Clarity is difficult in an evolving science. It is a permanent goal, a possible consequence of struggle. But it is never attained once and for all. As scientific advance brings some issues into focus, others arise that require clarification. Solving and creating problems are not two parts of science, but one.
A science may fall short of perfect clarity in different ways. One is relatively benign. A science may move forward, sideways, and backward as if in a fog that sometimes lifts a little and then resettles. The lack of visible landmarks makes progress difficult to gauge. Theories proliferate; models come and go. Whether theory replacement means theoretical improvement is often hard to determine. But a science enveloped by fog has at least one consolation. A fog does not foster the illusion of clarity; the lack of visibility is patent.
More insidious than the fog is the mirage. Fogs are seen for what they are. Mirages are trickier, engendering the mistaken conviction that things are as they seem. On the other side of the horizon there are some palm trees. Light reflected off a layer of clouds makes them seem to be where they are not. In “seeing” the mirage, one sees something real. One sees the trees, but they are not where they seem to be.1
Mirages, unlike hallucinations, are produced by things outside ourselves; they are not sheer fabrications arising from within. This explains why mirages can be jointly experienced by sensible people. Walking together in the desert, we share the same illusion, and this gives further credence to the idea that we are not making a mistake. Vision is a reliable guide to the location of objects; intersubjective agreement is an additional check against error. Mirages can be doubly hard to detect because they exploit this twofold assurance.
This book is about mirages. It is about a set of conceptual problems that have impeded the attainment of clarity in evolutionary theory. The mistakes I will discuss are not self-generated. They involve the perception of something real that is subject to a predictable sort of misapprehension. People misinterpret mirages in the pursuit of scientific knowledge just as they do in desert wandering.
Distinguishing reality from illusion is a characteristically philosophical undertaking. It also happens to be part of the ongoing activity of science itself. Indeed, the principal precedent for the kind of analysis I propose to conduct here is a work in evolutionary biology. In 1966, George C. Williams published Adaptation and Natural Selection. The field has not been the same since. This landmark in the study of evolution is a philosophical tour de force. It is a critique of fallacies and obscurities in the deployment of the concepts of fitness, natural selection, and adaptation. More important even than the specific conclusions that Williams reached is the analytic acuity that he established as a requirement in evolutionary theory.
One of Williams’ central insights was to see a mirage for what it is. According to the Darwinian interpretation that Williams defended, natural selection pits organism against organism in the struggle for existence. The characteristics that become common in such a selection process will be those that benefit the organisms that possess them. Yet characteristics that benefit individuals may also benefit the groups in which those individuals live. Some groups avoid extinction more successfully than others because they are composed of fitter organisms. But when this happens, we should not conclude that there is selection for group-beneficial characteristics. Selection works for the good of the organism; a consequence may be that some groups fare better than others. However, it does not follow that selection works for the good of the group. Williams’ own expression of this idea cannot be improved upon:
Benefits to groups can arise as statistical summations of the effects of individual adaptations. When a deer successfully escapes from a bear by running away, we can attribute its success to a long ancestral period of selection for fleetness. Its fleetness is responsible for its having a low probability of death from bear attack. The same factor repeated again and again in the herd means not only that it is a herd of fleet deer, but also that it is a fleet herd. The group therefore has a low rate of mortality from bear attack. When every individual in the herd flees from a bear, the result is effective protection of the herd (Williams 1966, p. 16).
A faster deer is less likely to be killed by a predator than a slow one. A fast herd of deer will therefore be less likely to go extinct than a slow one. The fitness of the group may simply be, as Williams says, a “statistical summation” of the fitness values of the individuals in it. This much is unproblematic. The mistake comes, according to Williams, if we think that selection acts for the good of the group, rather than for the good of the individual.
Just as a fast deer is fitter than a slow one, so the fast herd is fitter than the slow one. But according to Williams, the fitness of the group is a mirage. It is a reflection of something real—namely, the fitness values of the individuals in the group. It fosters an illusion—namely, that selection works for the good of the group, rather than for the good of the individual. This is an illusion that had tricked a number of evolutionists. Williams set himself the task of exposing the mistake.
In a desert mirage, there really are some palm trees and one is really seeing them. The mistake is to think they are where they seem to be. According to Williams, advocates of group selection made a similar error. Evolution by natural selection does have its causes. But by seeing the group rather than the individual as the unit of selection, these biologists located the causes in the wrong place. The survival and proliferation of groups is merely a reflection of causal processes at work elsewhere.
If one herd is fast and another slow, why not see the extinction of the second and the survival of the first as a case of group selection? In eliminating the slow deer, didn’t selection also eliminate the slow herd? One prominent reason that Williams gives, which has remained popular, is parsimony. The individual selection hypothesis, so it is said, is simpler than the group-level characterization. Several sorts of clarification are needed here. First, what makes the individual selection account the more parsimonious of the two? Second, why is the greater simplicity of one hypothesis a reason to think that it, rather than its competitor, is true? Finally, if we could with equal truth say that selection works against slow deer or that selection works against slow herds, then no methodological canon ought to advise us to think that one hypothesis is true and the other false. Parsimony is a reason for choosing between nonequivalent hypotheses; the substantive difference between group and individual selection remains to be clarified.
These and other questions must be answered before Williams’ critique of the idea of group selection can be accepted. My point here is a preliminary one: Williams identified a distinction of fundamental importance. It is one thing for groups to survive and go extinct because of group selection; it is quite another for this to happen in consequence of a process of individual selection. We will see in Part II that this distinction rests on the difference between group properties being causes and artifacts in a selection process.
Williams stressed the importance of distinguishing the mirage of group selection from the real thing. Few biologists explicitly dissented, but many advanced opinions about group selection that are implicitly at odds with it. For example, it is not uncommon to conceive of the difference between group and individual selection in terms of the analysis of variance. From this point of view, group selection exists when groups differ in fitness and individual selection occurs when the individuals in a group vary in fitness. An immediate consequence of this point of view is that Williams’ distinction cannot be drawn. If all the deer in one herd run at one speed and the deer in another run at another, slower, speed, then selection for fleetness will mean that all the variation in fitness is between groups. Williams refuses to conclude that the extinction of the slow herd and the continued existence of the fleet one must be chalked up to group selection. The view that rests on the analysis of variance idea, however, leads to precisely this conclusion.2
Another perspective has been influential in the units of selection problem, one at odds both with Williams’ distinction between artifact and cause and with the analysis of variance distinction between intra-and intergroup variation in fitness. It is frequently held that if the individuals in a group are relatives, then what might appear to be group selection may be redescribed as a case of kin selection, which is then claimed to really be a type of individual selection. So, for example, if all the individuals in a group were clones of each other, this would mean that the ensuing selection process should be viewed as a case of individual selection. But notice that in this circumstance, all the variation in fitness is between groups. Since the organisms in a group are carbon copies of each other, there is no within-group variation in fitness. The analysis of variance point of view concludes that this case must be an instance of group selection; the point of view that sees kin selection as a form of individual selection reaches precisely the opposite judgment.
It is easy to be confused by this proliferation of selection processes and by the very different definitions of them that biologists exploit. Group, kin, and individual selection need to be disentangled, their differences made clear.
Williams saw the importance of distinguishing a herd of adapted deer from an adapted herd of deer. When a fast herd survives and a slow one goes extinct, one can describe the former group as fitter than the latter. But the fact that this process can be “represented” by assigning fitness values to groups hardly shows that it is driven by group selection. Yet, while avoiding the mirage of group selection, Williams is, I think, taken in by the mirage of genic selection. He argues that since a selection process that issues in evolution can always be “represented” in terms of fitness values that attach to single genes, it is correct to think of the single gene as the unit of selection.3 The idea that the fitness value of a gene may be a mirage—a reflection of selection processes that occur at higher levels of organization—is swept aside. The crucial distinction first drawn with respect to the question of group selection is obliterated when a lower level of organization is considered.
There may be a principled reason for this. Parsimony has, after all, been invoked as a reason for casting the unit of selection at as low a level as possible. Other reasons favoring the single gene as the unit of selection have been advanced as well. Whether these show that there really is an oasis on the horizon or merely reinforce one illusion with another, we shall have to see.
Williams realized that the problem of the units of selection is not narrow and technical, but foundational. To think properly about the units of selection requires that we clarify the concepts of fitness, selection, and adaptation. It is for this reason that Williams ends his book with a plea. If we are to avoid the fallacies that have preyed upon evolutionary theory in the past, a precise vocabulary is needed for the scientific study of adaptation. Although the present book is frequently critical of the arguments that Williams advanced, his stress on the importance of logic and clarification has guided my inquiry. I have tried to take up what Williams began.
Because the questions are foundational, there is considerable scope for issues that are philosophical in nature. It is impossible to think about the units of selection controversy unless one thinks about causation, chance, explanation, and reduction. Williams’ book is built on a substantial body of doctrine concerning these issues, much of it implicit. I have tried here to contribute what a philosopher can—to uncover presuppositions and make them explicit.
Besides being guided by a set of problems that arose within evolutionary theory, I also have tried to write this book so that it connects with issues that are important within my home discipline, philosophy of science. Biologists are often surprised to learn how little Darwinism has influenced philosophy of science in the last one hundred years. They frequently think that the philosophical consequences of evolutionary theory must be so profound and unsettling that philosophers have chosen to stick their heads in the sand. Darwin, so it is said, continued the work that Galileo and Newton began of dethroning the human race from its position at the pinnacle of creation. No more can we think of ourselves as the reason that the universe is as it is, our standards of reason and morals having a timeless, unshakable authority. This message, says the biologist, fills philosophers with uncertainty; it even threatens to put philosophy out of business. If reason and morality have evolved, philosophers can no longer simply consult their heartfelt intuitions as a guide to the true, the good, and the beautiful. Perhaps the subjects of epistemology, ethics, and aesthetics ought to be taken over by evolutionary biology, which has managed to shake loose from the illusion that human beings are the measure of all things. No wonder, concludes the biologist I have in mind, that the message of Darwinism is a message that philosophy has chosen not to hear.4
To my ear, this explanation of why Darwinism has mattered so little in philosophy of science simply does not ring true. Philosophers now assimilate the fact of evolution with as little difficulty as the fact that the earth is not at the center of the solar system. It is accepted as an unproblematic part of the contemporary scientific world view. But this is scant reason for filling the pages of philosophical monographs with the news! Philosophers have no more reason to trumpet this piece of information than they have for announcing that the earth goes around the sun.
Philosophy of science in this century has been shaped by an interest in physics and mathematics (particularly mathematical logic). The reason is not that these subjects are scientifically important but that they have been connected by philosophers and scientists alike with matters of great philosophical weight. Einstein’s theories of special and general relativity have occupied center stage in philosophy of science for a very good reason: as philosophers, we care about issues of a priori knowledge, conventionalism, and about the general principles that permit radically different scientific theories to be compared and evaluated.
Evolutionary theory is undoubtedly of great scientific importance. But it remains for philosophers of biology to show why it has philosophical importance. We must show that by considering evolutionary theory, old problems can be transformed and new problems brought into being. It remains to be seen, I think, how radically the philosophy of science will be reinterpreted. I find the prospect tantalizing but the conclusion far from foregone.
In bringing biology and philosophy together, we must lose sight of neither. Skimming the surface of the biology will hardly do. One does not shift from the philosophy of geology to the philosophy of biology simply by changing one’s example of an inductive generalization from “all emeralds are green” to “all swans are white.” Nor can philosophers simply plunge into the details of biological debates, thinking that the science somehow matters for its own sake. Of course, it does matter for its own scientific sake. But as philosophers, the question of philosophical significance must always be paramount.
My remarks about Williams’ distinction between an adapted herd and a herd of adapted individuals were meant to give a flavor of the biological material that lies ahead in this book. It was an antipasto, not a full menu. I can provide a comparable foretaste of the philosophical ideas we will encounter by quoting a passage from Bertrand Russell:
All philosophers, of every school, imagine that causation is one of the fundamental axioms or postulates of science, yet, oddly enough, in advanced sciences such as gravitational astronomy, the word “cause” never occurs. . . . The law of causality, I believe, like much that passes muster among philosophers, is a relic of a bygone age, surviving, like the monarchy, only because it is erroneously supposed to do no harm. . . . No doubt the reason why the old “law of causality” has so long continued to pervade the books of philosophers is simply that the idea of a function is unfamiliar to most of them (Russell 1913).
Russell believed that causality was a mirage that a mature science can see through. We may talk of causes and effects in ordinary life and in a fledgling discipline, but as theory develops these concepts lapse from usage and are replaced by the idea of a mathematical function. Instead of laws of cause and effect, we find only equations.
Russell by no means had the last word on the subject; the role of causal concepts in science has continued to occasion philosophical debate. For example, a philosopher of physics—Patrick Suppes (1970, p. 5)—has commented as follows on Russell’s remarks: “Perhaps the most amusing thing about this passage is that its claim about the use of the word ‘cause’ in physics no longer holds. Contrary to the days when Russell wrote this essay, the words ‘causality’ and ‘cause’ are commonly and widely used by physicists in their most advanced work.”
In this book I tell a similar story about the concept of cause in evolutionary theory. From the writings of Darwin to articles in contemporary journals, ideas of causation have been very important in evolutionary biology. Yet they do not surface in any direct way in mathematics; the words “causality” and “cause” are not to be found in the algebra or diffusion equations of quantitative models but in the conceptual framework that motivates them and makes them intelligible.
Philosophers of science generally recognize that the legitimacy of a scientific concept is not to be decided on a priori philosophical grounds but rather by seeing whether science needs that concept to go about its business. The same attitude is appropriate when the concept is one that has been of traditional philosophical interest. If causation is an important issue in evolutionary biology, then philosophers of science must try to understand it, not dismiss it as unintelligible superstition. Much of this book aims at providing a detailed description of the conceptual structure and ontological framework that evolutionary theory deploys. There is much here that can be grist for the philosopher’s mill.
Because I have written this book with two audiences in mind, I should provide some guidance concerning its organization. Or, to put the point another way, I should give the reader a hint about what to skip. The chapters in Part I take up general questions about the structure of evolutionary theory, the concept of fitness, the nature of chance, the meaning of adaptation, and the sorts of explanation that the theory of natural selection provides. Some sections (1.1 to 1.5, 3.2, 4.1, 6.1, and 6.2) do not stray very far from discussing evolutionary theory, whereas others (2.1, 2.2, 3.1, 4.3, and 5.1 to 5.3) address more general philosophical issues.
In Part II, I plunge into the details of the units of selection problem. In Chapter 7, I provide a brief historical background and then attempt to identify a set of fallacies that have grown up in this biological problem area. Then, after characterizing in Chapter 8 what I think causation means in the theory of natural selection, I provide a positive account of what the idea of a unit of selection amounts to. In doing this, I think that a number of current research agendas in evolutionary theory can be seen as issuing from a unified cluster of questions about the mechanism of natural selection.
I must apologize to both biologists and philosophers for occasionally boring them with preliminaries. Biologists may roll their eyes when they read my few sentences about what the word “diploid” means. Philosophers will have their chance to yawn when they come to my remark that it is propositions, not concepts, that are tautologies. This may occasionally try the reader’s patience, but I hope that the exciting bits will come more frequently than the tedious ones.
If I were a biologist interested mainly in seeing why a philosopher has stuck his nose into the units of selection controversy, I would read only Section 3.2 of Part I and then look at Part II in its entirety. If I were a philosopher mildly curious about evolutionary theory, but more concerned to read about matters of general philosophical interest, I would read through Part I, skimming when the biological details get too boring, and then read Section 7.2 and Chapter 8 of Part II. And finally, if I were a reader of broad learning and wide interests, I would read the entire book, even the index, exclaiming loudly all the while about what a fine time I was having.