For most of the history of prejudice research, negativity has been treated as its emotional and cognitive signature, a conception that continues to dominate work on the topic. By this definition, prejudice occurs when we dislike or derogate members of other groups. Recent research, however, has highlighted the need for a more nuanced and “inclusive” (Eagly 2004) perspective on the role of intergroup emotions and beliefs in sustaining discrimination. On the one hand, several independent lines of research have shown that unequal intergroup relations are often marked by attitudinal complexity, with positive responses such as affection and admiration mingling with negative responses such as contempt and resentment. Simple antipathy is the exception rather than the rule. On the other hand, there is mounting evidence that nurturing bonds of affection between the advantaged and the disadvantaged sometimes entrenches rather than disrupts wider patterns of discrimination. Notably, prejudice reduction interventions may have ironic effects on the political attitudes of the historically disadvantaged, decreasing their perceptions of injustice and willingness to engage in collective action to transform social inequalities.

These developments raise a number of important questions. Has the time come to challenge the assumption that negative evaluations are inevitably the cognitive and affective hallmarks of discrimination? Is the orthodox concept of prejudice in danger of side-tracking, if not obstructing, progress towards social justice in a fuller sense? What are the prospects for reconciling a prejudice reduction model of change, designed to get people to like one another more, with a collective action model of change, designed to ignite struggles to achieve intergroup equality?

Despite downsides, it must, on balance, be good to reduce prejudice. Despite upsides, collective action can also have destructive outcomes. Improving intergroup relations requires multiple levels of analysis involving a broader approach to prejudice reduction, awareness of potential conflict escalation, development of intergroup understanding, and promotion of a wider human rights perspective.

I argue that Dixon et al. fail to maintain a careful distinction between the negative evaluation definition of “prejudice” and the implications of this definition for correcting the social ills that prejudice engenders. I also argue that they adduce little evidence to suggest that if prejudice were diminished, commensurate reductions in discrimination would not follow.

Dixon et al. accurately describe subtle mechanisms of discrimination that inhibit minorities' collective action in modern democratic societies. This commentary suggests that in contemporary non-Western societies, where ethnic conflicts are more violent, traditional overt forms of prejudice still exist and predict discrimination of ethnic and racial minorities. Thus, prejudice reduction models should and do improve intergroup relations in such contexts.

Here, I argue that Dixon et al. have overstated the prevalence of “benevolent” forms of prejudice; many stigmatised groups are currently the targets of overtly hostile evaluation and treatment by others (e.g., Muslims; immigrant groups). I also believe that the target article oversimplifies its presentation of prejudice researchers' primary theoretical and policy goals and that it overlooks important work in intergroup emotions.

Our research on non-religion supports the proposed shift toward more interactive models of prejudice. Being nonreligious is easily hideable and, increasingly, of low salience, leading to experiences not easily understood via traditional or contemporary frameworks for studying prejudice and prejudice reduction. This context affords new opportunity to observe reverse forms of interactive prejudice, which can interfere with prejudice reduction.

In line with Dixon et al.'s argument, I contend that prejudice should be understood in broadly political rather than in narrowly psychological terms. First, what counts as prejudice is a political judgement. Second, studies of collective action demonstrate that it is in “political” struggles, where subordinate groups together oppose dominant groups, that prejudice can be overcome.

The analysis offered by Dixon et al. fails to acknowledge that the attitudes that drive prejudice are attitudes that are constructed in particular contexts. These attitudes (e.g., toward men as childcare workers) can diverge strongly from attitudes toward the group in general. Social change is thus best achieved through challenging the requirements of roles and by changing group stereotypes.

A key argument of Dixon et al. in the target article is that prejudice reduction through intergroup contact and collective action work in opposite ways. We argue for a complementary approach focusing on extreme emotions to understand why people turn to non-normative collective action and to understand when and under what conditions extreme emotions may influence positive effects of contact on reconciliation.

Integrating a historical perspective into studies of prejudicial attitudes facilitates the interpretation of paradoxical findings of the kind cited in the target article. History also encourages research to move beyond the study of prejudice and to consider institutional and structural forces that maintain social inequities. Multilevel approaches can study these factors in both field and laboratory studies.

The target article challenges standard approaches to prejudice reduction, warning that they may inure people to inequality and deflect them from seeking collective solutions to it. I argue that the collective action approach has its own risks and limitations and that standard contact and common identity approaches may complement rather than work against it.

Three points that are implicit in Dixon et al.'s paradigm-challenging paper serve to make prejudice potent. First, prejudice reflects understandings of social identity – the relationship of “us” to “them” – that are shared within particular groups. Second, these understandings are actively promoted by leaders who represent and advance in-group identity. Third, prejudice is identified in out-groups, not in-groups.

Dixon et al. overlook the fact that contact predicts not only favorable out-group attitudes/evaluations, but also cognitions, affect, and behavior. The weight of evidence supporting the benefits of intergroup contact cautions against throwing the (contact) baby out with the bathwater. The goal to “ignite struggles” in pursuit of social equality, we argue, incautiously risks hurling us down the proverbial rabbit hole.

Dixon et al. have highlighted the importance of a political conceptualisation of intergroup relations that challenges individualising models of social change. As important as this paper is for the development of critical debates in psychology, we can detect at least three issues that warrant further discussion: (a) the cultural and historical conditions of structural inequality and its perception, (b) the marginalisation of post-colonial works on collective mobilisation, and (c) acknowledging the complex perspectives and politics of those targeted by prejudice.

We agree that promoting intergroup harmony “carries insidious, often unacknowledged, ‘system-justifying’ consequences” (sect. 4.1.3, para. 2) and identify several ways in which “benevolent” and “complementary” stereotypes, superordinate identification, intergroup contact, and prejudice reduction techniques can undermine social change motivation by reinforcing system-justifying beliefs. This may “keep the peace,” but it also prevents individuals and groups from tackling serious social problems, including inequality and oppression.

Developmental perspectives on prejudice provide a fundamental and important key to the puzzle for determining how to address prejudice. Research with historically disadvantaged and advantaged groups in childhood and adolescence reveals the complexity of social cognitive and moral judgments about prejudice, discrimination, bias, and exclusion. Children are aware of status and hierarchies, and often reject the status quo. Intervention, to be effective, must happen early in development, before prejudice and stereotypes are deeply entrenched.

Criticism of orthodox models of prejudice reduction is particularly relevant for lesbians, gay men, and bisexuals, particularly when considering stage models of coming-out. If social change is to be effected regarding endemic homonegativity and heterosexism, then it is argued that a radical rethink is needed to the understandable but misinformed desire to get us to like each other more.

Human behavior is guided by evolutionarily shaped brain mechanisms that make statistical predictions based on limited information. Such mechanisms are important for facilitating interpersonal relationships, avoiding dangers, and seizing opportunities in social interaction. We thus suggest that it is essential for analyses of prejudice and prejudice reduction to take the predictive accuracy and adaptivity of the studied prejudices into account.

This commentary focuses on Dixon et al.'s discussion on the dangers of employing prejudice-reduction interventions that seek to promote intergroup harmony in historically unequal societies. Specifically, it illustrates these dangers by discussing my work in Israel (now mentioned in Dixon et al.'s note 6) on the processes and practices through which reconciliation-aimed encounters between Jews and Arabs mitigate sociopolitical change.

A dual-audience signaling problem framework provides a deeper understanding of the perpetuation of group-based inequality. We describe a model of underachievement among minority youth that posits a necessary trade-off between academic success and peer social support that creates a dilemma not typically encountered by nonminorities. Preliminary evidence consistent with the approach is discussed. Such strategic agent perspectives complement the psychological approach put forth by Dixon et al., but with minimal ancillary assumptions.