Abstract
Recognition of the role played by emotions in negotiation is growing. This article synthesizes current research around four broad themes: moves and exchanges, information processing, social interaction, and context. The authors' review reveals that much of the research on this topic has focused on two key emotions, anger and happiness. More recently, negotiators have turned to other emotions such as guilt and disappointment, demonstrating that not all negative emotions have the same consequences, or activate the same regions of the brain. Focusing on social interaction, the authors note that negotiators may influence each others' emotions: whether negotiators converge to anger or happiness has different consequences for agreement. Researchers have broadened their examination of emotion by considering how external factors such as power, the number of negotiators, culture, and gender influence the impact of emotional expression. The authors also consider the function and impact of expressing authentic emotions, or choosing to use emotions strategically to gain an advantage — an issue that raises important ethical questions for negotiators. The article concludes with some practical implications of the research.
Introduction
How emotional expressions influence the course of negotiation and related interactions has become an increasingly popular topic in current research. Negotiation is a form of social exchange that pits the opposing motives of cooperating with and competing against one another. Most negotiators seek to reach an agreement with the other party; they also strive for an agreement that serves their own goals. This dual concern is reflected in a process that consists of both bargaining and problem solving.
Much of the research and practice literature has concentrated on ways to perform these activities effectively. In earlier writing, emotions were viewed largely as factors that impede performance, preventing parties from coordinating successfully. More recently, we have learned that emotions — particularly the expression of emotion within negotiation — can both help and hinder progress. Expressions may convey useful information about preferences; they can also signal dislike or malevolent intentions. Whether emotions move a negotiation forward or backward — or improve/threaten a relationship — depends on a variety of process and context variables. We explore these variables in more depth in this article.
The topic of emotions has been neglected in a literature that has emphasized strategy and information processing. This emphasis has been prominent in several dominant paradigms that have guided much of the research, including game and decision theory, behavioral approaches, cognitive framing/prospect theory, and the dual concern model. The role of such factors as motivation, trust, and identity, which would seem to have substantial emotional content, have been described mostly in terms of their possible impact on strategy. Motives have been construed in terms of their impact on relative and absolute or joint gains (Hopmann 1995; De Dreu, Weingart, and Kwon 2000). Trust has been defined as calculus‐based (anticipating the other person's behavior), knowledge‐based (information‐based predictability of the other person), and identification‐based (recognition of shared values; Lewicki and Stevenson 1997; Irmer and Druckman 2009). And identity has been defined in the context of constituent‐based representative role obligations (Bartunek, Benton, and Keys 1975; Wall 1975; Druckman 1994). Because of this emphasis on cognition, progress in developing theories or frameworks for understanding the role of emotions in negotiation has been slow. But significant growth in interest, as manifested by a spate of recent studies, bodes well for development of this topic.
We have organized this article around the broad themes that to date have guided the research on emotions in negotiation: namely, moves and exchanges (behavior), information processing (cognition), social interaction, and context. The research has also examined the role of negotiation at several levels: individual bargainers, dyads and groups, and the larger settings in which bargaining occurs. Taken together, research on the four perspectives has revealed much of what is known about emotions in and around the negotiation “table.” Finally, we identify gaps in knowledge discussed in a final section as questions raised for further research.
Behavioral Consequences of Emotional Expressions
Most of the research in this particular area has explored the effects of specific emotions — typically anger and happiness — on observed behavior during the negotiation process. The first wave of this research compared how expressions of anger by a negotiator can influence the other party's willingness to make concessions. Using programmed computer messages to convey anger to a negotiator, Gerben Van Kleef and his colleagues (Van Kleef et al. 2004) have consistently demonstrated that the strategic expression of anger elicits larger concessions from opponents. Consistent with the “affect‐as‐information” theories of emotion (Clore, Gasper, and Garvin 2001), a negotiator's greater willingness to make concessions to an angry opponent can be attributed to her belief that angry negotiators have higher limits.
Subsequent research has focused on identifying the boundary conditions for this effect. This research has shown that anger is most effective at eliciting concessions when factors such as negotiators' need for closure or time constraints (Van Kleef et al. 2004) increase pressure to close the deal. These results suggest that anger may be most effective when negotiators are concerned about whether they will reach a deal, and its impact may diminish as the ease of reaching agreement increases. Consistent with this interpretation, Van Kleef and his colleagues (2004) found that angry communications induce fear in negotiators. The link between expressed anger and fear implies that, when negotiators express anger, they increase the other party's anxiety.
While the relationship between expressed anger and concession making is mediated by fear and perceived threat (Nelissen et al. 2011; Sinaceur et al. 2011), it holds only when anger is expressed by a higher power negotiator; when it is expressed by a negotiator with less power, anger elicits reciprocal anger and fewer concessions (Lelieveld et al. 2012)
Expressed anger's impact is affected by whether it is directed at the task or the other negotiator. Roger Fisher and William Ury (1981) exhorted negotiators to focus on the problem, not on the person. This maxim implies that emotions expressed in relation to the task will be more effective than emotions directed at the person. Research indicates that this reasoning holds in relation to anger. When a negotiator expresses anger about the offer he or she has received, that anger elicits more concessions than when he or she directs anger at the counterpart because in the first case, the counterpart interprets the anger as indicative of high limits (Steinel, Van Kleef, and Harinck 2008; Lelieveld et al. 2011). Interestingly, this effect is reversed when negotiators express either happiness or disappointment: both of these emotions, when directed at the person (rather than the offer), elicit more concessions from the other party (Steinel et al. 2008; Lelieveld et al. 2011).
Negotiators may opt to intensify their own displays of anger because they anticipate a difficult or confrontational negotiation. Doing so enables them to elicit more concessions from their opponents (Tamir and Ford 2012). Finally, Fieke Harinck and Gerben Van Kleef (2012) found that expressing anger only has benefits when conflicts concern interests — when disputes concern values, expressions of anger trigger retaliation and escalate conflict.
In a recent review, Van Kleef (2009) and his colleagues summarized the conditions that influence the interpersonal effects of anger. Anger helps achieve favorable outcomes for both parties when:
it is directed at the task rather than the person;
it is viewed by the other as being justified;
the relationship between bargainers is interdependent;
the expression has informational value;
the bargainers take a strategic approach that encourages using the expression as information that can aid coordination, and;
the target of anger has few opportunities to deceive.
As noted above, negotiators may glean strategic information from the other party's verbal and nonverbal emotional expressions. Based on this summary of findings, the authors address the question: When does it pay to be angry? The answer is: when the parties are interdependent, when they use anger expressions strategically, when they direct anger at the issue not at the person, and when the anger is seen as being justified (see also Daly 1991).
A second negative emotion that has received research attention is disappointment. Disappointment, as a discrete emotion, appears to shape the other party's offers because it triggers his guilt (Lelieveld et al. 2011; Nelissen et al. 2011). This finding provides further support for the possibility that, in repeat prisoner's dilemma and ultimatum games as well as in divorce negotiations, guilt encourages higher levels of cooperation (Ketelaar and Au 2003; Wietzker et al. 2012). Recent research has identified a boundary condition to this effect: when expressions of disappointment do not elicit guilt, for example when disappointment is expressed by out‐group members, it leads to lower offers from counterparts (Lelieveld et al. 2013).
Similarly, in ultimatum games, negotiators who feel regretful show more prosocial behavior than those who feel disappointed (Martinez, Zeelenberg, and Rijsman 2011). Taken together, the research findings we have discussed so far suggest that to fully understand the impact of emotions in negotiations, not only do we need to differentiate between emotions with a similar valence (i.e., negative or positive), but we also need to consider whether negotiators are the targets of expressed emotion or are experiencing the emotion themselves.
So far, we have focused on the impact of emotional expressions on obtaining increased concessions from an opponent. Although this research informs us about the ways in which expressing such emotions as anger might affect value claiming, it tells us less about how emotional expressions affect value creation. It seems plausible that an opponent's emotions could provide indirect information about the relative priorities that she has assigned to specific issues, a critical element of the value creation process. Focusing on the relationship between value‐creating (integrative) behaviors and emotions, Davide Pietroni and his colleagues (2008) found that when negotiators display happiness in relation to high‐priority issues and anger in relation to low‐priority issues, value‐creating behaviors increase, but when they display the reverse pattern (anger about high‐priority issues, happiness about low priority‐issues), integrative behaviors decrease. Moreover, Elise Kalokerinos and her colleagues (2013) found that negotiators who suppress happiness at winning are rated more positively and more likely to be viewed as potential friends, in part because the suppression of happiness following victory conveys a desire to protect a counterpart's feelings.
This finding brings us to the role of positive emotions in negotiation. Negotiation researchers have, in recent years, paid considerably less attention to the consequences of positive emotions than to the consequences of negative emotions. Positive emotions, however, facilitate deal making: not only do they increase the likelihood that negotiators will reach a deal, but they also increase the likelihood that negotiators will be willing to interact again in the future (Kopelman, Rosette, and Thompson 2006). Employees who experience positive affect are more willing than those with neutral affect to implement a guaranteed contract but less willing to implement a contingent contract (Mislin, Campagna, and Bottom 2011). These findings suggest that positive emotions may help negotiators to focus on the future, facilitating agreements and strengthening ongoing relationships.
These findings suggest that temporal construal theory may give insight into the link between emotions and implicit time horizons. This theory suggests that viewing situations according to a distant time horizon (i.e., “taking the long view”) is more likely to trigger cooperation and creativity than considering the situation according to a more proximal time horizon (Henderson, Trope, and Carnevale 2006; Trope and Liberman 2010). They suggest that the different consequences of negative and positive emotions may be affected by a shift in negotiators' temporal horizons. Studies of international negotiation have also recognized that shifts in temporal perspective can affect outcomes. More cooperative (and creative) negotiation processes tend to produce more forward‐looking outcomes (Zartman and Kremenyuk 2005; Donohue and Druckman 2009). The question of concomitant emotions, however, remains unexplored in these contexts.
Linking emotion to decision frames, Peter Carnevale (2008) tested the proposition that positive affect can shift negotiators' reference points and reverse the well‐known framing effect. His research found that happy negotiators are more cooperative and make more concessions and more integrative offers to the other party. Importantly, he also reported that positive affect can reverse the framing effect, whereas negotiators who were not primed with emotions (were affect neutral) were more resistant to making concessions when their outcomes were framed as losses, negotiators primed with positive affect were more resistant to making concessions when their outcomes were framed as gains. Carnevale also highlighted the importance of emotion‐as‐information, the idea that negotiators use their own affective states to guide their behavior.
Carnevale's findings raise several interesting questions for future research. The first is: what are the links between affect and corresponding nonverbal behavior, the activation of neural systems, and negotiators' behavior? Although the link between cooperative/competitive strategic choices and neural activity has been explored in the context of prisoner's dilemma games, we do not yet understand how affective and neural systems shape strategic choices in the more ambiguous and uncertain context of negotiation. Daniel Druckman, Demetrios Karis, and Emanuel Donchin (1986) made a start along these lines with a study of the relationships between facial expressions, the P300 EEG waveform (an indicator of surprise), and shifts in concession strategy in a bargaining task.
Cognitive Perspectives
Research on the impact of emotions in negotiation has also focused on the relationship between emotions and information processing (e.g., Clore Gasper, and Garvin 2001). Obtaining information about opponents' priorities and preferences depends on skilled problem solving and judgments of authenticity. The former is a vigorous cognitive activity that can contribute to better, more integrative outcomes (Kressel et al. 1994). The latter involves interpreting the other party's intentions, which have been shown to be influenced by emotional expressions (Baron 1990). Both skills, known as decoding (diagnosing the other's intentions) and encoding (conveying impressions), improve with practice (Thompson 1990).
The need to interpret and respond to the other negotiator's intentions suggests that the strategic bargainer may be using Bayesian probability principles. This can be illustrated with anger. First, the negotiator ascertains whether the anger (or other expression) being conveyed has informational value. Then she asks how often such outbursts have occurred in the past (a priori probabilities). The next step is updating. She may ask whether the other party's expression is justified, whether it is intended to communicate information about limits, whether it is intended to convey information about the importance of the issue being discussed, or whether it indicates an impatience with the process or with herself, the target of the outburst. The answers to each of these questions can be estimated in terms of probabilities. They are the contingent probabilities that contribute to a decision about reciprocating the anger (escalation) or mollifying the other (de‐escalation): the former is a likely reaction to unjustified anger, the latter to justified anger. The bargainer's choice influences the chances that the process will move in the direction of an impasse or an agreement.
These judgments may be formed against a background of greater or lesser certainty. Larissa Tiedens and Susan Linton (2001) differentiated emotions associated with certainty, such as happiness and anger, from those associated with uncertainty, such as surprise, hope, and fear. This distinction involves whether, based on their experienced emotions, individuals are confident that they can (certain) or cannot (uncertain) predict what will happen next. A possible consequence of the distinction between certain and uncertain emotions is that predictions that negotiators make about the other party are influenced not only by whether an emotion is positive or negative but also by the degree of certainty associated with that emotion. For example, happiness — a certain, positive emotion — will create strong optimistic expectations that the other party is skilled, that a settlement is likely, and that cooperation is an appropriate strategy (e.g., Forgas 1998). Conversely, anger — a certain, negative emotion — creates the expectation of a difficult, competitive negotiation, resulting in disinterest and withdrawal (Knapp and Miller 1985; Forgas 1998; Van Kleef, De Dreu, and Manstead 2004). This greater certainty is likely to generate greater confidence about what the other party will do next and decrease negotiators' responsiveness to his actual behavior. We know less about how uncertain emotions influence negotiation, but it seems plausible that as the certainty associated with a specific negotiation decreases, negotiators will become less confident about what the other party will do. This may result in both increased responsiveness and strategic flexibility as well as increased vigilance and scrutiny of the other party's behaviors as negotiators strive to gain greater insight into their opponent's intentions.
Emotions are conveyed not only through speech but also through nonverbal behaviors. Starting with Charles Darwin's (1872) account of the processes of emotional expression in animals and humans, investigators have searched for the way in which different emotions are conveyed through speech and nonverbal behavior, particularly facial expressions. Robert Woodworth's (1938) listing of primary emotions (happiness, surprise, fear, sadness, anger, disgust/contempt, and interest) was the basis for studies designed to isolate expressions corresponding to each state (see Ekman and Friesen 1975). A question asked in many of these studies is whether these are universal emotions or cultural‐specific states (Ekman 1972). For negotiation researchers, an important question is: what is the connection between emotions and intentions?
Research on nonverbal indicators of deception has explored this connection (Ekman and Friesen 1974; DePaulo, Zuckerman, and Rosenthal 1980). Honest and deceptive intentions have been shown to be associated with such emotional states as confidence, stress, and interest (Druckman, Rozelle, and Baxter 1982). Each of these states has been found to be indicated by particular facial (and other bodily) expressions. For example, people who seek to deceive use gestures associated with confidence such as increased head shaking, rocking movements, and crossed hands; those who seek to evade will often indicate their stress by frequently gazing away from the other person; and, finally, honest intentions are often accompanied by expressions of interest or involvement and indicated by frequent leg movements and increased speaking frequency.
These correlational findings point to a connection between emotional states and particular intentions. They also suggest possibilities for diagnosis (decoding) and impression management (encoding) in negotiation and related types of social exchange. Several studies have examined how emotional expressions (both negative and positive emotions) can be used to deceive in negotiation (e.g., O'Connor and Carnevale 1997; Steinel and De Dreu 2004; Olekalns and Smith 2007).
More broadly, neuroimaging research suggests that emotions and cognitions are not distinct. These processes do not differ in kind — rather, decisions are influenced by an interplay between these processes. Both thinking and feeling occur within brain regions whose operations interact (Lindquist et al. 2012). With regard to negotiation, this means that intentional tactics, which may include evading or deceiving, combine elements of thought and feeling in an integrative rather than a sequential, competing, or additive fashion.
Social Interaction Perspectives
Emotional expression also serves important social functions and assists in the coordination of social action. For individuals, emotions facilitate survival; for groups, they facilitate social bonding and collaboration (Morris and Keltner 2000; Shiota, Keltner, and Hertenstein 2004; Keltner, Haidt, and Shiota 2006). Emotions influence interaction processes when negotiators use them to glean social information. Based on this idea, Van Kleef (2009) developed the emotions as a social information (EASI) model. Drawing on research from a variety of areas, he described how expressions influence observers' behavior by triggering inferential processes and/or affective reactions in them. According to Van Kleef's model, inferences and affective reactions are different processes that vary in relative predictive strength depending on both the observer's information processing — for example, witnessing someone expressing sadness to solicit help or happiness to encourage volunteers to contribute to a cause — and on such social‐relational factors as the type of interpersonal relationship, prevailing norms, and the way the emotion is expressed (directed toward the person or the situation).
Emotional expression is thus likely to play a role in the development of relationships between bargainers. Cooperative processes and mutually beneficial outcomes have been shown to improve relationships between parties (e.g., Druckman 1998; Olekalns and Smith 2005). Importantly, the expression of positive emotion has been found to be critical to forming and maintaining social bonds (Shiota, Keltner, and Hertenstein 2004).
Because emotions influence negotiators' behaviors and negotiation outcomes, presumably they would also affect such social outcomes as reputation and the ongoing relationship. For example, we would expect emotions to play a role in the development of trust or mistrust between negotiators. Research on the emotion–trust link has found that the expression of positive emotions builds trust among negotiators, suggesting that it is critical to problem solving (Dunn and Schweitzer 2005). Refining our understanding of this relationship, Meina Liu and Chongwei Wang (2010) found that, whereas expressions of compassion are linked to trust, expressions of anger are linked to distrust. Conversely, Joydeep Srivastava, Francine Espinoza, and Alexander Fedorikhin (2009) found a link between negative emotions, perceived unfairness, and the willingness to retaliate in experimental games.
A further consideration is whether the other party judges expressed emotions (specifically anger) to be strategic or authentic: strategically expressed anger reduces trust and elicits higher demands from opponents, whereas authentic anger increases perceived toughness and elicits lower demands from opponents (Côté, Hideg, and van Kleef 2013). Going in a slightly different direction, Kihwan Kim, Nicole Cundiff, and Suk Bong Choi (2014) found that a negotiator's perceptions of the other party's emotional authenticity is linked to his level of trust in the negotiator, as well as to his willingness to engage in future negotiations.
An open question is whether emotional expression mediates the trust–outcome relationship or whether trust and emotions provide two distinct paths to shaping negotiation outcomes. The interlocking relationships between trust and emotions may become even more important, when, as we discuss in the next section, negotiations are among three or more parties.
Erving Goffman's (1969) analysis of strategic interactions — which in effect bridges the social interaction and cognitive perspectives approaches to analyzing the impact of emotion — differs importantly from the way most research on negotiation and emotions has been conducted. His model, referred to as an expression game, focused on interactive dynamics between the subject (who makes an offer or demand) and the observer (who receives an offer or demand). These roles are separated in much of the research: the focus of analysis is usually on the person receiving information from another or on the person sending information to another as in buyer–seller concession making. Less attention has typically been paid to the interaction process in which observer–subject interactions consist of alternating moves, the one attempting to infer intent from the subject's expressions, while the other attempts to convey certain expressions. In this process, each person is in easy reach of both the observer and subject roles. Reversed roles are a feature of the interaction that occurs when parties attempt to mutually influence each other; the participant's sense of being more the subject or more the observer depends on whether he or she is persuading or analyzing during a particular episode in the process (see also Argyle, Lalljee, and Cook 1968; Pruitt 1995). Interchangeable roles require that bargainers use both skills in the course of negotiation, and analyzing and persuading are both essential for reading and conveying verbal and nonverbal emotional expressions in negotiation. Whether improved skill in interpreting expressions increases tactical proficiency in conveying intentions remains a research issue.
This approach assumes that the dyad or group is the unit of analysis and also takes a different methodological approach from much of the laboratory research in which half the interaction is controlled by the experimenter, for example, messages are often sent from computers or confederates. Field research shows that, in the field, anger and other emotions may play out differently than in laboratory settings. For example, negotiators' outcomes are influenced not just by whether the expressed emotions are positive or negative but also by linguistic patterns that evolve over time and also by the communication method used. In two studies of disputes that were conducted electronically, researchers found that expressions of anger halved the likelihood of settlement and that the reciprocation of anger predicted a failure to resolve the dispute (Friedman et al. 2004; Brett et al. 2007). Positive emotions, on the other hand, had no impact on the likelihood of reaching settlement in a buyer–seller dispute.
In another study, however, Mara Olekalns, Jeanne Brett, and William Donohue (2010) found that, in child custody disputes, the expression of positive emotions by wives, as well as the extent to which husbands “caught” these emotions, shaped outcomes. Agreement was reached when husbands converged to wives' high levels of positive emotion, whereas impasses occurred when husbands converged to wives' low levels of positive emotion. Similarly, hostage negotiations are more likely to conclude successfully when negotiators and hostage takers reciprocate positive affect (Taylor and Thomas 2008). Jointly, these findings suggest that positive and negative emotions may not mark two ends of an emotional continuum. They also suggest that the domain in which negotiations occur is linked to the relative efficacy of expressing positive or negative emotions.
Finally, negotiators may “catch” the emotion of the other party. “Emotional contagion” occurs when individuals experience others' emotion because of a general tendency to mimic and synchronize emotion (Hatfield, Cacioppo and Rapson 1993; Barsade 2002). Early experiments by Peter Carnevale and Alice Isen (1986) and by David Johnson (1971a,b) examined the impacts of socially induced affect on negotiation. Carnevale and Isen found that when positive affect was induced, negotiators used fewer contentious tactics and joint benefits improved. Johnson's studies showed that scripted communications of warmth made the person more likable but did not result in improved outcomes compared to a “cold” (angry) script. The most effective strategy was alternating between negative (acting cold) and positive (acting warm) emotions: negotiators compromised more and evinced a larger change in attitudes when faced with an opponent who alternated between showing anger or warmth than when faced with opponents who were consistently angry or warm throughout the interactions. This finding suggests that there may be an advantage to “fine‐tuning” one's expressions. The sequence of negative and positive expressions may also be important. Concession‐making studies have shown that creating expectations for toughness early coupled with a willingness to compromise later in the process can lead to better outcomes (Druckman, Zechmeister, and Solomon 1972). These findings have some implications with regard to negotiation ethics as well: is it ethical to display inauthentic or deliberately inconsistent emotion in order to induce concessions from the other party? The ethical implications should be considered in concert with further research on emotional displays in negotiation.
These effects may be due to increased trust. Consistent with findings that emotional contagion increases group cohesion and rapport (Sy, Côté, and Saavedra 2005), Roderick Swaab, William Maddux, and Marwan Sinaceur (2011) reported that trust mediated the relationship between mimicking the other negotiator's language and outcomes: linguistic mimicry during the early phases of the negotiation produced better outcomes for the mimicker. In an earlier study, Roderick Swaab and Dick Swaab (2008) found that making eye contact led to higher quality agreements for women but not for men: the visual contact increased comfort for the female negotiators but increased discomfort for the males. Together, these studies suggest that behaviors that increase either perceptions of trust or feelings of comfort lead to better outcomes. The findings also bridge the cognitive and social interaction perspectives of emotion. Verbal and nonverbal behaviors have diagnostic value for the strategic negotiator. He can use them to infer intentions that either facilitate or impede the negotiation process and have consequences on the outcomes. Negotiators may thus be able to strategically induce emotions in others, potentially enhancing their outcomes through the expression of emotion. (For a review of the more general literature on socially induced affect, see Druckman and Bjork 1994, chapter 10.)
Contextual Influences on Emotion
Many of the laboratory studies have failed to examine the ways in which emotional expressions are shaped by the contexts in which they are displayed. A contextualized view of emotions would complement the process view preferred by many negotiation researchers. Van Kleef's EASI Model is helpful in this regard because it provides a framework for research on the interpersonal effects of emotions, thus also providing a link between the cognitive and social interaction perspectives on emotion. By including such moderating variables as power, time pressure, and display rules in the model, Van Kleef has also incorporated contextual variables into the framework. Studies on conflict and negotiation have supported the model (e.g., Van Kleef, De Dreu, and Manstead 2006). Further, the meta‐analytic review conducted by Kristen Lindquist and her colleagues (2012) shows that brain states evoked by emotional expressions are sensitive to context. Different brain states occur when the same emotion is elicited under different circumstances. The corresponding brain state depends upon labeling the emotion as, for example, fear or anger and the social setting as a physical or social context.
Power is frequently a salient contextual variable in negotiations and thus provides an important lens through which to view expressions of anger. Social cognition research suggests that, because low power negotiators process information systematically (in greater detail) and are more attuned to the social consequences of their actions, they are likely to be more responsive to emotion displays than high power negotiators. Low power negotiators also concede more to others who express anger and also claim less value from angry opponents (Sinaceur and Tiedens 2006; Van Kleef, De Dreu, and Manstead 2006; Van Kleef and Côté 2007; Butt and Choi 2009). High power negotiators are also influenced by anger, but they respond both to their own and to the other party's emotional state. Responding to their own anger, high power negotiators are energized: they report feeling more focused and assertive and claim more value in the negotiation (Overbeck, Neale, and Govan 2010). In response to the other party's anger, they increase their demands when they believe that anger is unjustified (Van Kleef and Côté 2007). Finally, powerful negotiators set the emotional tone for a negotiation: their positive affect significantly influences the level of trust in the negotiation (Anderson and Thompson 2004).
Research on electronic communication sheds further light on the expression of emotion in negotiations and disputes (Hine et al. 2009). When negotiations occur electronically, they are more likely to be successful if negotiators express positive emotions and agreeableness. Critically, although expressions of agreeableness at any time during an electronic negotiation facilitate settlement, negative emotions affect success only when they are expressed in the second half of the negotiation (Hine et al. 2009). Studies of dispute resolution reported similar results. Two studies investigating eBay (the online auction service) disputes show that the expression of negative emotions and anger can delay and may prevent settlement (Friedman et al. 2004; Brett et al. 2007). This line of research highlights the importance of emotional tone as well as timing. When emotions are expressed and the extent to which others then converge to those emotions contributes to their impact on the outcome.
Because the expression of emotions is socially determined, we might also expect that the impact of emotional expression varies across cultures. For example, the emphasis on harmony and preserving face in Asian cultures suggests that negotiators from these cultures might be more reluctant to express negative emotions. Following this line of thought, Hajo Adam, Aiwa Shirako, and William Maddux (2010) found that expressions of anger elicit larger concessions from European American negotiators but smaller concessions from Asian and Asian American negotiators. Subsequent research by two of these authors also found that anger conveys greater toughness and threat when expressed by East Asians than when expressed by European Americans, perhaps because they occur less commonly in East Asian cultures (Adam and Shirako 2013). Expressions of anger also affect what negotiators do next: Chinese negotiators are more likely than American negotiators to respond to anger by increasing their use of persuasive arguments (Liu 2009).
Shirli Kopelman and Ashleigh Rosette (2008) have offered additional insight into the impact of culture‐based norms in their study exploring the issue of culture‐specific relationships between accepting offers and accompanying emotional expressions. They found that Asian negotiators were more likely to accept ultimatum offers made in the context of positive emotions than those made in the context of negative emotions; however, Israeli negotiators were indifferent as to whether an offer was accompanied by expressions of positive or negative emotions.
Like culture, gender could also determine which emotional expressions are seen as appropriate and consequently affect the impact of those emotions. In general, women are expected to both experience and express a greater range of emotions than men. The two exceptions to this general expectation relate to expressions of anger and pride, both of which are seen as more typical of men than women (Plant et al. 2000). Consistent with this view, men who express anger are more likely to obtain positive organizational outcomes than women who express anger (Gibson et al. 2009). Acting against expectation in general can have consequences: negotiators who switch strategies and thereby violate the expectations of the other negotiator influence that party's mood (Barry and Oliver 1996; Olekalns and Smith 2005).
The organizations to which they belong may also influence negotiators' emotional expressions in negotiation. Stephen Fineman looked in this direction, asking “In what ways do decisions unfold over time as a function of the way people feel, and change their feelings — about themselves, their projects and significant others? How, for example does anxiety, suspicion, love, and hate take decision making through various paths towards particular outcomes?” (1993: 217). The extensive literature on negotiation that occurs within and between organizations, and on boundary roles in particular (Walton and McKersie 1965; Adams 1976; Burke and Biggert 1997), has been more concerned with strategies than with emotional expressions.
Context also changes when individuals move from dyadic to multiparty negotiations. As is the case in dyadic negotiations, expressions of anger and negative emotions within multiparty negotiations decrease agreement. More importantly, negotiators who express anger are likely to be excluded from coalitions and hence achieve worse outcomes (van Beest, Van Kleef, and Van Dijk 2008; Huffaker, Swaab, and Diermeier 2011). However, if negotiators do form an alliance with an angry player then, as is the case in dyadic negotiations, angry negotiators obtain large concessions. Recently, Ilja van Beest and Daan Scheepers (2013) refined our understanding of the role played by anger in coalition formation: negotiators who receive an angry message from a preferred coalition partner show a cardiovascular response consistent with challenge, whereas those who received an angry message from a non‐preferred coalition partner showed a cardiovascular response consistent with threat. An interesting twist is that challenged, but not threatened, negotiators are more likely to find a new coalition partner.
Other studies have found that building perceived similarity, through linguistic convergence, increases agreement between coalition partners (Huffaker et al. 2011). When integrated with the positive emotion–trust link we described earlier, these findings suggest that positive emotions might strengthen alliances and enable negotiators to improve their outcomes (see also, Olekalns and Smith 2009). Our understanding of the role of emotion in multiparty negotiations is in its infancy. Because emotional expression can shape coalition formation and agreement, research designed to develop a better understanding of the role that emotional expressions play in multiparty negotiations would be useful.
Finally, recent research has investigated the longer‐term consequences of emotional expression. In their study, Gerben Van Kleef and Carsten De Dreu (2010) explored the long‐term consequences of expressing anger in a negotiation. They contrasted a spillover model, which would suggest that the target of anger would demand less in a subsequent negotiation with the same party, with a retaliation model, which suggests that a target would demand more in a subsequent negotiation. They found support for the spillover model: negotiators demanded less when they negotiated a second time with the same angry negotiators because they perceived that negotiator to be tough. Taking the idea of timing in a slightly different direction, Alan Filipowicz, Sigal Barsade, and Shimul Melwani (2011) compared the impact of expressing emotions consistently throughout a negotiation to the impact of making emotional transitions. They found that, compared to negotiators who are consistent in their emotional expressions, negotiators who “become angry” obtain better outcomes and also conveyed a more positive impression than negotiators who “become happy.” Similarly, Marwan Sinaceur and his colleagues (2013) reported that emotional inconsistency elicits greater concessions than emotional consistency.
New Directions for Research
To date, negotiation researchers have focused on a relatively narrow range of behaviors. Moreover, these researchers have not connected their research to the complex models of emotion that have been developed in other domains, nor have they sampled systematically across different dimensions of emotion. Negotiation researchers continue to concentrate their work primarily on two emotions, happiness, and anger. A few studies have examined the impacts of surprise on bargaining moves (e.g., Druckman, Karis, and Donchin 1986) and on the arousal of guilt when outcomes clearly favor oneself (Hegtvedt and Killian 1999). These and other emotional states have been studied in psychology more generally. One well‐known model, the circumplex model of affect (Russell 1980), differentiates emotions based on their valence (positive or negative) as well as their arousal level (active or passive). We encourage researchers to more systematically assess the relationship of the affect circumplex to negotiators' behaviors and outcomes.
More recent research has also shown strong links between emotions and the activation of different regions of the brain. The effects of mimicking generally have been shown to be associated with mirror neurons, which fire either when an individual acts or observes an action (Wicker et al. 2003; van der Gaad, Minderaa, and Keysers 2007). Similarly, emotions from different quadrants of the affect circumplex trigger activity in different regions of the brain: research shows that distinct brain regions activate depending on the valence and arousal of a specific emotion (Posner et al. 2009; Colibazzi et al. 2010). More recently, de Dreu and his coworkers have linked oxytocin to in‐group trust and cooperation in the face of intergroup conflict (De Dreu et al. 2010). These findings, which show different brain regions are activated by different emotions, underscore the importance of sampling emotions around the affect circumplex because they imply that emotions with similar valences may evoke distinctly different reactions. They further underscore the need to better understand the neurophysiological factors that drive or are influenced by our emotional experiences and reactions to others. (See Hyde et al. 2009 for findings on the influence of experience on structural brain changes.)
Related research has examined additional physiological aspects of negotiation. Two studies have found heightened arousal, which can be triggered by a fit between negotiators' agreeableness and the negotiating context, leads to better economic outcomes especially if negotiators feel positive about the negotiation (Dimotakis, Conlon, and Ilies 2012; Brown and Curhan 2013). Returning to our discussion of mimicry, research has found that negotiators who suppress their own emotional expressions are less sensitive to facial expressions of emotion in their counterparts than are negotiators who engage in mimicry (Schneider, Hempel, and Lynch 2013). This finding highlights a dilemma for negotiators who, on the one hand, may benefit from concealing their emotions yet, on the other hand, incur the costs of failing to recognize opponents' emotional reactions.
Many of the studies we reviewed were conducted in laboratories. An advantage of the laboratory controls is that the direction of influence — from the computer/actor to the subject — is clear. Causal inferences can be made with confidence. A disadvantage is that interactive dynamics are ignored, and the results are consequently less relevant to real‐world negotiations. Much of the research to date has made this trade‐off, favoring internal over external validity. A better balance between the two validities would materialize if researchers took on the challenge of field research. The interaction dynamics described by field researchers would complement the causal patterns inferred from the laboratory studies. Mood changes during the course of an interaction are fleeting and the multiple meanings of expressions can be vaguely defined — with such challenges, there is little wonder that the research to date has progressed slowly. Emotions are not easy to investigate.
Research questions that cross the four perspectives are particularly interesting. Examples include the following:
How are concessions (behavioral perspective) influenced by attributions of the other negotiator's intentions (cognitive perspective)?
How are interaction dynamics (social interaction perspective) shaped by the connection between decoding/interpreting and encoding/conveying expressions (cognitive perspective)? To what extent does Van Kleef's EASI model contribute to understanding this connection?
Which emotional expressions — and the corresponding negotiating behavior — are more or less influenced by such contextual variables as culture and gender?
And, what role do emotional expressions play in the development of long‐term relationships between negotiators? How is this connection — between emotions and social relationships — mediated by contextual variables?
Emotions in Practice
Research findings on emotions in negotiation have implications for practice, which will vary according to whether the negotiator's primary concern is strategic or relational. The former is a competitive or value‐claiming approach intended to secure better deals. The latter is a cooperative or value‐creating approach intended to enhance relationships and perhaps the durability of agreements. The research literature has explored the effectiveness of both these approaches to negotiation.
Relevant research, reviewed above, suggests that strategy is influenced by the timing of, and perceived authenticity of, expressions of anger as well as by the parties' interdependence and their relative power. The consistency of emotional expressions — for example, inconsistency or flexibility may elicit more concessions — is also relevant.
With regard to the relational approach, findings about the expression of positive emotions are relevant. Positive emotions encourage cooperation, a long‐term perspective, and value creation, particularly when displayed in relation to high‐priority issues. But anger and flattery can also bolster relationships when they are authentic and convey useful information about priorities. Negotiation training curricula should address how to convey anger to gain strategic advantage and happiness to improve relationships. What are the conditions under which the influence of these emotional expressions can be maximized?
A key to implementing either approach is the development of diagnostic and impression management skills, referred to also as decoding and encoding. It is important to distinguish between genuine and feigned emotional expressions. It is also important to convey the intended emotional expression, through both verbal and nonverbal channels, to obtain the desired impact on the negotiation process.
These skills improve with practice as illustrated by a role‐play exercise we designed for use in classes and workshops.1 The exercise is based on a historical event that occurred in 2001 when a U.S. Navy surveillance plane collided with a Chinese fighter jet about 70 miles off the Chinese island of Hainan. The two countries had different interpretations of this event: the Chinese government demanded an apology; the U.S. government refused to apologize for an incident that was not its fault. Those interpretations are the basis for a negotiation between a diplomat representing China (Mr([s]) Lin) and a counterpart representing the United States (Mr[s] Smith). The two role‐playing diplomats are provided with background on the crisis and briefing notes from their respective embassies. A thirty‐minute negotiation follows in two parts. The first part primes role players to express either genuine or false expressions of anger or flattery. The second part reverses that priming, providing students with experience in conveying the same emotions in both modes.
The encoding feature of the exercise places students in the role of actor. Improved encoding performance has been shown to increase the other negotiator's respect. The decoding feature provides an opportunity to detect intentions as false or genuine. Improved detection performance has been shown to increase the other negotiator's trust.
The need to move, often rather quickly, between these roles as the interaction unfolds creates an additional challenge. Goffman's (1969) idea of an expression game in which each role is in easy reach of the other captured this dynamic: role players learn how to alternate roles as receivers of information and as conveyors of impressions. In our simulation, the lessons are crystallized in debriefings in which students gauge their progress in developing these skills. Questions asked include: how accurately did they detect the other's intentions as genuine or false? How successful were they in disguising or revealing their own intentions? Were these challenges easier with expressions of anger or flattery? When agreements were reached, we probed whether the order of intention (genuine followed by false or vice versa) or type of emotion was the more important influence. These questions can also take the form of hypotheses to be explored in the laboratory, illustrating the synergistic relationship between research on emotions and the practice of conveying or detecting them.
Ethical issues are raised as well during the debriefings. The matter of manipulating emotions to achieve a desired response or concession is of particular concern. Although consonant with a strategic approach to negotiation, manipulation is a form of lying. By emphasizing the importance of detection, the exercise helps students develop sensitivity to the distinction between genuine and feigned expressions. Accurate detection — pointing out the inauthenticity of the expression — can alter the negotiating atmosphere. On the one hand, being caught in a lie could further deteriorate the relationship between negotiators, leading to impasses. On the other hand, it can discourage further manipulation on the road to agreement. Which direction the talks take following the revelation may depend on the effectiveness of the way a detector manages the communication. These encoding skills are also practiced during the exercise.
Conclusion
An examination of the research findings that we have discussed, taken together, suggests interesting sets of issues for continuing research within each perspective on emotions in negotiation. First, we can look forward to studies that examine other emotions such as surprise when expectations are disappointed, sadness when alternatives are unattractive and dependency increases, shame when face is lost, and interest or involvement when the stakes increase. Second, the interplay between emotions and cognitions presents an interesting array of research challenges: for example, the way that bargainers use affective information to develop or change strategies. Third, process dynamics call for further investigation. In particular, interactive processes such as those described by the expression‐game paradigm need to be better understood. A related issue is the coordination of emotional expressions: When does matching perpetuate impasses and when does it resolve them? Fourth, there is much yet to be learned about the contexts for emotional expression. Included in these contexts are the number of parties, organizational norms, and cultures. A question of interest is how these contexts shape the way emotions are expressed and read. The idea of cultural display rules may be relevant also for organizations and the institutional contexts within which negotiation occurs. These issues also have implications for practice and should be part of an ongoing agenda that aims to bridge research with training and practice.
NOTE
This exercise is available and can be obtained by writing to the authors.