Above and below left-right: Ideological narratives and moral foundations Jonathan Haidt, University of Virginia
Jesse Graham, University of Virginia
Craig Joseph, Northwestern University
October 8, 2008, draft 3
(Invited submission to Psychological Inquiry, special issue on political ideology, L. Martin, Ed.)
[8111 words in full document]
Author note: We thank Carlee Hawkins, Ravi Iyer, Selin Kesebir, Jason Kisling, Nicole Lindner, Leonard Martin, Dan McAdams, Gary Sherman, and Richard Shweder for helpful comments on earlier drafts. We thank our collaborators at YourMorals.org—Pete Ditto, Ravi Iyer, and Sena Koleva—for obtaining the data used in the cluster analyses. Contact information: Jonathan Haidt, Dept. of Psychology, University of Virginia, P.O. Box 400400, Charlottesville VA 22904, USA. Email: Haidt@virginia.edu.
Why do people vary in their views of human nature and their visions of the good society? Why do many people categorize themselves as “liberal,” “conservative,” “libertarian,” “socialist,” and so on? Some researchers try to answer these questions by starting with people’s self-identifications and then moving “down,” examining traits (such as openness to experience) that underlie and predict endorsement of an ideological label (see Jost et al., 2003, and Sibley & Duckitt, 2008, for reviews). In contrast, others find it more informative to move “up” from such labels, examining the network of meanings, strivings, and personal narratives that unite the individuals who endorse a label (e.g. Conover & Feldman, 1981; Geertz, 1964; Smith, 2003; Sowell, 1995, 2007).
These two approaches are quite obviously complementary. In this paper we attempt to integrate them by using two theories that were designed explicitly for such cross-level work: Dan McAdams' (1995; McAdams and Pals, 2006) three-level account of personality (1-dispositional traits, 2-characteristic adaptations, and 3-life stories), and our own Moral Foundations Theory (Haidt & Joseph, 2004; Haidt & Graham, 2007). In brief, we argue that the single dimension of left-right is indeed a useful construct that describes a network of Level 2 adaptations (such as right-wing authoritarianism) closely linked to Level 1 traits (such as openness to experience), but the study of ideology requires us to look at the Level 3 narratives of self and society that people construct and internalize as they develop, join groups, and share ideologies. Understanding these narratives may require moving beyond a single left-right dimension to better examine how specific ideologies provide meaning at both the individual and cultural levels. As we will elaborate in this paper, we view the "five foundations" of morality as Level 2 psychological constructs that people use in the construction of Level 3 narratives, including their individual life stories, and also the collective narratives that animate competing political ideologies.
Three Levels of Personality and Ideology
It's hard to argue with success, and the trait approach to personality has been extraordinarily successful, especially after having earlier been marginalized by the critiques of Walter Mischel (1968) and the "situationist" program (e.g. Ross & Nisbett, 1991). Today, the "Big Five" taxonomy is widely accepted as a valuable high-order model of personality, and there is evidence for a degree of heritability for most traits, including many related to political ideology (Bouchard, 2004; McCrae, 1996). Correlational analyses show that people's ratings on a simple left/right or liberal/conservative scale predict an extraordinary variety of other traits, behaviors, preferences, and interactional styles, most of which are related in some way to the tendency for liberals to score higher on measures of the Big Five dimension of Openness to Experience (Carney, Jost, Gosling, & Kiederhoffer, in press; Jost et al., 2003; McCrae, 1996).
Rather than arguing with success, an alternative response is to ask: "is that all there is?" As McAdams (1995) writes, "Reliable and valid trait ratings provide an excellent 'first read' on a person by offering estimates of a person's relative standing on a delimited series of general and linear dimensions of proven social significance" (p. 374). But the very generality of traits is, as McAdams also notes, also their greatest limitation in providing an understanding of individuals; they can only provide "a psychology of the stranger."
McAdams (1995) argues that there are, in fact, three qualitatively distinct "levels" of personality description. Level 1, the lowest level, is "dispositional traits" – global, decontextualized traits such as the Big Five or disgust sensitivity that can be measured with little regard for what else is going on in a person's life. Level 2 refers to "characteristic adaptations" which are, in contrast to Level 1 traits, contextualized and conditional. They include values, goals, attachment styles, defense mechanisms, personal and moral strivings (such as a desire to save the whales or serve Jesus), conditional patterns, and domain-specific skills and talents. These constructs are often empirically related to Level 1 traits – for example, religious strivings for spiritual purity might be stronger in a man who has a high score on disgust sensitivity than in his brother, raised in the same household, who is less sensitive to feelings of disgust. However, Level 2 adaptations are much more variable than Level 1 traits across life stages and contexts, and because they respond to experimental manipulations they are used as both independent and dependent variables in research. Finally, the third and highest level comprises what McAdams calls "integrative life stories." These are even more personal, idiosyncratic, and difficult to quantify. Level 3 centrally concerns identity, and more specifically identity as experienced in a narrative mode. At this level, we would examine the stories people tell themselves and others about how they came to hold the moral and political beliefs they currently hold. We would not expect these stories to be literally true as historical accounts, but we would expect them to influence a person's behavior, including political behavior such as voting and involvement in political movements.
Some psychologists may be skeptical that Level 3 really matters. Of course people tell themselves stories, but we psychologists know that such stories are often made up post-hoc (Nisbett & Wilson, 1977); they might be epiphenomena that can safely be ignored in the study of moral and political behavior. Yet even if such stories are generated post-hoc to justify the gut feelings that draw one to a particular cause, they may still have measurable effects on a variety of outcomes. For instance, Pennebaker has shown how writing about traumatic events in narrative form has both mental and physical health benefits (Pennebaker & Beall, 1986; Pennebaker, 2000). Text analyses of the words used in these narratives revealed that these benefits were predicted by increasing use of insight and causal words, indicating that participants were deriving narrative meaning from the events over the time course of the study (Pennebaker, Mayne, & Francis, 1997). Even in Haidt’s (2001) social intuitionist model, moral reasons that are generated post-hoc play an important role in influencing others, and are therefore necessary for understanding the spread of moral judgments through a population.
The psychological study of ideology is currently undergoing a resurgence (Jost, 2006), fueled by excellent integrative work on Level 1 and Level 2 constructs (Braithwaite, 1998; Jost et al., 2003; Sibley & Duckitt, 2008). There is, however, little recent work on ideology at level 3 (but see Hammack, 2008; Jensen, 1998). The main recent example we know of comes from McAdams and his students (McAdams et al., 2008), who recently collected stories by interviewing 128 highly religious adults about 12 important scenes in their lives. McAdams et al. then content-analyzed these scenes using both Moral Foundations Theory (MFT) and Lakoff’s (1996) “strict father/nurturant parent” model of moral and political psychology. In the next section, we describe MFT and some recent applications of it to moral psychology and the psychology of politics.
Moral Foundations Theory
Moral Foundations Theory was originally designed to analyze cultures, not individuals. It was not intended to be a trait theory, nor a theory about political ideology. Rather, it was created by two psychologists (Haidt & Joseph, 2004) who had worked with the anthropologist Richard Shweder on questions of morality and culture (see Shweder, Much, Mahapatra, & Park, 1997). We were both delighted by the variability of moral practices we read about in ethnographies. We had both tried to map out the moral domain in our fieldwork in Brazil and India (for Haidt) and in Egypt (for Joseph). We both agreed wholeheartedly with Shweder's dictum that "culture and psyche make each other up" (Shweder, 1990). Yet we also both recognized that the psyche was not a blank slate; it contained certain tools or building blocks, provided by evolution, which constrained and enabled the two-way co-construction of culture and psyche. We were influenced by Frans de Waal's (1996) account of these building-blocks—mostly emotional—in chimpanzees and other animals. We reviewed five works that took a "big picture" perspective on morality, including those by Shweder and de Waal, and we listed the virtues (or moral goods, or positive social appraisals) that appeared in any of these works. We did not aim to identify virtues that appeared in all cultures, nor did we try to create a comprehensive taxonomy that would capture every human virtue. Rather, we tried to identify the best candidates for being the psychological foundations upon which cultures create their moral systems.
We found five groups of virtues discussed by at least four of the five theorists. For each one, a plausible evolutionary story had long been told, and for four of them (all but Purity), there was some evidence of continuity with the social psychology of other primates. The five foundations are:
1.Harm/care: basic concerns for the suffering of others, including virtues of caring and compassion.
2.Fairness/reciprocity: concerns about unfair treatment, inequality, and more abstract notions of justice.
3.Ingroup/loyalty: concerns related to obligations of group membership, such as loyalty, self-sacrifice and vigilance against betrayal.
4.Authority/respect: concerns related to social order and the obligations of hierarchical relationships, such as obedience, respect, and proper role fulfillment.
5.Purity/sanctity: concerns about physical and spiritual contagion, including virtues of chastity, wholesomeness and control of desires.
The moral foundations are psychological systems that enable people to perceive actions and agents as praiseworthy or blameworthy, but we don't think of them primarily as individual-level traits. They are more like taste buds of the moral sense: everyone has them, yet moral "cuisines" differ around the world. Different cultures build upon the foundations in different ways, and what they build is everything we would call moral life: values, norms, virtues, vices, institutions, even religions (which of course draw on many psychological systems besides the five foundations). We therefore do not and cannot measure the foundations directly; rather, we measure the degree to which individuals endorse and value the culturally-constructed virtues and concerns built on one or more foundations. We created the Moral Foundations Questionnaire (the MFQ; Graham, Haidt, Nosek, Iyer, Koleva, & Ditto, 2008) to do just this, using abstract assessments of the moral relevance of foundation-related concerns, as well as endorsement of more contextualized moral judgments. The foundations as we measure them with the MFQ are therefore most assimilable to McAdams' Level 2 characteristic adaptations. Foundation scores do indeed correlate in meaningful ways with constructs at the first two levels, including low-level personality traits (e.g., scores on Purity/sanctity correlate r = .34 with disgust sensitivity), and more complex ideological constructs (e.g., scores on Authority/respect correlate r = .65 with Right-Wing Authoritarianism). But as we'll see, fully appreciating and understanding the varieties of moral experience will require integrating analyses at all three levels.
One of our goals in creating Moral Foundations Theory was to broaden the scope of inquiry in moral psychology (cf. Haidt, 2008). We wanted researchers to think about issues beyond Kohlberg's (1969) ethic of justice and Gilligan's (1982) ethic of care. But once we began taking this broader perspective, we immediately began to see the "culture war" in the United States in a new way. It seemed to us that on controversial issues such as abortion, gay marriage, and the death penalty, those on the left end of the ideological spectrum held moral values based primarily on the Harm and Fairness foundations, whereas those on the right had moralities based relatively equally on all five foundations (Haidt & Graham, 2007). We tested this hypothesis and found it to be true using a variety of measures, including the MFQ, content-analysis of liberal and conservative church sermons, and measures of people's willingness to violate taboos related to each foundation (Graham, Haidt & Nosek, 2008).
We have argued that the foundations are useful not just as another set of personality variables that correlate with political preferences, but as an explanatory framework with which to understand the meaning of moral debates in the culture war. For instance, the passions and intractability of the gay marriage controversy make more sense once you understand that the left side sees legalizing gay marriage as a straightforward way to reduce harm (to innocent victims) without hurting anyone else, while simultaneously increasing fairness (including issues of equality and rights). Using just the harm and fairness foundations, one simply cannot construct convincing arguments against gay marriage. The absence of good arguments based on harm and fairness leads liberals to conclude that conservatives are motivated by simple and immoral homophobia. Cultural conservatives, however, are more likely to see gay people as members of a different culture (attacking or infiltrating the heterosexual ingroup) who subvert gender roles (rejecting the authority of church, law, and tradition) while pursuing a carnal and hedonistic lifestyle (including “impure” sexual acts that trigger feelings of disgust). The opposition of these social conservatives may well be linked “downwards” to traits such as openness to experience, but it must also be linked “upwards” to the third level of meaning and identity. Consciously or unconsciously, opposition to gay marriage is related for some of these people to Judeo-Christian narratives of virtue, sex, and self-control, such as the story of Adam and Eve. We cannot truly understand the opposition to gay marriage using only Levels 1 and 2, for example by showing that conservatives are low on openness to experience and high on disgust sensitivity. There’s a lot more going on.
Many Settings on the Moral Equalizer
One way to think of the moral mind is to use the analogy of an audio equalizer with five slider switches. Each switch – Harm, Fairness, Ingroup, Authority and Purity – can be thought of as an independent parameter of moral functioning. If we imagine each moral volume switch going from 1 to 11, then there are 161,051 possible patterns of settings, in theory. But what about in practice? Are there a few major “presets,” or patterns of settings, that correspond to the major ideological positions?
To find out, we examined the database of survey responses we have collected at YourMorals.org, a website where over 25,000 people have taken the MFQ, and many of them completed additional scales related to moral or political psychology. Data reduction techniques (such as factor analysis) are usually used to group similar scale items together. But we wanted to group similar people together, so we performed cluster analyses on people’s averaged scores for each of the five foundations on the MFQ. We restricted the analysis to the 20,962 respondents who lived in the United States and had not grown up in another country. Our goal was to identify clusters of respondents based on their MFQ scores, and then to characterize the clusters on a number factors, including basic demographics, the Big Five, and a variety of moral and ideological measures.
In our first analysis, the two-step cluster procedure was permitted to determine the optimal number of clusters on its own. This analysis produced two clusters with a clear liberal/conservative split, both in terms of ideological self-placement and in terms of the patterns of MFQ settings we have found in previous studies (Graham, Haidt & Nosek, 2008). This provided support for a basic left-vs.-right view of morality that we ourselves have made use of (Haidt & Graham, 2007). But we wanted to see if that was the end of the story: if we looked for alternative clusterings, would the groups all fall along a single left-right dimension, or would other patterns emerge?
In subsequent analyses, we constrained the program to find three, then four clusters. The three-cluster solution revealed a group whose MFQ patterns looked different than liberal or conservative patterns (and was not a midpoint between the two), and the four-cluster solution revealed yet another distinctive, but interpretable, cluster. We focus here on the four-cluster solution.
If we looked only at the ideological self-placement of the people in these four clusters, we would interpret the clusters as representing adjacent regions of the one-dimensional left-right scale. Their self-ratings on our 7 point scale (where 1 = “very liberal” and 7 = “very conservative”) were 1.92 (Cluster 1), 2.63 (Cluster 2), 3.31 (Cluster 3), and 4.99 (Cluster 4). However, this linear pattern obscures a number of important non-linear effects, some of which are visible on the MFQ scores. Figure 1 shows a bar graph for each cluster representing the mean scores of the people in that cluster on each of the five foundation scores of the MFQ. These four graphs can be interpreted as four common pre-sets on the moral equalizer, just as commercial equalizer programs often have presets for playing rock, jazz, classical, and hip-hop. A close inspection of these MFQ scores, along with the other data we collected from these participants, indicates that the four clusters represent distinctive political and moral ideologies that go beyond left-right.
-- INSERT FIGURE 1 ABOUT HERE --
Cluster 1 is clearly the prototypical secular liberals we have described in previous publications (Haidt & Graham, 2007). People in this cluster had, on average, the highest scores on Harm and Fairness, and very low scores on Ingroup, Authority, and Purity. They had the highest scores on Openness to Experience and the lowest scores on Right-Wing Authoritarianism (RWA) and Social Dominance Orientation (SDO). People in this cluster (and in cluster 2) were twice as likely to describe themselves as atheists (13%) than were people in clusters 3 (6%) and 4 (7%).
Cluster 4 is clearly the prototypical social conservatives we have described elsewhere: they had the lowest scores on Harm and Fairness, and very high scores on Ingroup, Authority, and Purity. They had the lowest scores on Openness and the highest scores on RWA and SDO, as well as the highest frequency of religious attendance (40% reported attending a few times a month or more, compared to just 14% in Cluster 1).
Clusters 2 and 3 were not simply intermediate or “moderate” groups, dividing up the middle of the left-right spectrum. Cluster 2 is in a sense a noncontinuous hybrid of the “liberal” and “conservative” clusters: lower scores on Harm and Fairness, approximating those of the conservative cluster, and lower scores on Ingroup, Authority and Purity, approximating those of the liberal cluster. Moreover, almost 60% of self-identified libertarians are found in this cluster, a finding that is consonant with the fact that this cluster has the highest average score on Schwartz and Bilsky’s (1990) “hedonism” value, very low scores (like liberals) on condemning abortion, homosexuality, and other issues that matter to conservatives, and the lowest scores of the four clusters on condemning non-culture-war moral violations such as gambling and tax cheating. In other words, people in this cluster seem to have the moral volume turned down across the board. Consistent with our characterization of them as libertarians, their moral foundation settings seem to deny the general value of externally-imposed moral regulation of any kind.
Cluster 3 also seems to uniquely combine liberal and conservative aspects, resembling liberals with high scores on Harm and Fairness, but resembling conservatives with high scores on Ingroup, Authority and Purity. Approximately 59% of these respondents placed themselves in one of the three “liberal” categories on the seven-point scale, and another 20% described themselves as “neutral.” Yet on religious observance, they resemble the conservative cluster much more than the liberal cluster: 36.1% said they attended religious services at least a few times a month, compared to 39.6% for the conservative cluster and just 14.1% for Cluster 1. These facts suggest that this cluster might best be characterized as a “Religious Left” group, a tentative interpretation that is given some support by the fact that this group scored highest on Schwartz and Bilsky’s benevolence, tradition, conformity, security, and spirituality dimensions. Cluster 3 is, in a sense, the ideological opposite of Cluster 2. In cluster 3, participants seem to have the moral volume turned up across the board, preparing them to create a “thick” moral worldview in which people have many obligations to each other, both as individuals and as group members. Both of these clusters represent moral worldviews not captured by a single left-right dimension.
We are, of course, not the first to suggest a two-dimensional representation of ideology. While some have suggested that liberalism and conservatism are separate orthogonal dimensions (Kerlinger, 1984), the most common multidimensional finding is that of separate bipolar dimensions for economic and social issues (Knight, 1993), i.e., one liberal-conservative continuum for social issues and a distinct continuum for economic issues. This conceptualization is also currently finding mainstream appeal via popular outlets like politicalcompass.org. Two-dimensional conceptualizations of ideology (usually presented as a 2x2 table yielding four basic types or groups) are also common in values research (Feldman, 2003). Schwartz’s (1994) values matrix has one dimension for openness-conservatism and one for self-enhancement-self-transcendence; similarly, Braithwaite’s (1997) model has one dimension for valuing harmony and equality, and another for valuing national security and order. Rokeach’s (1973) separate dimensions for freedom and equality valuations yield the four ideological quadrants of fascism (low on both values), capitalism (high on freedom, low on equality), communism (high on equality, low on freedom), and socialism (high on both).
Our cluster analysis of moral foundation concerns revealed a similar kind of “four-square,” based on low/high contrasts of individualizing and binding concerns. These two dimensions map onto Rokeach’s dimensions fairly well, with the important difference that Rokeach’s “freedom” value has been inverted in the binding concerns of Ingroup, Authority, and Purity. This allows us to see the positive group-level moral concerns of social conservatives and the religious left as more than just a lack of “freedom” values—a characterization most of them would surely reject. Moreover, our cluster analysis reveals important moral ideological differences within one nation: these clusters are not the classic ideological opponents that terms like “fascist” or “communist” would suggest. Rather, they point to how much ideological variety the moral foundations can reveal within a single modern capitalist culture.
MFT sheds new theoretical light on these dimensional groupings with its bases in anthropological and evolutionary thought, and we think the foundations offer the most useful way to conceptualize and understand ideology from a moral perspective. Nevertheless, these cluster analyses are meant to be illustrative, not definitive. We present them to show how Moral Foundations Theory can be used to categorize people into groups and then help us make predictions about the moral values and ideals shared within those groups. But to understand the ideologies of these groups, we must go beyond the networks of correlated traits and adaptations that describe our clusters. We must move up to McAdams’ level 3 and try to find stories that our participants themselves would endorse.
Life stories, according to McAdams (2001, p. 101), are "psychosocial constructions, coauthored by the person himself or herself and the cultural context within which the person's life is embedded and given meaning." McAdams et al. (2008) collected such life stories from 128 highly religious adults and used them to test Moral Foundations Theory. They developed an extensive coding scheme to link themes and topics in the interviews to each of the five foundations, and then examined the frequency with which liberals and conservatives relied upon each of the foundations when narrating their lives. They replicated the basic pattern we have repeatedly found using the MFQ: self-ratings of politics (on a scale of 1= very liberal and 5 = very conservative) correlated negatively with use of the Harm/care (r = -.33) and Fairness/reciprocity (r = -.27) foundations, and it correlated positively with use of the Ingroup/loyalty (r = .35), Authority/respect (r = .43), and Purity/sanctity (r = .39) foundations. (All correlations were significant at p < .01, as were the standardized betas in regression analyses controlling for other demographic factors). McAdams et al. (2008, p. 987) help us to imagine how their subjects think about their own morality by summarizing the differences in this way:
When asked to describe in detail the most important episodes in their self-defining life narratives, conservatives told stories in which authorities enforce strict rules and protagonists learn the value of self-discipline and personal responsibility, whereas liberals recalled autobiographical scenes in which main characters develop empathy and learn to open themselves up to new people and foreign perspectives. When asked to account for the development of their own religious faith and moral beliefs, conservatives underscored deep feelings about respect for authority, allegiance to one's group, and purity of the self, whereas liberals emphasized their deep feelings regarding human suffering and social fairness.
If we want to understand the Level 3 narratives of people in our four clusters, we could follow the lead of McAdams et al. (2008) and ask people in each cluster to tell us their own personal stories. We could then content-analyze those stories and see if the patterns of moral foundation usage match the four graphs in Figure 1. But we suggest that there is another class of stories, much easier to obtain, that can be used to create links between Levels 2 and 3 in the study of political psychology: ideological narratives.
In The Political Brain, Drew Westen (2007) argues that successful political movements must have a “master narrative,” a story that explains the origins of our present problems and shows why the movement is the solution. He points out that coherent stories usually have an initial state ("once upon a time..."), protagonists, a problem or obstacle, villains who stand in the way, a clash, and a dénouement. These "ideological narratives," as we will call them, are clearly like life stories in some ways, but different in some ways too. Ideological narratives incorporate a reconstructed past and imagined future, often telling a story of progress or of decline, like the redemption and contamination narratives that McAdams finds are common in the individual life stories of adults in midlife (McAdams & Pals, 2006). But life stories cannot be shared; each person must have her own, and each person must be the first author of that story. Ideological narratives, in contrast, are successful only to the extent that large numbers of people accept the same ones (although they may edit their own versions to better complement their personal life stories). These ideological narratives are usually grander than life stories, often reaching back centuries or millennia for their "once upon a time," casting larger groups and forces as the actors, and justifying epic actions, reforms, and even violence as the way to reach the dénouement.
Ideological narratives have the great advantage that there is only a small number of major ones circulating in a society at any given time. Many versions can be found in books (such as the campaign biographies of presidential candidates) and on political web pages (such as nearly anything called a "manifesto," or even sometimes a mission statement). Some scholars and movement leaders have done us the favor of extracting them and condensing them down to just a few sentences. Here we present four such narratives and show how they match the moral foundations settings shown in the four graphs of Figure 1. We recognize that each of our four clusters contains its own diversity, and we can be sure that many members of each cluster would reject the narrative we associate with it. Nonetheless, we predict that a larger number of participants in each cluster would endorse the narrative, would endorse that narrative more than the other three narratives, and would prefer to have their ideology expressed in this way, as a story that makes claims about what is right and wrong, rather than simply having themselves described by a series of psychological traits.
Cluster 1: Secular Liberalism The sociologist Christian Smith (2003, p. 64) has observed that we are “animals who make stories but also animals who are made by our stories.” Smith describes a variety of high-order, often unconscious narratives that organize identity and moral judgment at both the individual and group levels. One of these he calls the “liberal progress” narrative:
Once upon a time, the vast majority of human persons suffered in societies and social institutions that were unjust, unhealthy, repressive, and oppressive. These traditional societies were reprehensible because of their deep-rooted inequality, exploitation, and irrational traditionalism... But the noble human aspiration for autonomy, equality, and prosperity struggled mightily against the forces of misery and oppression, and eventually succeeded in establishing modern, liberal, democratic, capitalist, welfare societies. [However,] there is much work to be done to dismantle the powerful vestiges of inequality, exploitation, and repression. This struggle … is the one mission truly worth dedicating one’s life to achieving.
Consistent with the first graph in Figure 1, the liberal progress narrative makes extensive use of the Harm foundation ("suffering," "misery," "oppression") and the Fairness foundation ("unjust," "inequality"). There is no mention of ingroup or nation, and no mention of purity or sanctity. Authority and tradition are mentioned only as the sources of harm and injustice.
Cluster 2: Libertarianism
For libertarians, the most important value, the good that may not be sacrificed to any other, is – as the name of this position implies – individual liberty. Libertarians are ever vigilant against infringements of liberty, even infringements motivated by the most sincere commitments to other worthy values, such as equality (see, e.g., Fried 2007). The novelist and philosopher Ayn Rand is one of the most iconic of libertarian thinkers. Rand did not speak for all libertarians, but she is revered by many, and her novels were naked ideological narratives. In these novels, “once upon a time” refers to the awful years of socialist oppression of the individual and worship of egalitarian mediocrity; the hero is a creative and rugged individualist who refuses to conform; and the dénouement is the restoration of freedom, which makes prosperity and happiness possible for everyone. In a 1964 interview with Playboy magazine (reprinted in Boaz, 1997) she stated her narrative goal frankly: "I seek to provide men—or those who care to think—with an integrated, consistent, and rational view of life.” She then described how her personal life story motivated her ideological story:
When I came here from Soviet Russia, I was interested in politics for only one reason—to reach the day when I would not have to be interested in politics. I wanted to secure a society in which I would be free to pursue my own concerns and goals, knowing that the government would not interfere to wreck them, knowing that my life, my work, my future were not at the mercy of the state or of a dictator's whim.
Knowing that Rand's father's pharmacy was confiscated by the Bolsheviks, that she was an outspoken atheist, and that she viewed the rapid rise of socialism (including Roosevelt's New Deal) with alarm, her rejection of the three binding foundations makes sense. Her extreme celebration of individualism also helps us understand her rejection of most liberal applications of the Harm/care foundation (she had a Nietszchean contempt for the weak, for the "leeches" of society) and most liberal applications of the Fairness/reciprocity foundation, especially those related to equality of outcomes (which communism had enforced horrifically). Like the people in our Cluster 2, Rand would likely have scored low on all five foundations on the MFQ (which does not at present capture the central libertarian virtue of unfettered liberty).
Cluster 3: The Religious Left
Jim Wallis, head of the Sojourners movement that has become synonymous with the religious left, lays out his vision and call to action by decrying both the left and right sides of the traditional ideological spectrum:
The religious and political Right gets the public meaning of religion mostly wrong—preferring to focus only on sexual and cultural issues while ignoring the weightier matters of justice. And the secular Left doesn't seem to get the meaning and promise of faith for politics at all—mistakenly dismissing spirituality as irrelevant to social change. I actually happen to be conservative on issues of personal responsibility, the sacredness of human life, the reality of evil in our world, and the critical importance of individual character, parenting, and strong 'family values.' But the popular presentations of religion in our time (especially in the media) almost completely ignore the biblical vision of social justice and, even worse, dismiss such concerns as merely "left wing." It is indeed time to take back our faith (Wallis, 2005, 3-4).
In a series of books, sermons and press releases, Wallis has spun a rich narrative of Christianity in which the “once upon a time” is the centuries before Jesus, the time of the prophets whose messages to humanity were broadcast with high settings on the Harm slider (e.g., Isaiah’s prophecy of a society free from poverty, calamity and hunger) and the Fairness slider (e.g., Amos’ and Micah’s lamentations about the injustices of wealth and power inequalities), setting the stage for the moral attention Jesus paid to the suffering poor and to oppressed minorities. But as the old and new testaments both make extensive use of all five foundations, so Wallis and Sojourners aim to right wrongs related to Harm and Fairness while embracing the group-centered foundations of Ingroup, Authority and Purity that secular liberals usually shy away from (e.g., crusades against pornography and sexualization in the media, family values, community bonding, and the importance of respect for traditions). Consistent with Cluster 3, this group combines the high liberal settings on Harm and Fairness with the high conservative settings on Ingroup, Authority and Purity. (We take this opportunity to note that foundations are foundations, not final structures. The religious Left and religious Right use each of the five foundations in different ways to construct two very different but very "thick" moral worldviews, both of which are despised by Libertarians.)
Cluster 4: Social Conservatism
Westen (2007) draws from several of Ronald Reagan's speeches to construct the master narrative that he says has guided the Republican Party from the early 1980s until today. We condense that story here:
Once upon a time, America was a shining beacon. Then liberals came along and erected an enormous federal bureaucracy that handcuffed the invisible hand of the free market. They subverted our traditional American values and opposed God and faith at every step of the way… Instead of requiring that people work for a living, they siphoned money from hard-working Americans and gave it to Cadillac-driving drug addicts and welfare queens. Instead of punishing criminals, they tried to “understand” them. Instead of worrying about the victims of crime, they worried about the rights of criminals…Instead of adhering to traditional American values of family, fidelity, and personal responsibility, they preached promiscuity, premarital sex, and the gay lifestyle...and they encouraged a feminist agenda that undermined traditional family roles... Instead of projecting strength to those who would do evil around the world, they cut military budgets, disrespected our soldiers in uniform, burned our flag, and chose negotiation and multilateralism… Then Americans decided to take their country back from those who sought to undermine it. (Westen, 2007, p.157-158)
This narrative is saturated with themes from all five foundations. It shows ingroup concerns with patriotism and America's enemies; authority concerns with traditional roles and the pervasive metaphor of "undermining;" and purity concerns with sexual morality. Harm is the foundation least in evidence (aside from "victims of crime"), although fairness concerns are pervasive, and pervasively different from liberal applications of fairness. Conservatives invoke fairness as reciprocity, particularly toward those who cheat or break the law; they are not concerned with equality of outcomes, which is so central in social justice movements.
Examining these four narratives together can yield insights into the shifting and often puzzling dynamics of American politics. One can see, for example, how the Republican Party forged an electoral majority by uniting two groups whose moral foundation scores and personality profiles are quite far apart. Republicans were traditionally the party of business, and they drew in Libertarians (or "laissez faire conservatives") more strongly by articulating a moral critique of the way liberals in the 1970s had implemented their Harm and Fairness concerns in pursuit of social justice, which often required heavy-handed regulation and intervention in business practices. But the big story is that the party of business captured most of the evangelical and Catholic vote through its forceful critique of a society that was disintegrating and going "down the toilet," and therefore desperately in need of a thicker morality based in large part on the three binding foundations of ingroup, authority, and purity.
Conclusion: Multi-Level Understanding and Empathy
Our objective in this paper has been to make a case for expanding the psychological study of ideology to incorporate the Level 3 ideological narratives that give individuals and groups a sense of meaning and purpose. As an example of this approach, we employed Moral Foundations Theory to explore the diversity of moral patterns in a large dataset. We found two patterns that perfectly exemplified the endpoints of the left-right dimension, but we also found two patterns that could not be neatly placed along it. We tentatively labeled these groups “Libertarian” and “Religious Left.” Despite working from a very large dataset with dozens of ideologically-relevant scales and variables, we came up against the limits of traditional Level 1 and Level 2 methodology. We were still engaged in McAdams’ (1995) “psychology of the stranger,” still fundamentally trying to understand moral ideologies from the outside. We turned to other sources in an attempt to understand – at Level 3 – how these people might see the world and integrate the many values, ideals, and policy preferences that sometimes seem inscrutable or downright evil to unsympathetic outsiders.
Psychologists now possess a huge store of knowledge (much of it reviewed in meta-analyses by Jost et al., 2003, and Sibley & Duckitt, 2008) regarding the personality traits and other Level 1 variables associated with ideology, especially conservatism. There is also a growing body of work on ideology’s interactions with Level 2 personal concerns and characteristic adaptations, such as attachment style (e.g., Weise et al, 2008), system justification processes (Jost, Banaji, & Nosek, 2004) and terror management theory (Pyszczynski, Solomon, & Greenberg, 2003). But psychologists have far less understanding of the Level 3 constructs that are so important in shaping opinions and attitudes and in drawing people into or out of political movements (Westen, 2007). Without becoming ethnographers, we could easily supplement our heavily questionnaire-based methodologies with other techniques that would offer at least the beginnings of a Level 3 understanding of networks of meaning. Methods that have proved fruitful for the study of ideology include multidimensional scaling of rating or sorting data (e.g., Wish, Deutsch and Biener, 1970), Q-method studies of conservatism (Brown 1970; Brown & Taylor, 1972) and of both liberalism and conservatism (Kerlinger, 1984), projective tests such as the TAT (e.g. Rothman & Lichter, 1996), and, of course, life narrative interviews (McAdams et al., 2008) and in-depth interviewing generally.
The approach to ideology we’re advocating is nothing new. Over 40 years ago, the anthropologist Clifford Geertz warned that progress in understanding ideology at Level 2 can make researchers forget that ideology must also be described at Level 3:
Whatever else ideologies may be—projections of unacknowledged fears, disguises for ulterior motives, [emotional] expressions of group solidarity—they are, most distinctively, maps of problematic social reality and matrices for the creation of collective conscience (Geertz, 1964, p. 220).
Geertz would have disdained an exclusive focus on aggregate quantitative data, correlation, and data reduction. He urged social scientists to see things “from the native’s point of view,” and to offer “thick descriptions” of informants' "experience-near" concepts – that is, words and ideas that a person would use naturally and effortlessly when talking about things that matter. "Love" is an experience-near concept; "object cathexis" is experience-distant and unrecognizable to the person in love. Applied to the study of ideology, "intolerance of ambiguity" and "low openness to experience" are experience-distant ways of describing what social conservatives might call a respect for the authorities, traditions, and order that makes it possible for people to live together in a dangerous and unstable world. Geertz argued that a full understanding comes only from moving back and forth between the experience-near and experience-distant perspectives. The challenge, he said, is to combine the two approaches
so as to produce an interpretation of the way a people lives which is neither imprisoned within their mental horizons, an ethnography of witchcraft as written by a witch, nor systematically deaf to the distinctive tonalities of their existence, an ethnography of witchcraft as written by a geometer (Geertz, 1984, p. 125).
Psychologists, being among the most politically liberal of academic fields (Redding, 2001), are at special risk for producing studies of conservatives that are "deaf to the distinct tonalities of their existence."
But listening to people telling stories about themselves, how they came to hold their views, and how they understand the story of our society turns research participants briefly from objects into subjects, invites the listener (or researcher) to do some perspective taking, and makes it easier for the listener (or researcher) to entertain hypotheses that go against her own ideological proclivities. We therefore believe that a three-level approach to the study of ideology will produce better science, deeper understanding, and perhaps even more civil politics.
Boaz, D. (Ed.) (1997). The libertarian reader. New York: Free press.
Braithwaite, V. (1997). Harmony and security value orientations in political evaluation. Personality and Social Psychology Bulletin, 23, 401-414.
Braithwaite, V. (1998). The Value Orientations Underlying Liberalism-Conservatism. Personality and Individual Differences, 25(3), 575-589.
Brown, S. B., & Taylor, R. W. (1972). Perspective in Concept Formation. Social Science Quarterly, 52(4), 852-860.
Brown, S. R. (1970). Consistency and the Persistence of Ideology: Some Experimental Results. Public Opinion Quarterly, 34(1), 60-68.
Bouchard, T. J. J. (2004). Genetic influence on human psychological traits: A survey. Current Directions in Psychological Science, 13, 148-151.
Carney, D. R., Jost, J. T., Gosling, S. D., & Kiederhoffer, K. (in press). The Secret Lives of Liberals and Conservatives: Personality Profiles, Interaction Styles, and the Things They Leave Behind. Political Psychology.
Feldman, S. (2003). Values, ideology, and the structure of political attitudes. In Sears, D. O., Huddy, L. & Jervis, R. (eds.), Oxford Handbook of Political Psychology. Oxford, UK: Oxford University Press.
Fried, C. (2007). Modern Liberty and the Limits of Government. New York: W.W. Norton & Company.
Geertz, C. (1964). Ideology as a cultural system. In C. Geertz (Ed.), The interpretation of cultures (pp. 193-233). New York: Basic Books.
Geertz, C. (1984). From the native's point of view: On the nature of anthropological understanding. In R. Shweder & R. LeVine (Eds.), Culture theory. Cambridge: Cambridge University Press.
Gilligan, C. (1982). In a Different Voice: Psychological Theory and Women's Development. Cambridge, MA: Harvard University Press.
Graham, J., Haidt, J., & Nosek, B. A. (2008). Liberals and Conservatives Rely on Different Sets of Moral Foundations. Manuscript submitted for publication.
Graham, J., Haidt, J., Nosek, B.A., Iyer, R., Koleva, S., & Ditto, P.H. (2008). Development and validation of the Moral Foundations Questionnaire. Manuscript in preparation, University of Virginia.
Haidt, J. (2001). The emotional dog and its rational tail: A social intuitionist approach to moral judgment. Psychological Review, 108, 814-834.
Haidt, J. (2008). Morality. Perspectives on Psychological Science, 3, 65-72.
Haidt, J., & Graham, J. (2007). When Morality Opposes Justice: Conservatives Have Moral Intuitions that Liberals May not Recognize. Social Justice Research, 20(1), 98-116.
Haidt, J., & Joseph, C. (2004). Intuitive Ethics: How Innately Prepared Intuitions Generate Culturally Variable Virtues. Daedalus, 12.
Hammack, P. L. (2008). Narrative and the Cultural Psychology of Identity. Personality and Social Psychology Review, 12, 222-247.
Jensen, L. A. (1998). Different habits, different hearts: The moral languages of the culture war. American Sociologist, 29, 83-101.
Jost, J. T. (2006). The end of the end of ideology. American Psychologist, 61.
Jost, J. T., Glaser, J., Kruglanski, A. W., & Sulloway, F. J. (2003). Political Conservatism as Motivated Social Cognition. Psychological Bulletin, 129(3), 339-375.
Jost, J.T., Banaji, M.R., & Nosek, B.A. (2004). A decade of system justification theory: Accumulated evidence of conscious and unconscious bolstering of the status quo. Political Psychology, 25, 881-919.
Kerlinger, F. N. (1984). Liberalism and Conservatism: The Nature and Structure of Social Attitudes. Hillsdale, N.J: Lawrence Erlbaum.
Knight, K. (1993). Liberalism and Conservatism. In Robinson, J. P., Shaver, P. R., & Wrightsman, L. S. (eds.), Measures of Political Attitudes. San Diego: Academic Press, 59-158.
Kohlberg, L. (1969). Stage and sequence: The cognitive-developmental approach to socialization. In D. A. Goslin (Ed.), Handbook of socialization theory and research. Chicago: Rand McNally.
Lakoff, G. (1996). Moral politics: How liberals and conservatives think. Chicago: University of Chicago Press.
McAdams, D. P. (1995). What Do We Know When We Know a Person? Journal of Personality, 63(3), 365-396.
McAdams, D. P. (2001). The psychology of life stories. Review of General Psychology, 5, 100-122.
McAdams, D. P., & Pals, J. L. (2006). A New Big Five: Fundamental Principles for an Integrative Science of Personality. American Psychologist, 61(3), 204-217.
McAdams, D.P. Albaugh, M., Farber, E., Daniels, J., Logan, R. L., Olson, B. et al. (2008). Family Metaphors and Moral Intuitions: How Conservatives and Liberals Narrate Their Lives. Journal of Personality and Social Psychology, 95, 978-990.
McCrae, R. R. (1996). Social consequences of experiential openness. Psychological Bulletin, 120, 323-337.
Mischel, W. (1968). Personality and Assessment. New York: Wiley.
Nisbett, R. E., & Wilson, T. D. (1977). Telling more than we can know: Verbal reports on mental processes. Psychological Review, 84, 231-259.
Pennebaker, J. W. (2000). Telling stories: The health benefits of narrative. Literature and Medicine, 19, 3-18.
Pennebaker, J. W., & Beall, S. K. (1986). Confronting a traumatic event: Toward an understanding of inhibition and disease. Journal of Abnormal Psychology, 95, 274-81.
Pennebaker, J. W., Mayne, T. J., & Francis, M. E. (1997). Linguistic predictors of adaptive bereavement. Journal of Personality and Social Psychology, 72, 863-71.
Pyszczynski, T., Solomon, S., & Greenberg, J. (2003). In the wake of 9/11: The psychology of terror. Washington, DC: American Psychological Association.
Redding, R. E. (2001). Sociopolitical diversity in psychology: The case for pluralism. American Psychologist, 56, 205-215.
Rokeach, M. (1973). The Nature of Human Values. New York: Free Press.
Ross, L., & Nisbett, R.E. (1991). The Person and the Situation: Perspectives of Social Psychology. New York: McGraw-Hill.
Rothman, S., & Lichter, S. R. (1996). Roots of Radicalism: Jews, Christians, and the Left. New Brunswick, NJ: Transaction Publishers.
Schwartz, S. H. (1994). Are there universal aspects in the structure and content of human values? Journal of Social Issues, 50, 19-46.
Schwartz, S. H., & Bilsky, W. (1990). Toward a theory of the universal content and structure of values: Extensions and cross-cultural replications. Journal of Personality and Social Psychology, 58, 878-891.
Shweder, R. A. (1990). Cultural Psychology: What is It? In Cultural Psychology: Essays on Comparative Human Development. Cambridge: Cambridge University Press.
Shweder, R. A., Much, N. C., Mahapatra, M., & Park, L. (1997). The "big three" of morality (autonomy, community, and divinity), and the "big three" explanations of suffering. In A. Brandt & P. Rozin (Eds.), Morality and health (pp. 119-169). New York: Routledge.
Sibley, C. G., & Duckitt, J. (2008). Personality and prejudice: A meta-analysis and review. Personality and Social Psychology Review, 12, 248-279.
Smith, C. (2003). Moral, Believing Animals: Human Personhood and Culture. New York: Oxford University Press.
de Waal, F. (1996). Good Natured: The Origins of Right and Wrong in Humans and Other Animals. Cambridge, MA: Harvard University Press.
Wallis, J. (2005). God's Politics: Why the Right Gets It Wrong and the Left Doesn't Get It. San Francisco: HarperSanFrancisco.
Weise, D. R., Pyszczynski, T. , Cox, C. R., Arndt, J., Greenberg, J., Solomon, S., & Kosloff, S. (2008). Interpersonal Politics: The Role of Terror Management and Attachment Processes in Shaping Political Preferences. Psychological Science, 19, 448-455.
Westen, D. (2007). The Political Brain: The Role of Emotion in Deciding the Fate of the Nation. New York: PublicAffairs.
Wish, M., Deutsch, M., & Biener, L. (1970). Differences in Conceptual Structures of Nations: An Exploratory Study. Journal of Personality and Social Psychology 16(3), 361-373.
Figure 1. Moral foundation patterns in four clusters.
Note. H=Harm, F=Fairness, I=Ingroup, A=Authority, P=Purity. Ns for each cluster are as follows: 5946 (cluster 1), 5931 (cluster 2), 6397 (cluster 3), 2688 (cluster 4). Error bars represent +/- 2 S.E.