​Chapter 4

There is a counterargument to everything said so far, and it comes from Jonathan Haidt, one of the most prominent figures in contemporary moral psychology. He would agree with much of the previous three chapters and then disagree at the point that matters.

In his Moral Foundations Theory, moral judgment is automatic and prior to reasoning. When a subject sees an act and declares it wrong, the declaration arrives first and the reasons come after. These arguments are post hoc rationalizations: the mind's press secretary drafting a statement to justify a decision the executive already made. Present someone with a moral scenario, record their judgment, and then dismantle every reason they offer. They will abandon each argument in turn but keep the verdict. "I can't explain why, but I know it's wrong." Haidt called this moral dumbfounding: the moment when the rational scaffolding collapses and the conviction stands there unsupported.

This observation is important. Moral reasoning is largely retroactive; it is a story told to an audience (including oneself) after the verdict has already been delivered. Haidt punctured the Enlightenment fantasy that moral agents are little Kants, calmly universalizing maxims before reaching a conclusion.

Haidt claims that the speed and automaticity of moral judgment reveal its evolutionary foundation. For him, moral intuitions are the products of natural selection: adaptive responses shaped over hundreds of thousands of years of social living. He identifies five (later six) moral foundations, corresponding to an adaptive challenge: care/harm, fairness/cheating, loyalty/betrayal, authority/subversion, sanctity/degradation, and liberty/oppression. These are innate modules, part of the species' cognitive hardware, activated by relevant stimuli.

The picture explains the universality of certain moral themes (every culture cares about fairness and has purity norms), the speed of judgment (hardware is faster than software), and moral dumbfounding (you cannot articulate the logic of a reflex).

Haidt reads the speed of moral judgment as evidence of its pre-rational nature. But speed is a property of any deeply installed program, regardless of origin. A pianist plays a Chopin étude at tempo without conscious deliberation: the fingers move before the mind can track them. This speed reveals the depth of training, not the presence of an innate piano module. A native speaker forms grammatically correct sentences without parsing syntax. The construct was installed early, reinforced relentlessly and has long since dropped below the threshold of conscious access.

Moral foundations do appear across cultures, but so do institutions with convergent needs. Every society requires cooperation (hence fairness norms), group cohesion (loyalty norms), and hierarchy to coordinate collective action (authority norms). They need to manage disgust to prevent disease (hence purity norms). The universality of the themes reflects the institutional pressure, not a common genetic blueprint. The content varies wildly because it is filled in by the local construct. If the foundations were truly innate modules, the content should be far more stable across cultures than it is.

When a subject says "I can't explain why, but I know it's wrong," Haidt sees a moral intuition operating below the reach of reason. We see a construct installed so deep that its justificatory layer has been stripped away. The rationale might have been there once, but the delivery mechanism has been forgotten; only the payload remains. The subject retains the conclusion without the argument, the way an adult retains a childhood fear of the dark without remembering the story that caused it.

Haidt's famous illustration makes the point: the case of Julie and Mark, a brother and sister who decide to have sex. They use contraception, enjoy the experience, agree never to repeat it, and keep it as a private memory. Haidt presents this scenario to subjects and asks: was it wrong? The overwhelming majority say yes. When pressed for reasons — harm? There is none. Consent? Both consented. Consequences? None, by stipulation. The reasons collapse one by one, and the subject is left with pure conviction. "I just know it's wrong."

This result is read as the care/sanctity foundations firing below conscious access. But there is another option. Incest is one of the most heavily policed prohibitions in human society since societies need controlled reproduction to maintain population. The taboo is reinforced universally, backed by disgust conditioning, legal sanction, religious prohibition, and a cultural horror so deep that even fictional incest provokes visceral reactions (the popularity of the theme in Greek tragedy and Game of Thrones trades on exactly this charge). The person who "just knows" that Julie and Mark did something wrong is reporting the output of a construct.

Subsequent research has complicated the dumbfounding claim. Royzman demonstrated that most subjects do not, in fact, run out of reasons and sit in silent conviction. They refuse to accept the stipulated conditions: contraception can fail, the secret will come out, psychological damage is inevitable. When experimenters controlled for this, dumbfounding nearly vanished. This finding is better evidence for the construct model than dumbfounding itself. The subject who invents harms is protecting a construct. Mercier and Sperber have argued that reasoning evolved to produce and evaluate arguments in social exchange; when one argument is refuted, the system produces the next, because persuasion is what it was built for.

If Haidt is right, then moral intuitions are part of the equipment, to be respected, perhaps refined, but trustworthy as a class. If the construct model is correct, then morality is an artifact of installation, and the speed and certainty with which intuitions arrive is evidence of the installation's success. That is the whole point of ontologization: the construct presents itself as something that was always there.

This does not imply that evolution plays no role. Human beings have a capacity for social learning, emotional contagion, coalition detection, and threat sensitivity that is shaped by natural selection. These traits are the hardware on which constructs run. But the hardware does not come pre-loaded with "loyalty to the nation" or "the wrongness of incest."

One may object that inbreeding avoidance is well-documented among animals, particularly primates, and the Westermarck effect (sexual aversion between individuals raised in proximity during early childhood) has empirical support. Grant all of this. The question we ask is not whether inbreeding avoidance exists but why it carries moral weight. Animals that engage in inbreeding do so without any sense of guilt or belief that they are violating a cosmic law. When inbreeding happens (which is frequent enough across various species that biologists consider it normal) it does not lead to any moral dilemma or psychological trauma.

Humans are among the few species that engage in sex primarily for pleasure rather than reproduction alone. The biological pressure that makes inbreeding costly (defective offspring) is largely bypassed by contraception, yet the taboo does not weaken in the slightest. A brother and sister who use contraception and produce no offspring have harmed no one, yet the reaction is as intense as if they had. This is the construct at work: a fact amplified into moral horror, and extended far beyond its logical limits. The taboo no longer tracks the biological risk; it has become an autonomous moral absolute.

Haidt's model, in the end, performs a familiar move: it naturalizes the construct. Labeling moral intuitions as "evolutionary" removes them from the question "who installed this and why?" central to the construct model's inquiry. The label "innate" serves as a more sophisticated version of "self-evident," and both have the same role: to close the inquiry before it reaches the foundation.

The evolutionary framework, in contemporary academic culture, is an ontologized construct: it is presented as the neutral bedrock of explanation, a fact requiring no further justification, while remaining one particular picture of the world among others. When Haidt grounds morality in natural selection, he does not escape the dependence of ethics on ontology — he exemplifies it. His moral psychology protects existing intuitions from scrutiny by granting them the authority of biological nature, which is treated as though it were not an authority at all but the way things are.

But grant him everything. Suppose the modules are genuinely evolutionary. What follows? Not immutability: the content of each foundation varies so wildly that the "innate module" is little more than an empty socket into which the local construct plugs its own wiring. Not authority: if evolutionary origin confers moral legitimacy, then every strategy that natural selection has retained is equally legitimate. Cannibalism among certain Amazonian and Melanesian peoples was a stable adaptive behavior, a means of acquiring protein and absorbing the spiritual force of the enemy. If Haidt's logic holds, this intuition deserves the same deference as any other foundation. There is no mechanism to distinguish the intuitions one wishes to keep from those he would rather forget.