Shame closely tracks the threat of devaluation by others, even across cultures
Daniel Sznycer, John Tooby, Leda Cosmides, Roni Porat, Shaul Shalvi, & Eran Halperin
Proceedings of the National Academy of Sciences, 2016

Abstract: We test the theory that shame evolved as a defense against being devalued by others. By hypothesis, shame is a neurocomputational program tailored by selection to orchestrate cognition, motivation, physiology, and behavior in the service of: (i) deterring the individual from making choices where the prospective costs of devaluation exceed the benefits, (ii) preventing negative information about the self from reaching others, and (iii) minimizing the adverse effects of devaluation when it occurs. Because the unnecessary activation of a defense is costly, the shame system should estimate the magnitude of the devaluative threat and use those estimates to cost-effectively calibrate its activation: Traits or actions that elicit more negative evaluations from others should elicit more shame. As predicted, shame closely tracks the threat of devaluation in the United States (r = .69), India (r = .79), and Israel (r = .67). Moreover, shame in each country strongly tracks devaluation in the others, suggesting that shame and devaluation are informed by a common species-wide logic of social valuation. The shame–devaluation link is also specific: Sadness and anxiety—emotions that coactivate with shame—fail to track devaluation. To our knowledge, this constitutes the first empirical demonstration of a close, specific match between shame and devaluation within and across cultures.

UCSB Press Release

Prominent theories of shame hold that shame is an inherently maladaptive emotion. Shame has an aversive phenomenology and is associated with debilitating conditions such as anxiety, depression and paranoid ideations. When shamed, people withdraw from the scene, destroy incriminating evidence, blame victims and third parties, and commit acts of aggression—even lethal aggression—over lost honor. The costs of shame are particularly apparent in research that contrasts shame with guilt. Guilt motivates wrongdoers to confess, to apologize, to make up to the victim, and to place more weight on the victim’s welfare. Shame looks bad in comparison.

But there is more to shame than costs. To know whether shame is adaptive or not—and to map its neuro-cognitive architecture—the question to answer is not so much Is shame costly?, but rather, What are the payoffs of having a normally developing human brain with the shame system compared to the payoffs of lacking this emotion (or having an impaired shame system)? In particular, is shame any good when others are likely to, or have already, devalued you? Recently, a number of functionally-minded researchers have argued that, in fact, shame is a well-designed system to adaptively deal with challenges of social life.  

Direct tests of the fit between shame and its probable target domain have not previously been conducted, however. In this research we tested the hypothesis that shame, although unpleasant (like pain), serves the adaptive function of defending against the social devaluation that results when negative information about you reaches others. For example, information that you are stingy, untrustworthy, or unfaithful, if it reaches others, tends to make those others less willing to trade off their own welfare in your favor. That is, it makes them devalue you. Behaviorally, they will be less likely to help you and more likely to benefit at your expense or exploit you. Information-triggered devaluation is a dangerous proposition for members of a social species like humans, in particular in the small-scale stateless environments in which our ancestors evolved. With no recourse to savings accounts and impartial police forces, the consequences of being devalued in such environments can be dire. Indeed, the costs of devaluation range from the mild to the lethal (around the world’s cultures, incorrigible wrong-doers are sometimes killed by other community members). This is true also among other non-human primates. This long-standing adaptive problem makes information that would cause others to lower their valuations of the individual a threat to fitness, and hence a selection pressure likely to have left its signature on the human neural architecture.

Considered in a functional light, the costliness of shame is more properly interpreted as the costliness of potential or actual devaluation, with shame operating to reduce or prevent those costs. (Shame may create genuine costs, of course, but, given that actual organisms inhabit a world with imperfect information and unavoidable tradeoffs, the same is true of every other biological organ). And to the disgraced individual himself shame may not be so bad after all. For example, withdrawal protects the shamed individual against aggression from an angry audience and may weaken the formation of common knowledge of the shameful act among the audience. Destroying incriminating evidence may be a legal offense in modern nation states but the benefits to the offender are obvious. Aggression may occur when injuring others is deemed a cost-effective deterrent against being devalued. Admittedly, aggression and the destruction of incriminating evidence are not socially desirable outcomes. Yet the amoral, mindless process that builds shame and other pieces of mental machinery—natural selection—acts relentlessly to enhance the survival and reproductive prospects of the individual and its component designs, not the aggregate welfare of groups (as the carnage and waste observed in the animal kingdom attest).

Similarly, contrasting the costs of shame against the benefits of guilt may not be highly informative. Although these two emotions sometimes fire together, their triggers, architectures, and functions appear to be different. The data suggest that the function of shame is to limit reductions in the weight placed on one’s welfare by an audience. By contrast, guilt seems to function to prevent or remedy events where one put too low a weight on the welfare of a valuable other (often unintentionally), independent of whether the other will know it. A pickup truck is not worse than a Mustang because it’s slower: Those vehicles simply have different functions. Likewise, the design specifications of shame and guilt are different, and so it is rather uninformative to contrast their (societal) goodness.

In past work we have proposed that shame is an emotion program designed to orchestrate motivation, behavior, physiology, and communication defensively, to minimize the likelihood and costs of information-triggered devaluation. We have called this the information threat theory of shame. The information threat theory predicts that the intensity of shame people feel concerning a given item of negative information will track the magnitude of devaluation that would happen if that item became known to others. Both under- and over-activation of shame are costly mistakes. The under-activation of shame would lead to maladaptive choices where (for example) the costs of the resulting devaluation exceed the benefits of the action that provoked the devaluation. Similarly, as with any defensive system, an over-activation of shame would entail diminishing or even negative returns. Decisions about actions must be made in advance of observing feedback about one’s actions, and so the anticipatory activation of shame for a given act should scale with the devaluation an audience would express for that act even when the actor does not know the audience’s actual reaction.

To find out if shame activates in lockstep with the threat of devaluation, we created two dozen brief fictional scenarios describing behaviors or traits that are likely to lead to devaluation: stinginess, infidelity, lack of ambition, and physical weakness, among others. These scenarios were chosen because the values they exemplify should be human universals. One group of participants (audience condition) was asked to report, for each scenario, how negatively they would view another person if those things were true of that person. A different group of participants (shame condition) was asked how much shame they would feel if those things were true of themselves. We conducted this study in the United States, India, and Israel.

As predicted, we observed a very close match between the intensity of shame felt by those imagining themselves in these scenarios, and the intensity of devaluation expressed by those evaluating individuals exhibiting those behaviors or traits. This was true in each of the three countries. On average, 50% of the variance in shame was accounted for by the devaluation expressed by same-country participants—a very large effect size. Recall that the shame and devaluation ratings were generated by different sets of participants. Thus, these correlations cannot be attributed to participants matching their devaluation and shame ratings.

A close match between shame and devaluation would be highly improbable if shame indeed was a maladaptive, pathological emotion. This match is expected, however, if shame was a defense designed to effectively but economically limit devaluation, and to move in lockstep with it. Intriguingly, shame also matched the devaluation of foreign participants. This between-culture match was also very close. On average, 41% of the variance in shame was accounted for by the devaluation expressed by foreign participants in the audience condition. This effect is a bit smaller than the within-culture match, but still it’s remarkably high.

The fact that shame tracks the devaluation of local and foreign audiences suggests that these two systems are informed by a common, universal architecture of valuation—a set of cognitive mechanisms drawing on a species-typical array of evaluative adaptations for mating, reciprocity, kinship, coalitions, and so on. Consistent with this, we also found very large agreement, both within and across cultures, in (i) how negatively participants in the audience condition would view the disgraced individual across the various scenarios, and (ii) how much shame participants would feel if they found themselves in these various scenarios.

Agreement on shame, devaluation, and the relationship between them across cultures, and not just within them, is noteworthy. Non-evolutionary views conceptualize cultures as being richly and arbitrarily different from each other. If this was true, then what people moralize and devalue and what makes people ashamed should be substantially different across cultures. Indeed, shame has long been argued to heavily rely on culture-specific schemas. A stark version of this is the distinction some anthropologists make between shame cultures and guilt cultures. The data we report say otherwise, however. The hypothesis of free cultural variation in values and emotion needs revision.

We also asked whether the match to devaluation is specific to shame or, instead, is a general feature of so-called negative emotions. To find out, we conducted follow-up studies with four between-subjects conditions: one audience condition assessing devaluation, and three emotion conditions asking subjects how much shame, sadness, and anxiety they would feel if the events described in the various scenarios were true of themselves. The shame, sadness, and anxiety ratings were highly inter-correlated, suggesting that these emotions often fire together. But shame was the only emotion that uniquely predicted audience devaluation. Thus, it appears that shame, specifically, is the emotion that deals with the threat of devaluation. In sum, the shame match to audience devaluation is close and specific, like the match of a key and its lock.

The scenarios we created were intended to tap devaluative threats that should be universal and species-typical. There are, however, actions that lead to devaluation in some cultures but not others (mixing meat and milk is wrong for Orthodox Jews, for example, but not for Christians). What happens when the scenarios tap culture-specific values, ones where the valuations of audiences are uncorrelated across cultures?

If shame is an evolved system designed to limit the likelihood and costs of devaluation, it should be tuned to the values of local audiences—the people you encounter in your social world, whose valuation-informed reactions will modify your welfare, your status, and, ultimately, your fitness. Given actions and traits for which the values of local and foreign audiences are uncorrelated, the theory predicts that shame will match the devaluation of local audiences, but fail to match the devaluation of foreign audiences.

By consulting anthropological and historical reports, we created scenarios that should elicit different degrees of devaluation, and shame, in India and the US. When the values of local and foreign audiences were uncorrelated, shame matched the devaluation of local audiences but not foreign ones, as the theory predicts. The match between shame and the evaluations of foreign audiences is found only for scenarios that represent evolutionarily recurrent devaluative threats—ones that tap an evolved, species-wide system for evaluating others.

That a single functional framework can begin to make sense of both universal and culture-specific aspects of an emotion suggests that the antinomy between elastic, idiosyncratic culture and a fixed, robotic biology is false. Instead, the data are more consistent with an evolved architecture of specializations featuring both unlearned content and open parameters sensitive to local social ecology.

Previous studies:

Earlier, related work provides evidence that individual and cultural differences in shame are lawfully distributed: Shame is higher among those individuals with fewer socially valued characteristics and in those places where there are few possibilities to form new (compensatory) relationships. See Cross-cultural differences and similarities in proneness to shame: An adaptationist and ecological approach by Daniel Sznycer, Kosuke Takemura, Andrew Delton, Kosuke Sato, Theresa Robertson, Leda Cosmides, & John Tooby in Evolutionary Psychology, 2012.

Past work also shows that shame governs offenders’ decisions whether to confess to their victims—the more a victim will devalue the offensive act, and the less the information about the act is likely to have leaked into the community, the less likely the offender is to voluntarily disclose his or her wrongdoing. See Regulatory adaptations for delivering information: the case of confession by Daniel Sznycer, Eric Schniter, John Tooby, & Leda Cosmides in Evolution and Human Behavior, 2015.

These papers are part of a more encompassing research initiative that is intended to place the study of emotion and motivation on a new, more rigorous foundation. The goal of this research is to produce detailed maps of the circuit logic of emotions and motivational systems, showing how the details of their problem-solving characteristics match the recurrent adaptive problems our ancestors faced. This program formed the core of a research program funded by a National Institutes of Health Director’s Pioneer Award (to LC) and a grant from the John Templeton Foundation (to JT and LC).

Discussions of this research program can be found in:

Cosmides, L. & Tooby, J. (2013). Evolutionary psychology: New perspectives on cognition and motivationAnnual Review of Psychology, 64, 201-229.

Tooby, J., Cosmides, L., Sell, A., Lieberman, D. & Sznycer, D. (2008). Internal regulatory variables and the design of human motivation: A computational and evolutionary approach. In Andrew J. Elliot (Ed.) Handbook of approach and avoidance motivation. pp. 251-271. Mahwah, NJ: Lawrence Erlbaum Associates

Tooby, J. & Cosmides, L. (2008). The evolutionary psychology of the emotions and their relationship to internal regulatory variables. In M. Lewis, J. M. Haviland-Jones & L. F. Barrett (Eds.), Handbook of Emotions, 3rd Ed. (pp. 114-137.) NY: Guilford.