Biologically caused and irreducibly social

John Searle (Reference Sieff and Leiby2004) argues that the mind can be causally reduced to the brain, but not ontologically reduced to the brain. The stuff of consciousness cannot be reduced to its causes without losing what it means to have consciousness. Commenting on Thomas Nagel’s question, ‘What is it like to be a bat?’ Searle (Reference Sieff and Leiby2004, 60) observed that we can know everything imaginable about bat neurophysiology, but we cannot know ‘What is it like to be a bat? What does it feel like?’ It feels like something to have consciousness that only one with consciousness can know, just as it feels like something to be a bat that only a bat can know. Consciousness has a first person, subjective ontology.

Compare consciousness to a setting sun or to how rational choice theorists think of trust, both of which have a third person, objective ontology. One can causally reduce a setting sun (to the earth’s rotation) and one can ontologically reduce it (to the earth’s rotation). There is one phenomenon, not two, and only an illusion is lost with the reduction (Searle Reference Sieff and Leiby2004). Or, rational choice theorists view ‘trust’ as a consequence of incentives. One can causally reduce rational trust to incentives, and one can ontologically reduce it to incentives for the content is no different than its cause. Rational choice theorists eviscerate the concept, turning trust into nothing but incentive-driven behavior (Mercer Reference Mercer2005; Eznack Reference Eznack2011; Rathbun Reference Rensmann2012).

Emotion is more like consciousness than a setting sun or rational trust. It has a first person, subjective ontology. Emotion can be causally reduced to the body (because nothing other than the body can cause emotion) but it cannot be ontologically reduced to the body (because it feels like something to have emotion). The content cannot be reduced to the cause. The mind cannot be reduced to the brain because it feels like something to have a mind, just as it feels like something to be a bat. If emotion were reducible to the body then it would have a third-person objective ontology and we could know what it feels like to be a bat. Or, one could in principle build a machine capable of experiencing guilt or build a toaster that could fall in love. The observation that emotion is ontologically irreducible checks the reductionist impulse that drives some neuroscientists to believe that because the brain causes the mind, the mind is (like free will) an illusion (Harris Reference Harris2012). Some psychologists (Barrett et al. Reference Barrett, Mesquita, Ochsner and Gross2007) detect reductionism whenever a scientist treats activity in certain brain areas (or changes in biochemical composition) as emotion.

Emotion cannot be reduced to ‘atoms of feelings’ (James Reference James1884, 11). For example, pain exists only if it is experienced (Searle Reference Sieff and Leiby2004). Even if scientists conclude that someone should be in pain, the pain does not exist if it is not experienced. The causes of pain are not identical with the experience of pain. One is not reducible to the other. Neurons are not red, painful, or sweet, though they can produce these properties (Feinberg Reference Feinberg2012, 19). Even in cases where an illusion causes pain (as in phantom limb syndrome), the experience of pain is real. That emotion is ontologically subjective does not preclude the objective study of it. Just as phantom limb pain is ontologically subjective, a scientific study of the illusion’s cause was the basis for the ingenious solution of using mirrors to remind the mind that the limb is gone (Chan et al. Reference Chan, Witt, Charrow, Howard, Magee, Pasquina, Heilman, Malcom and Tsao2007). Physical processes of emotion (e.g. firing of neurons) are different than emotional content (e.g. feeling pain), and only with a description of the content (or what one feels) can scientists know what they need to explain (Barrett et al. Reference Barrett, Mesquita, Ochsner and Gross2007, 374).

Emotion as a biologically caused and irreducibly social phenomenon is a central tenet of psychological constructivism, which contrasts with the ‘basic emotion’ view that has dominated psychology. Traditionally, psychologists created basic or core emotions (somewhat like primary colors) that in combination produce all other emotions. These basic emotions were thought to exist in nature, independent of human perception, and this conceptualization led to an endless (and tedious) taxonomy of emotions (Lindquist Reference Lindquist2013; Ellsworth Reference Ellsworth2014). Anger or fear was thought to have distinct, universal, physical characteristics. Yet, the hunt for the essence of anger or attempts to identify discrete physical correlates of discrete emotion has failed. Neuroscientist and psychological constructivist Lindquist (Reference Lindquist2013, 360) (see also Barrett Reference Barrett2013) observed that 100 years of psychological research ‘has yet to identify the discrete bodily, facial, behavioral, or neural basis of English emotion categories such as ‘“anger”, “disgust”, “fear”,“happiness”, and “sadness’”. Discrete emotion does not exist in the body because it emerges from ambiguous bodily sensations, external sensations (visual, tactile, auditory), as well as from culture and context (Boiger and Mesquita Reference Boiger and Mesquita2012; Lindquist Reference Lindquist2013; Hutchison and Bleiker Reference Hutchison, Emma and Bleiker2014, 491–514). How one experiences an emotion and what one views as an emotion changes with culture, language, and over time. In the late 19th century, ‘vicious delight’ in hurting animals was a normal American emotion (Ellsworth Reference Ellsworth2014, 22). Simply put, emotion categories are ‘socially constructed’ and influence the emotions one experiences (Lindquist et al. Reference Lindquist, Wager, Kober, Bliss-Moreau and Barrett2012, 125; Ellsworth Reference Ellsworth2014).

The belief that one’s own experience of emotion represents distinct bodily states helps to explain skepticism toward group-level emotion. Neuroscientist and psychological constructivist Barrett (Reference Barrett2009) complained that psychologists (and I would add political scientists) take too seriously the phenomenology of emotion, which involves the creation of psychological categories based on experience. Scholars often believe that their subjective experience of anger, for example, reflects a discrete entity (called anger) that can be mapped onto the brain or discovered in the body. This tendency can lead one to confuse psychological facts with physical facts (Barrett Reference Barrett2009). For example, because the experience of ‘feeling’ and ‘thinking’ is different, psychologists wrongly assumed the differences would be discovered in the brain. Or, political scientists assume that rational beliefs do not depend on emotion because thinking does not feel like an emotion (Mercer Reference Mercer2010). Phenomenology is taken too seriously when it leads to reification of psychological categories.

At the same time, political scientists skeptical of group emotion do not take phenomenology seriously enough. Because no group brain and body exists on which to map the experience of group emotion, skeptics think group emotion cannot exist. In the skeptic’s view, the experience is an illusion because it cannot have distinct physical correlates: no group body means no group emotion. But anger, love, pride, nationalism, and collective guilt are ontologically subjective entities. They are psychological (not physical) facts. Group emotion is real and others within (and often outside of) one’s culture recognize the experience as real. Because emotion is ontologically irreducible to the body, the absence of a group body is not a limitation of group emotion. The distinction between emotion and group emotion is conceptual, not physical. Only to material reductionists does the absence of direct physical correlates make something not real. Group emotion is as real as other categories that people create and agree upon such as normative structures, emotion, and identity.

Structure, emotion, and identity

What emerges from interacting individuals can be called properties of a group, and these properties can be called structures (Le Boutillier Reference Le Boutillier2003, 65). Group emotion is an ideational structure. Political scientists are accustomed to thinking of ideas as mental and emotion as physical, in which case emotion is not part of an ideational structure. Even scholars who might otherwise reject this mind/body dualism slip into an ideas vs. emotion dichotomy. Whereas seeing and hearing are distinct processes and one does not confuse one for the other, feeling and thinking are causally indistinct. As Barrett et al. (Reference Barrett, Mesquita, Ochsner and Gross2007, 390) put it, the distinction between cognitive events and emotional events ‘is probably phenomenological rather than causal’. We experience beliefs, thoughts, and memory as distinct from emotion even though they are not causally distinct. Psychologist Ellsworth (Reference Ellsworth2014, 24) (see also McDermott Reference McDermott2004; Stein Reference Tajfel and Turner2013) observed: ‘By now it has been amply demonstrated that cognition and emotion are largely inseparable’. Emotion and ideas are part of the same process and their distinction is overdrawn. Ideas can be stored in books or buried in the back yard, but so can emotion: emotion remains vibrant in Shakespeare’s Hamlet or in Henryk Górecki’s Symphony No. 3.Footnote ¹

Group-level emotion is no more dependent on a group body than ideational structures are dependent on a group brain. Although some ideas seem more dependent on emotion than others, for skeptics of group emotion the distinction between emotion and ideas must be categorical. If emotion must be reduced to individuals, then ideational structures containing even a whiff of emotion (such as norms or culture) must also be reducible to individuals. Separating the ideas part from the emotion part of culture, or the ideas from emotion in norms, or as two neuroscientists suggest, distinguishing cognition from emotion in the brain ‘might resemble trying to slice a cake into the flour and sugar that went into it’ (Zaki and Ochsner Reference Zaki and Ochsner2011, 23). The differences are conceptual not causal.

Culture, norms, and group emotion can be causally but not ontologically reduced to individuals. These structures are neither identical to, nor wholly autonomous from, the individuals who constitute them. Ideational structures are semi-autonomous (Le Boutillier Reference Le Boutillier2003; List and Spiekermann Reference List and Spiekermann2013). Emergence makes it possible to speak of group emotion (as well as norms and culture) as real without implying reification, which in this case means separating the structures from the agents that produce them (Wendt Reference Wendt1999, 146). What Young (Reference Zaki and Ochsner1990, 44) said of groups is equally true of group emotion: ‘Groups are real not as substances, but as forms of social relations’. Ideational structures are real without possessing an ability to think and feel. Social emotion is an emergent property of groups. How one experiences that emotion – what one feels – depends on one’s identity.

Psychologists Smith and Mackie (Reference Smith, Seger and Mackie2008, 436) capture what is distinct about a social identity approach to emotion when they note that ‘emotions pertain to an identity and not to a biological individual’. Who we are is what we feel. Identity and emotion depend on each other. Identification without emotion inspires no action for one does not care. Whereas indifference makes identities meaningless (and powerless), emotion makes them important. Pride in one’s group or hate of one’s enemy presupposes identities that one cares about. Or, whether one’s religious identity is important depends on how one feels – and sometimes how others’ feel – about that identity. Emotion makes identity consequential, and identity makes group-level emotion possible. They both depend on, but are not reducible to, individual bodies. As psychologist Smith (Reference Smith and Mackie1993, 303) neatly summarized: ‘Because the self is not limited by the skin, neither are emotions’. Identity, emotion, and consciousness (and culture, norms and group emotion) depend on the body without being reducible to it. Emotion goes with identity, not the body.

Social identity theory (SIT) explains group-level emotion. Psychologists Tajfel and Turner (Reference Tarrant, Branscombe, Warner and Weston1986) created SIT to better understand intergroup discrimination. Tajfel returned home to Poland after being captured by the Germans and found that his parents, brother, and other family members had been murdered in the Holocaust (Jahoda Reference Jahoda2008). The experience led Tajfel to conduct a series of now famous experiments to better understand discrimination and group identity. Even in the absence of power, material interests, and, as then imagined, without emotion, Tajfel found that categorization (or the creation of groups) created comparisons that triggered discrimination against out-groups. These experiments provided the basis for SIT, which posits that a desire to feel good about one’s own group explains why people in groups discriminate against people in other groups for no apparent reason.

Traditional psychological accounts of intergroup conflict exclude emotion (Parkinson et al. Reference Parkinson and Simons2005, 143). As late as 2008, two psychologists thought it ‘odd’ that the SIT literature paid almost no attention to emotion (Smith and Mackie Reference Smith, Seger and Mackie2008, 429). Putting emotion into SIT is easier than keeping it out (Mercer Reference Mercer2005; Sasley Reference Searle2011). Identification depends on a feeling of attachment. The more important the social identity, the more important is a positive view of one’s group. Researchers have found a strong relationship between positive in-group emotions and in-group identification, as well as a relationship between anger at an out-group and in-group identification (Smith and Mackie Reference Smith, Seger and Mackie2008, 433–34). The more positively one feels about one’s group the more one will identify with it (and vice versa); the angrier one is at an out-group, the stronger one’s in-group identification. How one feels about one’s group influences how one feels about other groups.

Mechanisms for group emotion

The validity of the argument for group-level emotion depends on showing how one can go from a personal identity to a social identity. Four mechanisms can explain the switch.

First, culture regulates emotion. Culture is an ideational structure that partly consists of emotion. Societal values provide criteria for evaluating events as good or bad. How one feels often depends on culturally framed interpretations (Crawford Reference Crawford2000; Parkinson et al. Reference Parkinson and Simons2005; Lebow Reference Lebow2008). In response to a rising tide of Afghan soldiers murdering their NATO counterparts, which was sometimes attributed to feelings of anger at being disrespected, the Afghan government issued a pamphlet for its soldiers on cultural differences: members of NATO do not mean to be offensive when they pat your back or behind, put their feet up on a desk when speaking to you, wink at you, ask about your female relatives, or expose their private parts when showering (Sieff and Leiby Reference Smith2012). Culture also influences cognition. Psychologists believed that cognition (unlike emotion) was immune to culture and thus highly generalizable. But as psychologists expanded the pool beyond white, educated, industrialized, rich, and democratic (or WEIRD) test subjects, they discovered that culture shapes cognition (Henrich et al. Reference Henrich, Heine and Norenzayan2010). Cultural neuroscientists find that ‘both the structure and the function of the developing human brain is shaped both by the environment and by cultural experiences’ (Chiao and Ambady Reference Chia and Ambady2007, 238). Culture changes the brain’s architecture. For example, divergent philosophical traditions lead to differences in neural activity (Chiao and Ambady Reference Chia and Ambady2007). Culture influences feeling and thinking, which provides one explanation for why feeling like a group is common.

Second, because people in a group are likely to interact most with members of their group, or at least care most about those interactions and use the in-group as their key referent group, group members are likely to influence each other. Experiencing group emotion is an expression of group identity, a reinforcement of that identity, and a way to maintain group boundaries. Emotion structures relationships. A group without emotion is a mere collection of autonomous individuals. Being a group member means experiencing and expressing similar feelings and a failure to do so is a signal that one is not really part of the group (Parkinson et al. Reference Parkinson and Simons2005). Group members dislike and sometimes reject group members who violate norms. The more one identifies with a group, the more one conforms to group attitudes and beliefs, and the more vigilant one is that others uphold these norms (Eznack Reference Eznack2011; Kesebir Reference Kesebir2012; McDoom Reference McFarland, Webb and Brown2012).

Being part of a valued social group generates esteem and meaning that is matched only by the esteem and meaning one gains by distinguishing one’s group from other groups (Wenzel Reference Wight2009). For example, music and rituals that are distinct to a group can contribute to a feeling of groupness. White power music known as ‘hatecore’ helps to sustain the white supremacist subculture. It rouses resentment and anger. A former leader of a hatecore band recalled his reaction after hearing the British skinhead group, Skrewdriver: ‘Listening to that music was an essential part of how we rallied around the idea of racism. It made me feel I was part of something greater, that I had purpose and that my race was something very special and was something I needed to defend’ (Dao and Kovaleski Reference Dao and Kovaleski2012). Emotion constitutes identity that then structures experiences within groups as well as between them (Tarrant et al. Reference Waytz and Epley2012b).

Third, emotion is contagious. As most people know and as psychologists confirm, other people’s emotion influences one’s emotion. This can happen because we use how others feel as evidence of how we should feel. My excitement at a decision can influence your appraisal and make you more excited. One study concluded that, ‘when two adults are both focused on a common object, their appraisals of that object and affective reactions to it often become calibrated’ (Parkinson and Simons 2009, 1081). In antagonistic situations, my alarm might trigger your excitement (or your anger might trigger my alarm). One can also catch emotion without being fully aware of how one caught it. Acting in synchrony with others – which includes marching, singing, and dancing with others, and might be extended to protesting, defending, or working closely with others on a task – is another mechanism for emotional contagion (Goodwin et al. Reference Goodwin, Jasper and Polletta2001; Wiltermuth and Heath Reference Wolf2009; Kesebir Reference Kesebir2012). Physical synchrony or even proximity might not be necessary for contagion. Researchers have identified what might be contagion effects among people using on-line social media, such as Twitter networks (Bollen et al. Reference Bollen, Gonçalves, Ruan and Mao2011). Exactly how contagion works remains unclear, though it seems likely that the pursuit of common goals, a close relationship, a common social identity, or synchrony makes contagion most likely (Parkinson and Simons 2009; Wiltermuth and Heath Reference Wolf2009; see also Ross Reference Sasley2014).

Fourth, events that have group-level implications elicit common group-level reactions. When group membership defines one’s interest, then one would expect events concerning the group to elicit common reactions (Smith Reference Smith and Mackie1993). A social identity means that others in one’s group are likely to share one’s interpretation of another group or event. This might explain why 69% of Greeks – responding to German-led economic austerity and German insults – felt German politicians were attempting to establish a Fourth Reich (Heyer and Batzoglou Reference Heyer and Batzoglou2012). The stronger the identification, the greater should be the impact of the group on the individual (Parkinson et al. Reference Parkinson and Simons2005). For example, traumatic events elicit shared emotional reactions within that group even when the experience is not first hand (Hutchison Reference Hutchison2010). Europeans expressed greater fear after September 11 when they categorized themselves as part of the ‘west’ (which put them into the same group as Americans) than when they identified themselves as ‘Europeans’ (Tarrant et al. Reference Tarrant, Calitri and Weston2012a). Another study found that British women responded differently to photos of the 7 July 2005 London bombings depending on which identity the experimenter cued: feelings of aggression and support for war against terrorism was common when British identity was salient, but not when gender was salient (Fischer et al. Reference Fischer, Haslam and Smith2010). When these British women were presented with details of Taliban treatment of women, aggression and support for revenge was most common when gender (rather than British identity) was salient. Different identities meant different interests. When one’s group identity defines one’s interest, then events that have group-level implications elicit common reactions.

Group emotion can be distinct from an individual’s emotion

How one feels as an individual cannot always be generalized to how one will feel as a group. For example, the literature on intergroup relations emphasizes the importance of empathy. In dyadic face-to-face settings, empathy leads to greater fairness (though in multiparty settings it can be viewed as unfair favoritism; Cropanzano et al. Reference Cropanzano, Stein and Nadisic2011). Taking an individual’s perspective leads one to view the individual as more similar to oneself; the more one thinks about another, the more similar the other becomes and thus the more sympathetically one views the other. People use their own feelings as a way to understand the mental states of others; this probably happens because we use the same brain processes to think of others as we do when we think of ourselves (Zaki and Ochsner Reference Zaki and Ochsner2011; Mercer Reference Monroe2013). The discovery that the natural hormone oxytocin contributes to feelings of warmth and trust, and vice versa, encourages the view that empathy provides one way to reduce intergroup conflict and to facilitate greater respect and understanding (Riling and Sanfey Reference Riling and Sanfey2011). The popular press dubbed oxytocin as a new love drug.Footnote ² Mirror neurons might also produce these feelings for another so that one feels another’s pain (Holmes Reference Holmes2013), though not literally. For example, the experience of empathic fear is different from the experience of fear and it engages different parts of the brain (Zaki and Ochsner Reference Zaki and Ochsner2011). And of course, mirror neurons cut both ways. When a person we perceive as selfish receives a mild electric shock, the only thing we feel is a little bit of pleasure (Haidt Reference Haidt2012).

Psychological studies on empathy typically focus on individuals not on groups, but social identities are crucial (Tarrant et al. Reference Waytz and Epley2012b). An individual who happens to be a member of a group relates to other individuals in ways that differ fundamentally from how a member of a group relates to other in-group and out-group members. The need for a positive social identity creates between group competition that can generate out-group derogation and discrimination when one needs to defend one’s social identity (Tarrant et al. Reference Waytz and Epley2012b). Studies have found that oxytocin makes people care more about their own group but not about out-groups, it makes people more willing to hurt other teams (in PD games) when doing so was the best way to protect their own group, and it makes men more willing (in hypothetical situations) to value in-group lives more than out-group lives (Haidt Reference Haidt2012). In one study, oxytocin motivated strong in-group favoritism among Dutch men with designated out-groups as Arab immigrants and Germans. The primary motivation was in-group favoritism, not out-group derogation (though derogation was sometimes evident). The authors found that the effect of in-group favoritism across the experiments was ‘strikingly similar’, which suggested to them that cultural norms and other between-group differences were unimportant (De Dreu et al. Reference De Dreu, Greer, Van Kleef, Shalvi and Handgraaf2011, 1264). Oxytocin’s effects are different at different levels of analysis. When one feels like a group, oxytocin promotes ethnocentrism.

Group emotion is often stronger than individual emotion

Group-level emotion can be more powerful than the individual experience of emotion because one experiences it as objectively true and externally driven, rather than as subjective and individually constructed (Smith et al. Reference Stein2007). Emotional consensus gives one confidence in the appropriateness of one’s feelings. Wolf (Reference Young2011, 118) observed that ‘many international reactions to respect or disrespect may be just as strong or even stronger than responses on the interpersonal level, depending on the intensity with which people identify with their nations’. It is not my feeling; it is our feeling. Group-level emotion means one is upset that one’s group has been insulted, not that one was singled out for insult. One feels this insult not on behalf of the group, but as a member of the group. In a person-on-the-street interview during massive Egyptian protests against an American film ridiculing the Prophet Mohamed, Samir Anwar said: ‘I am against those who produced, financed, and acted in the movie. I demand that they be publicly executed. They didn’t just insult the Prophet Mohammed. They insulted all Muslims’ (PBS Newshour 2012a). Group emotion means it is not my interpretation of an event that makes me angry; the event is ‘intrinsically angering’ (Smith et al. Reference Stein2007, 442).

An American atheist can imagine what it is like to be an Egyptian Muslim upset at a movie ridiculing the Prophet Mohamed, or an Ethiopian can imagine what it is like to be a Chinese nationalist hating Japan, but empathy is emotion once removed. It is hypothetical. Chinese nationalists are not hypothetically angry at Japan. They are angry. The People’s Daily editorialized: ‘No one would doubt the pulses of patriotic fervor when the motherland is bullied. No one would fail to understand the compatriots’ hatred and fights when the country is provoked’ (Johnson and Shanker Reference Johnson and Shanker2012; see also Gries Reference Gries2004; Löwenheim and Heimann Reference Löwenheim and Heimann2008; Hall Reference Hall2011). One’s group identity provides one’s emotional reality and that helps to distinguish individual from group-level emotion.

Group members should have broadly similar emotions

Londoners’ pride in their successful hosting of the 2012 Olympics and Para-Olympics, as well as in Team-Great Britain (GB) punching above its weight in the gold medal count, is an example of basking in reflected glory (Cialdini et al. Reference Cialdini, Borden, Thorne, Walker, Freeman and Sloan.1976) or even ‘ecstatic sociality’ (New Statesman 2012). Londoners were not feeling pride on behalf of the gold medal winners or on behalf of the people who organized the event. A British newspaper, The Independent, editorialized: ‘So the Union Jack has been rescued from the old connotations of vanished empire and has become a vibrant, colourful symbol of contemporary British identity’ Reference Zaki and Ochsner(2012). These feelings were not vicarious. Scotland’s First Minister introduced the term ‘Scolympians’ to praise Scottish achievements but not Team-GB. The term did not spread. The pro-independence movement hinges on feelings of Scottish identity and the anti-independence movement depends on dry economic arguments. The Independent’s editorial commented on a poll revealing declining support among Scots for independence and concluded: ‘The scenes of patriotic jubilation that have accompanied each Team GB win suggest something different. There is emotion too in the idea of a nation united’. The point is not that the London Summer Olympics ended the Scottish National Party’s drive for independence, but that, for a while, there was a surprising and broadly felt understanding in the United Kingdom of a new British patriotism. For many the experience was national, not individual.

The Libyan reaction to the terrorist attack on the American consulate in Benghazi and the murder of four Americans, including Ambassador Christopher Stevens, captures the feeling of collective embarrassment. In what reporters described as ‘an emotional meeting’, Libyan President Mohamed Magariaf apologized again to Secretary of State Clinton and said softly: ‘Madam Secretary, I also seize this opportunity to reaffirm that what happened on 11th of September towards these U.S. citizens does not express in any way the conscience of the Libyan people, their aspirations, their hopes or their sentiments towards the American people’ (Cooper and Myers Reference Cooper and Myers2012). Perhaps only political spin, though outside observers had a similar impression. Robert Malley of the International Crisis Group (and former director for Near East affairs at the National Security Council) reported from Tripoli that ‘from every conversation I have had with every Libyan here over the last three, four days, a sense of shock and dismay and embarrassment at what happened, at the killing of the U.S. ambassador’ (PBS Newshour 2012b). The embarrassment and dismay seem genuine and widespread (Hauslohner Reference Hauslohner2012).

Group-level emotion regulates behavior and attitudes toward both in-group and out-group members

Different identities have different implications for experiencing – or not experiencing – emotions such as guilt. Guilt can be a group-level emotion. One study (Doosje et al. Reference Doosje, Branscombe, Spears and Manstead1998) modeled on Tajfel’s minimal group paradigm found that even minimal group identification can trigger collective guilt: people who were told they were inductive thinkers and that inductive thinkers had systematically undervalued deductive thinkers experienced collective guilt (even though they themselves had not acted in a harmful way). Group-level guilt exists, though the relationship with identity is complicated. When high identifiers are presented with sufficiently compelling evidence that their country has tortured prisoners, so that neither denial nor a demand for more evidence is an available defense, they justify their group’s torture as morally appropriate. One study (Tarrant et al. Reference Tarrant, Calitri and Weston2012a, 516) concluded: ‘torture is more likely to be justified when it has implications for people’s social identity. Reminders of an in-group’s involvement in torture undermines the integrity of the group and group members respond to this threat in ways that help them to restore positive social identity’. Defending the group means identifying with the torturers, not the victims. When one does not feel like a state, or when one’s social identity is not in jeopardy, then one would probably view identifying with torturers as obscene.

Condemning torture committed by others is easy, but harder when part of one’s self is a torturer. Feeling guilty for what other individuals have done only makes sense if one feels collectively responsible for that behavior. It is the flip side of feeling pride, though one does not bask in reflected guilt, but (whenever possible) hide from it. The weaker one’s identification with a group, the more receptive one is to negative information about the group and thus the more likely one will experience guilt for a group’s behavior. The more one identifies with a group, the less receptive one is to negative information, which means one has nothing to feel guilty about (Doosje et al. Reference Doosje, Branscombe, Spears and Manstead2004). Germans who weakly identify with Germany experience greater collective guilt and Germans who strongly identify with Germany reject collective guilt using defensive strategies such as Holocaust denial (Rensmann Reference Riling and Sanfey2004). The more one feels like a state, the less likely one will feel guilt, with one exception. When the in-group is the source of negative information, then high identifiers are likely to feel greater guilt than low identifiers. Low identifiers might feel less guilt in this condition because they believe the in-group has already recognized its guilt, which lessens their own (Doosje et al. Reference Doosje, Branscombe, Spears and Manstead2004).

Though psychologists refer to ‘perspective taking’ as a way to better understand another’s feelings, interests, and beliefs, the process more closely resembles perspective giving. The perspective one takes from a group is whatever perspective one decides to give to the group. Although numerous studies find that perspective taking reduces support for an out-group when that group directly threatens an in-group, a recent study (Tarrant et al. Reference Waytz and Epley2012b, 975) found that even ‘in a relatively innocuous, everyday intergroup context, perspective taking elicits a negative reaction among some group members’. Individuals did not derogate others when they took another’s perspective; only group members did, and the stronger one’s identification with a group, the greater the number of negative traits one ‘took’ from the other group. SIT explains these findings. High identifiers need to view their group favorably. When asked to take the perspective of a rival group (either a rival college or a rival country), high identifiers derogated the out-group as a way of defending their social identity. Perspective taking would seem to work best on those who need it least.