Are These Traits Really Judged as Moral Virtues?
In what sense do these personality, mental health, and cognitive traits have a ‘quasi-moral status?’ There are at least four reasons to think they do – three from social psychology, and one from popular culture.
First, most people show a ‘just world belief’ that creativity, beauty, status, and wealth are merited by those who enjoy them, as both causes and consequences of moral virtue. People judge these traits as morally valenced and morally correlated.
Second, research using the Implicit Association Test shows that many dimensions judged in person perception are highly evaluative, and load on a common good/bad dimension that confounds moral goodness, likeability, pleasantness, status, racial similarity, and physical attractiveness.
Third, there is a powerful ‘halo effect’ around such traits, so they are judged as boosting the likely moral virtues of judged individuals. For example, defendants in criminal cases who are more physically attractive, high in occupational status, and wealthy are more likely to be acquitted or given lighter sentences by juries of their (often lower-status) peers. Conversely, information that a person is morally virtuous boosts ratings of their health and physical attractiveness. Some halo effects may reflect accurate inferences about genuinely correlated traits (‘true halo’), rather than perceiver bias (‘halo error’). In each case, people conflate moral virtues with personality traits, mental health, intelligence, and physical attractiveness.
Finally, the popular culture reason: people often attribute these quasi-moral traits in exaggerated form to mythical beings who have strong moral valences, such as gods, patriarchs, political leaders, movie characters, and comic-book super-heroes. In religion, believers typically credit benevolent deities with supernatural levels of the quasi-moral personality traits (intelligence, conscientiousness, agreeableness, and emotional stability), as well as the standard sexually-selected fitness indicators (size, strength, status, beauty, longevity). In monotheistic religions, these traits are bundled together; in polytheistic religions (such as ancient Egyptian, Greek, and Aztec pantheons), different super-normal traits are attributed to different deities. Contemporary fantasy films and comic books show the standard polytheistic pattern, with different super-normal quasi-moral traits attributed to different super-heroes.
Ever since Socrates, philosophy has tried to develop precise distinctions between theoretical constructs that are often empirically correlated. Most philosophers think in terms of necessary and sufficient conditions, not in terms of factor analysis. Thus, moral philosophers may balk at such flagrantly irrational conflations of moral goodness, social reputation, economic power, and sexual attractiveness. Indeed, they may be tempted to quote a cautionary verse from Ogden Nash: “It’s always tempting to impute / Unlikely virtues to the cute.”
But moral philosophers did not drive the genetic evolution of human virtues; ordinary folks did. If we are seeking a descriptive explanation for human morality, we should attend to the person-perception judgments that may have causally driven moral evolution in our species. Ultimately, it is an empirical question whether ordinary folks judge these traits to have a moral or quasi-moral status when making social and sexual judgments about others.
Are the Moral Virtues Really Sexually Attractive?
Research shows that many particular moral virtues are sexually attractive and relationship-stabilizing; these include kindness, empathy, niceness, honesty, and heroism. Most of these moral-virtue preferences are stronger when seeking a serious long-term partner than a short-term lover. Others, such as the preference for risky, pro-social heroism, may be stronger when females are seeking a short-term male partner.
The problem is that these studies so far cannot distinguish whether the moral virtues are preferred because they signal good genes, good parents, and/or good partners. As with most mate choice research, the first step is to demonstrate a preference that could drive selection for a signaling trait; the second step is to clarify why the preference exists – which often demands more nuanced, more experimental research methods. For example, if a certain moral virtue signals mainly genetic quality rather than parent or partner quality, it should be most preferred by women just before ovulation, within each monthly cycle, when it could be passed along to offspring. Likewise, ‘good genes’ moral virtues should be more preferred in short-term sexual liaisons and extra-pair copulations than in long-term relationships. Conversely, good parent and good partner virtues should be more preferred by women at less fertile cycle phases, and in longer-term relationships. Much more research is needed along these lines.
What About Cross-Cultural Differences in Moral Norms and Mate Preferences for Virtues?
The cross-cultural studies of mate preferences cited above raise the question: if some moral virtues are species-typical, sexually-selected indicators, why do we see large cross-cultural differences in some moral norms and behaviors? Here again, the virtue ethics framework helps clarify different levels of analysis. I suspect that individuals from every culture tend to value intelligence, mental health, and emotional stability as moral virtues in potential mates, but that radically different behaviors can be used to demonstrate these traits in different social, cultural, economic, and ecological contexts.
The two largest cross-cultural studies of mate preferences were coordinated by David Buss in 1989 and by David Schmitt in 2004. Buss and his collaborators asked 10,047 people from 37 cultures to rate and rank-order the desirability of several traits in a sexual partner. Among the top ten traits most desired by both men and women across almost all cultures were: kindness, intelligence, exciting personality, adaptability, creativity, chastity, and beauty. Each of these has at least quasi-moral status in many cultures. Schmitt and collaborators gathered data on 17,804 people from 62 cultures, and found similarly close links between morality and mate choice.
Across most cultures, sexual promiscuity, infidelity, and ‘mate poaching’ were predicted by low agreeableness and low conscientiousness. Also, many studies show that single’s ads across cultures often advertise and seek moral traits – especially kindness, generosity, honesty, fidelity, and capacity for commitment. Ideally, further research would examine cross-cultural preferences for moral virtues using more subtle, indirect, ecologically valid measures (such as revealed preferences in ‘fast dating’ parties for real singles), rather than stated preferences on questionnaires and in single’s ads (which may be biased by strategic self-presentation and adaptive self-deception).
Sexual Selection and Equilibrium Selection
An especially interesting, powerful, and neglected interaction may occur between sexual selection and group-level equilibrium selection. Many evolutionary games have multiple equilibria (states where each player is maximizing their individual payoffs given the strategies already played by others). Some equilibria are better for everybody (they bring net positive payoffs to everyone; they are ‘Pareto-dominant’); some equilibria are worse for everybody (‘Pareto-inferior’), but cannot be escaped easily because individuals who deviate from the equilibrium do even worse. Normally, natural selection alone is not very good at escaping from such Pareto-inferior equilibria to reach Pareto-dominant equilibria. Sexual selection may help, by conferring reproductive benefits on individuals who deviate from selfish, anti-social equilibria. This sexual payoff for virtue is functionally similar to the social-reputation payoffs for virtue modelled by other researchers.
However, standard social-reputation models create a second-order ‘free rider’ problem: who will altruistically take the trouble to punish the wicked and reward the virtuous? As research from behavioral game theory (such as on the Ultimatum Game) shows, most humans are emotionally compelled to impose this sort of ‘altruistic punishment’ of others who act selfishly. The question is, why? Most explanations for this sort of ‘altruistic punishment’ appeal to the dynamics of cultural evolution or social norms, without identifying any plausible individual fitness payoffs for punishing
the wicked.
By contrast, this mate choice model identifies selfish mate-choice incentives (such as good gene and good parent payoffs) for ‘rewarding’ the virtuously punitive with sexual relationships. That is, the effective imposition of punishment on anti-social others (at substantial risks and costs to oneself) should be seen as virtuous, socially status-enhancing, and hence sexually attractive – and recent research suggests that it is. Thus, individual socio-sexual payoffs for virtuous behavior can help solve the collective-action problems that pervade human social life. (Of course, pro-social punishers are likely to gain other costly signaling benefits, being viewed as more dominant, capable, and confident, and thus attracting more friends, allies, and support from kin.)
Most contemporary theories of moral evolution accept the importance of multi-level selection across the genetic, individual, and group levels – either implicitly or explicitly. Generally, group-level selection for prosocial behavior is what ‘breaks the symmetry’ between alternative equilibria in evolutionary games, to allow the evolution of genuine empathy and altruism. This model of sexual selection interacting with group-level equilibrium selection is a potent way that pro-social virtues can establish a genetic beach-head in an otherwise selfish population, long before group-level equilibrium selection can favor morally unified groups.
Sexual Choosiness as a Moral Virtue
In traditional sexual selection theory, from Darwin onwards, there is a crisp distinction between mate preference and preferred trait. However, if mate choice started to favor human sexual fidelity, chastity, and choosiness as moral virtues, the stage would be set for a new positive-feedback process. Just as poor taste in one’s friends and associates can reflect moral inadequacy, poor taste in one’s previous mates can. Thus, in our highly social species, given capacities for observing, remembering, and gossiping about other people’s sexual relationships, a conspicuously excellent capacity for mate choice could come under sexual selection as a preferred trait – perhaps even becoming a costly moral signal in its own right. The resulting evolutionary dynamics of such meta-selection (sexual selection for sexual choosiness) have not, to my knowledge, been explored.
Predictions of the Sexual Selection Model for Moral Virtues
Given the theoretical plausibility of so many moral evolution models, how can they be tested empirically? This sexual selection model makes many discriminating predictions. These often take an unusual form, since costly signaling adaptations have very different phenotypic and genetic features compared to other types of adaptations. In particular, many of these predictions concern individual differences in virtues – not a common research topic in evolutionary psychology or moral philosophy, which tend to focus on species-typical moral judgments and behaviors.
To test most of these predictions, one would need to develop measurement scales that can identify stable individual differences in particular kinds of moral virtues, and that fulfill standard psychometric criteria for reliability and validity. To discriminate between rival theories concerning the evolutionary origins and adaptive functions of specific human virtues, we need to assess the adaptive design features of each putative virtue in reliably quantitative ways. This will require much more psychometrically sophisticated approaches to virtue ethics – not just asking people to give answers to a few multiple-choice ‘trolley problems’ from moral philosophy.
Generally, sexually-selected virtues as quantified in this way should show most of the following twelve features. No one feature alone is strong evidence for the mate choice model. However, these features are highly discriminating when used in combination, as they would not be expected from most other models of moral evolution through other selection pressures (such as kin selection, reciprocity, group selection) – especially those that implicitly depend on stabilizing rather than directional selection. This list is intended mainly to help discriminate between different macro-models of moral evolution (such as sexual selection vs. kin selection, reciprocity, and group selection), although certain criteria are also helpful in discriminating different micro-models within the sexual selection framework (such as virtues evolved as good genes indicators, good parent indicators, or good partner indicators).
Genetic Features of Moral Virtues:
Positive heritability. If virtues are good genes indicators, they should prove genetically heritable in twin and adoption studies, with substantial additive genetic variance maintained by polygenic mutation-selection balance. Many studies report substantial heritability for various forms of anti-social behavior and its personality correlates, such as psychopathy, sensation-seeking, and disagreeableness. Moderate heritability for altruism, empathy, nurturance, and/or responsibility has been found in a few twin studies. Also, if virtues are costly and evolved under sexual selection, the genes underlying virtues should also show a distinctive life-history of heritability, and become more expressed only after sexual maturity, perhaps in response to sex hormones. This should lead to higher heritability of virtues in adults than in children, as has been found with intelligence. On the other hand, if virtues are mainly good parent and good partner indicators, they may show low heritability, though they should still show a distinctive life-history that reflects the shifting costs and benefits of producing good parent/partner indicators depending on whether one is sexually immature, mature but unmated, or mated securely.
Genetic inbreeding and paternal age effects. If virtues are good-genes indicators, heritable variation in virtues should reflect variation in overall mutation load. For example, the offspring of sibling or cousin marriages should show reduced virtue levels, due to the increased expression of harmful homozygous mutations (also known as inbreeding depression). Also, since mutation load in sperm increases gradually as men age, younger fathers should, all else being equal, sire more virtuous children. Such paternal age effects have been shown for various virtue-reducing mental illnesses, such as schizophrenia and autism; future research could use similar methods to investigate other virtues more directly.
Elusive molecular-genetic basis. Specific virtue-reducing alleles should be maintained mostly by mutation-selection balance. Thus, virtue-reducing alleles should be mostly of fairly recent evolutionary origin – recent, harmful mutations that have not yet been eliminated by sexual selection in particular breeding populations. Thus, despite the heritability of specific virtues as found in twin and adoption studies, it should be extremely difficult to find specific “virtue genes” that replicate across human groups using standard linkage and association studies. Rather than finding virtue genes, molecular geneticists should find mostly ‘vice genes’ – lineage-specific, evolutionarily transient, rare, recent mutations that decrease virtue, rather than common haplotypes that increase virtue.
Phenotypic Features of Moral Virtues:
Conspicuous courtship display. During courtship, individuals should conspicuously (if unconsciously) display virtues to the opposite sex. This could be measured across different time-scales, comparing courtship to non-courtship situations across different ovulation cycle stages, relationship stages, and social contexts. For example, ‘priming’ people to think about potential romantic situations increases their displays of conspicuous benevolence and heroism, as it does for displays of conspicuous creativity and conspicuous consumption. Good genes virtues should be selectively displayed more in short-term, opportunistic, or extra-pair mating; good parent/partner virtues should be displayed relatively more in long-term mating.
Condition-dependent costs and positive correlations with other fitness indicators. Virtues should incur a significant cost to produce, in energy, time, risk, or nutritional resources (technically, they should incur a significant relative marginal cost). Without the condition-dependence requirement, this feature sounds tautological, insofar as altruistic virtues always imply evolutionary costs. With the condition-dependence requirement though, we can make more specific falsifiable predictions. For example, individuals with higher genetic fitness or better phenotypic con
dition should be better able to bear the costs of conspicuous moral virtues as good genes indicators, so should more often display those virtues. Thus, good genes virtues should correlate positively with other well-established fitness indicators, such as physical health, mental health, longevity, fertility, body size, body symmetry, and intelligence. In particular, there should be genuine phenotypic correlations between good genes virtues, physical attractiveness, social status, charisma, etc. – not just stereotyped ‘halo effects’ in which more physically attractive people are seen as virtuous. By contrast, moral virtues as good parent/partner indicators may derive their temporal reliability (from early courtship to long-term relationship) not so much from condition-dependence, as from the endogenous stability and increasing heritability of personality traits across the life-span, the social-reputational costs of moral back-sliding, and the ever-looming threat of divorce for moral degeneracy.
Comorbidity among vices, developmental instabilities, and brain abnormalities. If different virtue-deficits (vices) reflect harmful pleiotropic mutations with partly overlapping effects, then vices should show positive genetic correlations (genetic comorbidity) with each other, especially as vices become more serious and extreme. Also, if vices reflect harmful mutations that impair normal neurodevelopment, then they should be associated with various standard brain abnormalities widely observed for other fitness-reducing behavioral traits such as mental illness and mental retardation: smaller cortical volume, larger ventricles, abnormal cortical lateralization, atypical localization of processing as observed in fMRI studies, and so forth.
Higher trait variance in males. In species that evolved with some degree of polygyny and some frequency of extra-pair copulation, the higher male variance and skew in reproductive success should favor a risk-seeking pattern of trait expression, such that male virtue levels show higher variance than female trait values. That is, there should be more conspicuously super-virtuous males (such as Gandhi or Martin Luther King, Jr.), but also more virtue-deficient males (such as Vlad the Impaler or Stalin). (The proclivity of conspicuously super-virtuous males to have covert sexual affairs with many females is evidence for this mate choice model, not evidence that they were unvirtuous).
Virtue Signaling Page 7