Comment: Moving (Further) Beyond Private Experience: On the Radicalization of the Social Approach to Emotions and the Emancipation of Verbal Emotional Expressions

The Emergence of the Social Approach to Emotions

About a century after the publication of Darwin’s landmark book, Ekman and colleagues embarked on an extensive program of research aimed at mapping the facial expressions of emotions across cultures (Ekman et al., 1969). Although the validity of the methods and conclusions of this research are subject to continuing debate, it laid the groundwork for future investigations of the interpersonal effects of (facial) emotional expressions. Toward the end of the previous century, several theorists began to draw attention to the communicative functions of emotions and nonverbal displays. For instance, Parkinson (1996) made the case that emotions are inherently socially constituted, Fridlund (1994) contended that nonverbal displays are emitted in the service of social goals, and others began to develop ideas about the interpersonal functions of emotions in social interaction (Frijda & Mesquita, 1994; Keltner & Haidt, 1999).

In the wake of these theoretical developments, a new empirical literature emerged that is chiefly concerned with the social effects of emotions—that is, the effects of one person’s emotional expressions on others. This research has uncovered robust effects of emotional expressions on observers’ affect, cognition, and behavior, which are thought to contribute to the coordination of social interaction. Interpersonal effects of emotional expressions have been documented in diverse social settings such as close relationships, group decision making, customer service, conflict and negotiation, and leadership (for a comprehensive review, see van Kleef, 2016). The accelerating accumulation of empirical findings on the social effects of emotions in turn inspired the development of new theoretical perspectives that can account for these findings by building on, integrating, and extending some of the field’s early propositions.

About a decade ago, emotions as social information (EASI) theory (van Kleef, 2009) was introduced to illuminate the mechanisms and contingencies that govern the social effects of emotions. A basic assumption of this theory is that emotional expressions have evolved at least in part because of their communicative functions in coordinating social exchange. Given that people lack direct insight into each other’s minds, they attend to verbal and nonverbal emotional displays that provide a window into expressers’ thoughts, feelings, motives, and intentions to inform adaptive social responses. EASI theory posits that emotional expressions inform observers’ behavior by eliciting affective reactions (i.e., reciprocal and complementary emotions and sentiments about the expresser) and by triggering inferential processes in them (i.e., inferences about the source, meaning, and implications of the expresser’s emotion). For instance, one person’s expressions of anger may evoke reciprocal anger and negative sentiments in another person at whom the anger is directed, which may in turn fuel antagonistic behavioral responses in that person; but they may also lead the target of the anger to infer they did something wrong, which may in turn fuel accommodative behavioral responses. Thus, behavioral responses to emotional expressions can be understood by considering the relative predictive strength of affective and inferential processes. EASI theory predicts that the relative prominence of inferential processes in predicting behavioral responses to other people’s emotional expressions increases to the extent that the focal person is more motivated and able to engage in thorough information processing and/or perceives the emotional expressions as appropriate; conversely, the relative predictive strength of affective reactions increases to the extent that the focal person is less motivated or able to engage in thorough information processing and/or perceives the emotional expression as inappropriate. Notably, EASI theory further posits that the social effects of emotions are qualitatively (though not quantitatively) similar across expressive modalities. That is, the interpersonal effects of emotional expressions are similar in terms of direction (but not necessarily magnitude), regardless of whether they are emitted via the face, voice, bodily postures, words, or symbols such as emoticons (i.e., the functional equivalence hypothesis; van Kleef, 2016, 2017).

More recently, the theory of affective pragmatics (TAP; Scarantino, 2017) was developed to further specify the communicative functions of emotional expressions. Scarantino distinguishes among three aspects of emotional expression—the expression itself (the act of expressing a particular emotion), communicative moves (what one does in expressing that emotion), and communicative effects (what one accomplishes by expressing that emotion)—with the focus being on communicative moves. Extrapolating insights from linguistic pragmatics (Austin, 1962; Searle, 1979) to affective science, TAP postulates that emotional expressions, as communicative moves, are means of expressing internal states (“expressives”), representing what the world is like (“declaratives”), directing other people’s behavior (“imperatives”), and committing to future courses of action (“commissives”). Importantly, like EASI, TAP highlights the social-communicative properties of verbal as well as nonverbal emotional expressions (Scarantino, 2019).

In the latest theoretical contribution to this field, Parkinson (2021) advocates a relation-alignment approach to emotion, according to which, emotions “serve to align people’s orientations to one another and to objects and events in their shared environment” (p. 8). In keeping with earlier arguments, Parkinson contends that emotional expressions (verbal or facial) are means of achieving social influence rather than mere representations of internal states (also see Crivelli & Fridlund, 2018; Fischer & Manstead, 2016; van Kleef et al., 2011). However, Parkinson goes a step further by questioning whether (non)verbal emotional expressions reflect internal emotional states in the first place. He proposes that emotional language may not have been designed to describe emotional experience, and that the primary function of emotional faces may not be to express emotions. Parkinson’s argument aligns with EASI’s assumption that emotions have evolved (in part) because of their interpersonal rather than their intrapersonal effects, and with TAP’s notion of emotional expressions as “imperatives,” but it represents a more radical stance by implying that emotional expressions may not reflect internal feelings at all (also see Fridlund, 1994). Finally, similar to EASI and TAP, Parkinson’s relation-alignment perspective acknowledges that emotions can be expressed verbally as well as nonverbally.

This short overview reveals two important developments. The first concerns a continuing shift in theoretical emphasis on the social-communicative rather than individual-level functions and effects of emotions—the radicalization of the social approach to emotion. The second concerns a growing awareness of the obvious fact that emotions can be expressed through multiple modalities, and that hitherto largely neglected expressive modalities (e.g., words) are equally “real” and worthy of study as more established modalities (e.g., the face)—the emancipation of verbal emotional expressions. I consider the implications of each development in turn.

The Radicalization of the Social Approach to Emotions

The shift in theoretical emphasis from intrapersonal to interpersonal effects of emotions resonates with a fast-growing body of empirical evidence that emotional expressions influence observers’ affect, cognition, and behavior across domains of life. The idea that emotions derive much of their functionality from the fact that they are expressed—and subsequently perceived by others—also fits with findings from the adjacent literature on nonverbal communication in nonhuman primates (Buttelmann et al., 2009; de Waal, 2009; Keltner & Haidt, 1999). At the same time, the theoretical coupling between emotional experience and emotional expression appears to be loosening. Whereas both EASI theory (van Kleef, 2009, 2016) and TAP (Scarantino, 2017) maintain that emotional expressions serve part of their communicative functions by providing a window into the expresser’s mind (e.g., their thoughts, feelings, motives, and intentions)—thereby allowing for effects through the communication of experienced emotional states (“expressives” in TAP; cf. Buck, 1985; Ekman & Oster, 1979)—as well as by calling for a particular behavioral response from the observer (“imperatives” in TAP; cf. Fridlund, 1994), Parkinson (2021) suggests that emotional expressions may not reflect internal feelings at all. This raises two important questions.

First, if emotional expressions do not in any way represent outwardly perceptible cues to a person’s internal emotional experiences, then what do they represent? Parkinson (2021) rightfully points out that the link between emotional experience and expression is not unequivocal. Indeed, the outward expression of emotional experiences can be upregulated or downregulated, and people may express emotions that they do not feel, for instance, to comply with display rules (Matsumoto, 1990) or to influence others (Côté & Hideg, 2011; van Kleef et al., 2011). But clearly there are also cases where people freely express the emotions they experience, thereby providing observers with direct access to their internal feelings. It seems untenable, then, to claim that emotional experience and emotional expression are completely divorced. Indeed, if they were, the meaning of the term “emotional expression” would become highly ambiguous. Theoretically, it also seems plausible that the coevolution of emotional experience and emotional expression reflects links between the two from which emotional expressions derive part of their social functions. For instance, the wrinkling of the nose in disgust, which may initially have served primarily or exclusively to reduce the inflow of potentially contaminated air into the lungs, over time took on a signaling function that allowed conspecifics to avoid the apparent source of the disgust. Coevolution of emotional experience and emotional expression implies that emotional expressions as we know them today have traces in intrapersonal functions of emotions, suggesting systematic (although imperfect) associations between experience and expression.

Second, it remains to be specified how the social effects of emotions would come about if they were not mediated at least in part by the communication of felt emotions. Surely, some established effects of emotional expressions on observers’ behavioral responses may have been driven by observers’ interpretations of emotional expressions as a call for behavioral adjustment (“imperatives” in TAP). Other effects, however, are more difficult to reconcile with the position that emotional expressions do not reflect emotional experience. Evidence that observers are able to infer others’ appraisal processes from their emotional expressions (de Melo et al., 2014; Scherer & Grandjean, 2008; van Doorn et al., 2015), for instance, would be more parsimoniously accounted for by a view that allows for emotional expressions to reflect emotional experiences that track appraisals of situations. Moreover, it seems likely that emotional expressions as imperatives derive their social impact from evolved associations between intrapersonal processes and interpersonal signals. For instance, expressions of anger may help to enforce behavioral adjustment in targets precisely because the experience of anger is associated with a tendency to aggress against the source of the anger, and people learn this association over time. That is, the power of anger expressions as imperatives derives from the fact that angry individuals are more likely than nonangry individuals to become aggressive and pose a threat to one’s well-being, so that expressions of anger imply increased likelihood of aggression and harm.

As a fervent advocate of the social approach to emotions, I see great value in emphasizing the communicative functions of emotional expressions, but I believe there are insufficient grounds for eradicating links between emotional expression and emotional experience. Moreover, doing so would require separate theoretical accounts of how and why emotional experiences and emotional expressions evolved, and would eliminate a major source of emotional expressions’ informational value and social impact. As a counterweight to the radicalizing social approach to emotion, I call upon scholars to develop more precise theoretical arguments about when, how, and why emotional experience and emotional expression do or do not correspond. In this respect, promising directions are provided by the blossoming literature on emotion regulation (Gross, 2014), and research on emotional authenticity in particular. This work shows that emotional expressions that correspond with internal feelings are perceived as more authentic and attract more favorable responses from observers than emotional expressions that do not correspond with internal feelings (Cheshin et al., 2018; Côté et al., 2013; Grandey et al., 2005; Krumhuber et al., 2007). In light of such findings, it seems implausible that emotional expressions are entirely independent from emotional experience. It would be more fruitful to conceptualize the link between experience and expression as a continuum, with correspondence increasing or decreasing as a function of individual and situational contingencies. Mapping such contingencies would allow for further theoretical integration of insights from the largely separate literatures on the intrapersonal versus interpersonal effects of emotions.

The Emancipation of Verbal Emotional Expressions

A second development, which is unfolding in parallel to the mounting traction of the social approach to emotions, consists of a shift in thinking about what counts as an emotional expression. Until recently, the predominant view—emanating from Ekman’s basic emotion theory (e.g., Ekman, 1992)—has been that emotional expressions must, by definition, be involuntary, and that facial displays are the pinnacle of emotional expression. Indeed, to date, many scholars continue to believe that nonverbal emotional expressions more accurately capture the essence of emotion than do verbal expressions. This view is becoming increasingly untenable in light of theoretical developments and accumulating empirical findings.

If one accepts the premise that emotional expressions shape social interaction by providing information to observers about the expresser’s feelings, thoughts, motives, and/or intentions (Keltner & Haidt, 1999; Scarantino, 2017; van Kleef, 2009), it follows that emotional expressions can have such effects regardless of the expressive modality involved, as long as they succeed in conveying the relevant information. Indeed, it takes only cursory observation of real-life interactions to realize that social coordination can be informed by facial displays (e.g., a frown or a smile), vocal expressions (e.g., a grunt or a hum), bodily postures (e.g., clenched fists or hanging shoulders), written communications or utterances (e.g., an angry remark or an expression of gratitude), or use of emotion symbols (e.g., a “smiley” or a “frowny”). Accordingly, dozens of studies in social psychology, organizational behavior, marketing, and management have documented qualitatively similar social effects of emotions across different expressive modalities (for a review, see van Kleef, 2016).

A common objection to treating verbal emotional expressions as bona fide forms of emotional expression is that verbal expressions are under voluntary control and can therefore be faked, in which case they would not represent underlying emotional experiences. This is certainly true, but the same holds for any other form of emotional expression. People can also deliberately and voluntarily produce nonverbal (e.g., facial or postural) expressions that do not correspond with their internal feelings. Granted, it may be easier to feign emotions via verbal expressions than via nonverbal expressions, but deliberately enacted facial emotional displays can nonetheless appear very authentic (e.g., when generated via “deep acting”) and have similar social effects as spontaneous facial emotional expressions (Côté et al., 2013; Grandey et al., 2005). Involuntariness and fakeability, therefore, cannot be criteria for deciding whether something constitutes a proper emotional expression. Indeed, the most recent formulation of TAP explicitly rejects the idea that verbal expressions do not qualify as emotional expressions on the grounds that the presumption of involuntariness is untenable (Scarantino, 2019).

In line with this development, Parkinson’s (2021) analysis, too, emphasizes that both verbal and nonverbal emotional expressions can contribute to relation alignment. He does note, though, that “facial movements are better equipped than words to present a continuous stream of activity that is dynamically attuned to ongoing events and other people’s own unfolding reactions to those events” (p. 4). Thus, Parkinson distinguishes verbal and nonverbal emotional expressions, but on very different grounds than was commonly done until recently. Parkinson’s view that—the above distinction notwithstanding—verbal and nonverbal expressions of emotions can in principle contribute to relation alignment in similar ways constitutes a next step toward the emancipation of verbal emotional expressions. The use of emotion words to achieve social influence is increasingly seen as integral to the mechanics of human emotion. This is an important development, because it brings the scientific study of the social functions and effects of emotions more in line with the daily reality of emotion expression.

Conclusion

We are witnessing a gradual transition toward a more interpersonal approach to emotion, where emotions’ purpose is increasingly sought in their interpersonal rather than their intrapersonal consequences. Although I welcome—and have actively contributed to—this development, I caution against radicalizing to an extreme position that precludes the possibility that emotional expressions derive a substantial part of their meaning, informational value, and social functions from their association with the adaptive intrapersonal processes with which they presumably coevolved. Rather than severing the theoretical links between emotional experience and expression, we must identify conditions under which such links are stronger or weaker. This is the path toward a unified theory of emotions that can account for their intrapersonal as well as their interpersonal effects.

The shift in focus from individual to social functions of emotions is accompanied by a growing awareness of the importance of verbal emotional expressions. The progressing emancipation of verbal expressions as integral parts of the emotion process is necessary to account for the fact that emotions are quite often expressed through other modalities than the face, voice, or body. With human communication increasingly taking place via social media that rely primarily on text-based communication, verbal expressions of emotion are becoming ever more prevalent in social interaction. Disregarding them as unnatural or unreal does not do justice to the ways in which emotions are expressed in everyday life and impedes an integrated understanding of the social functions of emotions.

Together, the increasing focus on the social effects of emotions and the growing appreciation of verbal emotional expressions pave the way for the development of a more complete understanding of the ways in which people leverage emotions to communicate with and influence each other. The toolbox of emotional influence is rich and varied, allowing people to match their choice of expressive modalities to situational affordances and constraints. The availability of verbal and nonverbal expression channels enables emotional communication in pretty much any social context, regardless of physical proximity and access to visual or acoustic cues. As such, emotional expressions may well be the most generic source of human social influence.