Everything you believe about what’s right and what’s wrong may have nothing to do with moral truth and everything to do with what helped your ancestors survive. The evolutionary debunking argument against moral realism–the view that there are objective moral truths independent of what anyone thinks, believes or desires–claims that even if objective moral truths exist, we have no good reason to think our evolved moral faculties are reliably connected to them. And if that is right, we face a choice about what to do with our moral beliefs from there.
This puzzle was the topic of my dissertation (‘Some Epistemological and Practical Challenges to Moral Realism: Evolutionary Debunking, Overgeneralization, and Afterward’). The dissertation itself examines what is best understood as a conditional argument: If robust moral realism is true, and if familiar facts about human psychology and evolution are taken seriously, then what, if anything, happens to our moral knowledge? And what do we do about our moral beliefs from there? This question unifies all four chapters of the dissertation.
The target throughout is robust moral realism. This is the metaethical view that there are irreducibly normative moral facts that are independent of any actual or idealized human attitudes and that do not enter into causal explanations. On this view, moral truths are neither projections of our evaluative practices nor reducible to natural facts, and they do not explain our moral judgments in the way that electrons explain vapor trails or pathogens explain illness.
That conception of moral facts is attractive for reasons like promising objectivity without relativism, normativity without reduction and a standpoint from which genuine moral error is possible. But it also invites a distinctive epistemological problem. If moral facts are causally inert, and if our moral psychology is the product of evolutionary forces that are indifferent to moral truth, then it becomes unclear why we should expect our moral beliefs to bear any reliable relation to those facts.
The dissertation explores that problem in stages. The first three chapters develop and refine evolutionary debunking arguments against robust moral realism, while the final chapter asks how agents should think and deliberate if those arguments succeed—what do we do with our moral beliefs if they are debunked by evolution?
CHAPTER ONE places evolutionary debunking arguments in a broader explanatory challenge to moral realism. The starting point is a familiar one that the philosopher Gilbert Harman made, namely that unlike scientific posits, moral facts do no explanatory work. When we explain why people hold the moral beliefs they do, we appeal to psychological dispositions, social learning, institutional incentives and emotional responses. Moral facts themselves never appear in the explanation. This explanatory silence is not, by itself, a refutation of moral realism, but it does make the realist vulnerable to genealogical critiques. If we can provide a complete causal explanation of moral belief that makes no reference to moral truth, then the epistemic standing of those beliefs demands further defense.
Evolutionary explanations provide exactly that kind of genealogy. The core idea is not that evolution directly selects for particular moral propositions, but that it selects for cognitive and motivational traits—such as norm sensitivity, punishment dispositions and cooperation-enhancing emotions—that reliably generate moralized judgments. These traits plausibly increased fitness by stabilizing social behavior, coordinating expectations and motivating costly forms of cooperation. Crucially, none of these advantages depends on the truth of the moral beliefs produced. Selection pressures favor beliefs that function well instead of beliefs that accurately represent stance-independent moral facts.
This observation motivates a standard form of evolutionary debunking argument. If our moral faculties were shaped by forces that are indifferent to moral truth, and if moral facts themselves play no causal role in shaping those faculties, then the reliability of our moral beliefs looks like a coincidence at best. At worst, we have reason to think those beliefs are systematically disconnected from the moral facts they purport to represent.
The chapter examines several ways of sharpening this worry, including attempts to frame it in terms of evidential sensitivity or counterfactual dependence. These strategies face familiar obstacles, especially when moral truths are taken to be necessary rather than contingent. But the underlying concern survives these technical complications. Even if some general moral principles are metaphysically necessary, the vast majority of our moral beliefs concern highly contingent matters: how to treat distant strangers, how to distribute social risks, how to weigh harms against benefits under complex institutional conditions. These are precisely the domains in which evolutionary pressures are least likely to track independent moral truth and most likely to distort judgment in predictable ways.
The chapter also challenges optimistic realist responses that appeal to reflective equilibrium or coherence-based reasoning as a corrective. If the initial set of moral judgments with which we begin is itself heavily shaped by evolutionary and cultural forces that favor parochialism, in-group bias, and aggression under threat, then coherence-seeking methods may simply refine those distortions rather than eliminate them. The problem is not inconsistency but contaminated input.
This sets the stage for the more focused arguments of the second and third chapters, which ask not merely whether evolutionary debunking arguments pose a challenge to moral realism, but which version of those arguments is strongest, and whether they can be dismissed as self-defeating or overgeneralized.
CHAPTER TWO turns from the general structure of evolutionary challenges to a more precise question: What, exactly, does evolution debunk? Much of the literature proceeds as if the answer were obvious—evolution shapes the content of our moral beliefs—but that assumption deserves scrutiny.
The most influential content-based debunking strategy is associated with the philosopher Sharon Street. In her view, evolutionary pressures have exerted a pervasive influence on our basic evaluative tendencies: tendencies to favor kin, to resent free riders, to respond negatively to unprovoked harm, and so on. These tendencies, once in place, heavily constrain the moral judgments we eventually endorse through reflection. The realist, then, faces what Street calls the Darwinian Dilemma. Either there is no relation between these evolved evaluative tendencies and stance-independent moral truths, in which case the truth of our moral beliefs would be an extraordinary coincidence, or there is such a relation, in which case the realist owes an explanation of why evolutionary forces would have selected for truth-tracking evaluative dispositions rather than merely fitness-enhancing ones.
The attraction of this argument is clear. It connects moral belief to evolutionary explanation without requiring strong claims about innateness or modularity, and it meshes well with contemporary moral psychology. But in the dissertation I argue that, from the perspective of a debunker, the content strategy is not the most effective one available.
The alternative is what I call capacity debunking, most prominently developed by Richard Joyce. Rather than focusing on the specific evaluative contents evolution might have shaped, the capacity approach treats evolution as having selected for a distinctive kind of cognitive framework: a capacity to think in irreducibly moral terms at all. In this picture, what mattered from an evolutionary standpoint was not whether our ancestors believed this or that moral proposition, but whether they possessed a way of representing certain actions as categorically required, forbidden or deserving of punishment. That representational framework itself—rather than its particular outputs—did the adaptive work.
The difference matters. Content debunking allows the realist some room to maneuver. Even if many of our evaluative tendencies were shaped by fitness considerations, the realist can try to identify a subset of judgments or principles that are insulated from those pressures or that arise through more truth-conducive processes of abstraction and argument. Capacity debunking makes this much harder. If the very conceptual machinery through which we form moral beliefs is an evolutionary artifact designed to promote coordination and compliance rather than truth, then every moral belief inherits the same epistemic liability.
Moreover, the capacity strategy places a lighter empirical burden on the debunker. Content debunking requires a fairly detailed story about which evaluative tendencies were selected and why. Capacity debunking requires only the modest and highly plausible claim that representing norms as categorical and authority-imposing was socially useful. Once that framework is in place, it can generate a wide range of first-order moral beliefs, many of which may be adaptively neutral or even maladaptive. Evolution need not fine-tune content if installing the capacity does most of the work.
The conclusion of chapter two is not that content debunking fails, but that capacity debunking poses the deeper challenge to robust moral realism. Much of the existing realist response literature is aimed at the former. If the latter is taken seriously, those responses address a weaker opponent than the one the realist actually faces.
CHAPTER THREE explores maybe the most potent objection to evolution debunking there is—an objection lodged repeatedly to me by a dissertation committee member who shall remain nameless. And it goes like this: If evolutionary debunking arguments undermine moral belief, why do they not also undermine belief in epistemic norms, prudential reasons or even ordinary empirical claims? If the answer is that they do, then the arguments appear self-defeating. If the answer is that they do not, then the restriction to morality appears arbitrary.
This style of objection—often framed as a companions in guilt argument—has serious intuitive force. Moral, epistemic and prudential reasons are normative, action-guiding and not reducible to causal explanations. If evolutionary genealogy discredits one, why not the others? The chapter argues that this objection rests on an overly coarse understanding of how debunking arguments work. The evolutionary debunking argument does not claim that any evolutionary influence on a belief-forming process defeats justification. Rather, it relies on a conjunction of claims: that evolution shaped the process in ways plausibly unrelated to truth, and that we lack independent reason to think the process is nevertheless reliable.
This gives rise to what I call the independent reason dilemma. Take reasons in any domain —morality, epistemology, perception—and then ask whether we have an independent reason to trust the relevant cognitive faculties, like our memory or our perception, i.e., one that does not itself presuppose their reliability.
If we do have such a reason, then evolutionary debunking does not get traction in that domain. For ordinary perception, for example, we may appeal to broad explanatory coherence, predictive success, and the indispensability of perceptual beliefs to successful action and inquiry. These considerations are not obviously undermined by the evolutionary genealogy of perception; indeed, they may be strengthened by it.
If we do not have such a reason in some domain, then evolutionary debunking arguments may well generalize to that domain. But this is not a reductio of the debunking strategy. It simply shows that the epistemic situation there is worse than we might have hoped. Overgeneralization is only a problem if we assume in advance that certain domains must be epistemically secure, regardless of the arguments.
Applied to morality, the point is straightforward. Robust moral realism posits causally inert facts that do no explanatory work, and it faces deep disagreement and evolutionary genealogy without any obvious independent confirmation of reliability. That combination is unusual. It is not obviously replicated in epistemology or perception. Where it is replicated, the debunker is entitled to accept the consequence.
The chapter examines several variants of the overgeneralization worry—analogical, entailment-based and practice-entanglement versions—and argues that none of them refute evolutionary debunking without substantive additional assumptions. The lesson is not that debunking arguments are easy or costless, but that they cannot be dismissed merely by noting that morality is not the only normative domain.
CHAPTER FOUR, the last chapter, shifts from metaethics to practical epistemology. Suppose the arguments of the previous chapters succeed, or at least generate serious doubt about robust moral realism. What attitude should agents adopt toward their moral beliefs?
One tempting answer is moral fictionalism. On this view, we should continue to use moral discourse because of its practical benefits—coordination, motivation, social regulation—while suspending belief in its literal truth. Moral claims become akin to idealizations in science or useful myths in politics.
The chapter argues that this position underestimates both the psychology of belief and the structure of moral practice. Belief is not a thin, voluntary attitude that can be neatly bracketed while leaving behavior unchanged. Contemporary cognitive science suggests that belief-like states are deeply integrated into inference, emotion and action. States that guide reasoning, trigger affective responses and structure expectations across contexts are, for all practical purposes, beliefs. Simply labeling them “fictional” at the meta-level does not prevent their cognitive uptake.
This creates a problem for fictionalism. The very features that make moral discourse effective—its automaticity, its emotional force, its resistance to opportunistic revision—are the features that make it belief-like. A community that genuinely treated moral claims as mere pretense would predictably lose much of what makes moral practice function. Conversely, a community that preserved full moral practice would almost certainly preserve moral belief, whether it acknowledged doing so or not.
Moral conservationism takes this psychological reality seriously. It holds that, even if robust moral realism is false or epistemically insecure, we have strong practical reasons to retain our moral beliefs rather than attempting to downgrade them to fiction. This is not because false beliefs are intrinsically good, but because the alternative is unstable and likely self-defeating.
The chapter also introduces a consideration of moral risk. If we abandon moral belief and moral realism turns out to be false, the cost is largely epistemic: We hold some systematically false beliefs that nevertheless structured valuable forms of cooperation and restraint. If we abandon moral belief and moral realism turns out to be true, the cost may be far greater: We risk acting wrongly in ways that matter morally. Given uncertainty at the metaethical level, and given the asymmetry of potential error, conservationism emerges as the more cautious and practically responsible stance.
The argument here is not Pascalian in the sense that it instead relies on ordinary decision-theoretic reasoning under uncertainty, combined with a realistic account of how belief and practice interact. When the costs of false negatives are plausibly higher than the costs of false positives, and when belief is not easily or cleanly optional, preservation rather than revision is often the rational response. So, regardless of whether evolution debunks the reliability of our moral beliefs, it fails to self-defeat and there are good reasons to preserve our moral beliefs and practices.