The Construction of Emotion in Interactions,Relationships,and Cultures

Emotion Construction in the Context of Interactions

The idea that emotions are constructed “in the moment” during ongoing interactions contrasts with models that propose that emotions are first and foremost “in the head.” On the one hand, constructivist models emphasize that emotions derive their different qualities in relation to different kinds of interactions, while more naturalistic models conceptualize emotions as innate, invariant, and universal programs (e.g., Ekman, 1984; Izard, 1992) that are prewired “in the head” before they are expressed in interactions. On the other hand, constructivist models articulate the social interaction as an important influence on meaning construction, whereas traditional cognitive approaches (e.g., Lazarus, 1991; Smith & Ellsworth, 1985) are agnostic about the sources of emotional construction. There is now a growing body of evidence for the central importance of interaction to emotion, mostly within the developmental literature.

Research starting from a dynamic systems approach (e.g., Fogel et al., 1992; see also Camras & Witherington, 2005) has shown that the development of infants’ emotions emerges over the accentuated accumulation of interactional sequences of sharing, attuning, proposing, and correcting, rather than being dictated by a set of prewired, discrete, and central emotion programs. These approaches build on the observation that young infants do not systematically show the invariant facial expressions in the appropriate emotion-eliciting contexts that discrete emotion theories predict. For example, infants frequently raise their brows (a prototypical expression of “surprise”; Ekman & Friesen, 1975) during the exploration of familiar objects (Camras, Lambrecht, & Michel, 1996). It appears that infants’ expressive behaviors at this early age are not associated with adult-like emotions; they rather signal undifferentiated positive or negative valence with different levels of intensity (e.g., Oster et al., 1992, as cited in Oatley, Keltner, & Jenkins, 2006). According to dynamic systems theories, infants develop differentiated emotions over the course of repeated interactions with the caregiver by “detecting dynamic invariants of emotion in jointly constructed social routines . . . rather than having [them] imposed from ‘outside’ on the regulation of ‘inner’ emotional programs” (Fogel et al., 1992, p. 136).

Viewing infant’s emotions as interactional constructions has consequences for how we understand infants’ communication of affective states. Infant’s signals of distress or pleasure do not communicate emotional messages in themselves, but are rendered emotionally meaningful in the context of the caregiver’s actions. Behaviors like smiling, fussing, and crying communicate to the caregiver that the interaction is going well or, conversely, that adjustments to the infant’s environment need to be made (Oatley et al., 2006). These behaviors are imbued with emotional meaning by the way the caregiver responds to them—either by maintaining or increasing stimulation in order to maintain positive states of the baby, or by reducing stimulation in order to discontinue bad feeling states. For example, mothers were found to maintain a baby’s positive state by mirroring the baby’s positive emotions and to discontinue negative states by ignoring or responding with surprise to a baby’s negative expressions (Oatley et al., 2006). These findings provide further support for the idea that emotions derive their qualities in relation to the interactions in which they occur.

Moreover, infants recognize emotions in their caregivers during interactions before they can infer them from simple images. This means that infants can process complex information as it naturally occurs in interactions before they can understand simplified referents to the emotion (Walker-Andrews, 2008). In particular, they can discriminate between anger, happiness, and sadness during real interactions with their caregivers. In contrast, young infants do not show any sign of understanding these emotions outside of particular interactive contexts. For instance, they do not recognize emotions when expressed in separate modalities (e.g., the face or the voice). Infant recognition of emotion is contingent on emotional expression in situ, that is, on the actual interactions in which infants engage early in life.

Research on emotion construction in adult interactions is fairly scarce; however, research on marital dispute provides an interesting exception. Gottman, Swanson, and Murray (1999) found, for instance, that some newlywed couples displayed emotional patterns that contained low-level reciprocal negative emotions (contempt, anger), with one spouse’s negative emotions predicting the other’s on the next interaction turn (even when controlling for the first spouse’s own emotions on previous turns). Other couples’ dynamics did not show the same reciprocity, because one of the spouses would have a higher “negativity threshold,” meaning that the expression of negative emotions of the other partner had to be much more pronounced in order to impact on the spouse’s behavior in the following turn (Gottman et al., 1999). One possible interpretation is that those who received too many messages of disapproval from their spouses stopped engaging in the interaction until the critical point of the negativity threshold was reached, at which conflict then occured at a much higher level of emotional negativity. This interpretation fits with the finding that a high negativity threshold predicted failure of the marriage 1 year later. The important point here is that emotions of the spouses mutually afford or constrain each other, and they do so even in the course of one single interaction.

There is also evidence for the construction of emotions during interactions with people with whom one is not already acquainted. In one study, undergraduates who were selected to be either high or low on social anxiety participated in a “getting to know you” interaction (Heerey & Kring, 2007). Nonanxious respondents were paired with nonanxious peers, or alternatively, with socially anxious counterparts. During the “getting to know you” stage, highly anxious participants talked more often about themselves, asked fewer questions, reciprocated genuine smiles more often with polite smiles, and sought more reassurance than the nonanxious participants in either type of dyad. Central to the current argument is that this emotional behavior also seemed to shape the emotions of the nonanxious participants who were paired with anxious conversation partners. These nonanxious participants were the only group who did not report an increase in positive affect as a result of the interaction. Moreover, they offered more empathy and support to their conversation partners than any other group in the study. Finally, both interaction partners in the socially anxious/nonanxious pair perceived lower quality of interaction, and fidgeted more than the partners of nonanxious pairs. Fidgeting, a behavioral sign of anxiety, tended to be initiated by the socially anxious partner, and appeared to be transmitted to the nonanxious partner. This study illustrates how emotions can be meaningfully described as an interactive pattern; moreover, describing emotions this way renders transparent how the emotions of one (e.g., nonanxious) partner in the interaction are constituted by the emotions of the other (e.g., anxious) person.

Illustrative of the way emotions are embedded in social interaction are also recent developments in techniques that are used to teach actors the expression of authentic emotions on stage. In contrast to the well-known “method acting” technique (Chekhov, 1953), in which actors are trained to perform emotions by retrieving emotional events from autobiographic memory, Meisner proposes that authentic emotions emerge in spontaneous interactions (Meisner & Longwell, 1987). Meisner recommends an exercise in which two actors go through cycles of rapid interactions. During these interactions, they are instructed to react immediately and spontaneously to their perception of the other actor by stating any observed changes. Over time, actors become increasingly sufficient at responding quickly to partners while remaining in the moment of the interaction. The result is an authentic portrayal of characters that are performed by actors who live “truthfully under imaginary circumstances” (Meisner & Longwell, 1987, p. 15). It appears that emotions constituted by spontaneous interactions between partners lead to more authentic performances than those emotions which are first conceived in the head and then put into reality.

In summary, there is support for the idea that emotions are constructed in the process of interactions. Infant emotions develop and are rendered meaningful in the interaction with caregivers; adults’ emotions are equally afforded by ongoing interactions with others: Partners construct their emotions in response to each other and the ongoing exchange between them, strangers quickly attune to the emotional behavior of their counterparts, and actors train to respond spontaneously on the stage in order to perform authentic emotions. Overall, these findings suggest that describing and studying emotions at the micro level of (dyadic) interactions does more than just increase ecological validity: An interactional approach offers a new, empirically grounded understanding of how emotions develop (not as innate maturating programs, but as self-organizing social processes), of how they are experienced and expressed (not as short-lived, but as ongoing interactional response systems), and of what they are (not discrete essences in the head, but social constructions).

The Mutual Construction of Emotions and Relationships

The experience and expression of emotions at one particular point in time is not only afforded and constrained by the ongoing interaction with others, but also by the history of the shared relationships as well as by future projections of where those relationships may go. Compared to research on emotional construction in interactions, research and theorizing on this topic is relatively extensive (e.g., Fischer, Manstead, & Zaalberg, 2003; Parkinson et al., 2005; Tiedens & Leach, 2004). In the present article, we argue that emotion itself derives its meaning and predictability from the relationship in which it takes place and which it shapes in return.

People strive for “common ground” when interacting with each other (Kashima, Klein, & Clark, 2007), and extended interactions in the context of committed relationships lead to the establishment of commonly shared meanings and values (e.g., Phillips, Bischoff, Abbott, & Xia, 2009). According to our argument, people who established shared meanings in relationships should also experience similar emotions. This appears to be the case: One of the most compelling examples of the ways in which emotions are grounded in relationships is the case of emotional convergence. A first demonstration of emotional convergence came from a study by Robert Zajonc and his colleagues, showing that married couples looked more alike after 25 years of marriage than they did at their first anniversary (Zajonc, Adelmann, Murphy, & Niedenthal, 1987). Moreover, students who were asked to judge the resemblance of these couples’ faces were better at matching couples who reported higher relationship closeness and marriage satisfaction. The authors postulated imitation as the underlying process: Relationship closeness might lead to empathy, and to the mimicking of emotions. Twenty-five years of mimicking one’s partner, and thus using similar facial muscles, would account for the resemblance between partners’ faces.

A recent set of studies has explicitly tested the assumption that emotions converge over time in relationships and are thus attuned to their relational engagements. Anderson, Keltner, and John (2003) studied the emotions of roommates and romantic couples; emotions were measured both by self-reports and facial coding, and were elicited in a number of different ways (conversations about topics with different valence, watching emotion-eliciting movies). In two longitudinal studies, convergence of emotions was higher after 6 or 9 months than at the beginning. Emotional convergence was related to how satisfied partners/roommates were with the relationship and how much they liked their partner/roommate, even when controlling for personality similarities. A third, experimental study showed that after 7 months of sharing a room, the emotions of roommates who had not known each other previous to sharing a room were more similar to each other than to randomly chosen students, even when roommates separately watched emotion-eliciting movies in order to prevent emotion contagion. Thus, living together may lead to similarity in emotional reactivity, and, whatever the mechanism, it develops in a relatively short time. The occurrence of emotional convergence makes a strong case for conceiving of emotion as a process that exceeds the individual mind and is shaped inherently by relational engagements.

Understanding emotions as grounded in relationships implies that emotions are not only shaped by shared relational contexts, but also shape the relationship in which they occur. Two recent studies about the role of gratitude for relational development by Algoe and her colleagues (Algoe, Gable, & Maisel, 2010; Algoe, Haidt, & Gable, 2008) demonstrated the importance of emotion for the development of both new and established relationships. In the first study, Algoe et al. (2008) followed the development of the relationships between little and big sisters of a sorority: Little sisters had just joined the sorority, while big sisters had been a member of the sorority since the previous year. During little-sister week, the big sisters anonymously surprised their little sisters, and the little sisters’ gratitude was measured. The extent to which little sisters experienced gratitude was associated with the little sisters’ appreciation of the relationship with their big sisters. One month later, the little sisters’ gratitude predicted the quality of the relationship and the amount of time spent together as reported by the big sisters. In the second study (Algoe et al., 2010), both partners of 67 heterosexual couples kept a daily diary on their own and their partners’ thoughtful actions, their emotional reactions (e.g., gratitude) towards their partners’ actions, and their relationship satisfaction. Thoughtful actions of the partner predicted the experience of gratitude, which in turn lead to higher relationship satisfaction for both partners the following day. In both studies, emotional experience shaped the course of the relationships, either in terms of making the establishment of new relationships more likely or in terms of advancing the (short-term) development of established relationships.

Finally, a longitudinal study of marital interaction and satisfaction has equally demonstrated the importance of emotional interaction for long-term relationship development. Gottman and Levenson (2000) invited 73 married heterosexual couples into the lab and asked them to discuss both nonconflictual daily experiences and a continuing disagreement in their marriage. The couples were then repeatedly contacted over a period of 14 years to assess their marital status and satisfaction. After 14 years, 21 of the couples had divorced. Several observations about these couples are worth noting: First, emotions during interaction could predict divorce with 93% accuracy. Second, couples who displayed more negative emotions such as contempt, anger, or stonewalling during conflict interactions were more likely to get divorced quickly. This pattern appeared to create an emotional climate that lead to a fast deterioration of the relationship. Third, those couples who displayed more positive affect and less anger during the discussion of everyday matters were less destructive in their behavior during the discussion of disagreements in their marriage. The quality of a couple’s relationship is thus clearly related to the actual emotions of the partners during various interactions.

In summary, emotions are constructed in the context of relationships with others. A number of studies have shown the mutually constitutive character of emotions and relationships: On the one hand, people who share a relationship with each other—from committed relationships to roommates in college—become emotionally more similar over time. On the other hand, both positive emotions (e.g., gratitude) and negative emotions (e.g., contempt) are closely linked with the development of relationships, such that the former help form and maintain relationships whereas the latter lead to a fast parting of the ways. It is hard to imagine a relationship that is not defined by and constitutive of the emotions that are experienced and expressed within it.

The Cultural Construction of Emotion

Interactions and relationships are always framed by the prevalent ideas, meanings, and practices of how to be a person and how to relate to others that are also referred to as cultural models (e.g., Holland & Quinn, 1987; Markus & Kitayama, 2003). Cultural models inform the person’s central values, goals, and concerns, and thus constitute the background against which appraisals are formed. For example, when the dominant cultural models emphasize the importance of meeting social norms, failure to do so will most certainly elicit shame; shame in those cultural contexts may then be a common emotional experience. Cultural models include practices, in addition to meanings; these practices may increase or decrease the likelihood of appraisals and emotion as well. For instance, practices that focus attention on instances in which the individual fails to conform to norms or expectations will elicit shame. Examples of such practices are the clear structuring of interactions by politeness norms and classroom routines that allow for the collective monitoring of norm-inconsistent behavior (e.g., communal self-criticism or hansei in Japanese classrooms; Lewis, 1995). Finally, cultural models constitute the context in which emotional behavior takes place, and, as such, also guide and inform emotional responses. Whether fighting or yelling are effective anger responses depends on the meaning these behaviors have, and their putative or real social consequences in the relevant cultural context.

Differences in emotions between European American and Japanese cultural contexts may serve as an example for the cultural construction of emotion. Patterns of emotions in European American contexts can be understood from the prevalent cultural model in these contexts, which emphasizes an independent self that is free from others, autonomous, unique, and focused on the maintenance of high self-esteem and the pursuit of its own goals. On the other hand, the emotional patterns prevalent in Japanese contexts can be understood from the dominant cultural model in those contexts, which emphasizes an interdependent self that is embedded in relationships, connected with others and focused on maintaining harmony by adjusting to environmental demands (Boiger, Mesquita, Tsai, & Markus, 2012; Heine, Lehman, Markus, & Kitayama, 1999; Markus & Kitayama, 1991, 1994; Morling, Kitayama, & Miyamoto, 2002; Triandis, 1995).

First, several studies have suggested that the perception of emotion in others is constructed according to the culturally different models of the self in relation to others: While emotions in the Western groups are judged to be a feature of the individual, Japanese participants construct an individual’s emotion from the relatedness between different people’s emotions. Thus, in one perception study (Masuda, Ellsworth, Mesquita, Leu, & Veerdonk, 2008), respondents were asked to rate the feelings of a target person (Jon) who was surrounded by several others, whose emotional expressions varied independently of the target person. When Westerners judged Jon’s emotions, they only used the information of his facial expressions. On the other hand, Japanese considered the expressions of the surrounding faces when rating Jon’s emotional experience: Jon was not judged to be as happy when the others were angry or sad compared to when the others were also showing smiling faces. Similarly, in another set of studies, Uchida, Townsend, Markus, and Bergsieker (2009, Studies 3, 4a, and 4b) investigated American and Japanese participants’ perception of emotions in successful Olympic athletes. Americans were more likely to see emotions when athletes were portrayed alone rather than in the context of others. Japanese, on the other hand, inferred more emotions when athletes were portrayed in relational contexts than when portrayed alone. Across studies, North American groups thus judged emotions to be a feature of an independent individual, whereas Japanese constructed emotion by assuming relatedness between people’s emotions.

Furthermore, the most prevalent emotions in a given cultural context appear to be the ones that fit the culturally preferred relationship arrangements (Mesquita & Leu, 2007). Kitayama, Mesquita, and Karasawa (2006) found that socially engaging emotions, such as friendly feelings or shame, were experienced more intensely in the Japanese interdependent than in the American independent cultural context; the opposite was true for socially disengaging emotions, such as pride and anger. Engaging emotions underline and reinforce the relatedness between people and thus fit the goals of social harmony relevant for interdependent selves. This is obviously the case with friendly feeling, but also the emotion of shame signals attentiveness to social rules and the intention to conform to those rules. Socially disengaging emotions, on the other hand, tend to signal and affirm the independence of an individual, and thus fit the goals of personal achievement and autonomy that are relevant in independent contexts. Data from an experience sampling and a vignette study of 22 very diverse emotional events converged: In both studies, engaging emotions were experienced or perceived more intensely than disengaging emotions in Japanese contexts, whereas the reverse was found in European American contexts (Kitayama et al., 2006).

One reason for the different construction of emotional lives across cultures may be the different emotional situations that people encounter. We showed in a recent set of studies (Boiger, Mesquita, Uchida, & Barrett, 2012) that the frequency of interpersonal situations of anger and shame varies between the U.S. and Japan: Situations that were more likely to elicit culturally condoned emotional experiences (i.e., anger in the U.S., shame in Japan) were also judged to happen more frequently. While anger-eliciting situations appeared to be “up-regulated” in the U.S. cultural context, the same situations appeared to be “down- regulated” in the Japanese context. The opposite picture emerged for shame: Situations that elicited stronger feelings of shame were found to be more frequent in Japan and less frequent in the USA. Established differences in the prevalent emotions that people experience across cultures can thus, at least in part, be explained by the different social situations that people encounter across cultures.

Finally, there is evidence that the experience of emotions themselves is constructed differently in different cultures. This means that, across cultures, similar kinds of interactions elicited different patterns of emotional responses. In one interview study (Mesquita, Karasawa, Banjeri, Haire, & Kashiwagi, 2010), Japanese and American respondents reported in detail on situations in which they felt offended, which was associated with the emotion of anger in both groups. However, participants’ accounts of their emotional experience differed substantially between the two cultural groups: In North American contexts, offense situations were framed as threats to the individual’s autonomy and self-worth and as rejection by the other person, and were to be solved by reaffirming the self and by getting back at the other person. In Japanese contexts, on the other hand, offense episodes tended to be interpreted as threats to the relationship that call for a better understanding of the other person’s motives and for an interpretation of the situation from a generalized other’s perspective. The appropriate actions in these situations were to keep one’s calm by doing nothing or by moving away. These differences in appraisals and actions can be understood from differences in the prevalent cultural models: Maintenance of high self-regard, assertiveness, and aggression fit well with an American independent model that foregrounds individual interests and promotes the maintenance of self-esteem. On the other hand, focusing on the perspective of the offender, staying calm, and doing nothing is consistent with the Japanese cultural model that emphasizes self-scrutiny and the maintenance of relationship harmony.

In summary, emotions are differently constructed across cultures according to different cultural models of self and relating. A number of studies have shown that the most prevalent emotions and the most prevalent ways of perceiving, experiencing, and expressing emotions can be understood from the different ways of interacting and relating to which people are committed. It thus appears that, as people embody different cultural models, their emotions come to differ according to the different interactions and relationships that they engage in.

Interdependencies between Contexts of Construction

Only few studies have explicitly investigated interdependencies between interactions, relationships, and cultural meaning systems as contexts of construction. The construction of emotions in interactions and relationships has mainly been studied against the backdrop of a North American cultural model. However insightful these studies are about the role of interactions and relationships for the construction of emotions, they do not allow us to draw conclusions about the way that cultural models come into play. Similarly, research on cultural differences in emotion has yielded insights into the different ways in which emotional lives are shaped culturally, but it has rarely examined how these differences are constructed during real-life interactions within ongoing relationships.

A rare example of the interplay between contexts of interactions, relationships, and cultural systems in the construction of emotion is a cross-cultural study on mother–child interactions in Germany and Japan. Trommsdorff and Kornadt (2003) found that, when responding to disobedient children, German mothers were quick to infer that the child was acting on purpose (“he wants to make me angry”, p. 296), therefore reacting with anger. An escalating interaction ensued: The mother would insist on compliance, to which the child reacted with protest. This, in turn, affirmed the mother’s initial appraisal that the child’s misconduct was intentional and led her to assert herself even further or threaten the child with punishment. The child felt rejected and threw a tantrum, crying and protesting aggressively. The mother reacted with despair, lost control, and punished the child. The conflict ended unresolved, leaving both the mother and the child angry, hurt, and expecting that the conflict would repeat itself in the future. A very different interaction pattern was observed in Japanese mother–child dyads: Japanese mothers were more likely to interpret disobedience empathetically from the child’s perspective (“the child is just a child, is too much absorbed in playing, is too tired”; Trommsorff & Kornadt, 2003, p. 296), and to react in an accommodating manner. Following misconduct, Japanese mothers repeated their demands in a friendly manner, feeling pity for the child. The children reacted with irritation to the continued demand, which led to disappointment and regret about the situation on the mother’s side. Concerned about jeopardizing relational harmony and the desired feeling of one-ness (ittaikan), both parties started making partial concessions: The mothers comforted and distracted the children, partially giving in to their initial demands. The children in return complied, at least partially. Finally, the conflict resulted in a compromise, harmony was reinforced, and both parties expected future interactions to turn out positively. These interactions had long-term effects on children’s relational patterns: Escalation of interactions in early childhood predicted the level of empathy-based altruism and aggression 9 years later (Kornadt & Tachibana, 1999, as cited in Trommsdorff & Kornadt, 2003). These interaction patterns can be understood as necessary socialization experiences for an independent German self that learns to assert itself and to have its needs met, in comparison to an interdependent Japanese self that learns to accommodate and to place relational harmony over individual desires. This example vividly shows the intricate interplay between contexts of emotion construction: Mother–child emotional interactions were constructed based on prevalent cultural meanings, were shaping relational contexts in situ as well as over time and, therefore, maintained culturally prevalent ways of emoting in the long run.

The cultural shaping of people’s emotional lives is, however, not limited to early age or to interactions with caregivers, as was illustrated in a recent study on the acculturation of immigrants’ emotions. De Leersnyder, Mesquita, and Kim (2011) compared the emotional patterns of Turkish immigrants in Belgium and those of Korean immigrants in the USA with the average emotional patterns of their respective cultures of settlement. Participants reported emotional events from their own lives that varied along the dimensions of valence and social engagement (Kitayama et al., 2006). They subsequently reported how they had felt by rating their feelings on a scale of 20 different emotions. The degree to which immigrants’ emotional patterns were similar to the culture of settlement was calculated for each Turkish and Korean participant by comparing their individual profiles with the average profile of their respective host culture. The results confirmed the idea that the interactions with others shape an individual’s emotions: The emotional pattern of immigrants became more similar to those of the culture of settlement with increasing time spent in the new culture. Moreover, engaging in social interactions with members from the new culture played a major role: Social contact with mainstream others was an important predictor for emotional similarity between immigrants and their culture of settlement. Although interactional processes were not explicitly investigated, it might well be the case that as immigrants engage in relationships with members from their new culture, they renegotiate their emotional meanings and in consequence their emotional patterns.

In sum, a few studies portrayed vividly the interdependencies between interactions, relationships, and cultural models for the construction of emotion. Throughout the course of the lifespan, people’s emotional experiences are shaped by the interactions and relationships with others. These interactions and relationships, and therefore the ensuing emotional experiences, are structured by and constitutive of the cultural contexts in which they occur.

Conclusion and Future Directions

There is evidence for the social construction of emotion in three embedded contexts: in-the-moment interactions, relationships, and cultural contexts. Studies of children–caregiver and adult dyadic interactions provide support for the idea that emotions are constructed in the process of interaction. Moreover, these interactions are embedded in the context of relationships, which shape and are shaped by the emotions of the people committed to these relational engagements. This process of construction unfolds within a rich environment of cultural meanings and practices (i.e., cultural models) that render certain emotional themes, meanings, or actions salient for the emoting self. Although many of the reported studies have investigated emotional construction at only one level of analysis, the combined findings make a strong claim for a social constructionist perspective of emotions; moreover, the few studies that have studied emotional processes across contexts support the idea that emotions are constructed at the aggregate level of the combined contexts.

The idea that emotions are socially constructed responses to environmental changes is of course not new. Social constructionist accounts of emotion have been issued repeatedly over recent decades, often as part of an ongoing and heated debate between naturalistic and constructionist perspectives on emotion (e.g., Averill, 1980; Harré, 1986; Hochschild, 1979; Lutz, 1988; Ratner, 1989). However, while social constructivist views argued against the naturalistic assumptions of other emotion theories, they continued to treat emotions as unitary entities that, once formed, remained stable. In contrast, the approach outlined here conceives of the social construction of emotions as an ongoing process. We suggest that it is this responsiveness of emotional processes to dynamically changing and complex social environments that enables people to successfully navigate the variety of social contexts in which their lives take place.

What might a future agenda of emotion research that starts from a perspective of sociocultural construction then look like? We suggest that research on the social construction of emotion would benefit from moving beyond differential outcomes of construction, as reported or observed in different contexts, and should conceive of emotional construction as an ongoing, interactive process that unfolds within relational and cultural contexts. This would, for example, entail studies on emotions in real-time interactions, investigating how emotions afford each other in complex social contexts, while considering and comparing the meanings that may be relevant across cultural contexts. One of the reasons why this kind of research has been scarce to date may be the methodological and analytical difficulties associated with studying interactive processes. However, in recent years a number of tools have been developed that make it possible to handle the complexities in interaction processes and contextualized emotions by achieving a level of parsimony that renders findings scientifically useful. For instance, the Rapid Couples Interaction Scoring System (RCISS; Gottman, 1996) or the Specific Affect Coding System (SPAFF; Gottman, 1996) have been successfully used as tools to reduce complexities in interactions and to identify emotional patterns during interactions (e.g., Gottman & Levenson, 2000). Automated text analysis programs, such as the Linguistic Inquiry and Word Count (LIWC; Pennebaker, Booth, & Francis, 2007), allow for a fast analysis of people’s use of emotion language in written or transcribed text and have shown the power of synchronization of interactions for relational development (e.g., Ireland et al., 2011). Moreover, powerful statistical tools are now available that make it possible to model sequential processes across contexts, by specifying, for instance, the types of people (across cultures) that experience different types of emotion sequences in different kinds of situations (classification model for the study of sequential processes and individual differences therein [CLASSI]; Ceulemans & van Mechelen, 2008). In summary, while researching emotion processes within a range of contexts remains a challenge, recent advances make this kind of research more feasible.

As psychological research moves from establishing that emotions are socially constructed to unveiling how emotions are constructed, it may borrow some inspiration from other disciplines, such as sociology, anthropology, and linguistics (see, e.g., Lutz & White, 1986; Turner, 2009; Turner & Stets, 2005; Wierzbicka, 1999), that have suggested several mediating processes. Symbolic interactionist theories, for example, emphasize the importance of congruency between self-perception, cultural norms, and the responses of others for the construction of emotion. Positive emotions are assumed to be experienced when others respond in ways that are congruent with one’s self-perception; negative emotions follow when responses and self-view are incongruent (Turner, 2009). Language also plays a central role in the construction of different emotions across cultures (Wierzbicka, 1999). By using the culturally available emotion concepts, people may highlight and encourage certain experiences and expressions rather than others; in addition, the use of emotion concepts may guide regulation by invoking different evaluations, norms, and implications that are culturally associated with the concept.

Finally, a caveat is in order: Starting from the perspective that emotions occur in and are shaped by social and cultural contexts does not preclude similarities in emotional experience across contexts. Taking a perspective of extreme relativism is not only contradictory to the finding that cross-cultural similarities in emotion do exist (Mesquita, 2001; Mesquita & Frijda, 1992), but is also of little heuristic value. A perspective of emotion as being a socioculturally constructed, dynamic, and interactive process, however, safeguards against oversimplifications of what emotions are and how they are experienced, perceived, and communicated.