When researchers presented Israeli settlers and Palestinian refugees with hypothetical peace deals, they expected material sweeteners to help. The sweeteners had the opposite effect. Among participants who regarded the disputed issues, such as land, Jerusalem, and the right of return, as sacred, adding substantial financial compensation increased anger, disgust, and reported support for violence. In the vocabulary of rational choice, making an offer objectively better made it subjectively worse. The deal did more than fail to satisfy. It insulted.

That result, published by Jeremy Ginges, Scott Atran and colleagues in 2007, captures a puzzle that moral psychologists have pursued for three decades. Standard economic and decision-theoretic models assume that goods are commensurable: some exchange rate, however steep, links any two things a person cares about. Sacred values appear to violate this assumption. They do not behave like unusually strong preferences. They behave like something structurally different: a refusal to enter the market at all.

Philip Tetlock defines sacred values as those a moral community treats as possessing infinite or transcendental significance, which rules out comparison or trade-off with secular goods. The research programme that followed has moved from vignettes to fMRI scanners to interviews on active front lines. Its findings bear directly on Joshua Greene's dual-process account of moral judgment. They also matter in practice for anyone trying to model human values, whether across a negotiating table or inside the objective function of an artificial agent.

Tetlock's sacred value protection model builds on Alan Fiske's relational models theory, which holds that human social life runs on a small set of cognitive templates. Market pricing governs exchange through ratios and prices. Communal sharing governs relationships of identity and unconditional obligation. A taboo trade-off occurs when someone applies market logic to a domain governed by communal norms, such as pricing a child, a vote, a homeland, or a friendship. The violation is categorical, not merely quantitative.

The experimental signature is striking. In Tetlock's studies, participants evaluated a hospital administrator deciding whether to spend $1 million on a child's liver transplant or on general hospital operations. Choosing the budget provoked outrage, and the outrage peaked when the administrator deliberated at length before choosing it. Lengthy deliberation over a tragic trade-off, one child's life against another's, was instead read as conscientiousness. Participants treated merely contemplating the forbidden exchange as moral contamination.

Participants exposed to such scenarios also engaged in moral cleansing, expressing greater willingness to volunteer for causes linked to the violated value. Jonathan Baron and Mark Spranca, working on what they termed protected values, documented three complementary features. The first is quantity insensitivity: the number of lives or acres at stake barely shifts judgment. The second is heightened sensitivity to harms caused by action rather than inaction. The third is outright denial that any trade-off exists. None of these patterns fits a steep but finite utility curve.

Yet sacredness proves less absolute under pressure than its holders claim. Baron and Sarah Leshner found that protected values weakened when participants were prompted to generate counterexamples or confront extreme stakes. Societies also trade lives against money constantly, through road safety budgets, regulatory cost-benefit analysis, and the statistical value of a life. This works as long as the exchange is never named as such. Tetlock calls this rhetorical reframing: taboo trade-offs survive by being disguised as routine or elevated to tragic.

This suggests that sacred values function partly as social commitments rather than private preferences. Declaring a value beyond price signals trustworthiness to one's community: a person who will not consider selling a friend can be relied upon not to. On this reading, the refusal to calculate is not a cognitive failure. It is the moral content itself. Openly weighing a sacred good against money reveals that one's relational commitments were contingent all along. In such domains, treating a value as commensurable can itself be an act of betrayal.

Takeaway

Some values are defined less by how much we care about them than by our refusal to put them on a scale. Sometimes the act of weighing is itself the violation.

If sacred values constitute a distinct mode of cognition, the difference should show up in the brain. Gregory Berns and colleagues tested this in a 2012 fMRI study. Participants were offered real money to sign statements contradicting their own beliefs, ranging from mundane preferences to propositions about religion, national identity, and morality. Statements they refused to renounce at any price were classified as sacred, which gave the study a behavioural criterion rather than relying on self-report alone.

Processing these sacred statements was associated with greater activity in the left temporoparietal junction and left ventrolateral prefrontal cortex. Both regions are implicated in semantic retrieval and rule-based reasoning. Notably, the pattern did not centre on circuitry classically linked to subjective valuation and cost-benefit integration. The authors concluded that sacred values are processed as deontic rules, meaning stored propositions about right and wrong, rather than as entries in a utility calculation.

This finding refines Greene's dual-process model. Greene associated characteristically deontological judgments with fast, emotionally charged responses to personal harm, and contrasted them with effortful utilitarian reasoning in dorsolateral prefrontal regions. Berns's data suggest that at least some deontological commitments work by retrieving rules rather than by triggering raw emotional alarm. The automatic system may therefore be less a single emotional reflex than a mix of processes, including well-rehearsed knowledge of norms.

Work by Clara Pretus, Nafees Hamid, Atran and colleagues extended the approach to a high-stakes population: young men of Pakistani origin in Barcelona, some sympathetic to a jihadist organisation. Willingness to fight and die for sacred values was associated with reduced activity in dorsolateral prefrontal regions and related networks linked to deliberately weighing costs. A follow-up study induced social exclusion experimentally. Exclusion increased willingness to fight for previously non-sacred values and shifted their neural processing toward the sacred profile. Sacredness, it appears, is not fixed. It can be manufactured.

The usual cautions apply with particular force. Samples are small. Inferring a cognitive function from activity in a brain region is notoriously unreliable. And a hypothetical willingness to fight is not combat. Still, behavioural markers and neural profiles line up across very different populations, which is hard to dismiss. This supports the claim that sacred and instrumental values are not points on a single continuum. Instead, they appear to draw on partly separate computational strategies: the brain looks up rules for one and computes prices for the other.

Takeaway

Sacred values may not be our strongest preferences but a different kind of mental object altogether: rules retrieved rather than prices computed, and capable of expanding under social threat.

The 2007 findings had a second half that receives less attention. Where material incentives backfired, symbolic concessions worked. Some hypothetical deals included the adversary recognising a sacred value: an Israeli acknowledgement of Palestinian historical suffering, or a Palestinian recognition of Israel's right to exist. These deals reduced opposition and support for violence, even with the material terms unchanged. Atran, Robert Axelrod and Richard Davis heard similar positions in interviews with senior political leaders on both sides.

These results underpin Atran's devoted actor framework. Ginges and Atran found that support for violence in defence of sacred values was largely insensitive to the probability of success, so it behaved more like duty than strategy. Field research with combatants confronting the Islamic State in Iraq linked willingness to make costly sacrifices to two factors. One was identity fusion, in which personal identity merges with a group and its sacred values. The other was the perception that an adversary's strength was spiritual rather than material. Deterrence calibrated to material capability may therefore systematically misjudge such opponents.

Several practical strategies follow. Negotiators can trade symbolic goods, such as apologies, acknowledgements, and ceremonial gestures. These cost little materially but carry weight within the relational model the other side invokes. Atran and Axelrod also advocate reframing: reinterpreting what a sacred value actually requires, as when religious authorities reinterpret obligations in ways that permit coexistence. Postponement can help too. Parties can defer sacred questions while building trust on secular ones, without either side ever being asked to name a price.

The implications extend to machine ethics. Most formal approaches to value alignment represent preferences as single numerical scores or learned reward functions. That architecture presupposes the very commensurability sacred values reject. An AI system that offers compensation where acknowledgement is called for, or treats a community's commitments as adjustable parameters, risks reproducing the backfire effect at scale. More faithful alternatives exist, if computationally awkward ones: strict priority orderings, hard constraints, and explicit models of relational context.

Greene's own normative position complicates the picture. His deep pragmatism treats intuitive moral rules as useful within groups but often obstructive between them, and it favours a common currency of consequences for disputes between groups. Sacred values research partly vindicates this concern, since sacredness entrenches conflict. It also warns that ignoring sacredness is strategically self-defeating. And the same mechanism that makes territorial claims intractable also protects prohibitions against torture and slavery. The question is not whether to hold sacred values, but which ones deserve the status.

Takeaway

In disputes over sacred values, what is acknowledged can matter more than what is offered. Respect may purchase what money cannot.

Sacred values research shows that moral cognition is not a single valuation engine running at different intensities. It includes at least one mode that treats exchange as a category error, relies on rule-based rather than valuation-based processing, and grows more rigid when offered incentives. Commensurability, the foundational assumption of much economics and decision theory, turns out to be a psychological achievement rather than a given.

For philosophical theory, this complicates two positions: the consequentialist hope for a common currency of value, and the deontological claim that inviolable rules reflect rational insight. Sacred values look more like socially calibrated commitments. They can be cultivated, hardened by exclusion, or softened by recognition. Their moral standing must be argued case by case, not inferred from how intensely they are held.

For practice, the lessons span diplomacy, policy design, and AI alignment: model what people refuse to trade, not just what they want. The most consequential values in any dispute may be precisely those that never appear on the bargaining table.