Journal of Personality and Social Psychology © 2011 American Psychological Association 2011, Vol. 101, No. 3, 541–556 0022-3514/11/$12.00 DOI: 10.1037/a0023545 Understanding Emotional Transitions: The Interpersonal Consequences of Changing Emotions in Negotiations Allan Filipowicz Sigal Barsade and Shimul Melwani INSEAD, Singapore The Wharton School, University of Pennsylvania Research on the interpersonal functions of emotions has focused primarily on steady-state emotion rather than on emotional transitions, the movement between emotion states. The authors examined the influence of emotional transitions on social interactions and found that emotional transitions led to consistently different outcomes than their corresponding steady-state emotions. Across 2 computer-mediated nego- tiations and a face-to-face negotiation, participants negotiating with partners who displayed a “becoming angry” (happy to angry) emotional transition accepted worse negotiation outcomes yet formed better relational impressions of their partners than participants negotiating with partners who displayed steady-state anger. This relationship was mediated through 2 mechanisms: attributional and emotional contagion processes. The “becoming happy” (angry to happy) emotional transition as compared with steady-state happiness was not significantly related to differences in negotiation outcomes but was significantly related to differences in relational impressions, where perceivers of the “becoming happy” emotional transition gave their partners lower relational impression ratings than perceivers of steady-state happiness. Keywords: emotional transitions, emotions, interpersonal functions of emotions, negotiations, social interactions A rich and growing body of research shows that emotions have here on the largely unexplored phenomenon of emotional transi- interpersonal functions; that is, our own emotions not only influ- tions and how they influence social interactions. ence ourselves but also influence the emotions, thoughts, and An emotional transition is a movement between two or more behaviors of the people with whom we interact (Frijda & Mes- affective, or emotional states. Specifically, we examine whether quita, 1994; Goffman, 1959; Keltner & Haidt, 1999; Morris & the transition to a particular emotion (e.g., the transition from Keltner, 2000; Oatley, 2004). Research in this area has demon- happy to angry, or “becoming angry”) has different outcomes from strated the interpersonal functions and impact of both generalized its steady-state counterpart (e.g., steady-state anger, where one positive and negative affect, as well as a range of discrete emo- begins angry and stays at the same level of anger). In the context tions, such as anger, jealousy, joy, and fear, on various social of interpersonal negotiations, in which individuals communicate interactions and behavioral outcomes (Keltner & Buswell, 1997; about their opposing preferences, we examine whether and how Morris & Keltner, 2000; van Kleef, De Dreu, & Manstead, 2004a). emotional transitions in one’s counterpart influence one’s own To date, work in this area has focused mainly on the influence of negotiation behavior and relational impressions. In three experi- steady-state emotions. However, during social interactions, people ments, we compared both the “becoming angry” (happy to angry) do not necessarily stay in one steady emotional state; they can and transition and the “becoming happy” (angry to happy) transition to do move between emotional states based both on their own intra- their corresponding steady-state emotion displays, steady-state an- psychic changes and in response to environmental stimuli (Frijda, ger and steady-state happiness. 1993; Lazarus, 1991; Thagard & Nerb, 2002). Hence, we focus How Emotional Transitions Influence Social Interactions This article was published Online First May 30, 2011. Past research on the interpersonal outcomes of discrete emotions Allan Filipowicz, Organisational Behaviour Area, INSEAD, Singapore; has found that steady-state emotions influence perceivers’1 behav- Sigal Barsade and Shimul Melwani, The Wharton School, University of ior via both inferential and affective processes. For instance, when Pennsylvania. The study was funded by a research grant from the Alliance Center for confronted with anger, perceivers might cognitively seek to un- Global Research and Development, jointly at INSEAD and The Wharton derstand the cause of the displayer’s anger, as well as actually School, University of Pennsylvania. We gratefully acknowledge the con- experience anger themselves through emotional contagion (Bar- tributions of the Wharton Behavioral Lab, our confederates, Kimberly Rust sade, 2002; Hareli & Rafaeli, 2008; Morris & Keltner, 2000; Van and Caleb Green, Marina Milanova for her help in conducting Study 3, and Kleef, 2009). In keeping with this work and, more broadly, with our video coders. Correspondence concerning this article should be addressed to Allan Filipowicz, INSEAD, 1 Ayer Rajah Avenue, Singapore, 138676. E-mail: 1 We refer to the person expressing the emotion as the displayer and the [email protected] person observing that emotion as the perceiver. 541 542 FILIPOWICZ, BARSADE, AND MELWANI other multiprocess models (e.g., De Dreu, Baas, & Nijstad, 2008; emotional contagion differences that come with emotional transi- E. R. Smith & DeCoster, 2000), we posit that emotional transitions tions versus steady-state emotions are predicted to generate differ- influence perceivers’ behavior through the same inferential and ent attitudinal, emotional, and behavioral responses on the part of affective mechanisms, but with differing consequences than the perceiver. steady-state emotions. Emotional displays signal the displayer’s thoughts and inten- tions about the perceiver (Fridlund, 1994; Morris & Keltner, 2000; Emotional Transitions in a Negotiations Context van Kleef et al., 2004a, 2004b). When faced with the behavior of Because there has been much research examining the social another, including emotional displays, people seek to make sense of this by making inferences or attributions about the cause of that functions of steady-state emotions in a negotiation context, we behavior. Specifically, perceivers may make either dispositional have chosen to study the influence of emotional transitions in this attributions, looking to the displayer’s stable and inherent charac- domain. Negotiations—or social interactions between two parties teristics as cause for the emotional display, or situational attribu- to resolve opposing preferences “with the goal of reaching agree- tions, attributing the displayer’s actions to elements in the ongoing ment” (Carnevale & Pruitt, 1992, p. 531)—are a natural arena for interaction, including the perceivers’ own behaviors (Jones & the observation of affective influences (Barry & Oliver, 1996; Davis, 1966; Kelley & Michela, 1980; Reeder & Brewer, 1979). Carnevale & Isen, 1986). There is some preliminary empirical We propose that perceivers of a steady-state emotion will make evidence indicating that the negotiation setting is one in which increasingly dispositional attributions. In this case, because the transitions can differentially influence outcomes. For example, displayers’ behaviors stay the same over the course of the inter- transitions in a negotiator’s behavior, both from less cooperative to action, perceivers may consequently believe that because they are more cooperative (i.e., offering small, then larger concessions), unable to influence the displayer, the displayer’s emotions must be and vice versa, have been found to elicit higher levels of cooper- the product of a dispositional inclination. For example, when ation from the negotiator’s counterpart (Harford & Solomon, 1967; exposed to steady-state anger, perceivers will be more likely to Hilty & Carnevale, 1993). Furthermore, in Rafaeli and Sutton’s attribute the anger to the displayers’ dispositions, labeling them as (1991) qualitative study, both criminal interrogators and bill col- angry or hostile people (van Kleef et al., 2004a). However, due to lectors were observed using good-cop/bad-cop transition tactics, the very nature of the change in emotion, perceivers of an emo- which included a strong emotional component, in attempts to gain tional transition are predicted to make increasingly situational compliance. attributions. As such, when confronted with an emotional transi- As most research on (steady-state) emotions in negotiations has tion, perceivers are more likely to attribute the displayers’ emotion focused on anger and happiness (e.g., Sinaceur & Tiedens, 2006; change to something in the situation, including their own behavior, van Dijk, van Kleef, Steinel, & van Beest, 2008; van Kleef et al., rather than the displayers’ personality. Thus, a transition to an 2004a, 2004b), we examine transitions between those two emo- emotion can lead to different attributions about the cause of the tions. We hypothesize that the attribution and emotional contagion emotion, and therefore different behavioral responses, than a mechanisms described above will cause perceivers of a “becoming steady-state display of that emotion. angry” transition (from happy to angry) to cede more in the Emotions can also influence perceiver behavior via affective negotiation than perceivers of steady-state anger. processes such as emotional contagion (Hatfield, Cacioppo, & The reason for this, from an attributional perspective, is that Rapson, 1994). Emotional