Emotion processes in social and political contexts: the case of affective polarisation

ABSTRACT

Democratic societies have to face the challenge of increasing affective polarisation – a strong, emotionally fuelled, opposition between adherents of different political parties, or conflicting views and attitudes in different social groups. Social and psychological sciences can contribute to a better understanding of this phenomenon and the special role of emotion theory and research in this respect. In their invited contribution, Bakker and Lelkes note the lack of conceptual precision of the term affective and invite emotion scientists to fill this notion with a theoretical structure that encourages systematic research across different problem areas and cultures. Two invited comments on this article support this initiative but favour an approach that privileges discrete emotions. In his comment, Von Scheve argues that intentionality and the associated action tendencies are essential to understand the emotional basis of polarisation. Halperin et al., based on a conceptual analysis and their own empirical research, they highlight the multi-faceted emotion of disappointment, covering both negative and positive aspects, as the most promising candidate. Finally, I consider the potential role of mixed or blended emotions as well as the potential contributions of appraisal theory, especially with respect to appraisal bias and ensuing emotion dispositions and action tendencies.

KEYWORDS:

• Affective polarisation
• appraisal bias
• emotion disposition
• blended emotions

Several decades ago, Lord et al. (1979) published an empirical study on “biased assimilation and attitude polarization” based on the following premises:

People who hold strong opinions on complex social issues are likely to examine relevant empirical evidence in a biased manner. They are apt to accept “confirming” evidence at face value while subjecting “disconfirming” evidence to critical evaluation, and as a result to draw undue support for their initial positions from mixed or random empirical findings. Thus, the result of exposing contending factions in a social dispute to an identical body of relevant empirical evidence may be not a narrowing of disagreement but rather an increase in polarization. (p. 2098)

Lord and associates tested these assumptions and predictions in two groups of participants with opposite opinions on the issues of capital punishment, finding clear support for their prediction.

Up to now, the Lord et al. (1979) article has been cited over 6000 times. The authors probably did not envisage that this study would turn out to be a precursor of a larger phenomenon, affective polarisation, threatening the fundamental structure of democracy that is rapidly becoming a central issue in the political, social, and psychological sciences. Iyengar et al. (2019) illustrated this phenomenon as follows “Americans increasingly dislike and distrust those from the other party. Democrats and Republicans both say that the other party’s members are hypocritical, selfish, and closed-minded, and they are unwilling to socialize across party lines” (p. 129) and provide an extensive overview of the relevant literature at that point in time. Since then, the literature on this issue has grown exponentially (see Iyengar et al., 2019; Webster & Albertson, 2022) and is no longer limited to the USA or to adherence to political parties but is recognised as an important phenomenon in many parts of the world, concerning many different issues in addition to political partisanship.

In an earlier theory section in this journal, the relevance of the affective polarisation topic was already briefly mentioned (Scherer, 2022). Given its great importance and the rapidly growing research on this phenomenon, it seems timely to devote a full theory section to this topic to examine the theoretical underpinnings of what has been mostly applied research on political issues. The invited article in this section has been contributed by Bert Bakker and Yptach Lelkes, two political psychologists who have been active in this domain for many years.

Putting the affect into affective polarisation

In their contribution “Putting the affect into affective polarization” (2024), Bakker and Lelkes point to the lack of a solid theoretical foundation in the current literature, requiring a cogent conceptualisation and definition of the phenomenon and, in particular, the integration of relevant work from emotion science. The authors then review the different definitions of polarisation found in this area, with a particular emphasis on social identity theory (Tajfel & Turner, 1986) and intergroup emotion theory (Mackie et al., 2008) which have strongly influenced much of this work, especially in the context of a political two-party system.

Bakker and Lelkes then go on to review the role of emotions in the affective polarisation literature, both in terms of positive emotions for the political in-group, such as love, enthusiasm, or pride, and negative emotions for the political out-group, such as anger, hate, or worry. Much of this work has focused on the role of emotion processing and media communication in the elicitation and changing intensity of these emotions. Importantly, Bakker and Lelkes point out that these emotional effects are not limited to individual reactions but may have repercussions on system-level dynamics (as discussed in this journal’s theory section on interpersonal coordination, Butler, 2022).

Turning to definitional and methodological issues, Bakker and Lelkes (2024) call for moving beyond the feeling thermometer, a visual analog scale (0–100) to measure how positively individuals feel about a given group, individual, issue, or organisation, which is frequently used in election surveys. They suggest to adopt a dimensional approach using both valence and arousal (and possibly other dimensions, see Fontaine et al., 2022; Schiller et al., 2024) to allow a more fine-grained assessment of the affective state, including the use of physiological measures in addition to self-report.

Bakker and Lelkes (2024) conclude with a plea to emotion researchers to contribute to the important research on affective polarisation with their specific expertise and mention a number of specific issues such as the interplay of conscious and non-conscious instantiations of polarised affective states, the role of context, and the use of interventions in regulation efforts.

The role of discrete emotions in affective polarisation

In his invited commentary, Christian von Scheve (2024) endorses the general theoretical approach proposed by Bakker and Lelkes with respect to social identity theory and group emotion theory but suggests some important extensions. In particular, von Scheve argues for a clear definitional distinction between affect as a somewhat diffuse mental experience that lacks intentionality, and emotions, which are in large part characterised by specific action tendencies. Although affect can be modelled with a dimensional approach mapping the experience into a multidimensional space, discrete emotions are distinguished in part by specific action tendencies. Von Scheve argues that as the literature on affect polarisation clearly involves intentionality, a discrete emotion model is required as a theoretical basis and guide for empirical work in this domain.

In addition, von Scheve (2024) suggests that group-based emotions, the central phenomenon in affective polarisation, should be clearly designated as political emotions in the sense of Szanto and Slaby (2020). The latter authors argue that the political sphere is essentially an emotional sphere in which major concerns of citizens such as freedom, power, autonomy, or recognition – in sum, individual well-being – are negotiated between social groups. Importantly, these emotions are not just experienced individually but jointly felt because of shared evaluations of the central political concerns. In addition (on p. xx), “Different from basic and non-conceptual affective stances, discrete emotions carry the weight of culturally derived and linguistically circumscribed prototypical and intentional relations to matters of concern and import” (von Scheve & Slaby, 2019). Von Scheve argues that given the importance of intentionality and cultural connotations, a dimensional model could not do justice to the complexity of differentiation required to understand affective polarisation. Thus, while anger and fear are similar with respect to negative valence and high arousal, the respective action tendencies and their social import are very different. This becomes particularly clear when considering their role in the processes of social cohesion or exclusion.

In a second invited commentary, Eran Halperin and his collaborators (2024) also plead for a discrete emotions rather than a dimensional approach to affective polarisation, emphasising the important differences, despite similar positions in a two-dimensional space, between hatred vs. contempt, anger vs. fear, guilt vs. shame or empathy vs. hope. Specifically, the authors focus on the difference between hatred and disappointment driving affective polarisation. While hatred is often evoked as the main driver of polarisation, the authors argue that, until now, there is not sufficient evidence for this claim and, more importantly, that hatred is too narrowly focused on negative valence with a very restricted range of action tendencies. For this reason, hatred does not correspond to the many facets of emotions likely to be involved in affective polarisation, particularly the positive aspects highlighted by Bakker and Lelkes (2024).

Instead, Halperin et al. (2024) propose to study disappointment as a mediator of affective polarisation, as a more appropriate direction to examine the ambivalent nature of ideological intergroup relationships. The assumption is that attitudes towards the out-group are not necessarily entirely negative. For example, people may believe that there are shared goals and values such as the quest for the common good and the respect for democratic values. Disappointment is experienced when these expectations are not met, which can then lead to a variety of other negative emotions and action tendencies such as frustration or anger. Essentially, the argument is that disappointment should be the focus of affective polarisation research and potential interventions as it is at the root of different emotional responses and may even help to conceive effective interventions by fuelling optimistic hopes of eventually finding a common ground to achieve positive outcomes. The authors urge researchers to work towards such interventions to bridge the ideological divide and enhance social and democratic stability.

Mixed emotions in affective polarisation

The three contributions in this theory section provide, in addition to a rather comprehensive review of the relevant literature, a large gamut of options for the further development for advancing with the conceptualisation and clarification of the theoretical structure and provide valuable suggestions for future research on interventions for affective polarisation. This provides an ideal basis for further enlarging the scope of theoretical development and empirical studies in this domain that can serve to bridge the affective gap.

One of the central issues raised by Halperin et al. (2024) is the choice of the specific emotion involved in the phenomenon. They argue for the major role of disappointment, which can, in turn, give rise to further emotions like anger and sadness or resignation. It is important to also consider an important role of mixed emotions or emotion blends, produced by a rapid succession of different emotions in certain situations or to complex appraisal patterns of events that produce a blend (Scherer, 2009a; Watson & Stanton, 2017).

Empirical evidence shows that such blends occur much more frequently than one might expect based on the bulk of the literature that usually concentrates on single emotions. For example, Smith and Ellsworth (1987) showed that a majority of students reported experiencing complex blends of two or more emotions after receiving their examination grades. Scherer and Ceschi (2000) found that airline passengers, whose baggage was lost upon arrival, generally reported mixed emotions, containing various elements of anger, resignation, worry, statistically accounted for by differential appraisals of personal importance and general attitudes towards airlines. About half of the participants show an anger constituent and there is a tendency of overall intensity of the emotion blends to increase with the rising intensity of the anger constituent.

The important role of the anger family (rage, irritation, disgust) in mixed emotions was confirmed in a large-scale study with over 6000 participants in English-, German- and French-speaking countries who used the Geneva Emotion Analyst App to report their emotional experiences (Scherer & Meuleman, 2013). Participants reporting irritation, rage, or disgust also reported higher proportions of mixed emotions with anxiety, fear, sadness, and despair. This may be of interest in the light of Halperin et al.’s (2024) discussion of the consequences of disappointment. In consequence, consideration of mixed emotions and emotion blends in the affective polarisation domain might offer some new perspectives on understanding the nature of the affect.

Appraisal processes in affective polarisation

The three contributions in this theory section provide, in addition to a rather comprehensive review of the relevant literature, a large gamut of options for the further development for advancing with the conceptualisation, the clarification of the theoretical structure and valuable suggestions for future research on interventions for affective polarisation. This provides an ideal basis for further enlarging the scope of theoretical development and empirical studies in this domain that can serve to bridge the affective gap.

As mentioned above, many results on emotion blends in studies of real emotion experiences could be explained by complex patterns of underlying appraisals. Halperin et al. (2024) mention the appraisal theory-based work on disappointment by Van Dijk (1999) but so far, appraisal theories of emotion have not been seriously considered in research on affective polarisation. Because the appraisal of situations and events is commonly considered as the determinant of the nature of the ensuing emotion, a serious consideration of this factor may provide an important enrichment of work on this social phenomenon. In fact, current appraisal theories provide concrete hypotheses that are directly applicable to the issue of the emotional quality of in-group/out-group relationships, as the published predictions of individual appraisals can be easily extended to prevailing appraisal tendencies in groups of individuals.

Table 1 shows the predictions made in a similar fashion by different appraisal theorists. Applying this to the affective polarisation paradigm, one would predict, for example, that in-group members would experience rage, anger, or irritation if they appraise the assumed action tendencies of out-group members on a particular political decision as new and unusual, as well as illegitimate or immoral, that would most likely have a negative impact on the well-being of the in-group, requiring urgent action. Furthermore, appraisal theorists would assume that the out-group members are entirely responsible for these deleterious effects. However, they expect that they, as a group, still have some control and even power for intervention or that they can at least adjust to the new situation. Emotion blends are likely to occur if some appraisal dimensions do not correspond to this theoretical profile. Of particular importance for the issue of affective polarisation is the appraisal criterion of compatibility with social values or standards, legitimacy, or moral convictions, as illustrated by a number of recent publications in this domain (Garrett & Bankert, 2020; Simonsen & Bonikowski, 2022; Tappin & McKay, 2019). This criterion is also of particular importance for cross-cultural comparisons.

Table 1. Examples of theoretically postulated appraisal profiles for different emotions.

Appraisal criteria	Joy/happiness	Anger/rage	Fear/panic	Sadness
Novelty	High	High	High	Low
Intrinsic pleasantness	High	Open	Low	Open
Goal significance
Outcome probability	High	Very high	High	Very high
Conduciveness/consistency	Conducive	Obstructive	Obstructive	Obstructive
Urgency	Low	High	Very high	Low
Coping potential
Agency/responsibility	Self/other	Other	Other/nature	Open
Control	High	High	Open	Very low
Power	High	High	Very low	Very low
Adjustment	High	High	Low	Medium
Compatibility with standards/value relevance/legitimacy	High	Low	Open	Open

Note: Adapted from Table 29.2 in Ellsworth and Scherer (2003).

The appraisal of the different criteria as part of the emotion process is generally quite realistic, ensuring that the emotion-generated action tendency is functional and adaptive. However, this is not always the case. Unrealistic or biased appraisal carries the risk of maladaptive reactions and behaviour tendencies. For example, there is strong empirical evidence that an appraisal bias of systematically underestimating one’s coping potential in challenging situations can lead to persistent emotion dispositions such as habitual anxiety or sadness and thus constitute a major risk factor for depression (Scherer, 2021; Scherer et al., 2022).

It seems promising to examine the role of appraisal bias and emotion dispositions in affective polarisation. Highly polarised individuals may possibly have an appraisal bias with respect to the personalities, attitudes and intentions of the out-group members as well as to the negative effects of any actions on their part for the members of the in-group. This may lead to an emotion disposition of anger or anxiety (or a blend of both). Similarly, they may show an appraisal bias of extreme standards of morality, possibly based on religious beliefs, producing emotion tendencies of contempt, disgust, or rage. Bakker and Lelkes (2024) proposal to consider arousal is very pertinent in this respect. It is likely that a high level of arousal in the case of emotions of negative valence can produce a stronger tendency towards aggressive action tendencies.

Empirically measuring the appraisal profiles and appraisal biases prevalent in an in-group would allow to predict the affective nature of the polarisation with greater precision, as well as to render verbally labelled emotional indicators more comprehensible. The downside is, of course, that the measurement instruments have to be more extensive and sophisticated in comparison to the simple valence-based “feeling thermometer”. In addition, it would be useful to develop a standard instrument to be used by different research groups in different countries. The advantage, however, might be an in-depth understanding of the underlying affective structure of the polarisation, allowing the prediction of action tendencies. Such detailed analyses might provide a very useful guide for the development of appropriate interventions. In developing such interventions, it may be worthwhile to consider the results of the work by Lord et al. (1979) on attitude polarisation, mentioned at the outset – merely presenting empirical evidence to combat appraisal bias may even strengthen the polarisation.

In addition to research investigating discrete emotions as a basis of affective polarisation, more general approaches regarding processes of implicit emotion dynamics may also provide an interesting perspective on the phenomenon of affective polarisation. “Polarization” describes a reciprocal process of mutual reference between two parties, in which responses of one party trigger opposing responses of escalating extremity in the other party. A possible basis to theoretically understand and explain these escalating tendencies is spelt out in the idea of affective-motivational counter-regulation (Rothermund, 2011): It has repeatedly been shown that experiencing or anticipating an emotion or motivation of positive or negative valence induces an attentional bias towards the opposite valence, in particular when the original event is salient and strong (Rothermund et al., 2008, 2011; Schwager & Rothermund, 2013a, 2013b, 2014). These processes may typically serve an adaptive purpose in that they guarantee balance and prevent impulsive escalation.

This functionality may only hold, however, as long as the opposing tendencies and perspectives become integrated in a joint representation. If social groups no longer adhere to the conception of a superordinate common in-group identity or define the superordinate category in a way that excludes the out-group (Mummendey & Wenzel, 1999), counter-regulation may actually foster polarisation rather than balance. To prevent affective polarisation and to support balance and integration in social contexts, it might thus be of paramount importance to highlight attributes that foster identification with a superordinate common in-group identity, which includes both groups, and allows previous opponents to see themselves as part of a larger social entity that also comprises the other party.

Interdisciplinary research on affective polarisation

Before concluding, one conceptual/definitional issue needs to be mentioned. The contributions in this special theory issue, have consistently addressed the role of emotions in affective polarisation. However, emotions, sensu strictu, are defined as bounded episodes of component processes of limited duration, triggered by events or situation changes (Scherer, 2005, 2009b). In contrast, the emotional element in affective polarisation, while possibly generated by specific emotions, is more durable and closer to what is sometimes called an emotional attitude. Ever since Dewey (1894), this term has been frequently used in the literature although there is no consensual definition. Most social psychologists adopt some version of the classic tripartite theory of attitudes (Rosenberg & Hovland, 1960), the ABC model – Affect (emotion, feeling), Behaviour (action tendencies), and Cognition (beliefs). This model also fits in the context of affective polarisation – in-group members have certain beliefs of the attitudes and intentions of the out-group members, to which they often react in an emotional fashion, possibly developing action tendencies of withdrawal or aggression. There is a huge literature of research on attitudes and attitude change, often in a social or group context (see Albarracin & Johnson, 2018), which is directly relevant to affective polarisation but not often referred to in this literature. In consequence, it would be desirable for the affective gap in affective polarisation to be filled by a closer collaboration between different disciplines such social and political sciences, social psychology, and emotion science.

Interdisciplinary theorising and research on affective polarisation is all the more urgent and important given the increasing role of digital social media in the process of social organisation and political behaviour (Harel et al., 2020; Törnberg, 2022; Yarchi et al., 2021). Recently, Fischer and Lelkes (2023) have addressed this issue concluding that “ … anger can lead people to process and consume information in biased ways. But it can also lead people to share emotionally laden content, which extends the polarising effects of such media to others. In combination, these two mechanisms create a polarising feedback loop that reinforces people’s prior beliefs, attitudes, and opinions, making it more difficult for partisans to compromise with or, in some cases, to even accept the decency of their political opponents". (p. 416).