect if the Spencerianf ittest are surviving; wiggle room, redundancy, and slop are what you expect if the Darwinian fit are surviving. Both are likely present, but the point is simply that it is hard to know ahead of time whether any particular feature is actually an “engineered” adaptation or not.
Students of human evolution have repeatedly pointed out that it is unwise to assume that any particular feature is an adaptation, specifically arisen by natural selection, regardless of how useful it seems today, in the absence of strong supporting evidence.15 Use does not explain origin, since any trait may have multiple uses, which may assume different degrees of importance in particular contexts. This is readily visible in cultural evolution, where (despite the limitations of the analogy to organic evolution) origins are often known and can easily be shown to be different from later primary uses—for example, gunpowder for entertainment, and the Internet as a means of decentralizing computers in the event of nuclear attack. The features indeed found new uses: killing people efficiently and downloading pornography.
Aristotle was right about the saw being made for a specific purpose, but the saw was a carefully chosen cultural feature. If he had chosen something as mundane as clothing, whose purposes include warmth (but we dress even when it’s hot out), taboos (certain body parts shouldn’t be seen by others), aesthetics, physical protection or comfort, and the communication of a social identity, his error would have been obvious. Old features have multiple uses; some of them may be new, and they may affect our perception of what the feature is primarily used for, which may be quite different from how the feature got started.
The paleontologist Stephen Jay Gould challenged the assumption that any specific biological feature has an origin in natural selection for any one of its particular properties, calling it “Darwinian fundamentalism.”16 Adaptation is more readily seen than established, and living organisms can be surprisingly good at making do with what they have. We know of ways that adaptive, non-adaptive, and even maladaptive features can evolve. The choice of whether to see crafted machinery in nature, as scholars since the Enlightenment have tended to, or bricolage, that is, genetic elements cobbled together into a stable functional state, as modern molecular geneticists do, is an intellectual choice, neither right nor wrong. They are divergent approaches, both of which can be reconciled to evidences of the history of life.
Indeed, this is an intellectual choice that transcends Darwinian evolution, for “adaptationism” goes back to classical times, and to the intellectual themes of natural theology—seeing the wisdom of God in the contrivances of living forms—that Darwin studied in college. We can study what a feature does, and we can study how it got there, but to ask what it is for is to decorate the scientific question with a lot of metaphysical accessories that it just doesn’t need. To ask what it is for is to assume that there is a reason for it—a deterministic, selective regime for the feature; a particular optimal solution to a problem. But actually, there may be no reasons for some things, just naturalistic causes and uses, and a lot of random noise; life may be more like clothes than like saws.
The fundamental contribution of population genetics to evolutionary theory is its ability to reduce evolution to the transformation of gene pools, and to reduce the transformation of gene pools to a small number of processes, with mathematically predictable effects. This was accomplished by the middle of the twentieth century and came to be known as the “synthetic theory” of evolution. Mutations make new alleles for populations; Darwinian natural selection makes populations different in ways that track the environment, and result in a fit between the gene pool and its surroundings; genetic drift makes populations randomly different, not tracking the environment; and gene flow or interbreeding makes two gene pools less distinct and more homogeneous. Two things were sacrificed, however: bodies and species. By exchanging bodies for genotypes and species for gene pools, midcentury biologists deferred two important questions of physiology for future generations. First, what is the relationship between genotypes and bodies; how reliable a predictor of the latter is the former? And second, how do animals come to identify one another as a part of the same species?
The great evolutionary biologists of the mid-twentieth century evaded these problems by defining them out of existence. By reducing species to gene pools, so they could be mathematically formalized, we made animals essentially automatic outgrowths of their genotypes. By failing to problematize the body itself, then, we failed to problematize the origin of adaptive novelties, the things that allow us to survive and reproduce; it was kept as an article of faith that genetic changes somehow create new bodies. In its most extreme version, Richard Dawkins famously argued that genes are the only significant evolutionary units, and bodies themselves are simply “gigantic lumbering robots.”17
One of the cardinal tenets of Darwinism is the continuity between the patterns and processes that differentiate animal varieties or breeds from one another, and those that differentiate animal species from one another. And Darwin was certainly mostly right about that: higher taxonomic categories (like orders of mammals) generally have the same kinds of differences that lower taxonomic categories (like genera of African monkeys) have—in body form, coloration, behavior, chromosome number and shape, DNA sequence—but more of them, and to a greater extent. The differences among breeds of pigeons or dogs or cows are fewer and smaller, but are the same kinds of differences as those that differentiate pigeons from doves, dogs from bears, and cows from antelopes. There is one point of departure, however. Animals of the same species recognize others of their species as potential mates or competitors for mates. Sometimes they try to mate and breed with other species and fail, but usually they don’t even try. Cats mate with other cats, not with dogs or cows.
There is some biological unit, call it a species for the sake of simplicity, within which animals see each other as potential mates or competitors for mates, and outside of which they don’t. In three dimensions, they constitute clusters of reproductively compatible organisms; in four dimensions, they are diverging lineages.18 Consequently, the production of new evolutionary lineages, or speciation, must entail the development of mate-recognition systems:19 in the case of flies, doing the right dance or having the right pheromone; in the case of chimpanzees, pink swellings of the female genitalia; in the case of humans, looking good. The diverse things that turn people on—power, fame, a great body, flattery, fantasy, erogeny—just aren’t meaningful to a chimpanzee. And swollen pink genitalia don’t work on us. At least on me, anyway.
The classical assumption of theoretical population genetics is that the accumulation of difference somehow causes the multiplication of lineages, that differences of genetic quantity eventually translate into differences of evolutionary quality. Population genetics showed how to model the transformation of a gene pool, but there is more here than just transformation, there is multiplication. At some point populations of animals become so different from one another, or so different in particular ways, that they become separately evolving lineages. By the 1940s, evolutionary biologists had begun to examine the process of diversification itself, in genetic, geographic, and temporal dimensions.20 And as we will see in chapter 4, by the 1980s mainstream evolutionary biologists were appreciating the limitations of the reductive definition of evolution as “changes in gene frequencies through time” and had come to acknowledge that even a minimal definition of evolution had to incorporate diversification, the production of new species, and not merely the transformation of old ones.21
Philosophically, this entails recognizing that a species is not a class of animals defined by the possession of common attributes, but an elemental unit of animal history composed of interrelated parts. The analogy would be to the composition of your body as made up of just cells, but different from the contents of a flask of cells in a biology laboratory. The cells of your body compose you by virtue of their organizational, relational, or epigenetic aspects,22 in spite of being genetically identical to one another; organisms make species likewise by virtue of their relationships to one another. That is to say, a cell begins and ends, reproduces and interacts with an environment. So does a body. And so does a species. It begins, it goes extinct, it speciates, and it occupies an ecological niche. And it is composed of organisms that relate to one another in specific ways, analogous to the way in which organisms are composed of cells that relate to one another in specific ways—differently from the contents of a large flask. Cell biologists had long acknowledged these hierarchical relationships in the natural world;23 and the idea of hierarchy yields a valuable alternative to the reductive view that holds organisms to be “just” genotypes, evolution to be “just” changes in gene frequencies, and species to be “just” gene pools.
It’s a sloppy central concept, though, the species—reminiscent of “culture” in anthropology and “gene” in genetics. It can mean different things in different contexts, and is most applicable only among the most familiar kinds of creatures—namely, sexually reproducing animals. And yet, it clearly represents something real, a natural unit of animals partaking of a common gene pool, with a genetic, ecological, and historical existence separate and distinct from other comparable units. It is this knowledge, what kind of animal you are, that establishes the limits of the gene pool, and circumscribes a species in space and time—at least in theory. In practice, it is always a bit more complicated, with biological issues like intermediate states of interfertility, and cultural concerns sometimes trumping evolutionary genetics.24
Paleontology works with less information than the study of living species does, without physiology or social behavior or genetics, with the principal exception being the study of DNA from recently extinct animals. But paleontology does have one set of data that the study of living species lacks, namely, time depth. This allows it to ask questions that would be otherwise invisible. How rapid is the process of speciation, relative to the duration of the species? What is the role of unpredictable and non-adaptive processes, such as mass extinctions, on the history of life?25
The latter question actually hits at some existential philosophical questions. Are we here for a reason? Is there something special about our species? Here again, there are epistemic choices. On the one hand, Stephen Jay Gould argued that the history of life was full of randomness, like history is. If Hitler hadn’t invaded Russia, you might be reading this passage in German, or not at all. If the dinosaurs hadn’t died out 65 million years ago, primates probably wouldn’t have evolved, and again, you wouldn’t be here. That suggests that our existence as a species is historically precarious and not in any sense inevitable. On the other hand, some biologists point to the ubiquity of parallel evolution in nature. Given that flight arose in insects, reptiles, birds, and mammals, these biologists ask, Why wouldn’t intelligence evolve eventually in another group of species?26
On the third hand, a lot of species haven’t evolved—species with three hands, for example, or telekinesis or invisibility or the Midas touch. Is it really true that if we wait long enough, eventually a species will arise that will shit gold ingots? No, your imagination is not the limiting factor in evolution. The fact that something has never evolved is not a good guide to whether it could ever evolve, and the fact that something evolved once may not be a good guide to whether it could arise again. This view of course is also a bucket of cold water on exobiology, which presumes that life could/did evolve elsewhere, and that the evolution of an intelligent technological lineage, which took 3 billion years to happen here just once, would happen somewhere else, in some kind of recognizable form.27
Watch the skies! (For the weather, not for space aliens—because the weather is all that’s up there.)
To return to reality, what about the nature of species—are they stable through time, or constantly changing? That was the question posed by paleontologists Niles Eldredge and Stephen Jay Gould in a series of papers in the 1970s and 1980s. Although they gave it a highfalutin name, punctuated equilibria, it was rooted in querying a basic assumption about the nature of species.28 Are they constantly adapting to constantly changing environmental circumstances, or do they remain more or less as they began, until another new and slightly better-adapted descendant species comes to supplant them? In either case, we are describing the same set of data, animal A alive at one time, and animal B, very similar, alive at a later time. Obviously we have to connect the dots, but what is the geometry of the connection? Are A and B representatives of different species? A straight line from A to B would imply not just that A evolved into B, but that it did so in a particular way, gradually and by indiscernible increments. The alternative is that species A was stable through time, and its successor, species B (assuming they were different species), was also stable through time, but that the descent of B from A was brief relative to the longevity of species A and B. A good analogy might be the nine-month human gestation relative to an eighty-year life span: assuming that you remain a single entity from cradle to grave, it took a relatively very short time to make you.
What the punctuated equilibria controversy highlighted is that the nature and pattern of ancestral and descendant relationships are not discovered; they are imposed. (A common misconception about punctuated equilibria is that it purports to explain the “gaps” between higher taxa, like reptiles and mammals, or between fish and tetrapods, or between whales and other artiodactyls. Those skeletal transitions are known, to greater or lesser extents, in the fossil record in genera such as Morganucodon, Ambulocetus, and Tiktaalik, but that’s not what punctuated equilibria is about—it’s about the nature of the relationship between two closely related forms, and the nature of species.)
Patterns of similarity among living beings are most plausibly explained by a process of common descent. Nevertheless, living species constitute a trivially small and non-random subset of all species. Extinct species, however, do not leave a trail of descent; they leave a trail of similarities, which must be transformed into a narrative of descent.29 The properties of an organism may fossilize, but the relationships between organisms do not; they have to be inferred. So, is an organism in species A literally the ancestor of an organism in species B? Alas, we can probably never know, but what we do know suggests that it’s pretty unlikely. Of course, lots of individual organisms don’t themselves reproduce. Moreover, since (1) patterns of ancestry are invariably inferred, not discovered, and (2) the sampling of extinct species is very poor, it follows that we rarely, if ever, can discover a particular species that is literally the ancestor of another, much less that an individual in one species is an ancestor of an individual in another species. Instead we say, “This evolved into that” when we really mean, “Something rather like this evolved into that.” The statement (“This evolved into that”) is a shorthand; it is precise without necessarily being accurate.
Substituting precision for accuracy, with embarrassing results, is not altogether unknown in science. For decades, cell biologists had been trying to count the number of chromosomes in a human cell that is about to divide. It was, however, rather like trying to count the strands in a bowl of spaghetti. It was clearly a number in the high forties—but rather than say, “a number in the high forties,” they went with a particular number in the high forties, namely, forty-eight, because in 1923, the most respected researcher in the field said that’s what he thought the number was. And biology textbooks from the 1920s through the 1950s routinely told students that there are forty-eight chromosomes in a human cell. Researchers, knowing the precise answer, routinely convinced themselves that they could see all forty-eight chromosomes. But the precise answer was not the correct one, because in 1956, with technological improvements, scientists began to see only forty-six chromosomes in each normal human cell.30
And remember, all they were doing is counting.
The issue here is the relationship between things, as distinct from the properties of the things themselves. If A looks a lot like B, and lived earlier than B, and you are committed to a naturalistic explanation for the history of life, then it is certainly quite reasonable to infer that something like A evolved into something like B, even if A itself didn’t exactly evolve into B itself.
That’s boring. It was even boring to write. But “A evolved into B” is an origin myth. It is a narrative about the relationship between A and B, extrapolated from their properties and relative chronology.31 It is a narrative about ancestry and descent, which humans are always interested in, because narratives of ancestry and descent tell them who they are and where they fit in, in a world of close relatives, distant relatives, and strangers.
Those narratives are always important and meaningful. What do the actor Kirk Douglas and the anthropologist Ashley Montagu have in common? They both tried to put a little bit of distance between themselves and their ancestry by renaming themselves in less “ethnically marked” ways.32 In a world where your ancestry may be held against you, you may need to create a new ancestry for yourself. In early Christian communities the desire to establish Jesus as a true King of Israel seemed to necessitate tracking his descent from the biblical King David. And different Christian communities tracked that ancestry in different ways. Consequently, two of the Gospels we now have, Matthew and Luke, do precisely that—they track the ancestry of Jesus back to King David—but they do so in different numbers of generations, and with almost entirely different names.33
Narratives of ancestry are invariably mythic, for a simple statistical reason. Every ancestor had two parents; the number of your ancestors in every generation increases exponentially. Barely 300 years ago, you had thousands of lineal ancestors. To make sense of such chaos, what human groups do is to privilege certain ancestors over others, The fact that you are a lineal descendant of George Washington is far more important than the fact that you are also a lineal descendant of thousands of his contemporaries, who aren’t very important, or at least not as important as he is. And frankly, the chance that what little DNA you actually share with your lineal ancestor George Washington was actually his best DNA, is pretty small.
Ancestry, then, is an origin myth. It takes the world of biological data and emphasizes some things, invents others, and relates the present to the past in a meaningful way. Each way of doing so is constrained by cultural rules—and evolution, being a scientific origin myth, is constrained by the assumptions of naturalism, empiricism, and rationalism that bound modern science. And of course, there are other ways of understanding ancestry than the biological or scientific; and these may intersect in weird ways.
For example, no sensible person thinks that molecular genomics yields any support to biblical literalism or creationism, but consider these two sets of facts. First, genomics is a different kind of science, in which money is often at stake, because it is highly corporatized. Second, the facts it produces are also consequently a different class of facts—they are bio-cultural facts. Now you have the tools to make some sense of recreational ancestry testing.
Recreational ancestry testing is a thriving business. A company such as “rootsforreal.com” can tell you if you have the Y chromosome of Moses. To wit:
The priestly caste of the Cohanim are thought to have the same Y chromosome as the biblical Moses, because Aaron, Moses’ brother, founded this priesthood, whose duties traditionally pass from father to son. The Cohanim Y type identified in groundbreaking analysis by the team of Prof. David Goldstein and colleagues agrees with the biblical tradition, and a simple Y test using our database search can confirm whether a Cohen male indeed carries the Cohen Y type.34
This is independent of the fact that Moses is as much a mythic character as King Arthur and Odysseus, although if anyone claimed to be able to test whether you have the DNA of wily Odysseus, you would think they were crazy. Especially if you consult your Bible, and learn from Genesis 5 of the patrilineage connecting Adam and Noah, from Genesis 11 of the patrilineage connecting Noah and Abraham, and Exodus 6, which extends the patrilineage to Moses and Aaron. Yes, this is not just the Y chromosome of the Lawgiver, but the Y chromosome of Adam as well.
Shhhhh. Don’t tell the creationists.
What on earth is going on here?
It’s about the significance and marketability of ancestry. Let us add some more bio-cultural facts. First, scientists are more willing to accept Exodus literally than Genesis. I’m not sure why. Second, people tend to be genetically more similar to people with the same surnames (in this case, Cohen or its derivatives) than to random people. Third, people who consider themselves to be part of the priestly lineage in Judaism are disproportionately named “Cohen” or a derivative. Fourth, the story is relatively innocuous; hence we can label the consumer product as “recreational.” Fifth, the Jews a have a complicated demographic history, even a mythic one, with ancient origin stories from Palestine, Egypt, and Babylonia. Sixth, sure, there might be other ways of explaining the data, but this interpretation—that the Y-chromosome configuration held by most people named Cohen is the descendant of the Y chromosome of the original high priest Aaron, who had the same Y chromosome as Moses, because they were brothers—might be true.
There is, of course, real science at work here.35 The initial paper was published in Nature, the leading science journal in the world, and actually began, “According to biblical accounts, the Jewish priesthood was established about 3,300 years ago .” My hat is off to anyone with the chutzpah to start a paper in Nature, “According to biblical accounts . . .” But there is DNA sequencing, and plausible analysis. Now it could be that they have discovered the Y chromosome of Moses and Adam; or alternatively, it could be that they have discovered that a sample of Jews with similar surnames tend to be genetically alike, and that happens to be the Y-chromosome configuration that most of them have by virtue of the complexities of Jewish demographic history. Of course, if it really were the latter, who would be interested in buying the test at $300 a pop?
So, never mind that the leading science journal in the world published a paper that begins with the assumption that the biblical characters are real (although perhaps without their participation in the miraculous plagues, manna from heaven, and parting of the sea)—and nobody batted an eye. Imagine if, instead of Moses and Aaron, they had actually claimed to have discovered the Y chromosome of Noah and Abraham—in which case the Science Police would have rung alarms in every conceivable forum. (And they actually could have made that claim from their data, given the biblical genealogical connections noted above.) The point is that this is about business and mythology and genomics simultaneously, and you can’t disentangle them. We like to think that genetics or genomics is an uncultural, purely objective scientific view of ancestry, but it isn’t—as the 1995 classic, The DNA Mystique: The Gene as Cultural Icon, by the sociologist Dorothy Nelkin and historian Susan Lindee famously explored. This is science, all right, but it is very cultural science; for this is about ancestry, and ancestry involves the privileged relations among people. And it is precisely those relations that are constructed and anthropological, not given by nature.36
Each of us has an inheritance from our ancestors. That inheritance, however, is complex, consisting of both organic (living cellular matter) and non-organic (traditions, silverware) features. The organic heritage bounds and differentiates us as individuals: everyone’s DNA is slightly different. It also bounds and differentiates us as a species: every species’s DNA is slightly different. In between the organism and the species, however, our biological patterns and distinctions are far more subtle.
Groups of humans are similar to, yet different from, their neighbors. We aren’t like our neighbors; we do things differently, in a more civilized, sensitive, spiritual way. We define them in opposition to ourselves, we don’t like their ways, or their mode of speech, and yet we trade with them, in a pinch we may rely on them, we may even fall in love with them. It’s a peculiarly human way to think: we imagine others as similar or different according to shifting, situation-specific, historically produced ideas about what kinds of similarities and differences matter. Certainly the most fundamental idea of similarity and difference resides in the decision about who is a member of our family and who is not. That decision creates the available choices for sexual and marriage partners, given the broad taboo on having sex with a nuclear family member. And yet, that decision about who is a part of our family is subject to extraordinary levels of flexibility, as social anthropologists have documented extensively.
In other words, we need to know who is a member of our family and who is not, but because the family is constituted from compromises between lineal blood relationships (parenthood, generally speaking) and legal bonds (marriage, residency, and adoption), the boundaries of the family are often quite fuzzy. So we sharpen them with our special rules, which may not map particularly well onto our genetic relationships, but at least we now know what to do.
And the same problem recurs at a higher level. Our family is “especially close relatives,” who are segregated by definition from a broader category, “relatives.” And yet “relatives” is not an unproblematic natural category either, since biologically, we are all related. Somehow we also have to decide that a second cousin is a relative, but a twentieth cousin is not. Or even more arbitrarily, that one twentieth cousin (sharing your last name or a critical bit of your genome) might be a relative, and another twentieth cousin is not. These are units built up from nature/culture—the family, the kin-group, the race, the nation, the species—bounded in part or in varying degrees by natural properties, and in part by imaginary fences.
Historically, narratives of human origins have incorporated narratives of human diversity (the former presumably explaining the latter), but these scientific narratives of contemporary difference have always been co-produced by the author’s social and political circumstances. Thus, Darwin’s Descent of Man (1871) is far more a text of Victorian social prejudices than his earlier Origin of Species (1859), which famously omitted all but the most oblique reference to people, and is consequently far more readable all these years later.
The cultural aspects of ancestry, even in evolution, come out in another interesting way. Since the cessation of gene flow classically implies a new species and a new evolving lineage, it is classically assumed that, above species level, gene pools can only diverge from one another, since they can’t get more similar through gene flow or interbreeding. There might be some superficial similarities emerging when distinct lineages cope with certain environmental challenges in convergent ways, but the fact is that bats can’t mate with birds, and dolphins can’t mate with sharks. Consequently the most famous image of evolution is as a tree, its branches ever diverging from one another.37
That is a useful image for macroevolution. For microevolution, however, we must look to another part of the tree—to its root system, Roots, unlike branches, are not always separating from one another. Roots may often fuse with one another, to create a connected network whose individual paths may be very difficult to delineate. They’re like populations of organisms, evolving somewhat separately, but still connected by gene flow. While they become distinct in minor ways, nevertheless like Michael Corleone trying to escape from the Mafia, they keep getting pulled back in.
There is an important difference between the two systems. A group of distinct species who are each other’s closest relatives is a clade; and a network of subspecific populations is a rhizome. Within a clade, there is a simple answer to the question, Which are really the closest relatives? The closest relatives are the species that shared a common ancestor most recently. But for a rhizomatic network, which may resemble a train trellis or a capillary system more than the branches of a tree, there is no simple answer, since sharing recent common ancestry is not the only variable; it gets combined with how much and how recently there has been interbreeding with other parts of the network.
In principle, there may actually be no answer to the “closest relatives” question in a system that isn’t constantly diverging, as species are. In practice, however, you can program a computer to answer a different question—Which are most similar?—and draw a rhizomatic system as if it were a tree. The results might then look like they had a great deal more evolutionary validity than they actually do.
Thus, a population genetics project might pose a question about whether, “for example, the Irish are more closely related to the Spaniards or to the Swedes.”38 And they can get an answer. But that answer will be dependent upon who is actually taken to represent the nationalities in question (are we sampling the real Swedes of today, or the Swedes we imagine of 500 years ago?), their demographic expansions and contractions, and the particular algorithms used to construct the tree—as well as the nature and extent of gene flow, and the divergences that actually frame the question. The idea that a tree would represent only the last of these is at best a very hopeful one.
A parallel problem exists when the “closest relatives” question is applied to things like languages and human artifacts. The issue is the imposition of a tree-like structure on histories that are basically not tree-like, an altogether too-common practice, often concealed by appeals to evolution and technology.
But the bigger problem remains its application to human populations, and the casual interpretation of the resulting tree of statistically generated similarity as a phylogenetic tree of history. And perhaps the most unusual situation exists when we can’t tell whether the units we are clustering are species (in which case we might well be reconstructing relationships of descent) or subspecies and local populations (in which case we probably aren’t). If we “split” the human fossil lineages, we make it look as if we are indeed dealing with species, and the cladistic analysis ought to work: a tree ought to be a good approximation of a branching history. But if we “lump” those fossils, all bets are off—because we might be dealing with the Irish-Spanish-Swedish problem, except over much larger ranges of space and time.
The fossils recently discovered at Dmanisi, Georgia, suggest that we are indeed dealing with strongly rhizomatic relationships in the human fossil record, back to nearly 2 million years ago.39 Several anthropologists had suggested this over the years—Earnest Hooton invoking the metaphor of a capillary system; Franz Weidenreich and, later, Frederick Hulse invoking the train trellis; and others invoking a root system or mesh net40—and it looks like they might just have been right. How you allocate the fossils taxonomically is how you begin to make human evolution into a story—whether you narrate human evolution as linear, with very few species, culminating in our own; or as bushy, with many species, and all but one having gone extinct. The line and the bush—and their intermediates, bushy lines—are each narratives of human evolution, and understanding that narrative aspect is central to thinking clearly about human evolution.
Human evolution, then, is a theory of kinship—or a set of theories about kinship—and is not fully accessible through zoology. Theories of human relatedness and descent at all levels are bio-cultural theories, not strictly natural ones.