Moral Foundations Theory is a framework in moral psychology proposing that human moral reasoning does not rest on a single value like fairness or harm prevention, but on several distinct intuitive “foundations” that evolved to solve different social problems. Developed by Jonathan Haidt and colleagues, MFT describes morality as a pluralistic set of values shaped by evolutionary pressures and edited by culture, which helps explain why people across the political spectrum and around the world can look at the same situation and reach sharply different moral conclusions. The theory has become one of the most widely cited models in social psychology over the past two decades, influencing research on political polarization, public health messaging, consumer behavior, and even artificial intelligence.
The Foundations Themselves
MFT originally identified five core moral foundations, each tied to a recurring challenge in human social life. Care/Harm is rooted in the need to protect vulnerable members of a group, especially children. Fairness/Cheating reflects the pressures of reciprocal cooperation and the desire to punish free-riders. Loyalty/Betrayal arose from the demands of living in coalitions, where sticking with your group mattered for survival. Authority/Subversion tracks hierarchical relationships, including respect for leaders and traditions. And Sanctity/Degradation is connected to feelings of disgust and the impulse to treat the body, certain objects, or certain practices as sacred or contaminated.
Each foundation comes with a built-in emotional trigger. Care is felt as compassion; violations register as cruelty. Fairness shows up as a sense of justice or outrage at cheating. Loyalty manifests as pride or fury at betrayal. Authority involves respect or feelings of insubordination. Sanctity is experienced as reverence or disgust. The theory treats these emotional responses as fast, automatic intuitions that arrive before reasoned argument, a position drawn from Haidt’s broader “social intuitionist” model of moral judgment.
A sixth foundation, Liberty/Oppression, was added later to capture people’s resentment toward those who dominate them and restrict their freedom. This foundation helps explain both left-wing opposition to corporate or governmental overreach and right-wing resistance to regulation and taxation. It cuts across the political spectrum in a way the original five do not always manage.
Why Liberals and Conservatives Disagree About Morality
The most influential finding to come out of MFT research is the asymmetry in how liberals and conservatives weight the foundations. Across multiple studies using different methods, liberals consistently showed greater endorsement of the Care and Fairness foundations compared to the other three, while conservatives endorsed and used all five foundations more equally.1PubMed. Liberals and conservatives rely on different sets of moral foundations In other words, it is not that conservatives lack compassion or that liberals lack loyalty. The difference is in emphasis: liberals build their moral worldview primarily on two pillars, while conservatives spread their moral weight across a broader base that includes loyalty, authority, and sanctity.
This finding reshapes how we think about political disagreement. From MFT’s perspective, a liberal who opposes military spending and a conservative who supports it may both be acting on sincere moral convictions, just different ones. The liberal foregrounds harm to vulnerable populations, while the conservative foregrounds loyalty to country and respect for military authority. Neither side is being dishonest about their values. They are speaking different moral languages, and the frustration of political debate often comes from each side failing to recognize the other’s moral vocabulary as legitimate.
Moral Reframing and the Art of Persuasion
If people on opposite sides of the political aisle rely on different moral foundations, one practical question follows immediately: can you persuade someone by speaking to their values rather than your own? The answer appears to be yes. Research on “moral reframing” finds that political advocates overwhelmingly default to arguments grounded in their own moral values, not the values of the people they are trying to convince. When arguments are reframed to match the moral foundations of the target audience, they become more persuasive. This pattern held across diverse issues including same-sex marriage, universal health care, military spending, and language policy.2PubMed. From gulf to bridge: when do moral arguments facilitate political influence?
Reframed arguments work because they increase the perceived alignment between a policy position and the listener’s own moral commitments. A separate experiment conducted in South Korea found that moral reframing promoted intuitive, fluent processing and a sense of self-affirmation, which together enhanced how persuasive the messages felt. The effect came not from making people think harder about the arguments, but from making the arguments feel more natural and self-consistent to the audience.3Communication Research. How Moral Reframing Enhances Political Persuasion: The Role of Processing Fluency and Self-Affirmation
The practical implication is straightforward. If you want to convince a conservative to support environmental protection, framing the issue around sanctity (keeping nature pure, not polluting sacred landscapes) may land better than framing it around harm. If you want to convince a liberal to support military funding, framing it around care for soldiers and their families may be more effective than invoking patriotic duty. This is not manipulation; it is translation.
Vaccine Hesitancy and Public Health Messaging
Researchers have tested MFT-informed messaging in public health contexts, and the results are more complicated than the persuasion literature might suggest. In government communication aimed at reducing vaccine hesitancy, the foundations of Authority and Liberty showed the largest positive effects on vaccination rates, with suggestive evidence for a small positive effect from Purity messaging. But Loyalty-based framing actually had a negative effect on vaccination rates, and Harm/Care messaging, the foundation most people would intuitively reach for when promoting vaccines, showed no significant effect at all.4PubMed Central. Using moral foundations in government communication to reduce vaccine hesitancy
A cross-sectional survey in Great Britain during the COVID-19 pandemic added another layer of complexity. People who strongly endorsed the authority foundation and those who more strongly endorsed the liberty foundation tended to be more vaccine hesitant. To a lesser extent, lower endorsement of care and higher endorsement of sanctity also predicted general vaccine hesitancy.5PubMed Central. Evaluating the relationship between moral values and vaccine hesitancy in Great Britain during the COVID-19 pandemic: A cross-sectional survey At first glance this seems contradictory: authority-based messaging helped increase vaccination in one study, while authority endorsement predicted vaccine hesitancy in another. The difference likely reflects context. Whether someone respects authority in general is separate from whether a specific authority-framed message resonates with them. People who value authority may resist vaccines when they perceive public health mandates as overreach, but respond positively when messaging invokes trusted institutions or expert guidance rather than compulsion.
The takeaway for public health campaigns is that moral foundation appeals are not a universal key. Their effects depend on which foundation is invoked, how it is framed, and who the audience is. A blanket “protect your loved ones” message (pure care framing) may feel emotionally compelling to the people crafting it, but the evidence suggests it does surprisingly little to change behavior in vaccine-hesitant populations.
Climate Change and Environmental Behavior
MFT has also been applied to understanding why people differ in their willingness to take personal action on climate change. Research on voluntary public good provision found that Care and Fairness are strong, robust predictors of pro-environmental behavior and willingness to contribute to climate change mitigation. The group-focused foundations, however, matter much less. Loyalty contributes positively only when you ignore the fact that climate change is a global public good rather than a local one.6Ecological Economics. Moral Foundations and Voluntary Public Good Provision: The Case of Climate Change
The logic here maps neatly onto the structure of the problem. The individualizing foundations (Care and Fairness) apply to all people regardless of group membership, which fits a crisis whose benefits from action extend to everyone on the planet. The binding foundations (Loyalty, Authority, Sanctity) are oriented toward protecting the in-group, whether that is family, community, or nation. Since climate mitigation helps people far beyond any in-group, those foundations have less direct motivational power. The group-focused foundations do still matter indirectly by fostering general environmental concern, but they are not the primary moral engines driving climate-related behavior.
How Moral Foundations Are Measured (and Why That Is Contested)
The primary tool for measuring moral foundations has been the Moral Foundations Questionnaire (MFQ), a self-report survey asking people to rate the relevance of various moral considerations and their agreement with moral statements. A newer version, the MFQ-2, was designed to address limitations of the original. Validation work on the MFQ-2 broadly supports a six-foundation structure, demonstrates good reliability, and predicts most external validity scales at meaningful levels.7Personality and Individual Differences. Moral Foundations Theory: Validation and replication of the MFQ-2 The MFQ-2 also needed higher-order group factors for individualizing and binding foundations, and the loyalty foundation split into separate factors for loyalty to a group versus loyalty to a nation, hinting that the original five-category scheme may be somewhat idealized.
That hint becomes louder in studies that push harder on the original MFQ’s factor structure. One large-scale analysis found that the binding foundations do not hold together as cleanly as the theory predicts. In that work, the original five foundations became seven: purity and sanctity split apart, loyalty divided into loyalty to clan and loyalty to country, and authority narrowed into a hierarchy factor. These seven foundations replicated with good fit across samples from the United States, Australia, and China, involving over nine thousand participants total.8PLOS ONE. Remapping the foundations of morality: Well-fitting structural model of the Moral Foundations Questionnaire
Other psychometric work has been even more critical. A reanalysis of the MFQ using data from British voters found that rather than five or six distinct factors, only three meaningful clusters emerged: traditionalism, compassion, and liberty.9PubMed. Reanalysing the factor structure of the moral foundations questionnaire And a cross-national study testing measurement invariance across 27 countries found it difficult to specify MFQ items in a five-factor model that would hold up consistently across diverse populations.10PubMed. Testing Measurement Invariance of the Moral Foundations Questionnaire Across 27 Countries A Brazilian validation study similarly arrived at just two factors: binding morality and individualizing morality.11Psicologia: Teoria e Pesquisa. Validity evidence of a Brazilian version of the Moral Foundations Questionnaire
The measurement picture, in short, is messier than the theory’s clean five-or-six-foundation framework suggests. The broad distinction between individualizing and binding moral concerns holds up robustly. The specific number and boundaries of the foundations within those two clusters remain genuinely debated.
Competing Theories
MFT is not the only game in moral psychology, and its critics have proposed alternatives that challenge the framework at its roots. The Theory of Dyadic Morality (TDM), developed by Kurt Gray and colleagues, argues that all moral judgment ultimately reduces to perceived harm between an agent and a patient. TDM contends that what MFT treats as distinct foundations (loyalty, purity, authority) are actually varieties of perceived harm rather than independent moral dimensions. Studies supporting this view found that moral cognition in both liberals and conservatives is rooted in a harm-based template, and that harm is the most accessible, organizing, and translatable form of moral content. TDM’s proponents argue that past research has substantially exaggerated moral differences between liberals and conservatives.12PubMed. The Unifying Moral Dyad: Liberals and Conservatives Share the Same Harm-Based Moral Template The theory argues explicitly against the idea of modular “foundations” while still embracing moral pluralism through varieties of values and the flexibility of what counts as harm.13PubMed. The Theory of Dyadic Morality: Reinventing Moral Judgment by Redefining Harm
A different challenge comes from Morality-as-Cooperation (MAC), which uses game theory to predict that each distinct type of cooperative problem in human social life should give rise to a distinct moral domain. MAC identifies more moral domains than MFT does, and a self-report measure developed for it, the MAC Questionnaire, has been directly compared to the MFQ in terms of psychometric properties.14Journal of Research in Personality. Mapping morality with a compass: Testing the theory of ‘morality-as-cooperation’ with a new questionnaire Where MFT grounds its foundations in evolutionary psychology and emotional intuitions, MAC grounds its domains in the mathematics of cooperation, which generates a more specific and testable set of predictions about which moral values should exist and why.
These competing perspectives keep the field honest. MFT’s strength has always been its practical utility: it gives researchers, communicators, and policymakers a workable vocabulary for moral diversity. Its weakness, from a scientific standpoint, is the uncertainty around whether the specific foundations it names are truly distinct psychological systems or convenient but somewhat arbitrary groupings of a more continuous moral landscape.
What Happens in the Brain
Neuroimaging studies have started to look at whether the foundations have distinct neural signatures or blur together at the level of brain activity. One functional neuroimaging investigation found that neural activity could distinguish social norm violations from moral foundation violations, and could partially separate individualizing from binding foundations. But the picture was not clean. Authority violations showed a different neural basis from other binding foundations (loyalty and purity), which challenges the idea that the binding foundations form a unified cluster in the brain. Purity violations were persistently associated with amygdala activity, while moral foundation violations more broadly tended to engage what neuroscientists call the default network.15Social neuroscience. A functional neuroimaging investigation of Moral Foundations Theory
The amygdala link for purity violations makes intuitive sense: the amygdala is heavily involved in processing disgust and threat, and purity-related moral judgments (about bodily contamination, sexual taboos, or spiritual defilement) often carry a strong disgust component. The broader engagement of the default network for moral violations generally aligns with the idea that moral reasoning draws on the brain’s systems for social cognition and self-referential thought. But the neural evidence is still too young and too sparse to confirm that each foundation has its own dedicated circuitry. The more conservative interpretation is that the brain processes moral content along a few broad dimensions rather than five or six neatly separable channels.
Culture, Religion, and the Foundations
MFT was built partly to explain cross-cultural moral diversity, and the cross-cultural evidence is both its strongest selling point and an ongoing source of complication. Cultures do vary substantially in which moral orientations they emphasize, shaped by history, religious beliefs, social ecology, and institutional structures like kinship systems and economic markets.16PubMed Central. Universality and Cultural Diversity in Moral Reasoning and Judgment MFT captures this variation well at a broad level: more collectivist societies tend to place greater weight on loyalty, authority, and sanctity, while more individualist societies lean toward care and fairness. But the measurement problems discussed earlier complicate the picture, since the questionnaire items that work in one culture may not carry the same meanings in another.
Religion adds another dimension. A study of moral foundations among religious people in New Zealand found that people who endorsed certain moral foundations tended to identify with distinct religious orientations, and that not all religious people share the same moral profile.17PLOS ONE. Moral Foundations Predict Religious Orientations in New Zealand This is an important corrective to the assumption that religiosity simply equals high sanctity and authority. Different religious orientations pull on different foundations, so the moral psychology of a deeply devout person depends on the kind of devotion, not just its intensity.
Twin studies have also added a biological dimension to the cultural story. MFT predicts that each foundation reflects a heritable trait, and research using common pathway models has found support for strong heritability across the five foundations.18European Journal of Personality. Testing heritability of moral foundations: Common pathway models support strong heritability for the five moral foundations This does not mean moral values are fixed at birth. It means that genetic variation contributes to the differences in how strongly people respond to each foundation, with culture, upbringing, and personal experience layered on top.
Mining Morality From Text and Training AI
MFT has become a go-to framework for computational social science, where researchers want to detect moral content in large volumes of text. The extended Moral Foundations Dictionary (eMFD), a tool built from annotations by a large sample of human coders, outperforms earlier dictionary-based methods for extracting moral intuitions from text corpora.19PubMed. The extended Moral Foundations Dictionary (eMFD): Development and applications of a crowd-sourced approach to extracting moral intuitions from text Another tool, MoralStrength, uses a moral lexicon and embedding similarity to predict which foundation a piece of text activates, reaching much higher accuracy than earlier approaches.20Knowledge-Based Systems. MoralStrength: Exploiting a moral lexicon and embedding similarity for moral foundations prediction
These tools let researchers analyze moral framing at scale in news articles, social media posts, political speeches, and advertising. They can track how the moral language around an issue shifts over time, or how different media outlets frame the same event through different moral lenses. The applications for understanding polarization, propaganda, and public discourse are substantial.
More recently, MFT has been applied to studying the moral reasoning of large language models. A framework called PRIME evaluates the moral priorities of AI systems across multiple ethical dimensions, including the moral foundations.21arXiv. The Convergent Ethics of AI? Analyzing Moral Foundation Priorities in Large Language Models with a Multi-Framework Approach The question here is whether AI systems have implicit moral biases baked in from their training data. If a language model’s training corpus overrepresents text grounded in care and fairness (as English-language media tends to do), the model may systematically underweight loyalty, authority, or sanctity in ways that matter when these systems are used for content moderation, counseling, or policy recommendation. MFT gives researchers a vocabulary for identifying and measuring those biases, even if the theory’s own boundaries remain under debate.