The many faces of emotional contagion

Explaining affective linkage: An affective process theory

The core tenet of the affective process theory (APT) is that each mechanism for affective linkage must flow through the emotion process while (a) starting from one of the three visible steps and (b) travelling through a valid pathway illustrated in Figure 1. As such, a heuristic for uncovering the many mechanisms is to start one-by-one with each externally visible point in the emotion process model, and then to move through all possible pathways that include the source’s emotional experience and eventually reach the recipient’s emotional experience. ¹ This results in 10 distinct mechanisms of affective linkage. After a discussion of the emotional appraisal theory in which the APT is grounded, each of the 10 mechanisms are described in detail and integrated with existing treatments in the literature.

What determines shared vantage point

The affective process theory’s concept of shared vantage point (SVP) attempts to integrate into a single framework a wide range of previously studied moderators for emotional contagion. These moderators, in turn, help to determine when we should expect SVP to be high or low. In outlining how SVP fits with these moderating factors, it is important to emphasize that SVP is about the temporary situation in which we find ourselves—that is, whether we are on the same side of the same situation. This construct is not merely a matter of social closeness, or indeed anything else about the enduring relationship between the two parties. Although much of the past work on moderators has examined the personal relationship between the contagion source and recipient, according to APT this factor has only an indirect influence. That is, our relationships can influence our basic assumptions about whether we are on the same side of a situation, but it is the situation itself that influences whether we experience convergent, divergent, or complementary linkage. The same person can bring out happiness, sadness, or anger in us, as events change during our fleeting moments together.

Sharing a vantage point can come from either structural or motivational factors. That is, two people can choose the same vantage point or it can be thrust upon them. The range of moderators that influence SVP are summarized in Table 1. Structural factors in organizational settings include being in the same corporate division, working on the same team, and holding job roles that face the same risk of layoffs. Totterdell (2000) found greater linkage for cricket teams when playing defense instead of offense, which involves greater social coordination—and, thus, greater shared stimuli. Other structural factors include being in a competitive versus cooperative role. Competition structurally limits the ability to share a vantage point with another person. Lanzetta and Englis (1989) demonstrated convergence in the contagion of smiles and grimaces when participants believed they were cooperating, and complementarity when they believed they were competing. According to van Kleef, de Dreu, and Manstead’s (2010) emotions as social information (EASI) theory, there is a complementary nature to the emotions elicited in competitive settings. Notably, negotiators’ negative displays elicit greater concessions, a sign of fearful appeasement, whereas happy displays elicit lesser concessions, a sign of resolute hopefulness—suggesting a complementarity to parties’ affective linkage. Power differences create structurally different vantage points, with a notable asymmetry: Hatfield et al. (1994) argued that lower power individuals pay more attention to the other party and, therefore, are stronger recipients of contagion than higher power individuals. APT argues that the linkage typically experienced by lower power parties is complementary, because they are particularly aware that their vantage is not shared. Empirical results support this notion (e.g., Sy et al., 2005; van Kleef et al., 2008; but cf. Hsee, Hatfield, Carlsson, & Chetomb, 1990). Finally, familiarity with a situation can structurally influence SVP, because the familiarity can make the other person’s experience seem more self-relevant and shared. Hawk, Fischer, and van Kleef (2011) found that previous experience with an embarrassing task led to greater contagion when viewing other embarrassed participants.

OPEN IN VIEWER

Table 1.

Factors that increase shared vantage point (SVP).

Type	Variables	Sample references
Structural factors	Job roles that expose employees to the same versus different work environment, for example, divisions, teams, functions, etc. Job roles that require more versus less social coordination Cooperation versus competition Low versus high power differences Familiarity with each other’s stimuli	Hatfield et al. (1994); Hawk et al. (2011); Lanzetta & Englis (1989); Totterdell (2000); van Kleef et al. (2008)
Motivational factors	Liking Empathy Personal similarity, for example, sharing attitudes Social in-group versus out-group	Hatfield et al. (1994); Howard & Gengler (2001); McHugo et al. (1991); Parkinson (2011); van der Schalk et al. (2011)

Motivational factors depend on an active choice to take another’s vantage point, and these factors largely stem from social closeness and what contributes to it. When we care about another person, we literally have a sense of including the other in the self (Aron, Aron, & Smollan, 1992)—which is a matter of appraising the world from their same vantage point. Research on empathy establishes that we more strongly take the perspectives of cherished others (Davis, 1983; Hatfield et al., 1994). Liking also leads us to pay greater attention to other people’s affect and to take their cues more seriously (Parkinson, 2011). Along these lines, it has been a consistent and long-standing observation that affective linkage is convergent and more intense when the recipient feels closer to the source. William James (1890/1984) noted over a century ago that we feel congruent emotions with some people yet differing emotions from others. Hatfield et al. (1994) argued that contagion is most common when individuals “love, like, or identify with others or share their goals” (p. 9). Indeed, they used the term countercontagion when individuals experience discrepant states, with the implication that an opposing relationship deliberately drives people’s feelings apart. The case of completely opposing linkage within hostile relationships involves extremely low levels of SVP. For example, individuals experience the German concept of Schadenfreude by delighting at the misfortune of the other, and can likewise feel resentment at the other’s success. Most relationships fall somewhere in the more ordinary continuum of liking. Convergent linkage is greater when receivers like senders—even when liking is randomly assigned through believing the party had offered a token gift (Howard & Gengler, 2001). Convergence is also greater when participants assume similarity with the other person. For example, McHugo, Lanzetta, and Bush (1991) found that convergence with a politician’s emotional expressions varied based on whether the participant shared the same political attitudes. Finally, being a member of the same social group increases the level of SVP. Ingroup members may be both more likable and presumed to be more similar. Van der Schalk et al. (2011) showed that self-reported emotions were more convergent with those of ingroup members and more complementary with those of outgroup members.

Although structural and motivational factors are theoretically distinct from each other, they can also interact. Our relationships can influence the situations that we craft together versus separately from other people. Being similar to another person can lead to greater liking—a motivational explanation—but it can also indicate a realistic estimate about whether we face the same situations from the same vantage point. The same is true of group membership. As an unfortunate example, in many parts of the world, membership in a shared ethnic group can determine where people fit with each other in terms of common life situations, for example, as a proxy for competing over resources or how they can expect to be treated by authority figures.

Ten mechanisms for affective linkage

This section outlines and distinguishes the ten mechanisms for affective linkage that are identified by Affective Process Theory (APT). These are illustrated in Figure 2 (stimulus-based processes), Figure 3 (expressive cue-based processes), and Figure 4 (post-emotional behavior-based processes), and summarized in Table 2.

OPEN IN VIEWER

Figure 2a.

Emotional contagion via the shared stimulus pathway (Mechanism 1).

OPEN IN VIEWER

Figure 2b.

Emotional contagion via the imitated stimulus pathway (Mechanism 2).

OPEN IN VIEWER

Figure 2c.

Emotional contagion via the empathetic-through-stimulus pathway (Mechanism 3).

OPEN IN VIEWER

Figure 3a.

Emotional contagion via the emotion recognition pathway (Mechanism 4).

OPEN IN VIEWER

Figure 3b.

Emotional contagion via the consequences-through-recognition pathway (Mechanism 5).

OPEN IN VIEWER

Figure 3c.

Emotional contagion via the imitated cues pathway (Mechanism 6).

OPEN IN VIEWER

Figure 3d.

Emotional contagion via the empathetic-through-recognition pathway (Mechanism 7).

OPEN IN VIEWER

Figure 4a.

Emotional contagion via the behavioral consequences pathway (Mechanism 8).

OPEN IN VIEWER

Figure 4b.

Emotional contagion via the behavioral imitation pathway (Mechanism 9).

OPEN IN VIEWER

Figure 4c.

Emotional contagion via the empathetic-through-consequences pathway (Mechanism 10).

OPEN IN VIEWER

Table 2.

Mechanisms for affective linkage.

	Stimulus	Expressive cues	Consequences
Forward process	1. Shared stimulus	Cue recognition 4. Emotion recognition Emotional responses: Hatfield et al. (1994) Emotional interpretation: Hareli & Rafaeli (2008) 5. Consequences-through-recognition Emotions as social information: van Kleef et al. (2010) Social comparison: Barsade, (2002); Gump & Kulik (1997) Social appraisal: Hareli & Rafaeli (2008); Parkinson (2011)	8. Behavioral consequences Interpersonal emotional regulation: Mayer et al. (2000) Emotional labor: Côté (2005) Artifacts: Rafaeli & Vilnai-Yavetz (2004)
Imitative process	2. Imitated stimulus	6. Imitated cues Primitive contagion: Hatfield et al. (1994); Barsade (2002) Entrainment: Kelly & Barsade (2001)	9. Behavioral imitation
Empathetic process	3. Empathetic- through-stimulus	7. Empathetic-through-recognition	10. Empathetic-through-consequences

Note. Text after bullet points denotes examples of previous related theoretical accounts that cover the mechanism outlined, including special cases.

1. Shared stimulus (Figure 2a). In the shared stimulus mechanism, emotional linkage takes place when individuals share the same initial stimulus. In this case, SVP is high. Employees showing up to the same workplace each day might be exposed to the same job characteristics, physical environment, and senior management. This mechanism can link people who do not have direct contact with each other—or, indeed, who may not even know the other party exists. In a historical example, the 1938 radio broadcast of Orson Welles’s War of the Worlds created a wave of panic among independent listeners who believed the fictional story was actually taking place. On a more day-to-day level, we share many work events with coworkers, performing many of the same tasks, using the same tools, surrounded by the same colleagues, and in the same physical plant.

In this mechanism, emotions do not pass or spread directly from an individual source to a recipient, but rather each individual comes to an emotion on their own. In response to their shared stimulus, the two parties might interpret the stimulus similarly or differently, in which case their linkage is convergent or divergent, respectively. In organizational settings, even if colleagues do not register events similarly at first, they tend to converge over time to shared interpretations—which are often managed actively by leaders, but can emerge just as strongly over the water cooler.

2. Imitated stimulus (Figure 2b). In the imitated stimulus mechanism, people imitate each other’s stimuli, and then register them as in the shared stimulus (Mechanism 1). The distinction here is that the common stimulus does not occur independently to each person, but rather the recipient needed to imitate it. For example, consider a company that has a policy of casual dress on Fridays. This policy is a shared stimulus and exists independent of employees’ choice. However, if some employees started wearing platform-heeled shoes on casual dress Fridays and others joined in this trend, then the platform heels are an imitated stimulus. Imitated stimuli require people to make a choice to expose themselves to the same stimuli, and this choice allows the stimulus to spread and reach multiple people. Other examples of imitated stimuli involve following fashions that differentially influence affective states, such as viewing comedies versus tragedies, exercising with yoga versus kick-boxing, consuming psychoactive substances such as opium versus methamphetamine, or taking part in outdoor activities such as walking versus mountain climbing. Within organizational settings, employees can follow cultural rites and other normative behaviors, such as having an open-door policy, eating with colleagues instead of alone, and telecommuting. Norms influence us to imitate workplace stimuli, and affective events theory argues that a wide range of workplace stimuli are emotionally evocative (Weiss & Cropanzano, 1996). As with the shared stimulus mechanism, imitated stimulus has a high shared vantage point (SVP), and affective linkage is convergent or divergent depending on whether the source and recipient have similar interpretations. For example, imitating the trend for platform-heeled shoes instead of comfortable shoes could create a wave of irritation for some individuals, but others might find their newly increased physical height from the heels to be empowering.

3. Empathetic-through-stimulus (Figure 2c). As discussed before, the empathetic process can begin with a stimulus, expressive cues, or postemotional behavior. In the empathetic-through-stimulus mechanism, a partner becomes aware of an emotionally evocative event, and then imagines how that event could make another person feel (Hawk et al., 2011). This act of imagining another person’s feelings then rubs off on the imaginer. Empathetic processing typically results in high SVP—which leads to convergent or divergent linkage. As discussed before, the two exceptions are when parties’ emotions diverge due to inaccuracy when attributing the other party’s emotions or when they are complementary due to the recipient’s consumption value of the source’s emotion, with the latter involving low SVP.

4. Emotion recognition (Figure 3a). The fourth mechanism for affective linkage, emotion recognition, involves a critical tool for group living: deriving information from the emotional expressions of others. So-called basic theories of emotion argue that emotion’s whole purpose is its power as an interpersonal communication device (e.g., Ekman, 1972; see also Parkinson, 2012). Bühler’s (1934/1990) organon model (Scherer, 1988b) argues that emotions can communicate in three different ways: (a) a symptom of the speaker’s authentic and spontaneous internal state, and thereby expresses emotions, intentions, and attitudes; (b) an intentional signal, thereby serving as an appeal to produce a reaction; and (c) a symbolic statement that conveys deliberately what might otherwise be spoken in words. In each of these cases, a source’s expressive cues become stimuli that feed into the recipient’s emotional registration and experience.

Emotional displays are central to the social function of emotion (Fischer & Manstead, 2008). We engage in backtracking (Elfenbein, 2007), which allows us to use our awareness of another person’s cues to infer their symptoms, intentional signals, and symbols (see also Hareli & Hess, 2010; Hareli & Rafaeli, 2008; Parkinson, 2011). As such, emotional cues in organizations offer information that we use for sensemaking (Maitlis, Vogus, & Lawrence, 2013; Weick, Sutcliffe, & Obstfeld, 2005). The backtracking process uses perceivers’ implicit knowledge of appraisal theory to reconstruct events based on the emotional expressions produced (see also MacCann & Roberts, 2008; van Kleef et al., 2004).

The notion that emotion recognition contributes to affective linkage is strongly supported on a theoretical basis even while empirical evidence is indirect. For example, support can be inferred from the finding that contagion increases under the factors that tend to increase emotion recognition accuracy, such as greater attention—even minimal, implicit attention—and greater individual sensitivity to others’ states (Hatfield et al., 1994; Verbeke, 1997). Indeed, Scherer (1988b) argued we should shift the dependent measure of emotion recognition research from the attributions of others’ states to their effects on the judge’s own state.

Emotional expressions are powerful stimuli, privileged to grab our attention, which makes emotion recognition replete in personal and organizational life. Even so, emotional expressions feed into the registration process just like any other stimulus, for “hot” or “cold” interpretation. On a “hot” level, seeing an angry supervisor can make a supervisee fearful. On a “cold” level, negotiators subscribing to an inverse affect hypothesis can analyze their counterpart’s disappointment as a sign they have reached a favorable deal (L. L. Thompson, Valley, & Kramer, 1995). Emotion recognition can span from automatic to deliberate. It can occur quickly and outside of perceivers’ awareness, but alternately it can result from careful analysis. We become particularly aware of emotion recognition’s conscious basis when it is difficult to answer the questions represented by appraisal theory’s checklist without more information, for example not knowing whether to be enthusiastic or discouraged with additional task demands until receiving clarification about what the task actually involves.

It is important to point out that emotion recognition can be rife with inaccuracy—in spite of our natural ability to recognize other people’s emotional expressions, even under carefully controlled situations that involve high intensity expressions and explicit instructions that participants should pay attention to them, average accuracy rates rarely exceed 80% (Elfenbein & Ambady, 2002). In real-life situations, we have contextual clues to assist with emotion recognition, as well as full-channel presentations that include facial expressions, vocal tone, and body language, which aid in recognizing others’ emotions (Kudesia & Elfenbein, 2013). However, we have drawbacks as well: the lack of perfect attention as well as the challenge of display rules (Ekman, 1972) that attempt to hide expressions for the sake of social goals. Accurate emotion recognition is particularly challenging across cultural boundaries, given that cultural groups tend to have subtle differences in their expressive style (Elfenbein, 2013).

The mechanism of emotion recognition is classified as low SVP—because the linkage source’s expressive behavior is itself the recipient’s stimulus—which means that emotion recognition produces complementary linkage. As discussed before, even if emotion recognition produces a congruent state between the source and recipient, this results from an emotion spiral or a metaemotion, rather than from convergent linkage.

5. Consequences-through-recognition (Figure 3b). The fifth mechanism, consequences-through-recognition, involves the emotion recognition pathway with an additional detour. In this case, the information provided from another person’s emotional cues is analyzed outside of an emotional pathway, but this analysis itself becomes registered for emotional meaning. Rather than catching an emotion, a perceiver thinks through its implications, and these implications are themselves evocative. Hatfield and colleagues give the example of another person’s happiness or sadness triggering our memories about times we have felt happy or sad (Hsee et al., 1990). In this case, it is our own memories that instill emotion, rather than the partner’s expressive cues themselves. In terms of the process model, the linkage recipient’s postemotional behaviors mediate the pathway from their emotion recognition to registration.

An influential theory emphasizing the first steps of this pathway is the emotions as social information (EASI) model (van Kleef et al., 2010; van Kleef, Homan, & Cheshin, 2012). In van Kleef et al.’s (2004) research on negotiations, partners infer that a negotiator expressing anger has reached the maximum willingness-to-pay, whereas a happy negotiator still has room to give. Extending the EASI model, this inference can then itself be registered emotionally. After recognizing that a negotiation partner is angry, the recipient may infer that the partner has reached their financial limit, and this inference itself may then produce disappointment or pity. Likewise, in a workplace setting, a colleague’s chronic anger might lead an employee to realize they need to avoid that colleague, and the thought of distancing makes them sad. These are examples of complementary contagion.

There is a special case of the consequences-through-recognition mechanism that has been emphasized in the literature as social comparison (Barsade, 2002), also called social appraisal (Parkinson, 2011). In this case, people infer information about their own predicament from the expressive cues of other people sharing the same predicament, particularly in ambiguous situations when people aren’t sure how they should feel. Note that this is a special case with high SVP and yet it involves recognizing others’ emotions, backtracking to understand the meaning behind these expressions, and then appraising this meaning for oneself. In a classic developmental study, toddlers were only willing to crawl across a glass visual cliff if their mothers showed happy versus fearful displays (Sorce, Emde, Campos, & Klinnert, 1985). This form of social referencing (Campos & Sternberg, 1981) allows us to make sense of our environment by using others’ emotional expressions as information for the self. When this information is emotionally evocative, social referencing leads to affective linkage. As an example, participants in Latane and Darley’s (1969) classic study on social influence who were in the presence of calm confederates while smoke entered the room reported that they believed the situation wasn’t dangerous, and presumably it was this belief that kept them calm. Examples of social comparison abound in organizational settings. Novice steelworkers who see their colleagues willing to work at great heights—even joking and playing around—infer that the situation is not dangerous and they shouldn’t be scared (Haas, 1977). These examples follow the consequences-through-recognition mechanism versus shared stimulus, because the linkage source’s expressive cues are recognized, and then the recipient appraises their own analysis of the situation.

As with emotion recognition, the detour mechanism of consequences-through-recognition is a process with low SVP, in that the behavior of the linkage source is itself the recipient’s stimulus. As such, with the exception of the special case for social comparison—which can produce convergent or divergent linkage—consequences-through-recognition normally produces complementary linkage.

6. Imitated cues, also known as primitive contagion (Figure 3c). The form of affective linkage that has received the most attention is imitated cues. Through a two-part process highlighted by Hatfield et al. (1994), the expressive cues of others invoke behavioral mimicry that leads, in turn, to affective experience. These primitive processes evolved as a form of social glue, where we first automatically mimic and synchronize facial expressions, vocal tones, and body movement to be in rhythm with those around us (Allport, 1937; Bernieri, Reznick, & Rosenthal, 1988; Chartrand & van Baaren, 2009; Hatfield et al., 1994; Kelly & Barsade, 2001). Second, the mimicker receives afferent feedback from their own expressive cues (Zajonc, 1998), which brings their affective state in line with the expression they imitated.

In establishing evidence for the first step, numerous studies demonstrate that individuals spontaneously mimic the facial expressions of others, and even that they imitate their vocal tones and postures (e.g., Blairy, Herrera, & Hess, 1999; Gross & John, 1995; Laird et al., 1994). Communication accommodation theory argues that we adapt to others’ behavioral styles along multiple dimensions to ensure effective communication (Giles, Coupland, & Coupland, 1991). For example, even during brief interactions, individuals tend to mimic their interaction partner’s linguistic patterns (Pennebaker, Mehl, & Niederhoffer, 2003). This mimicry is particularly strong for well-liked interaction partners (Lakin, Jefferis, Cheng, & Chartrand, 2003; van Baaren, Holland, Kawakami, & van Knippenberg, 2004), which feeds into theories that mimicry serves as a communication act (Bavelas, Black, Lemery, & Mullet, 1987) that signals unity (Tickle-Degnen & Rosenthal, 1987). Newer work emphasizes the automatic biological basis of this social process via mirror neurons (Iacoboni, 2009), even though mimicry also has a deliberate basis in social intentions (Hess & Fischer, 2013).

In establishing evidence for the second step of the imitated cues mechanism, research shows that mimicry feeds back into emotional experience (Hatfield, Hsee, Costello, & Weisman, 1995). For example, inserting images of smiling people into a videotape of a comedy routine increased levels of reported enjoyment, but not when participants were prevented from moving their facial muscles (Bush, Barr, McHugo, & Lanzetta, 1989). This idea of feedback is not new—as early as 1872, Charles Darwin argued that feedback from skeletal musculature should affect emotional experience. William James (1890/1984) likewise argued that people infer their own emotions by sensing their muscular, glandular, and visceral responses. Tomkins (1962) introduced the facial feedback hypothesis, arguing that emotional experience actually results primarily from proprioceptive feedback due to muscular contraction, although he later argued this was sufficient but not necessary (Tomkins, 1980). Zajonc’s (1998) efference theory emphasizes that we receive temperature feedback from facial muscles. However, it is worth noting that feedback effects have also been demonstrated for vocal expressions and posture, for which muscular and temperature mechanisms are not applicable (Hawk, Fischer, & van Kleef, 2012; cf. Halberstadt, Winkielmann, Niedenthal, & Dalle, 2009). Even without the primal nature of afferent feedback, we can still mundanely experience afferent feedback based on years of conditioned associations from the opposite direction.

Overall, due to the disproportionate research focus on imitated cues, authors making the point that it is an incomplete account of emotional contagion have often had to shout. Empirical evidence does support the primitive contagion mechanism—for example, positive mood transfer is greater the more recipients mimic the smiles of senders (Howard & Gengler, 2001). However, other empirical evidence highlights the limits of mimicry’s explanatory power. Notably, the most methodologically sophisticated studies provide minimal evidence for facial feedback (Manstead, 1988), and associations are often nonsignificant between mimicry and reported mood (Gump & Kulik, 1997; Hess & Blairy, 2001). Even the positive findings are not large enough in magnitude to explain the large effects of contagion (Parkinson, 2011).

Note that there are variations on the imitated cues mechanism. One variation is embodied cognition (Hawk et al., 2012; Niedenthal, Barsalou, Winkielman, Krauth-Gruber, & Ric, 2005; Ponari, Conson, D’Amico, Grossi, & Trojano, 2012). After mimicking expressive cues, there is feedback not only directly to emotional experience but also to emotion recognition, and this emotion recognition then influences emotional experience. In another variation, feedback is not automatic, but could be mediated by one’s own expressive cues being emotionally registered as if they were any other kind of stimulus from the outside. This is grounded in the James–Lang theory of emotion (Lang, 1994), with an emphasis that we undergo cognitive appraisal of our own cues and are not strictly influenced by them automatically. This pathway is grounded also in Bem’s (1967) attribution theory, whereby people can view their own behavior more generally as if from the outside to infer their own internal states.

Continuing the analysis of shared vantage point (SVP) that underlies the affective process theory, it is worth noting that imitated cues is a special case among the 10 mechanisms. This is because it is the only mechanism that actually bypasses the emotional appraisal process—so there is no place for analyzing whether or not partners share a vantage point. As such, contagion via imitated cues should always have high SVP, because there is no chance to deviate from what is presumed to be a default in light of mimicry serving to promote social bonds. Interestingly, van der Schalk et al. (2011) showed that primitive autonomic reactions to outgroup versus ingroup members were simply a quantitatively lesser reaction rather than a qualitatively different reaction. However, for self-reported reactions—which appear to go through the appraisal process and thus involve a vantage point—the ingroup had more convergent contagion and the outgroup had more complementary contagion. Overall, research shows that, instead of the potential for the qualitative shift based on SVP—which distinguishes convergent, divergent, and complementary linkage—in the case of mimicry, the shift is typically quantitative. The same motivational factors described in the previous section on moderators help to determine the magnitude of mimicry. Social closeness appears to produce greater mimicry—including liking (Gump & Kulik, 1997; Likowski, Mühlberger, Seibt, Pauli, & Weyers, 2007) and ingroup membership (Bourgeois & Hess, 2008; van der Schalk et al., 2011). Although the imitated cues mechanism should always have high SVP, due to the lack of opportunity to appraise a vantage point, it is not necessarily convergent in every case. People might experience divergent contagion if they imitate cues that are different from the cues they would personally have used. Notably, individuals across cultures tend to use subtly different dialects to express their emotions (Elfenbein, Beaupré, Lévesque, & Hess, 2007), and imitating cues from a different dialect may produce states that are less convergent—or even divergent—compared with imitating cues that are more familiar. In attempting to understand the association between mimicry and moderating factors, it is worth noting that the relationship between mimicry and social closeness is bidirectional—synchrony itself, even when randomly assigned, promotes rapport (Lakin et al., 2003; van Baaren et al., 2004; Wiltermuth & Heath, 2009).

7. Empathetic-through-recognition (Figure 3d). The empathetic process, as described before, can begin with a stimulus, expressive cues, or postemotional behavior. In the empathetic-through-recognition mechanism, a recipient views the source’s emotional expressions, imagines what the other party must feel, and through this imagination develops into an emotional experience. Given that empathetic processes typically have high SVP, affective linkage is convergent or divergent. However, the source and recipient can have complementary emotions in the low SVP case of empathy errors or a consumption value for the source’s emotion.

8. Behavioral consequences (Figure 4a). In the behavioral consequences mechanism, a source’s postemotional behaviors become emotionally evocative stimuli for the recipient. For example, a disgruntled employee might steal equipment or skip work, and leave a colleague feeling anxious about lacking the tools of their trade or sad they have to miss a child’s hockey game to finish a project. Another employee might be pleasantly surprised by a new development and pay-it-forward to a colleague who later feels gratitude. Whereas emotion recognition uses other people’s nonverbal and verbal expressive cues as stimuli, the behavioral consequences mechanism uses other people’s outward behaviors of any kind. Linkage recipients might not even be aware the source person exists—they are simply responding to a stimulus, and that stimulus is the behavioral trace of another person who acted under an emotional influence. In organizational settings, the behavioral consequences mechanism looms large when people see each other’s work outputs. The quality of work that team members produce can be affected by their enthusiasm, pride, or disengagement, and team leaders or other stakeholders can have favorable or unfavorable affective reactions to the work.

This mechanism bridges the affective linkage literature with extensive work on interpersonal emotion regulation (e.g., Mayer, Salovey, & Caruso, 2000) and emotional labor (e.g., Côté, 2005), whereby people shape their behavior deliberately to influence others’ emotions (e.g., Côté & Hideg, 2011). Emotion regulation and emotional labor can be mechanical, without necessarily resulting from the actor’s authentic affective states, but often it does—particularly when employees use self-regulation strategies such as reappraisal and suppression to force themselves to regulate (e.g., Grandey, 2003). For example, an enthusiastic service provider may try to instill enthusiasm in others, or a nervous police officer might encourage citizens to calm down or to intimidate them into fear. People have an intuitive sense of the pathways in the emotional process model, and they act accordingly to influence others. For example, placing crosses next to road accidents can instill fear in motorists by reminding them of others’ sadness, and thereby encourage safe driving. The behavioral consequence mechanism applies whether or not the interpersonal effects are deliberate—what might appear as emotion regulation can be simply a matter of expressive behavior, for example if the roadside crosses were strictly for remembrance rather than the hope no one else should suffer the same loss. It is important to emphasize that interpersonal emotion regulation can be directed toward selfish or prosocial goals (Niven, Totterdell, & Holman, 2009).

The behavioral consequences mechanism also bridges affective linkage with research on organizational culture, and organizational artifacts in particular. Rafaeli and Vilnai-Yavetz (2004) emphasize the extent of artifacts in organizational life, from dress codes to logos, the appearance of stores, the norms for queues, and even the color of a bus. People interpret these artifacts in terms of their usefulness, aesthetics, and symbolism, and these evaluations can invoke emotional reactions. Any kind of behavioral trace can affectively influence employees, customers, and suppliers. Sometimes these artifacts are deliberate, such as Mary Kay cosmetics’ famous pink Cadillac for top salespeople, but deliberation isn’t necessary.

Given that shared vantage point (SVP) is low in the behavioral consequences mechanism, linkage is complementary. This is because the other person is itself the stimulus, and analysis of their behavior feeds into the recipient’s appraisal process.

9. Behavioral imitation (Figure 4b). In the behavioral imitation mechanism, a recipient imitates the source’s postemotional behaviors, and then the act of engaging in these behaviors becomes emotionally evocative. For example, in a culture of violence, imitating the violence of an angry neighbor cycles back into anger—or could make the imitator feel guilty. As described before, the James–Lang theory of emotion (Lang, 1994) emphasizes how we interpret our own behaviors to determine our emotions—with their famous metaphor of feeling fear after seeing a bear because you realize you are running.

Behavioral imitation technically has low SVP, because the stimulus being appraised is the recipient’s own behavior, not the behavior of the source. That said, there is a greater than usual likelihood that the resulting complementary contagion will produce an emotion cycle with convergent states. Through conditioned responses, the imitation of a person’s behavioral trace of anger has a better-than-average chance of making another person angry, just as imitating a person’s enthusiasm-driven behavior has a better-than-average chance of making another person enthusiastic. Further, similarity is enhanced because the recipient made an active choice to imitate the source’s behavior. However, individuals can also experience noncongruent affect under behavioral interpretation to the extent that they have different self-perceptions when interpreting their own behavior vis-à-vis the emotions that would have caused the source to behave in the way the recipient imitated (Bem, 1967). For example, some people feel silly or ashamed when reacting the same way as the people all around them.

10. Empathetic-through-consequences (Figure 4c). The final mechanism, empathetic-through-consequences, takes place in a two-step process when someone attempts to reconstruct what another person was feeling based on the behavioral traces they leave behind, and then they come to experience an emotion based on what they imagined. Sometimes we ask ourselves: what could have made someone act this way? The answer to that question can provoke an emotion in us. For example, witnessing graffiti could lead to pity when imagining someone’s feelings about lacking opportunities and hope for the future, and witnessing counterproductive work behaviors could lead one to empathize sadly about what makes a colleague so disgruntled. On a positive side, one can vicariously experience pride and hopefulness when witnessing the results of community engagement, such as a public garden, even in the absence of witnessing the instigators or what sparked their inspiration. This mechanism is distinct from behavioral consequences (Mechanism 8), in that it is not about one’s own reaction to the graffiti, but one’s reactions to the feelings that one imagines the graffiti writer experienced. As with the other mechanisms involving empathy, SVP is high and the empathetic-through-consequences mechanism typically produces convergent or divergent linkage, unless there is an empathy mistake or consumption value to imagining the other party’s emotion.

Taking these 10 mechanisms together, APT demonstrates that there is a multiplex set of pathways that can produce a link between the affective states of two or more people. Although this long list of 10 mechanisms can seem unwieldy, what unifies them is one specific moderating factor that integrates past work on separate moderators of emotional contagion—such as social closeness, competition versus cooperation, personal similarity, and ingroup versus outgroup membership. In particular, APT argues for the need to examine the extent to which the linkage source and recipient have a high versus low shared vantage point (SVP) in interpreting the emotionally evocative stimuli in their environment. In the case of high SVP, linkage is convergent when they interpret the stimulus similarly and divergent when their interpretations differ. In the case of low SVP, the linkage source is literally the recipient’s stimulus, and linkage is complementary.

The previous analysis shows that, according to APT, most of these 10 mechanisms will typically produce congruent states. This is because most involve a high likelihood that individuals share the same vantage point. In only few of the mechanisms is the linkage source actually the recipient’s stimulus, notably emotion recognition, recognition-through-consequences, and behavioral consequences. Within high SVP mechanisms, linkage can be convergent or divergent—but convergence becomes the norm because people who share work environments and other social and cultural influences tend over time to converge in their interpretations of the world around them. That said, although emotion recognition is only one of the 10 mechanisms, its ubiquity in everyday life makes it arguably the most frequent source of linkage in organizational settings. As such, complimentary contagion may indeed be the most prevalent, even though past research designs have paid relatively less attention to it. The mechanisms that tend to produce complementary states have often not been labeled as “contagion” in the existing research literature, which exacerbates the lesser scholarly attention. Further, we tend to see congruence in most contagion research due to the focus on positive affective states. Positive affect is relatively less differentiated into distinct emotional categories (Ekman, 2003; but see Sauter, 2010) and, as such, it has less room to reveal complementary influences—for example, anger with fear, guilt with forgiveness, and pride with envy. Because positive affect is a typical reaction when witnessing a positive contagion source, research designs that emphasize positive affect offer less opportunity to distinguish among the three types of contagion.

It is worth noting that the pathways illustrated in Figure 1 are based on our current understanding at a time when new findings are actively expanding. Notably, evidence has accumulated that displaying specific emotional expressions can enhance the recognition of those same emotional states (Neal & Chartrand, 2011; Niedenthal et al., 2005; however, see also Blairy et al., 1999). As additional pathways become established between the emotion processes, additional mechanisms can emerge from the APT.

Reconciling the affective process theory with existing models

The APT brings together into one theoretical umbrella a number of distinct mechanisms that have been treated across the literature, and it is important to reconcile these accounts for the sake of cumulativeness. Hatfield et al.’s (1994) influential treatment focused on primitive mimicry, and other researchers have built on and elaborated this model. For example, Kelly and Barsade (2001) distinguished between primitive mimicry and behavioral entrainment, where the latter involves mutual versus unidirectional synchrony. Additional influential accounts emphasize other mechanisms, notably social comparison, emotional interpretation, and empathy, which are reconciled in what follows.

1. Social comparison and social appraisal. Other than primitive mimicry, social comparison—also called social appraisal—is the most broadly discussed mechanism for emotional contagion (Barsade, 2002; Bartel & Saavedra, 2000; Hareli & Rafaeli, 2008; Parkinson, 2011; Totterdell, 2000). As discussed before, social comparison is the special case under consequences-through-recognition where people assume high shared vantage point and use compatriots’ emotional cues as their stimulus. That is, not knowing how one should feel, one looks for guidance from the cues and behaviors of others who are in the same predicament. These ideas are grounded in Festinger’s (1954) and Schachter and Singer’s (1962) classic work on social comparison. Following from theories of cognitive appraisal, social appraisal allows us to change our interpretation of the environment after backtracking and analyzing other people’s interpretation—so that we register the meaning of an expression actively rather than simply mimicking it (Parkinson, 2011). For example, someone else’s anxiety makes a shared event seem worse than previously realized (Parkinson & Simons, 2009).

2. Emotional interpretation. Hareli and Rafaeli (2008) articulated emotional interpretation as the process of using others’ emotional displays as information, and reacting with whatever emotion is appropriate for the situation. This mechanism maps onto emotion recognition.

Relatedly, Hatfield et al. (1994) discussed a mechanism of conditioned and unconditioned responses to others’ emotions, which map onto consequences-through-recognition. Over time we start to associate the way other people feel with their postemotional behaviors and, consequently, with our own emotions. This sequence becomes overlearned, and we jump from perceiving another person’s expressions to feeling our own resulting emotion. Hatfield et al. (1994) gave an example of abused children being sensitive to their parents’ expressive cues that suggest they may strike. In the unconditioned case, this is a matter of having an implicit understanding of appraisal theory—which might lead to fear in response to others’ anger. In the conditioned case, however, this is a matter of creating new associations based on learning about unique individuals’ behaviors. Likewise, Kelly and Barsade (2001) also discussed vicarious affect as the process of learning from other people’s emotional expressions and, consequently, developing if-then contingencies for understanding the social world—which extends social learning theory (Bandura, 1977, 1986) to the realm of emotion. In the special case when this learning is itself emotionally evocative, vicarious affect is a form of affective linkage: the empathetic-through-cues pathway if appraised from the perspective of the person vicariously observed, and the emotion recognition pathway if appraised with respect to the self.

3. Empathy. Hatfield et al. (1994) and Barsade (2002) discuss the role of empathy as a mechanism for emotional contagion. The APT incorporates the empathy mechanism, and elaborates three different types, based on the visible cue that sparks the empathetic response—that is, witnessing a stimulus, expressive cues, or postemotional behaviors.

Other mechanisms that are not explicitly about affective linkage can also be extended and reconciled within the APT. Hareli and Rafaeli (2008) discussed a concept of drawing inferences—for example, regarding a partner’s social status, competence, and credibility. In some cases, these inferences can themselves be appraised emotionally, in which case it is a matter of the consequences-through-recognition mechanism. This is akin to van Kleef et al.’s (2010) emotions as social information (EASI) model, in which others’ emotions provide usable information, and this information can be—but is not necessarily—emotionally evocative. Another notable example of concepts that can be reconciled within APT are interpersonal emotion regulation and emotional labor, which fit within the behavioral consequences mechanism in the special case when the linkage source was inspired by their own emotions to try to influence the emotions of others. On a related note, Kelly and Barsade (2001) discuss affective impression management and intentional affect induction as processes of using our emotional expressions to influence others.

Testing the affective process theory

There are two main strategies to test the APT. The first is to test its integrating construct of shared vantage point (SVP) and the second is to attempt to test the 10 individual mechanisms. The former is realistic and ripe for new avenues of research, whereas the latter is a tall order that is unlikely to bear great fruit. This is because, as discussed next, the individual mechanisms tend to interact with each other and occur in tandem, in an instant, and largely outside of participants’ awareness. Even so, for theoretical purposes it is worthwhile to distinguish these 10 distinct processes.

Testing the role of SVP has the potential to provide a unifying framework for understanding emotional contagion. The importance of the SVP construct is the primary testable proposition that APT yields, namely that it determines whether linkage is convergent, differential, or complementary.

In observational studies, it should be possible to code for shared vantage point, or to infer it from a situation. In one of the few observational studies of real-time affective linkage in an organizational setting, Scherer and Ceschi (2000) studied airline passengers who lost their baggage. The researchers found that different emotions created different types of linkage between the passengers and airline employees. It is possible to interpret their results by arguing that these emotions sent a signal about the extent to which the passenger and employees were able to see the situation from the same vantage point. On a structural level, the airline employee truly does not share the customer’s vantage point of having lost the luggage, which suggests a default of complementary contagion. Indeed, there were inverse reports of sadness by the two parties. In the case of anger, the two parties converged, but this is also likely to be complementary through an anger spiral, inasmuch as a customer’s anger would likely be an accusation of unfairness directed at the airline. However, in the case of humor, passengers’ and employees’ affect converged, and with humor it is possible to share a laugh about the situation together. Future work should test these kinds of inferences directly.

In experimental studies, it is possible to manipulate a shared vantage point directly. Given that the APT proposes the concept of SVP in order to encompass the moderating factors previously proposed—namely, social closeness, competition, similarity, and group membership—it is important to study SVP using manipulations that are not merely variations of these individual elements and that can be orthogonal to them. Participants could be labeled as having explicitly similar or differing roles with respect to key events taking place in the study. In addition, to test the overarching role of SVP, previously studied moderating factors can be manipulated, and SVP can be measured and tested as a mediator for the influence of these other factors.

In attempting to demonstrate the role of SVP, it is crucial to distinguish convergent from complementary states. As discussed in the previous lines, the frequent focus on positive emotion tends empirically to confound the effects of these distinct types of linkage, in that positive emotion does not have as specific sets of complementary categories. This makes it somewhat challenging to interpret much of the existing evidence in terms of how it fits with APT. For this reason, it is helpful for researchers to expand their work more extensively into negative affect, particularly using discrete emotional states (e.g., Hawk et al., 2011). This will make it clearer when affective linkage is complementary.

In terms of attempting to distinguish the 10 individual mechanisms of affective linkage, this pursuit may be unrealistic and any complete test would be futile. Affective linkage flows through almost any possible setting in organizational life, constantly and through every channel possible. It shines through even the most impoverished forms of interpersonal contact. Cheshin, Rafaeli, and Bos (2011) showed that emotional contagion took place across virtual teams communicating only via text. The likelihood is that multiple mechanisms for linkage occur in tandem, so they can never fully be disentangled. In organizational settings, a complicated event—for example, the announcement of layoffs—can serve as a shared stimulus, can inspire empathetic processing on behalf of colleagues, and can invoke emotions while watching others pack their belongings. Further work might elaborate on the types of occasions when we use certain linkage mechanisms versus others. ² Complex interplay is particularly likely for naturalistic studies of ongoing relationships, for which the effects of emotional convergence have been documented to last over time (Goodman & Shippy, 2002; A. Thompson & Bolger, 1999).

Given the complexity of linkage, the specific mechanism generally remains untested in empirical work (Howard & Gengler, 2001). Few studies have conclusively distinguished between mimicry and social appraisal (cf. Hess & Blairy, 2001; Parkinson & Simons, 2009), let alone distinguishing among 10 mechanisms. The processes involved in linkage occur so implicitly, and—as illustrated in Figure 1 using rounded outlines—flow through internal processing that is outside of access to researchers. Most likely this processing is also largely outside of access to participants themselves, given that people have difficulty reporting about their internal affective processing (Brackett, Rivers, Shiffman, Lerner, & Salovey, 2006; Matthews, Zeidner, & Roberts, 2007). Contagion can occur automatically, without the participant knowing its origin (Blairy et al., 1999; Neumann & Strack, 2000). As such, the evidence suggests that attempts to distinguish mechanisms using self-report may be limited. Another challenge is that linkage processes tend to be fast. Totterdell et al. (1998) found that convergence among colleagues was so quick in a longitudinal design that it always occurred during single time periods. Empirically, even viewing pictures of facial affect at less than 500 ms can be enough to elicit congruent affect (Wild, Erb, & Bartels, 2001). Taken together, there should be pessimism about designing empirical work that can distinguish among all 10 mechanisms.

However, some moderating factors also help to distinguish the individual mechanisms from each other. Some of the mechanisms can be ruled in or ruled out via specific research designs that focus on specific steps in the emotion process. In terms of the stimulus, for example, Totterdell’s (2000) research controlled for shared hassles, which accounts for the three stimulus-based mechanisms. In terms of emotion recognition, Totterdell (2000) also measured perceptions of partners’ moods, thereby controlling for the two recognition-based mechanisms. Emotion recognition can also be influenced by experimental manipulations, notably instructions that inspire more versus less motivation to read other people’s emotional states (Ickes, Gesn, & Graham, 2000). Within the two recognition mechanisms, consequences-through-recognition requires an analysis of the other party’s situation that should be at least partly thwarted by cognitive load. The four mechanisms that begin with cues can be made more influential in designs where emotional cues are deliberately provided (e.g., Barsade, 2002). The imitated cues mechanism can be specifically ruled in or out in certain research designs. Parkinson (2011) measured mimicry and controlled for it, which allowed him to demonstrate that affect linkage occurred above and beyond it. Mimicry can also be prevented with techniques such as facial Botox or stage makeup (Havas, Glenberg, Gutowski, Lucarelli, & Davidson, 2010; Neal & Chartrand, 2011). Another way to rule out mimicry is research on vocal tones, in which quiet participants cannot mimic the senders (Neumann & Strack, 2000)—although work on embodied cognition (e.g., Halberstadt et al., 2009) might suggest participants could engage in facial micromovements that mimic vocal cues without producing sound. The empathetic mechanisms can be highlighted with instructions for perspective taking (Howard & Gengler, 2001), and can also be highlighted with a greater chance for participants to experience the kind of meaningful social interaction that enhances empathy. Cognitive load, likewise, can reduce the influence of empathetic mechanisms—to the extent these require analysis of the other party’s predicament (Davis et al., 1996). Within the empathetic mechanisms, Hawk et al. (2011) distinguished empathy-through-recognition from empathy-through-stimulus. They had participants do the embarrassing task of dancing in front of a camera, and in some cases participants merely knew about the situation, whereas in other cases participants watched embarrassed others doing the same task. Finally, the mechanisms that flow through behavior can be inadvertently thwarted in experimental designs that have strong situations limiting behavioral flexibility.

Some moderating factors can influence several classes of mechanisms, without necessarily implicating a specific one—notably, whether the linkage recipient has any direct exposure to the source. In most mechanisms, there is some kind of direct personal contact with the source, but this is not necessarily the case for shared stimulus, imitated stimulus, or any of the three mechanisms involving postemotional behavior. Indeed, the recipient may not even know that the source exists. As such, researchers can prevent awareness of the other party to document that contagion must take place through one of these mechanisms. Affective linkage without personal contact is increasing in the modern era of mass communication. Interestingly, studying the magnitude of affective linkage may be limited in helping to disentangle its cause, as there is little evidence that its strength differs across the mechanisms (Parkinson, 2011).

Future research directions for affective linkage

Research on affective linkage continues actively, and is moving in multiple directions that are relevant to organizational life.

Practical relevance of affective linkage

Emotional contagion has captivated researchers in organizational behavior for its relevance to the workplace. To the extent that the field has witnessed an affective revolution (Barsade, Brief, & Spataro, 2003), the spread of affect from one person to another looms large. The deliberate linkage of affective states—whether these states are ultimately the same or different—is critical to the study of leadership. Some have gone so far as to say that “put simply, management’s job has become the management of emotion” (Rafaeli & Worline, 2001, p. 107), or that the modern leader is “an emotional manager” (Pescosolido, 2002, p. 584).

Discussions of the social influence of emotion—and particularly emotional contagion—emphasize its theoretical as well as practical relevance to organizational life (for reviews, see Barsade, 2002; Elfenbein, 2007; Totterdell, 2000). The social functional theory of emotion argues that emotion and its expression evolved to solve the problems of group living and serves a crucial role in relationship formation, maintenance, negotiation, and repair (Fischer & Manstead, 2008; Morris & Keltner, 2000). Emotions generate readiness to respond to life’s challenges and opportunities (Frijda, 1986), and in organizational settings this means that the emotions we catch from other people help influence how we complete our tasks.

From the perspective of affective process theory, on a practical level it is important to analyze the extent of shared vantage point among colleagues. Managers can ask: to what extent are people on the same side of the table? What events and environmental conditions do they tend to face together? What are the “hot buttons” to which a lot of people are reacting? These are the seeds of larger scale emotional contagion, which iterates and bubbles up into organization-wide emotional culture (Barsade & O’Neill, 2014). Even when people share the same vantage point, how do they interpret what is going on around them? Do they converge in their interpretation or do factions vary in their reactions? These questions driven by APT speak to the importance for managers to help with sense making and sense giving. Each of the dimensions of emotional appraisal is like a question that we continually ask ourselves—how can managers anticipate these questions and provide authentic information to help people answer them? For example, the feeling that there is a lack of control can create a culture of fear, but managers witnessing fear around them can help to share data in an effort to reassure employees that leaders or the employees themselves are in control. The feeling that events are unfair can create a culture of anger, but managers witnessing anger can help to explain the reasoning behind changes in policies. Even the primary appraisal dimension of whether an event is positive or negative can be managed actively, as managers communicate to what extent an apparent setback could also represent a new opportunity (Jackson & Dutton, 1988).

Low shared vantage points abound in organizations, in that we are constantly working in close quarters and reacting to each other’s cues and behaviors. Managers can ask: how do people tend to influence each other? Ordinary levels of conflict can emerge whenever people try to coordinate their work, but are there more pressing incidents of chronic troublemaking, bullying, and redoing faulty work? Is there instead a culture that encourages the kind of open communication that allows people to resolve their challenges? It is important to remember that we have a lay theory about how our behavior influences other people’s emotions (MacCann & Roberts, 2008; Niven et al., 2009). As such, we do a great deal deliberately to create complementary contagion, with the sense that instilling emotions in others can further our goals on an instrumental basis.

Conclusions: Ten for the road

In introducing the affective process theory (APT), this paper presents the first systematic attempt to examine and enumerate the range of mechanisms by which the affective states of two or more people can be connected. APT integrates the mechanisms for linkage that have received the most attention in the literature—namely, primitive mimicry, social comparison, emotional interpretation, and empathy—and attempts to move from a list to a single theoretical framework for finding, distinguishing, and elaborating these multiple mechanisms. In doing so, it introduces the concept of shared vantage point (SVP), and demonstrates how SVP is the key moderator that distinguishes among convergent, divergent, and complementary linkage processes. When we share the same vantage point in appraising an emotionally evocative stimulus, we experience convergent or divergent contagion—convergent when we interpret the stimulus in the same way and divergent when our idiosyncratic interpretations differ. When we do not share the same vantage point, there is affective linkage when other party is itself the stimulus for us to interpret, and this leads to complementary contagion. Previous moderating factors that have been studied for emotional contagion can be divided into structural factors such as similar job roles, having power differences, or being in competition, as well as motivational factors such as liking, personal similarity, and group membership. These moderating factors help to predict SVP, which is itself what explains the type of affective linkage.

The APT further uses a rule-governed theoretical process that starts with process models of emotion and elaborates 10 distinct mechanisms for emotional linkage. This analysis is intended to return emotional contagion to its original aspiration as a family of phenomena (Hatfield et al., 1994), in spite of the lopsided research attention that has been paid to the single process of primitive mimicry. Pathways that are relatively more mundane are likely to have greater predictive power (Parkinson, 2011), and yet they lack the appealing surprise factor. Opening up the possibility of a multiplex set of explanations for the intriguing phenomenon of affective linkage can assist not only in theorizing about affect in the workplace, but also can connect affective linkage to theories across the breadth of organizational psychology.