Chapter 25: Natural Selection & Teleology

Aristotle distinguished four types of causes: material, formal, efficient, and final. The final cause — the telos or purpose — was central to his understanding of nature. For Aristotle, to explain why birds have wings is to cite the purpose of flight. This teleological mode of explanation dominated natural philosophy for two millennia.

The Scientific Revolution expelled final causes from physics. Francis Bacon declared that final causes were “barren virgins” that produce nothing. Descartes reduced the physical world to mechanism. Newton’s laws made no reference to purposes. But biology seemed to resist this mechanistic turn. The living world appeared designed, and design seemed to require a designer.

“In crossing a heath, suppose I pitched my foot against a stone, and were asked how the stone came to be there; I might possibly answer, that, for anything I knew to the contrary, it had lain there forever... But suppose I had found a watch upon the ground... the inference, we think, is inevitable, that the watch must have had a maker.”— William Paley, Natural Theology (1802)

Darwin’s great insight was that a mindless process — natural selection — could produce the appearance of design. The key idea is disarmingly simple: organisms vary in their traits; some of these variations are heritable; organisms with traits better suited to their environment tend to survive and reproduce more successfully. Over vast stretches of time, this process produces organisms exquisitely adapted to their environments — organisms that look designed but are not.

But did Darwin eliminate teleology from biology, or did he naturalize it? This question remains contested. Some philosophers (e.g., Rosenberg) argue that Darwin showed teleological language in biology to be strictly eliminable — it is a useful shorthand but does not track any real explanatory category. Others (e.g., Millikan, Neander) argue that Darwin made teleology respectable by grounding it in natural history rather than divine intention.

Natural selection can be decomposed into three necessary and jointly sufficient conditions, often called the “Lewontin conditions” after Richard Lewontin’s influential 1970 formulation:

When all three conditions are met, evolution by natural selection occurs: the frequency of traits associated with higher fitness increases over generations. Crucially, this is an algorithmic process. As Daniel Dennett emphasized, natural selection is “substrate-neutral” — it operates wherever the three conditions are met, whether in biological organisms, computer programs, or cultural practices.

“Darwin’s dangerous idea: the algorithmic level is the level that best accounts for the speed of the antelope, the ## of the eagle, the shape of the orchid, the diversity of species, and all the other occasions for wonder in the world of nature.”— Daniel Dennett, Darwin’s Dangerous Idea (1995)

The elegance and generality of this structure have led some philosophers to characterize natural selection as a schema or framework for explanation rather than a specific empirical law. Elliott Sober has argued that natural selection is best understood as a force that acts on populations, analogous to the forces of Newtonian mechanics. Just as gravity and electromagnetism are forces that change the state of physical systems, selection, drift, mutation, and migration are forces that change allele frequencies in populations.

This “force” interpretation has been influential but also contested. Statisticalists like Denis Walsh, Tim Lewens, and Andre Ariew argue that natural selection is not a cause but a statistical consequence of individual-level events. On their view, selection does not cause changes in population structure; rather, it describes or summarizes the aggregate effects of individual births, deaths, and reproductive events. The debate between causalists and statisticalists remains one of the liveliest in the philosophy of biology.

One of the most persistent criticisms of natural selection is the charge that it is tautological — that “survival of the fittest” is a mere tautology if “fittest” is defined as “those that survive.” If so, the theory would be unfalsifiable and explanatorily vacuous.

Karl Popper himself endorsed this criticism in the 1970s. In his autobiography Unended Quest(1974), he wrote:

“I have come to the conclusion that Darwinism is not a testable scientific theory, but a metaphysical research programme — a possible framework for testable scientific theories.”— Karl Popper, Unended Quest (1974)

However, Popper later retracted this view. In a 1978 lecture, he acknowledged that natural selection is testable and that his earlier criticism had been mistaken. The key to resolving the tautology charge lies in recognizing that “fitness” can be defined independently of actual reproductive success. Fitness is a propensity — a dispositional property of organisms in particular environments — that can be measured or estimated independently of the actual outcome. An organism can be fitter than its competitors yet fail to reproduce due to bad luck (being struck by lightning, for instance).

The propensity interpretation of fitness, developed by Susan Mills and John Beatty (1979), characterizes fitness as the expected number of offspring, not the actual number. This breaks the alleged circularity: fitness is a causal property of organisms that explains (but is not defined by) reproductive success. The theory of natural selection then becomes a substantive empirical claim: organisms with higher fitness propensities tend, on average, to leave more offspring.

Moreover, specific selectionist hypotheses are clearly testable. The hypothesis that “the peppered moth evolved melanism because darker coloration provided better camouflage against predators in industrial environments” is a specific, falsifiable claim that can be (and has been) tested experimentally.

Natural selection requires entities that vary, are heritable, and differ in fitness. But whatare these entities? The “units of selection” debate has been one of the most contentious in the philosophy of biology, with candidates ranging from genes to organisms to groups to entire species.

The Gene’s-Eye View (Dawkins)

Richard Dawkins’s The Selfish Gene (1976) argued that the gene is the fundamental unit of selection. On Dawkins’s view, organisms are merely “vehicles” or “survival machines” built by genes to propagate themselves. What really matters in evolution is the differential replication of genes. George Williams had anticipated this view in Adaptation and Natural Selection (1966), arguing that group selection was almost never a significant evolutionary force and that individual-level selection was best understood in terms of gene-frequency change.

“We are survival machines — robot vehicles blindly programmed to preserve the selfish molecules known as genes.”— Richard Dawkins, The Selfish Gene (1976)

The Organism as Unit of Selection

Many biologists and philosophers resist the gene-centered view, arguing that it is organisms, not genes, that are the primary targets of natural selection. It is organisms that live or die, reproduce or fail to reproduce. Genes are selected only because they are packaged in organisms that interact with their environments. As Ernst Mayr insisted, natural selection is fundamentally about the differential survival and reproduction of whole organisms in ecological contexts.

Group Selection (Sober & Wilson)

The revival of group selection by Elliott Sober and David Sloan Wilson in Unto Others (1998) challenged both the gene-centered and organism-centered views. Sober and Wilson argued that selection can operate at multiple levels simultaneously — a framework known as multilevel selection theory. Groups of cooperators can outcompete groups of defectors, even if defectors outcompete cooperators within each group. The key insight is that selection at different levels can push in different directions, and the outcome depends on the relative strength of within-group and between-group selection.

The debate remains unresolved. Some argue that gene-level, organism-level, and group-level descriptions are merely different perspectives on the same underlying process (this is the “bookkeeping argument”). Others insist that the different levels correspond to genuinely different causal processes. The philosopher Samir Okasha, in Evolution and the Levels of Selection (2006), provides the most comprehensive philosophical treatment, arguing that the levels-of-selection question is not a single question but a family of related questions that require different answers in different biological contexts.

Adaptationism is the view that natural selection is the primary, or even the sole, important force shaping biological form. Adaptationists assume that most traits of organisms are adaptations — products of natural selection for the trait’s current function. The adaptationist strategy in biology is to identify the selective advantage of a trait and construct a “just-so story” explaining how natural selection produced it.

The most famous critique of adaptationism came from Stephen Jay Gould and Richard Lewontin in their landmark 1979 paper “The Spandrels of San Marco and the Panglossian Paradigm.” Gould and Lewontin drew an analogy between biological traits and the spandrels of the Basilica of San Marco in Venice. Spandrels are the triangular spaces formed where arches meet; they are not designed features but necessary by-products of architectural constraints. Yet the spandrels in San Marco are beautifully decorated, giving the appearance of having been designed as spaces for mosaics.

“The adaptationist programme... regards natural selection as so powerful and the constraints upon it so few that direct production of adaptation through its operation becomes the primary cause of nearly all organic form, function, and behaviour... We fault the adaptationist programme for its failure to distinguish current utility from reasons for origin.”— Gould & Lewontin, “The Spandrels of San Marco” (1979)

Gould and Lewontin identified several alternative explanations for biological traits that adaptationism ignores:

• •Structural constraints: Some traits are by-products of the physical or developmental constraints on organisms, not direct products of selection.
• •Genetic drift: Random changes in allele frequencies can produce traits that have no selective advantage.
• •Pleiotropy and linkage: A gene may be selected for one effect but produce other effects as by-products.
• •Exaptation: A trait that originated for one function (or for no function at all) may be co-opted for a new function. Gould and Elisabeth Vrba coined the term “exaptation” in 1982 to describe this phenomenon.

The adaptationism debate remains unresolved, though most contemporary philosophers of biology adopt a moderate position: natural selection is a powerful and important force, but it operates alongside other evolutionary mechanisms, and not every trait is an adaptation. The challenge is to develop rigorous methods for distinguishing genuine adaptations from spandrels, exaptations, and products of drift.

The most influential attempt to naturalize teleology through evolutionary theory comes from the tradition known as teleosemantics, developed primarily by Ruth Garrett Millikan and Karen Neander. The central concept is that of a proper function.

Millikan’s account, developed in Language, Thought, and Other Biological Categories(1984), defines the proper function of a trait in terms of its evolutionary history. A trait has the function F if and only if it was selected because it performed F in the past. The function of the heart is to pump blood because hearts were selected for pumping blood — that is, ancestral organisms with hearts that pumped blood survived and reproduced more successfully than those without.

“A proper function of a device is a function that its ancestors performed and that helped account for the proliferation of the genes responsible for producing it.”— Ruth Millikan, Language, Thought, and Other Biological Categories (1984)

This etiological account of function has several attractive features. It distinguishes function from mere effect: the heart produces sounds, but producing sounds is not its function because hearts were not selected for sound production. It also allows for malfunction: a defective heart that fails to pump blood still has the function of pumping blood, because its function is determined by its evolutionary history, not its current performance.

Karen Neander (1991) refined the etiological account and applied it specifically to the philosophy of biology, arguing that function attributions in biology are implicitly historical claims. When a biologist says “the function of hemoglobin is to transport oxygen,” this is a compressed claim about the selective history of hemoglobin.

Critics have raised several objections to the etiological account. Robert Cummins (1975) argues that functional explanation in biology does not require evolutionary history; we can explain what a trait does without knowing why it was selected. Larry Wright (1973), though sympathetic to etiological accounts in general, worried about cases where a trait’s current function differs from its historical function (exaptations). And the “swamp creature” thought experiment — imagine a creature that spontaneously assembles by lightning, molecule-identical to you — poses a challenge: its organs have no evolutionary history, so on Millikan’s account they have no functions, which seems counterintuitive.

The tension between teleological and mechanistic explanation in biology reflects a deeper philosophical puzzle. Mechanistic explanation proceeds “bottom-up,” from parts to wholes, from causes to effects. Teleological explanation proceeds “top-down,” from functions to structures, from goals to means. Both modes of explanation seem indispensable in biology.

Consider the question: why do mammals have hearts? A mechanistic explanation would describe the developmental biology of the heart — the gene regulatory networks, the signaling cascades, the cell migrations that produce a functioning heart during embryonic development. A teleological explanation would say: mammals have hearts in order to pump blood. The mechanistic explanation tells ushow the heart develops; the teleological explanation tells us why it exists.

Ernst Mayr attempted to resolve this tension by distinguishing “proximate” from “ultimate” causation. Proximate causes are the immediate mechanistic causes of a biological phenomenon (the gene networks that build a heart). Ultimate causes are the evolutionary reasons why the phenomenon exists (natural selection for blood circulation). Both types of explanation are legitimate and complementary, not competing.

Contemporary philosophers of biology have largely accepted Mayr’s distinction, though with important refinements. The new mechanistic philosophy (Machamer, Darden, and Craver, 2000) provides a framework that accommodates both mechanistic and functional explanation within a single account. On this view, biological mechanisms are identified by their functions, and explanations of mechanisms cite both their components (mechanistic explanation) and their roles in larger systems (functional explanation).

The Modern Synthesis of the 1930s and 1940s — the integration of Darwinian natural selection with Mendelian genetics by figures like Ronald Fisher, J.B.S. Haldane, and Sewall Wright — provided the theoretical foundation for twentieth-century evolutionary biology. But since the 1990s, a growing number of biologists and philosophers have called for an Extended Evolutionary Synthesis (EES) that incorporates phenomena not adequately captured by the Modern Synthesis.

The EES emphasizes several processes that the Modern Synthesis either ignored or marginalized:

• •Developmental bias: The developmental process constrains and channels evolution, influencing which variants are produced and hence which are available for selection.
• •Phenotypic plasticity: Organisms can modify their phenotypes in response to environmental conditions without genetic change. This plasticity can facilitate evolutionary change by allowing populations to persist in new environments.
• •Niche construction: Organisms modify their own environments, thereby changing the selection pressures that act on themselves and other species. Beaver dams, earthworm soil modification, and human agriculture are classic examples.
• •Epigenetic inheritance: Heritable changes in gene expression that do not involve changes in DNA sequence can be transmitted across generations.

The philosophical significance of the EES is that it challenges the gene-centered, selection-centered picture of evolution. If developmental bias, niche construction, and epigenetic inheritance are important evolutionary forces, then the role of natural selection may be less central than the Modern Synthesis assumed. The EES does not reject natural selection, but it embeds it within a richer causal framework. Whether the EES represents a genuine paradigm shift or merely an extension of existing theory remains a matter of vigorous debate.

• •Dawkins, R. (1976). The Selfish Gene, Chapters 1–3.
• •Gould, S.J. & Lewontin, R. (1979). “The Spandrels of San Marco and the Panglossian Paradigm.”
• •Millikan, R.G. (1984). Language, Thought, and Other Biological Categories, Chapters 1–2.
• •Sober, E. & Wilson, D.S. (1998). Unto Others, Chapters 1–3.
• •Dennett, D. (1995). Darwin’s Dangerous Idea, Chapters 2–3.
• •Okasha, S. (2006). Evolution and the Levels of Selection, Chapters 1–2.
• •Neander, K. (1991). “Functions as Selected Effects,” Philosophy of Science 58(2).