Abstract
Contributions to this special section on positive emotions are summarized and integrated within a framework for organizing theory and research on particular emotions. Emotions are conceptualized as evolved strategies for coping with crises and opportunities, elicited by situational and appraisal antecedents–with phenomenological, physiological, expressive, behavioral, and emotivational goal components. Within this framework, theories are compared, inconsistencies and gaps in knowledge are identified, and issues in emotion theory are discussed.
Frameworks for Theory and Research on Positive Emotions
The articles in this special section illustrate the fruitfulness of investigating particular patterns of positive emotion. This burgeoning research domain also presents a challenge: how can the rapidly accumulating collection of ideas and data best be organized, understood, and assimilated?
Extending the approach taken by Sauter (XXXX), Table 1 presents a framework to summarize and structure a number of the authors’ contributions and identify gaps in knowledge and possible research directions. The table includes key information, but is not a comprehensive review of everything in the articles or the literature on positive emotions, both of which are beyond the scope of this short commentary. I have also included results of relevant research from our lab, on appraisal patterns (Roseman, 2013) and characteristic responses (Roseman, King, Nugent, & Gordon, 2013) distinguishing four positive emotions discussed in this special section (relief, hope, love, and pride) and another positive emotion often studied by researchers (joy), along with theoretical material about their functions.
Special section content within a framework for studying proposed positive emotions.
Note. Boldface indicates special section contributors clearly referenced empirical support. Footnotes refer to articles in this special section (secondary sources); these articles list authors of original research and theories (primary sources). aSauter (XXXX). bKringelbach and Berridge (XXXX). cStellar et al. (XXXX). dArmenta et al. (XXXX). eValdesolo et al. (XXXX). fRoseman, Antoniou, and Jose (1996). gRoseman, King, Nugent, and Gordon (2013). hCohen-Chen et al. (XXXX). iMatsumoto and Ekman (2008). jRoseman (2011).
The rows of Table 1 list hypothesized emotions, and columns show causes (antecedent events and appraisals), components (phenomenological, physiological, expressive, behavioral, and emotivational), and functions. The emotivational component consists of goals that people want to pursue as part of experiencing the emotion (Roseman, 2011), such as learning more about something when feeling interest (Sauter, XXXX), and alleviating the suffering of vulnerable persons when feeling compassion (Stellar et al., XXXX). Comparing entries within the table suggests potentially important questions.
What is the Most Accurate Specification of Emotions Described Differently in Different Theories?
For example, two articles present awe as an “epistemological” emotion, which functions to enhance information processing (Sauter, XXXX) and learning (Valdesolo, Shtulman, & Baron, XXXX). In contrast, Stellar et al. (XXXX, p. XXX) view awe as a self-transcendent emotion, functioning to “help individuals negotiate their own status vis-à-vis another by highlighting their subordinate position,” thereby promoting social cohesion. Of course awe might be both an epistemological emotion and a self-transcendent emotion, with emotivational goals of both cognitive and social accommodation.
Consistent with the epistemological view, note the striking similarities between awe and the emotion of surprise. According to Reisenzein and Meyer (2009, pp. 386–387), surprise is elicited by unexpectedness; is associated with a facial expression whose “full-blown” form includes “eyebrow raising, eye-widening, and mouth-opening / jaw drop”; is manifest in orienting to the surprising event, “investigative activities,” and “the reallocation of processing resources to the unexpected event”; and whose function involves “schema update.” Stellar et al. (XXXX) also cite research suggesting the experience of awe can be positive or negative in valence, which is an attribute that is rare and uncharacteristic of emotions other than surprise.
Should Each of These States Be Considered Discrete Emotions?
According to Sauter (XXXX, p. XXX), extant evidence on nonverbal signals suggests that pride, amusement, relief, awe, and interest may be the most recognizable positive emotional states, and in that sense “the most likely candidates for potentially basic positive emotions, equivalent to the set of negative emotions that are reliably communicated via nonverbal signals (e.g., fear, disgust, anger).”
If some positive emotion states are not distinct emotions, what else might they be? One possibility is combinations of emotions. For example, awe might be a combination of fear and joy (Konečni, 2005), or (given the correspondences noted above), surprise, fear, and joy. Relevant to the role of fear within this mix, Stellar et al. (XXXX) cite research linking awe to piloerection, which Benedek and Kaernbach (2011, p. 327) suggest may be associated with “the threatening aspect of being moved” (emotionally). Compassion might involve a combination of love and sadness. Sauter (XXXX) notes that studies have found some facial features of sadness in compassion expressions, and that compassion expressions are often mistaken for sadness or confused with love and gratitude expressions. Sauter also cites research indicating that elation is caused by unexpected positive events, and facially expressed by smiles with open mouth, raised brows, and widened eyes; could elation be a mixture of joy (manifest in smiling) and surprise (corresponding to its other specified features)?
A second possibility is that some emotion states are types or variants of others. For example, Sauter (XXXX) cites data showing vocal expressions of love and gratitude are often identified as compassion, suggesting all three vocalizations may communicate a general prosocial affiliative state. Might gratitude and compassion then be variants of love, akin to Harlow’s (1971) “infant love” (which for Harlow encompassed love of any young child for the mother) and “maternal love”? In light of extant research results, Feldman (e.g., 2012) has proposed that oxytocin pathways are among the mediators of maternal, romantic, and peer attachments. Thus there may be commonalities in physiological mechanisms involved in different types of love (though physiological data on oxytocin in attachment to parents is scarce and oxytocin’s effects may vary depending on environment, sex, and other factors; e.g., Hammock, 2015).
A third possibility is suggested by Kringelbach and Berridge’s article (XXXX). They review evidence showing that brain substrates of wanting are distinct from those of liking. Wanting is typically thought of as a motivational construct, linked to instrumental behavior; liking is more of an emotional response, associated with pleasurable feeling. Emotions (such as joy, pride, fear, and anger), more often than motives (e.g., hunger, thirst, need for achievement), have been linked to characteristic facial expressions. This may be relevant to evidence regarding the states of desire and interest. According to Sauter’s review (XXXX, p. XXX), “it cannot currently be concluded that desire is associated with a nonverbal signal in any modality.” As noted in Table 1, interest has been associated with contradictory facial expressions (e.g., lips pressed together or parted; eyelids closed or widened), and its forward lean and forward head movements could be manifestations of information acquisition and processing motives. Should wanting generally, and interest (desire for interaction with certain types of stimuli) and sexual desire in particular, be considered motives rather than emotions?
The distinction matters if motives and emotions have different properties. In addition to greater likelihood of having nonverbal signals, emotions seem more general and preemptive compared to motives (Roseman, 2008; Tomkins, 1970). Regarding generality, motives (e.g., hunger, thirst, and sexual drive) appear more likely to be activated by relatively specific conditions (e.g., food and water deprivation, erotic stimuli) and their biological substrates, and they direct behavior toward relatively specific ends (e.g., obtaining nutrients, water, sexual activity). In contrast, emotions may be elicited by contingencies applicable to any motive. For example, attaining any rewarding state (including food when hungry, water when thirsty, and sexual activity when aroused) can elicit joy. Fulfillment of any motive, if attributed to another person, can elicit affection for that person. Attainment of any motive attributed to the (individual or extended) self can result in pride. Although both motives and emotions can engender instrumental behavior, the emotivational goals of emotions (e.g., to continue rewarding events in joy, to maintain closeness in love, to obtain recognition for whatever accomplishments or attributes are positively valued in pride) may be broader and more variably fulfilled than the goals of motives.
With regard to preemptiveness, though qualified by the relative intensity of a motive versus an emotion on a particular occasion, emotions may tend to take precedence over motives (Leeper, 1970; Tomkins, 1970). Frijda (2007, p. 123) also distinguishes between emotions and “concerns” (his term for constructs such as motives, needs, and goals). Appraisals of concern relevance (e.g., of events) give rise to emotions. Emotions have “control precedence” (pp. 28–29): focusing attention on emotion-relevant stimuli and tending to interfere with “other pursuits” (though Frijda does not specify them, the goal-directedness of “pursuits” suggests they fall within the concern domain). We can find positive emotion examples of preemptiveness in research by Shaver, Schwartz, Kirson, and O’Connor (1987): describing experiences of love, participants reported being preoccupied with the beloved person, and also forgetful and distracted. Describing experiences of joy, participants reported seeing only the bright side of things and feeling invulnerable. While mild positive affect may be associated with cognitive flexibility (Isen, 2008), high levels of positive emotion may lead to inattention to dangers and difficulty inhibiting risky behaviors (Gruber, Mauss, & Tamir, 2011).
What Can Be Learned From Conceptualizing Emotions as Strategies?
Looking across rows in Table 1 suggests that the components of a given emotion may be understood not as isolated phenomena, but as interrelated responses that have coevolved, with each emotion forming a (typically nonconscious) strategy for coping with a particular type of crisis (specific negative emotions) or opportunity (specific positive emotions; Roseman, 2011). For example, as shown in Table 1, the hypothesized strategy formed by the responses of love involves moving toward someone in physical or social space, increasing contact and interaction with a person whose attributes or actions are appraised as causing satisfaction of motives (affiliation, sex, approval, etc.).
The emotivational component of an emotion provides goals that motivate instrumental action to aim for outcomes consistent with the coping strategy corresponding to the emotion. In love, the goal of maintaining closeness and connection with another person motivates behavior that establishes and strengthens interpersonal relationships (within which satisfying contact and interaction can be regularized and facilitated).
The behavioral component suggests particular actions likely to achieve emotivational goals and implement the strategy corresponding to the emotion. For example, hugging and stroking are behaviors that can increase closeness and connection between people.
The expressive component comprises signals that com-municate an emotion (e.g., Buck, 1984) and its behavioral and motivational components (Frijda, 1986), thereby prompting perceivers to act in ways that fit with the corresponding strategy. For example, nonthreatening gaze, smiling, and soft speech can communicate love and invite reciprocation. (The specific signals used to communicate an emotion often have additional adaptive functions in the situational context within which the emotion occurs, as when the gaze of lovers allows and facilitates interpersonal interaction, or the expanded posture of pride attracts others’ attention upon occasions of relatively likely social reward; cf. Frijda, 1986, pp. 11–29.)
The thoughts and feelings of the phenomenological component make important aspects of a situation salient and cue retrieval of other experiences of the emotion and associated information, thereby influencing automatic and effortful action. This is consistent with evidence for emotion-congruent memory (e.g., as discussed by Levine & Pizarro, 2004; and effects of discrete emotions on judgments, e.g., as demonstrated by Lerner & Keltner, 2000). In love, appreciating and thinking we belong with someone focus attention on the partner’s desirable qualities and mutual compatibilities, which may increase memory for similar thoughts and experiences. Feeling drawn to the person gives priority and urgency to the emotivational goals and behaviors of love, which can strengthen the relationship.
The physiological component comprises the biological substrate of each of the other components. Increases in oxytocin and activity of the dopamine reward system (De Boer, van Buel, & Ter Horst, 2012) may contribute to love’s feeling of being drawn to someone, and its salience and urgency.
Other emotions in Table 1 can be similarly analyzed. If the strategy that corresponds to awe is to enhance information processing (Sauter, XXXX) or facilitate the development of new worldviews (Valdesolo et al., XXXX), we can examine whether there are
characteristically present emotivational goals (e.g., uncertainty reduction and explanation-seeking) whose pursuit tends to implement the strategy;
behaviors (e.g., exploration) likely to achieve those goals;
nonverbal displays that communicate the emotion to others, prompting them to act in ways consistent with the expresser’s coping strategy (e.g., providing information in response to awe-struck looks) or directly contribute to the strategy (e.g., by increasing information intake);
thoughts (e.g., of uncertainty) and feelings (e.g., of humility) that make awe elicitors salient and direct cognition and instrumental action toward coping with them.
We can also examine the extent to which people feeling awe pursue goals related to self-diminishment and the hypothesized function of subordination to a group; enact behaviors instrumental to those goals (self-deprecation? compliance?); display nonverbal signals related to the strategy (submission gestures?), etc.
If compassion functions to encourage care of people in need (Stellar et al., XXXX), then
the emotivational goal of caring for vulnerable others is consistent with that strategy;
helping is behavior instrumental to that goal;
patting touch may communicate sympathy and encourage recipients to accept care;
thoughts of similarity and connection to others may overcome obstacles to care-giving;
midbrain periacqueductal gray (PAG) activity may contribute to feeling others’ pain and to parental nurturance behaviors (Simon-Thomas et al., 2012).
Researchers can also study whether there is similar coherence among responses related to the hypothesized function of promoting nonkin cooperation.
If the strategy corresponding to gratitude is to promote nonkin reciprocity (Stellar et al., XXXX) and strengthen social bonds (Armenta, Fritz, & Lyubomirsky, XXXX), then
seeking to repay the benefactor (Armenta et al., XXXX) is an emotivational goal which seems consistent with that strategy;
helping the benefactor is instrumental to that goal;
verbal expressions of thanks communicate gratitude and may prompt additional prosocial action by the benefactor (Stellar et al., XXXX);
thoughts about connectedness and feelings of humility (Armenta et al., XXXX) may encourage and give priority to reciprocating.
We can also investigate whether self-improvement is integral to the various components comprising the response syndrome of gratitude (Armenta et al., XXXX).
Are All Responses Associated With an Emotion Part of a Coherent Strategy?
What if the goal of self-improvement were not related to the strategy of promoting reciprocity and strengthening social bonds? That might suggest weaker links between gratitude and self-improvement, but it would not invalidate observed relationships. Motivation for self-improvement may be a frequent effect of gratitude, or of one or more of gratitude’s components, such as feelings of humility (Armenta et al., XXXX). It could be a second-order effect of the appraisal pattern that elicits gratitude, with perceptions of the benefactor’s generosity evoking an emotional state of elevation, which in turn inspires self-improvement (Armenta et al., XXXX). Similarly, prosocial behavior might be an effect of awe. Piff, Dietze, Feinberg, Stancato, and Keltner (2015) found evidence that awe triggers the sense of a small self, which in turn leads to greater prosocial behavior.
Can Gaps in the Framework Be Filled?
Empty and partially filled cells in Table 1 suggest the need for research. For example, we may try to identify nonverbal expressions of hope, action tendencies of elation, emotivational goals of sensory pleasure, and the physiological substrates of the differentiated emotional responses in the table. Some empty cells (e.g., antecedents of interest, physiology of sexual desire) can be filled from existing research not reviewed in this special section.
This framework may help make sense of the rapidly accumulating knowledge about positive emotions presented in and beyond this special section. A prototype approach to defining emotions suggests that (a) the more components of prototypical emotions a particular state is found to have, the more likely it is to be regarded as an emotion; (b) the more distinctive features that comprise a state (e.g., as shown in Table 1), the more likely it is to be regarded as a distinct or “discrete” emotion; and (c) the more that two states overlap in features, the more likely they are to be considered variants of a single emotion. For example, the differentiable observed nonverbal displays of amusement, awe, interest, and relief lead Sauter (XXXX) to suggest them as the most likely candidates for potentially basic positive emotions. The overlap in features between contentment and sensory pleasure, and confusions in distinguishing between them, raise questions about their status as distinct emotions.
Beyond perfect coherence
It should be acknowledged that this approach, though it seems useful, is not definitive. Suppose there is no nonverbal expression of hope. Could hope still be considered an emotion (Cohen-Chen, Crisp, & Halperin, XXXX) based on its having other emotion components and properties, which form a strategy for coping (with situations of potential goal attainment)? Suppose there are no distinct brain networks corresponding to the different positive emotion states distinguished in Table 1 (see Lindquist, Wager, Kober, Bliss-Moreau, & Barrett, 2012). Would that mean the empirical relationships among responses documented by the articles in this special section don’t really exist, or are not based in brain and bodily response?
Insofar as there are patterns among responses, identifying them adds to scientific knowledge. Debating whether a given state is or is not a discrete emotion can advance the field insofar as the classification facilitates prediction. Proposing that a particular state is a positive emotion prompts us to look for (in addition to positive hedonic tone) the likely presence of nonverbal signals, action tendencies, and emotivational goals, together with related properties such as subjective urgency and control precedence. I say likely presence because relationships among variables in science very rarely hold under all conditions. The authors whose work appears in this special section are to be commended for contributing to what we know about the causes, components or correlates, and functions of such a rich variety of states.
Footnotes
Declaration of Conflicting Interests
The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
