Running Head: MORAL CONFLICT AND COMPLEXITY

Moral conflicts pose a serious challenge to the field of conflict resolution. When they turn destructive, they can pull thoughts, feelings, and actions toward a narrow, negative state that becomes self-organizing and self-perpetuating, thus trapping the dynamics of the conflict. On the other hand, constructive conflict dynamics over difficult issues are thought to be characterized by more complex movement between different states of thinking, feeling and acting. Originating from a new dynamical-systems theory of intractable conflict, this research offers important insights into the basic dynamics of destructive versus constructive moral conflicts. Fifty-nine dyads were asked to reach consensus on a polarizing topic on which their opinions differed. Results from the study provide strong support for the basic propositions of the theory. Moral Conflict and Complexity 3 Moral Conflict and Complexity: The Dynamics of Constructive versus Destructive Discussions over Polarizing Issues . “You can't shake hands with a clenched fist.” Indira Gandhi Conflicts over morally-laden issues such as the current war in Iraq, the Israeli-Arab conflict, abortion rights, and the death penalty tend to be divisive, emotionally draining, and have the potential to quickly move down a destructive path and become intractable. When such conflicts intensify, they seem to generate a strong ‘pull’ that captures the parties’ emotions, thoughts and actions and launches them into a polarized destructive dynamic, often against their initial intentions. Gaining systematic understanding of these dynamics, and of the conditions that can help to determine more constructive versus more destructive, intractable dynamics in moral conflicts is vital to both scholars and practitioners alike. Kriesberg (2005) has stressed three aspects that distinguish more intractable from more tractable conflicts: their high levels of persistence, destructiveness, and resistance to resolution. But is there a way to understand these conflicts more fundamentally, beyond mere description? We suggest that such an understanding can be achieved through investigating basic differences in the attractors or stable patterns of such conflicts as they evolve over time. The study presented here tests the basic propositions from a new dynamical-systems model of intractable conflict, which characterizes such conflicts as stable attractors for malignant social relations (Coleman, Bui-Wrzosinska, Vallacher, & Nowak, 2006; Coleman, Vallacher, Nowak, & Bui-Wrzosinska, 2007; Vallacher, Nowak, Bui-Wrzosinska, & Coleman, 2006). It addresses two basic questions: 1) What are the underlying dynamics that distinguish constructive versus destructive responses to difficult moral conflicts, and 2) What basic parameters determine qualitative differences in these dynamics over time? This paper also presents a set of methodological innovations for conflict research which constitute our Difficult Conflict Laboratory; a venue for measuring the emotional, cognitive, and behavioral dynamics of disputants engaged in difficult moral conflicts, over time. Moral Conflict and Complexity 4 This paper has 5 sections: 1) a brief overview of the dynamical-systems approach to the study of intractable conflicts, 2) a discussion of the main parameters in our study, 3) a description of the methods employed or developed for the research, 4) a presentation of our findings, and 5) a discussion of the implications of this study for future research on conflict. Intractable Conflicts as Attractors: The collapse of complexity Conflicts between people and between groups are typically characterized by both incompatibilities and compatibilities on a variety of different issues, but this does not necessarily result in destructive dynamics and intractability. To the contrary, the complexity and nuance of relationships with both common and conflicting goals most often prevents the progression toward intractability and enhances the likelihood of constructive resolution. Here, each party may lose on some issues but gain on others. However, we suggest that it is the collapse of such relational and multi-issue complexity that promotes conflict intractability. When conflicts escalate and evolve, previously unrelated issues can become linked and mutually interdependent, diminishing the likelihood of finding a solution that satisfies all the concerns. That’s because the activation of issue A triggers all the others to which it is linked. For example, if a small slight occurs between neighbors with a history of conflict, it is likely to bring to mind all the past provocations, injustices, and conflicts of interest that had occurred between them. Thus, the parties are likely to respond disproportionately to the magnitude of the triggering event. Even if the instigating issue is resolved, the activation of other issues can serve to maintain or even intensify the conflict. A reduction in complexity of the perception of conflict issues has been identified in psychological research on escalation. Factors such as stress, anxiety, and emotional intensity can impair cognitive processes, promoting a simplified and polarized view of an otherwise complex and nuanced situation (Conway, Suedfeld, & Tetlock, 2001; Pruitt & Kim, 2004; Winter, 2007). Under heightened threat to one’s safety, for example, people’s cognitive processes tend to promote overly simplistic, rigid, black-and-white perceptions, thoughts and Moral Conflict and Complexity 5 judgments (Osgood, 1983). When in conflict, a high level of political thinking (Rosenberg, 1988)—which correlates with cognitive complexity, integrative complexity, tolerance for ambiguity, and moral development—is associated with reduced vulnerability to the influence of emotions on destructive orientations toward conflict (Conway, et al., 2001; Golec & Federico, 2004). We suggest that there is a clear link between the loss of issue complexity and the development of strong attractors for destructive conflict. In generic terms, an attractor refers to a subset of potential states or patterns of change to which a system’s behavior converges over time (e.g., Schuster, 1984). In a system governed by attractor dynamics, even very different starting states tend to evolve toward the subset of states defining the attractor. For instance, social relations are typically complex and multi-dimensional, with various mechanisms operating at different points in time, in different contexts, with respect to different issues, and often in a compensatory manner. The alignment of distinct relational elements into a single dimension (i.e. the most central and salient conflict issue), establishes positive feedback loops such that the issues have a mutually reinforcing rather than a compensatory relationship. This would be particularly likely with moral conflicts, which often involve a clash of incompatible worldviews that can be central and defining for individuals and groups (Pearce & Littlejohn, 1997). Here, all events that are open to interpretation are ultimately construed in a consistent fashion and promote coherent patterns of thought, emotion and behavior regarding other people in the conflict. The common state toward which diverse thoughts and behaviors converge represents a fixed-point attractor for the system. Even an unambiguous event that runs counter to the attractor can over time be assimilated to the attractor. A peaceful overture by a member of the outgroup, for instance, may be seen as insincere or as a trick if there is strong sense of antagonism. Thus, attractors represent particularly strong and coherent patterns of thinking, feeling, and acting when in conflict. Attractors in essence attract; pulling thoughts, feelings, actions, Moral Conflict and Complexity 6 norms, and even policies and institutions into a state that can become self-organizing and selfperpetuating (Coleman, 2006; Coleman et al., 2007). In conflict, this is seen when the psychosocial realities and interactions of disputants collapse into destructive, coherent us-versusthem dynamics that are largely self-organizing and thus resistant to change. If a strong, destructive fixed-point attractor captures the state of an interpersonal or social system during conflict, then any deviation from this attractor will result in the system activating its mechanisms to return to the attractor. This is particularly likely if the relational system lacks or has lost any attractors for neutral or positive forms of interaction. So while the severity of conflict may be related to the intensity or amount of violence between people or groups, the intractability of conflict may be defined in terms of a dominant fixed-point attractor for negativity. Other scholars have found similar attractor patterns with other types of difficult social dynamics. Gottman’s (Gottman, Murray, Swanson, Tyson, & Swanson, 2002) research on marriage and divorce has found consistently that the relational phase-space of couples whose relationships end in divorce tend to be characterized by relatively strong fixed-point attractors for negative emotions and relatively weak attractors for positive emotions. Similarly, Losada (Losada, 1999; Losada & Heaphy, 2004) has found that the emotional and behavioral patterns of low-functioning business strategy groups are best characterized as fixed-point attractors, whereas high-functioning groups tend to display more complex, chaotic patterns of emotions and behaviors. The current research builds on prior theoretical work on conflict and attractors (Coleman et al., 2006; Coleman et al., 2007; Vallacher et al., 2006) and on the research of Gottman and Losada to test the relationship between intractable and tractable moral conflicts and fixed-point versus high-complexity attractors, respectively. Model Parameters Research by Gottman (Gottman et al., 2002), Losada (Losada, 1999; Losada & Heaphy, 2004) and others (Conway et al., 2001; McClintock 1977; Tetlock, 1991) has begun Moral Conflict and Complexity 7 to identify a set of basic parameters that seem to distinguish between fixed-point versus more complex types of psycho-social attractors in social relations. They include the ratio of positivity-to-negativity of emotions (Bales, 1950; Gottman et al., 2002), levels of integrative complexity (Conway et al., 2001; Tetlock, 1991), the balance between internallyand externally-focused references in groups (Losada, 1999; Losada & Heaphy, 2004), and the balance between inquiry acts (questioning) and advocacy acts (positioning) in groups (Losada, 1999; Losada & Heaphy, 2004). Below, we elaborate on these and other parameters included in this study. Emotional Dynamics Previous research has identified the central importance of the ratio between positive and negative emotions over time in predicting problems in social relations. Positive emotions have been associated with broader and more complex thinking and action tendencies (Fedrickson & Branigan, 2001; Losada, 1999). Experimental laboratory research has found that positive emotions broaden a person’s momentary thought-action repertoire, whereas negative emotions narrow the repertoire. Also, positive emotions have been found to lessen and downgrade the lingering effect of negative emotions. (Fedrickson, 2001, Fredrickson & Branigan, 2005). On the other hand, more negative emotions have been found repeatedly to be associated with destructive conflict dynamics such as ineffective communication, aggression, and violence (c.f. Gurr 1970, e.g. Coleman, Goldman & Kugler, in press; Lindner, 2001, 2002; Margalit, 2002, Barki & Hartwick, 2004; Jehn & Benderesky, 2003; Yang & Mossholder, 2004). Gottman’s (2002) research on marriage and divorce has found consistently that the patterns of couples whose relationships end in divorce (or a stable state of misery) tend to be characterized by relatively strong fixed-point attractors for negative emotions and relatively weak attractors for positive emotions. Similarly, Losada (Losada 1999; Losada & Heaphy, 2004) has found that the emotional patterns of low-functioning business strategy groups are Moral Conflict and Complexity 8 best characterized as fixed-point attractors, whereas high-functioning groups tend to display more complex, chaotic patterns of emotions over time. Both researchers found that higher levels of responsiveness to other people’s emotions was connected to more positive and complex patterns and more constructive and innovative processes. Similarly, Mayer and Salovey (1997) state that emotionally complex people among other things, evidence more interpersonal adaptability and more empathic tendencies. We propose that the dynamical emotional patterns of disputants engaged in discussions over difficult moral issues which allow for more positive emotions in conflict in addition to negative emotions will lead to more constructive conflict processes and outcomes. Hypothesis 1a: More destructive moral conflict discussions will evidence dynamics that are emotionality restricted to negativity (high-degrees of negativity, low-degrees of positivity), whereas more constructive moral conflict discussions will evidence patterns which are more complex and balanced (more positivity or more balanced degrees of negativity and positivity). Hypothesis 1b: More constructive moral conflict discussions will evidence more emotional responsiveness and adaptation to the other party’s emotional experiences. Cognitive complexity. Following Steufert & Steufert (1978), cognitive complexity refers to the degree to which a potentially multidimensional cognitive space is differentiated and integrated, often in an iterative fashion. A person in a state of high cognitive complexity would employ differentiation and integration in his or her information processing – in other words this person would tend to both differentiate situations and events in terms of their various aspects and from different points-of-view, and be able to integrate this information into a more global level of understanding. These complimentary processes are central to cognitive complexity as 1 A diverse area of research is devoted to the study of cognitive complexity. In this paper we use the term cognitive complexity to include related concepts like integrative and conceptual complexity (c.f. Suedfeld, et al., 1992). Moral Conflict and Complexity 9 emphasized by many scholars in the area (Koo, Han, & Kim, 2002; Lee & Truex, 2000; Suedfeld & Coren, 1992; Suedfeld & Leighton, 2002; Suedfeld, Tetlock, & Streufert, 1992; Woike & Matic, 2004). Therefore a complex understanding of a moral conflict would be characterized by the capacity to see different aspects and points of views on the issues and to integrate them into an informed opinion. This would be in direct contrast to more automatic, simplistic, black-and-white thinking. Cognitive complexity, not being a simple function of mental ability, is subject to change depending on the situation and general environment in which the person is functioning (Suedfeld & Coren, 1992). In a variety of studies across different contexts, scholars have found that more aggressive and competitive acts are associated with lower levels of cognitive complexity. (Conway et al., 2001; Koo, Han, & Kim, 2002; Suedfeld & Leighton, 2002, Winter, 2007). These findings lead to the following hypotheses: Hypothesis 2: More destructive discussions over moral conflicts will evidence more restricted and limited cognitive exploration of the problem space of conflicts (lowdegrees of integrative complexity), whereas more constructive discussions will evidence more complex patterns (high-degrees of integrative complexity). Learning Several scholars have has suggested that learning is critical to more constructive forms of conflict resolution (c.f. Bar-Siman-Tov, 2003; Kegan, 1994). Levy (1994) defines learning as “a change of beliefs (or the degree of confidence in one’s beliefs) or the development of new beliefs, skills, or procedures as the result of the observation and interpretation of experience. (1994, p. 283). It generally involves a change in one’s mode of thinking, beliefs, or values (Tetlock, 1991). In the discourse of moral conflict, shifts in peoples basic understanding of the issues, history, and parties is paramount (Pearce & Littlejohn, 1997). This leads to our third hypothesis: Moral Conflict and Complexity 10 Hypothesis 3: More destructive discussions over moral conflicts will evidence less learning by the parties than more constructive discussions. Behavioral dynamics A good deal of research on negotiations in conflict has emphasized the important role of social motives in determining more constructive versus destructive processes and outcomes (c.f. De Dreu, Beersma, Steinel &Van; Kleef, 2007; De Dreu, Weingart & Kwon, 2000). This approach is based on the work of McClintock (1977; Messick & McClintock, 1968), but closely resembles to two related theories of conflict: the theory of cooperation and competition (Deutsch, 1949, 1973); and the dual concern model (Pruitt & Rubin, 1986). Negotiation research on social motives typically involves the distinction between proself and prosocial motivations and their associated behaviors. Proself motives are those in which the negotiation is seen as a competition and in which individualistic goals dominate. Prosocial motives include concern for others and typically results in more collaborative processes with the goal of mutual satisfaction. Research suggests that prosocial motivation is associated with less selfish and contentious behaviors and the achievement of higher joint outcomes than proself, which evidence more other-oriented cooperative behaviors (see De Dreu, et. al., 2007). Furthermore, there is some evidence that social motives in conflict situations influence how people process information: Prosocial motivation encourages people to see the world in broader categories whereas proself motives result in more black-and-white thinking and can freeze cognitive schemas (Carnevale & Probst, 1998). Investigation of the psycho-social dynamics of disputants engaged in moral discussions over time requires us to move away from dichotomous distinctions between proself and prosocial orientations and to begin to conceptualize the conflicting but often concurrent tendencies people typically display over time when addressing issues that affect both themselves and others. In other words, we are interested in how both orientations – proself and prosocial operate over time in relation to one another to affect the related Moral Conflict and Complexity 11 behaviors (self-focused versus other-focused) and conflict dynamics of the parties involved. We suggest that having the capacity to employ both orientations in a somewhat balanced fashion will result in a more complex set of behaviors and more constructive dynamics than displaying either orientation alone. On the other hand, we would expect that a strong attractor for proself behaviors alone will be associated with more destructive conflict dynamics. As Lax and Sebenius (1986) have suggested, the ability to mange tensions between cooperative and competitive impulses in a flexible and adaptive way is important as cooperative and competitive elements are entwined in most conflicts and are difficult to separate. This, of course, is highly likely in addressing moral conflicts which often have serious implications for both individuals and for the larger community. Similar results have been found in research on the associated behaviors of inquiry (exploring the other’s concerns and point-of-view) and advocacy (advocating for one’s own needs, interests, and positions; Losada, 1999; Losada & Heaphy, 2004; Senge 1990). For instance, Losada found that balancing inquiry and advocacy is the key to effective action and high performance in teams and a crucial skill for advanced problem-solving, whereas pure advocacy has been found to be strongly associated with more destructive social dynamics (Losada, 1999; Losada & Heaphy, 2004). These findings lead to the following hypotheses: Hypothesis 4a: More destructive discussions over moral conflicts will evidence more referential acts restricted to the self (high-degrees of self-referential acts/low-degrees of other-referential acts) than more constructive discussions. Hypothesis 4b: More destructive discussions over moral conflicts will evidence more behavioral acts restricted to advocacy (high-degrees of advocacy acts/low-degrees of inquiry acts) than more constructive conflicts. Rumination and Delayed Emotional Reactions It is not only the experience of negative emotions during discussions that drive conflict, but also the recall of and dysphoric rumination about such experiences that can Moral Conflict and Complexity 12 motivate the perpetuation of aggressive responses to conflicts and feed intractability (Berkowitz, 1993). Research has shown that especially negative emotional events are often well retained with respect to memories of the event itself (Christianson, 1984; Christianson & Loftus, 1987, 1990, 1991; Coleman, et al., in press; Margalit, 2002; Yuille & Cutshall, 1986, 1989). Research on rumination (defined as self-focused attention particularly on one’s own negative mood) shows an increase after negative emotional events and suggests that rumination also intensifies the experience of negative emotions and intentions regarding aggressive retaliation as time passes (e.g. Bushman, 2002; Bushman, Bonacci, Pedersen, Vasquez, & Miller, 2005; Konecni, 1974; Lyubomirsky & Nolen-Hoeksema, 1995). Thus, we propose the following related hypotheses: Hypothesis 5a: Participants who experience more destructive conversations over moral conflicts will show more negative and less positive emotions after a week has elapsed than participants who experienced more constructive conversations Hypothesis 5b: Participants who experience more destructive conversations over moral conflicts will evidence more negative rumination regarding the discussion than participants who experience more constructive conversations. Hypothesis 5c: Participants who experience more destructive conversations over moral conflicts will show more hostile behavioral intentions after a week has elapsed than participants who experienced more constructive conversations. Methods: The Difficult Conflicts Laboratory Building on the pioneering laboratories of both Gottman and Losada, we established our Difficult Conflicts Laboratory to enable us to conduct research on real conflicts in real time. In particular, we were interested in developing our capacities to collect dynamical data on disputants’ emotions, cognitions, and behaviors over time, as they engaged in difficult discussions over typically polarizing moral issues. A central component of the current study Moral Conflict and Complexity 13 was a minimally-facilitated discussion over a moral issue by two individuals who held opposing opinions on the topics. Conflicts over moral issues that are important to people are potentially intractable and can lead to polarization and animosity towards other-minded individuals and groups (Pearce & Littlejohn, 1997). Dynamical data collected during the conversations and data from several questionnaires were the basis for the analysis. Participants 118 participants were matched into 59 dyads based on their opposing views on a moral issue. Participants were recruited from a wide variety of courses at a large Northeastern University in the USA and were each paid $25 for their participation. They varied widely in age, gender, ethnicity, nationality and educational background. The demographic differences of the sample had no significant influence on any of the variables or scales of the study. Procedure Each participant took part in a two hour research process. Upon agreeing to participate in this study, participants first received an online questionnaire which assessed their opinion and degrees of concern for different moral topics: death penalty, euthanasia, affirmative action, abortion and global warming. After completing the questionnaire, participants were invited to attend a 1 1⁄2 hour session in our lab. Two participants, whose opinions differed on a certain topic and who had medium to high concern about that topic, were matched together and attended each session. The study was described as: “Difficult Conversations: Discussing Sociopolitical Issues in Today’s Society” and participants were told that they would engage in several activities including a discussion about a socio-political topic with a fellow student. The participants were not informed that the other student in her or his discussion held an opposing view on the topic of discussion. During the lab session, participants were initially asked to discuss the moral issue and to try to come up with a position statement on which they both agreed and which would be anonymously shared with a Dialogue Forum of (the University) on (the topic under Moral Conflict and Complexity 14 discussion). The discussion was audio-recorded and a trained facilitator was present. However, the facilitator was instructed to not intervene nor speak unless the situation escalated and became unsafe. Immediately after the discussion, participants responded to an online-questionnaire to assess various scales, among other things their constructivedestructive experiences of the discussion. They were then instructed to listen and code their tape-recorded conversation by indicating their moment-to-moment feelings (described below). After this coding activity, participants were debriefed, thanked and paid $25. One week after this session participants received a link to another online-questionnaire via email. This questionnaire investigated their delayed reactions (delayed emotions, rumination and delayed behavior) to the conversation. Measures and Instruments Topic Opinions. Participant’s opinions on different socio-political topics were assessed using questions from opinion poll research (death penalty Jones, 2006; Euthanasia Carrol, 2007; Afirmative Action Swim & Miller, 1999; Global Warming Nisbet & Myers, 2007; Abortion – Davis, Allan & Smith, 1972). Three additional questions for each topic investigated their concern for the topic. Differences in the choice of topic had no significant influence on any of the other variables in the study. Constructive vs. destructive conversations. The degree to which a conversation was constructive or destructive was determined by the participants’ responses to several self-report scales, summarized in Table 1. These questions were included in the online 7-point Likert scale questionnaire, which participants answered directly after the conversation. Six subscales were constructed using Deutsch’s (1973) definitions of destructive versus constructive conflict: satisfaction with the process (3 items, α=.869), satisfaction with the outcome (3 items, α=.821), satisfaction with the relationship / character of the relationship (3 items, α=.821), perception of the other regarding trust, affection and friendliness (3 items, α=.880), perception of cooperation during the conversation (6 items, α=.868), and desirability of future Moral Conflict and Complexity 15 interactions with the partner of the conversation (2 items, α=.648). A scale of general satisfaction for each individual was computed using all items of the subscales (20 items, α=.948). The scores on the general satisfaction scale ranged from 1.11 to 7 with a mean of 5.56 and standard deviation of 0.99. A general satisfaction score for each dyad was determined using the mean of the general satisfaction scale of the two people from each dyad. Emotional Complexity. While participants listened to tape-recordings of their own conversation, they were asked to manipulate the mouse from a computer to indicate how their emotional experiences changed over time during the conversation. They were instructed to move the mouse to position a curser on a black screen with a white circle in the middle to either the left side of the screen (to indicate very negative emotions), the middle of the screen (neutral emotions) and the right side of the screen (for very positive emotions). A program named the mouse paradigm, developed by Vallacher and Nowak (Vallacher, Nowak, & Kaufman, 1994; Vallacher & Nowak, 1994; Nowak & Vallacher, 1998) registered the position of the mouse for every second of the recorded conversation. The analysis of this data was based on the percentage of negative, neutral and positive emotions that participants experienced during the conversation, which were computed from time series data gathered with the mouse paradigm. A ratio for emotions (% of positive emotions divided by % of negative emotions) was used to categorize participants into six equal groups form very low ratios to very high ratios (referred to emotional ratios) Cognitive Complexity. Each of the individual participant’s levels of cognitive complexity was assessed after the discussion in accordance with a standard measure of integrative complexity (c.f. Suedfeld, et al., 1992). Included in the questionnaire collected after the conversation, was the following open ended question: ‘Please take a few minutes to think about the topic that you just discussed and write down all the thoughts which occur to be relevant to you.’ The answers were coded following the coding plan of Baker-Brown, Ballard, Bluck, de Vries, Suedfeld, and Tetlock (1992). The statements were coded by two Moral Conflict and Complexity 16 independent and blind coders. Their initial inter-rater reliability was: Cohen’s Kappa = .834. In case of disagreement the score was discussed by the raters and either an agreement was reached or the conversation was not coded. The coding plan suggests 7 levels of integrative complexity. On level 1 no differentiation and no integration is shown. Level 3 indicates differentiation between different aspects of the topic, however no integration. Differentiation and integration is the characteristic in level 5-7. Additionally in level 7, a global statement should be offered by the author of the statement. Levels 2-5 represent intermediate levels. The mean score of participants was 3.1 with a standard deviation of 1.6. Learning. Participant’s learning was assessed with a six-item self-report scale developed for this study, which was included in the post-discussion questionnaire (Cronbach’s Alpha .748). Behavioral Complexity. Two behavioral components of the dyads were investigated in this study: Speech acts of inquiry vs. speech acts of advocacy and proself vs. prosocial referential acts. Two trained coders blind to condition listened to the 5 most constructive and the 5 most destructive conversations (determined through the general satisfaction measures of the dyads) and coded the conversations according to a coding plan developed by the investigators. The mouse paradigm was used for the coding of behaviors, however, unlike the self-coding of emotions, the behavior data was not measured on a continuous scale but as categorical data in 3 categories: advocacy behaviors (position statements)-neutral-inquiry behaviors (exploratory questions), and prosocial behaviors (other-oriented references)neutral-proself behaviors (self-oriented references). The percentage of proself/prosocial references and inquiry/advocacy acts were generated and computed into ratios ( % of prosocial divided by % of proself; % of inquiry divided by % of advocacy) for the analysis. Delayed emotions. After one week participants were asked to indicate on a 7-point Likert scale the extent to which they felt a set of positive and negative emotions when reflecting on the conversation at that moment. Moral Conflict and Complexity 17 Rumination. One week after the discussion participant’s rumination over the conversation was measured with nine items from the rumination scale by Sukhodolsky, Golub and Cromwell (2001; 7-point Likert scale). The nine items were divided into two subscales: for rumination without (α=.871) and with negative connotation (α=.838). Delayed behavioral intentions. In the follow up questionnaire two short scenarios were described and participants were asked to indicate on 7-point Likert scales how likely they would be to show certain behaviors. The first scenario put the participant in a situation in which he/she met the partner from the conversation in the hallway. In the second scenario, they found themselves in a situation in which they had a class with their conversation-partner and were assigned to a workgroup with them. The behavioral options ranged from friendly and open behaviors to avoidance to more unfriendly behaviors. Results The main objective of the analysis was to investigate whether more destructive conflict dynamics over difficult moral issues evidenced more simple, coherent attractor dynamics for thoughts, feelings and behaviors than more constructive conversations. Figure 1 illustrates these differences, which are described in more detail below. INSERT FIGURE 1 HERE Emotional Dynamics The hypothesis that constructive conversations would evidence more balanced and positive emotions than destructive conversations, which would be more restricted to negative emotional dynamics, was supported. The percentage of negative emotions was negatively correlated with the degree of satisfaction (r=-.545), positive emotions showed a positive correlation (r=.390) and neutral emotions were not correlated. Further the higher the emotional ratio the more satisfied the participants (r=.573) were. The biggest increase in satisfaction was found for the ratios around 2-3.5. Interestingly, lower and higher ratios did not significantly influence the level of satisfaction (see Figure 2). Therefore conversations Moral Conflict and Complexity 18 which allowed participants to experiences at least some degree of positive emotions above a threshold were rated more constructively. INSERT FIGURE 2 HERE It was further hypothesized that participants in constructive conversations would evidence more responsiveness and adaptation to the other participant’s emotional experience. Cross-correlations (lag 0) between the time series data for the participants of each dyad disclosed that the emotions of dyads in constructive conversations were more synchronized than those in destructive conversations as they showed higher cross correlations (Correlation with general satisfaction scale or dyad: r = .219, p<.05; ANOVA after median split on general satisfaction scale of dyad F=5.48; p<.05). Cognitive Complexity As hypothesized, participants who experienced a constructive discussion showed significantly higher scores in integrative complexity when writing about the discussed topic after the conversation (r=.535; p<.001) than those with destructive experiences. An ANOVA comparing the two groups of participants after a median-split on the scale for general satisfaction confirmed this finding (F=21.69; p<.001). Furthermore, the score for integrative complexity was positively correlated with the percentage of positive emotions (r=.336, p<.01) and with the emotional ratio (r=.468, p<.001). Therefore along with the experience of positive emotions during the conversation, more satisfied participants were able to write more nuanced statements on the topics by elaborating on different approaches to the issues before taking a position, than more dissatisfied participants (see Figure 1). The causal direction of this relationship is difficult to establish from these data, but the association of higher levels of cognitive complexity and more constructive conflict is strong and clear. Learning Consistent with Hypothesis 3, we found that participant’s level of learning was higher in constructive conversations (Correlation with general satisfaction: r= .590; p<.001; ANOVA Moral Conflict and Complexity 19 after median-split along general satisfaction: F=23.36, p<.001). Looking at the emotional experience during the conversation we found that learning was somewhat negatively correlated with the percentage of negative emotions (r=-.238; p<.05), but not correlated with the percentage of positive or neutral emotions. The correlation with the emotional ratio showed the highest effect size (r=.311; p=.01). While negative emotions were associated with low levels of learning, participants in constructive conflicts were able to learn more from the other party and therefore gain a more elaborate understanding of the topic (see Figure 1). Behavioral Dynamics Proself/Prosocial referential acts. It was hypothesized that more constructive conversations would show more balanced patterns of proself/prosocial referential acts, whereas more destructive conversations would be more restricted to proself referential acts. This hypothesis was fully supported. For this analysis, the 5 most constructive and 5 most destructive conversations were analyzed resulting in an N of 20. A T-Test showed that the 5 most constructive conversations evidenced significantly more prosocial referential acts (t=2.47, p<.05) and significantly less proself referential acts (t= 3.0, p<.01) than the five most destructive conversations over the time of the discussions. Whereas in the destructive conversations the percentage of proself referential acts was significantly more than the prosocial referential acts (t=4.64; p<.001), this difference was not significant for the constructive conversations (t=-0.81; p>.05) and hence the proself-prosocial ratios also differed significantly when comparing those two groups (t=-4.06, p<.001). As shown in Figure 1, participants in constructive conversations made use of self-referential acts, however they were able to counterbalance those with other-referential acts, showing more balanced and complex behavioral dynamics. Speech acts of inquiry and advocacy. Analyzing the percentage of inquiry and advocacy of the 5 most constructive and 5 most destructive conversations (N=20) yielded following results: Whereas participants in constructive and in destructive conversations Moral Conflict and Complexity 20 evidenced a similar percentage of advocacy acts (no significant difference) the level of inquiry was significantly higher in constructive conversations (-2.21; p<.05). The inquiryadvocacy ratio was also significantly different when comparing those two groups (t=-2.31, p<.05). Figure 1 illustrates that in constructive conversations although participants strongly advocated for their position, they were also open to exploring the other party’s position. Delayed Reactions When recalling the conversation after one week, participants who had experienced constructive conversations felt more satisfied (r=.358; p<.01), grateful (r=.210; p<.05), and comfortable (r=.194; p<.05) and less angry (r=-.195; p<.05), disappointed (r=-.491; p<.01), irritated (r=-.250; p<.05), sad/depressed (r=-.306; p<.01), disturbed (r=-.392; p<.01), helpless (r=-.381; p<.01), and disrespected (r=-.197; p<.05). There were no significant differences found between constructive vs. destructive conversation participants on neutral ruminations, however the more constructive the conversation the less participants showed negative rumination after one week (r=-.291; p<.01). Regarding behavioral intentions, we found that the more constructive a conversation was the less participants wanted to ‘greet and pretend I am in a hurry’ (r=-267; p<.01) when meeting the partner in the hallway and the less they intended to ‘do nothing and wait and see what happens -if the partner turns out to be nice, it’s fine, if not I give him / her a hard time’ (r=-.230; p<.05) when having to work together with the partner for a class. Thus, more destructive conflicts resulted in more negative emotions, ruminations and behavioral intentions after a delay of one week. Discussion Moral conflicts have the potential to divide our world or to move us forward toward a more humane place. In this study our main interest was to better understand the basic, underlying dynamics which distinguish more constructive from more destructive moral conflicts. Building on a dynamical systems theory of intractable conflict, we hypothesized that in destructive moral conflicts the psycho-social realities and interactions of disputants Moral Conflict and Complexity 21 would collapse into negative, coherent us-versus-them dynamics. Accordingly, we hypothesized a decrease under these conditions in the level of complexity of participants’ emotional, cognitive and behavioral functioning. Constructive conflicts, on the other hand, were predicted to evidence more complex, balanced attractor dynamics. The findings from our study provide strong support for the hypothesized differences in dynamics. In comparison to more constructive moral conversations, destructive conversations evidenced fixed-point attractor dynamics, which were more restricted to negative emotions, to lower levels of complexity in understanding of the issues, and to primarily advocacy behaviors and self-referential acts. More constructive conversations on the other hand showed patterns that allowed for more ratios with more positive and less negative emotions, for more nuanced understanding of the topics, for both advocacy and inquiry behaviors, and for otherreferential acts in addition to self-referential acts. These findings suggest that the relative balance of these emotions (positive and negative), cognitions (differentiation and integration), and behaviors (self-oriented and other-oriented, advocacy and inquiry) represent more complex dynamics in moral conflicts. Thus, dynamical complexity should be recognized as a critically important parameter determining more constructive versus destructive conflicts. In addition, more constructive moral conversations showed more openness and responsiveness in terms of higher levels of learning, more inquiry and a higher level of synchronization between the parties’ emotions. Differences in the patterns were still found after one week. At that point destructive conversations were associated with negative emotions, negative rumination and unfriendly behavioral intentions. Even though we cannot make causal conclusions from this study we were able to show that constructive and destructive conversations differ fundamentally in terms of the level of complexity-simplicity of their emotional, cognitive and behavioral dynamics and patterns. The university student sample and correlational and mostly self-report natures of the study are its main limitations. However, having now established important differences in Moral Conflict and Complexity 22 these dynamics, the next step will be to conduct a series of studies that systematicallymanipulate the parameters in the study (emotional ratios, cognitive complexity, andbehavioral ratios) to test their role in triggering qualitative differences in moral conflictdynamics over time. We see such research as theoretically meaningful and practicallyimportant. Furthermore, it is important to point out that the data for this study was collectedfrom conversations over issues of genuine real-world importance to the participants. Suchresearch is rare in our field and compliments other laboratory findings based on simulations,role-plays, and games. The research presented here represents a first foray into what we hope to be asubstantial, long-term program of research in our Difficult Conflicts Lab. Gaining systematicunderstanding of the basic dynamics differentiating constructive versus destructiveapproaches to difficult conflicts is vital to both scholars and practitioners in the fields ofconflict resolution and peace building. Our goal of course is not to oversimplify suchconflicts, but to better understand their essence. As Vaclav Havel once said:Simple answers which lie on this side of life’s complexities are cheap. However,simple truths, which exist beyond this complexity and are illuminated by it, are worthyof a lifetime’s commitment.ReferencesBaker-Brown, G., Ballard, E. J., Bluck, S., de Vries, B., Suedfeld, P., & Tetlock, P. E. (1992).The conceptual/integrative complexity scoring manual. In C. P. Smith, J. W. Atkinson,D. C. McClelland, & J. Veroff (Eds.), Motivation and Personality: Handbook ofthematic content analysis (pp. 401-418). New York: Cambridge University Press.Bales, R. F. (1950). A set of categories for the analysis of small group interaction. AmericanSociological Review, 15, 257 – 263.Barki, H., & Hartwick, J. (2004). Conceptualizing the Construct of Interpersonal Conflict.International Journal of Conflict Management, 15, 216-244. Moral Conflict and Complexity 23 Bar-Siman-Tov, Y. (Ed.). (2003). From conflict resolution to reconciliation. New York:Oxford University Press.Berkowitz, L. (1993). Aggression: Its causes, consequences, and control. New York:McGraw-Hill. Bushman, B, J. (2002). Does venting anger feed or extinguish the flame? Catharsis,rumination, distraction, anger, and aggressive responding. Personality and SocialPsychology Bulletin, 28, 724-731. Bushman, B. J., Bonacci, A. M., Pedersen, W. C., Vasquez, E. A., & Miller, N. (2005).Chewing on it can chew you up: Effects of rumination on triggered displacedaggression, Journal of Personality and Social Psychology, 88, 969-983.Carnevale, P.J., & Probst, T. (1998). Social values and social conflict in creative problemsolving and categorization. Journal of Personality and Social Psychology, 74, 1300-1309.Carrol, J. (2007). Public Divided Over Moral Acceptability of Doctor-Assisted Suicide. But,majority still supports doctors legally helping terminally ill patients die. In Gallup.Retrieved September 16, 2008 from http://www.gallup.com/poll/27727/Public-Divided-Over-Moral-Acceptability-DoctorAssisted-Suicide.aspxChristianson, S. A. (1984), The relationship between induced emotional arousal and amnesia,Scandinavian Journal of Psychology, 25, 147-160.Christianson, S. A. & Loftus, E. F. (1987), Memory for traumatic events. Applied CognitivePsychology, 1, 225-239.Christianson, S. A. & Loftus, E. F. (1990), Some characteristics of people’s traumaticmemories, Bulletin of the Psychonomic Society, 28, 195-198.Christianson, S. A. & Loftus, E. F. (1991), Remembering emotional events: The fate ofdetailed information, Cognition and Emotion, 5, 81-108. Moral Conflict and Complexity 24 Coleman, P. T. (2003). Characteristics of protracted, intractable conflict: Towards thedevelopment of a meta-framework I. Peace and Conflict: Journal of PeacePsychology, 9, 1-37. Coleman, P. T. (2006) Conflict, complexity, and change: A meta-framework for addressingprotracted, intractable conflicts III. Peace and Conflict. Journal of Peace Psychology,12, 325-348. Coleman, P. T., Bui-Wrzosinska, L., Vallacher, R., & Nowak, A. (2006). ApproachingProtracted Conflicts as Dynamical Systems: Guidelines and Methods for Intervention.In A. Schneider & C. Honeyman (Eds.), The Negotiator’s Fieldbook (pp. 61-74).Chicago: American Bar Association Books.Coleman, P. T., Goldman J., & Kugler, K. (in press). Emotional Intractability: Gender, Anger,Aggression and Rumination in Conflict. International Journal of Conflict Management.Coleman, P. T., Vallacher, R., Nowak, A., & Bui-Wrzosinska, L. (2007). Intractable conflictas an attractor: Presenting a dynamical model of conflict, escalation, and intractability.American Behavioral Scientist, 50, 1454-1475.Conway, L. G. III, Suedfeld, P., & Tetlock, P. E. (2001). Integrative complexity and politicaldecisions that lead to war or peace. In D. J. Christie, R. V. Wagner, & D. D. N. Winter(Eds.), Peace, conflict, and violence: Peace psychology for the 21st century (pp. 66-75). Upper Saddle River, NJ: Prentice Hall.De Dreu, C. K. W., Beersma, B., Steinel , W., & Van Kleef, G. A. (2007). The psychology ofnegotiation: Principles and basic processes. In A.W. Kruglanski & E.T. Higgins(Eds.), Social psychology: Handbook of basic principles (2nd edition, pp. 608-629).New York: Guilford. Moral Conflict and Complexity 25 De Dreu, C. K. W., Weingart, L. R., & Kwon, S. (2000). Influence of social motives onintegrative negotiation: A meta-analytical review and test of two theories. Journal ofPersonality and Social Psychology, 78, 889-905. Deutsch, M. (1949). A theory of cooperation and competition. Human Relations, 2, 129-151.Deutsch, M. (1973). The resolution of conflict: Constructive and destructive processes. NewHaven, CT: Yale University Press. Fredrickson, B. L. (2001). The role of positive emotions in positive psychology: The broaden-and-build theory of positive emotions. American Psychologist, 56, 218-226.Fredrickson, B. L., & Branigan, C. (2001). Positive emotions. In T. J. Mazne & G. A.Bonanno (Eds.), Emotions: Current Issues and Future Directions (pp. 123-151). NewYork, NY: Guilford Press.Fredrickson, B. L., & Branigan, C. (2005). Positive emotions broaden the scope of attentionand thought-action repertoires. Cognition and Emotion, 19, 313-332.Gandhi, I (n. d.). Retrieved February 1, 2008 form http://www.quotedb.com/quot.Golec, A., & Federico, C. M. (2004). Understanding responses to political conflict:Interactive effects of the need for closure and salient conflict schemas. Journal ofPersonality and Social Psychology, 87, 750-762.Gottman, J. M., Murray, J. D., Swanson, C. C., Tyson, R., & Swanson, K. R. (2002). Themathematics of marriage. Cambridge, MA: MIT Press.Gurr, T. R. (1970). Why men rebel. Princeton, NJ: Princeton University Press.Jehn, K.A., & Bendersky, C. (2003). Intragroup conflict in organization: a contingencyperspective on the conflict-outcome relationship. Research in organizationalbehaviour, 25. 187-242.Jones, M. J. (2006). Two in Three Favor Death Penalty for Convicted Murderers. Publicdivided over death penalty or life imprisonment as better punishment. In Gallup. Moral Conflict and Complexity 26 Retrieved September 16, 2008 from http://www.gallup.com/poll/23167/Two-Three-Favor-Death-Penalty-Convicted-Murderers.aspxKegan, R. (1994). In Over Our Heads: The Mental Demands of Modern Life. Cambridge,MA: Harvard University Press. Konecni, V. J. (1974). Self-arousal, dissipation of anger, and aggression. Personality andSocial Psychology Bulletin, 1, 192-194. Koo, J., Han, J., & Kim, J. (2002). Integrative Complexity of South-North KoreanCorrespondences: A Time-Series Analysis, 1984-1994. Journal of Conflict Resolution,46, 286-304.Kriesberg, L. (2005). Nature, dynamics, and cases of intractability. In C. A. Crocker, F. O.Hampson, & P. Aall, (Eds.), Grasping the Nettle: Analyzing cases of intractableconflict (pp. 65-98). Washington, DC: United States Institute of Peace Press.Lax, D. A., & Sebenius, J. K. (1986). The Manager as Negotiator: Bargaining for Cooperationand Competitive Gain. N.Y.: Free Press.Lee, J., & Truex, D.P. (2000). Cognitive Complexity and Methodical Training: Enhancing orSuppressing Creativity. Proceedings of the 33rd Hawaii International Conference inSystem Science.Levy, J. F. (1994). Learning and Foreign Policy. Sweeping a Conceptual Minefield.International Organization, 48, 279-312.Lindner, E. G. (2001). Humiliation and the human condition: Mapping a minefield. HumanRights Review, 2, 46-63.Lindner, E.G. (2002). Healing the cycles of humiliation: How to attend to the emotionalaspects of ‘unsolvable’ conflicts and the use of ‘humiliation entrepreneurship’. Peaceand Conflict: Journal of Peace Psychology, 8, 125-138.Losada, M. (1999). The complex dynamics of high performance teams. Mathematical andComputer Modelling, 30, 179–192. Moral Conflict and Complexity 27 Losada, M., & Heaphy, E. (2004). The role of positivity and connectivity in the performanceof business teams: A nonlinear dynamics model. American Behavioral Scientist, 47,740–765. Lyubomirsky, S., & Nolen-Hoeksema, S. (1995). Effects of self-focused rumination onnegative thinking and interpersonal problem solving. Journal of Personality andSocial Psychology, 69, 176-190. Margalit, A. (2002). The ethics of memory. Cambridge, MA: Harvard University Press.Mayer, J. D. & Salovey, P. (1997). What is emotional intelligence? In P. Salovey & D.Sluyter (Eds.), Emotional Development and Emotional Intelligence: Implications forEducators (pp. 3-31). New York: Basic Books.McClintock, C.G. (1977). Social motivation in settings of outcome interdependence. InDruckman, D. (Ed.), Negotiation Behavior (pp. 49–77). Beverly Hills : Sage.Messick, D.M., McClintock, C.G. (1968). Motivational bases of choice in experimentalgames. Journal of Experimental Social Psychology, 4, 1–25Nisbet, M. C., & Mayers, T. (2007). The polls-trends: twenty years of public opinion aboutglobal warming. Public Opinion Quarterly, 71, 444-470.Nowak, A. & Vallacher, R. R. (1998). Dynamical social psychology. New York: GuilfordPublications.Osgood, C. E. (1983). Psycho-social dynamics and the prospects for mankind. Urbana:University of Illinois.Pearce, W.B., & Littlejohn, S.W. (1997), Moral conflict: When social worlds collide, Sage,Thousand Oaks, CA.Pruitt, D., & Kim, S. H. (2004). Social conflict: Escalation, stalemate, and settlement. NewYork: McGraw-Hill.Pruitt, D. G., & Rubin, J. (1986). Social Conflict: Escalation, Stalemate and Settlement. NewYork, NY: Random House. Moral Conflict and Complexity 28 Rosenberg, S.W. (1988). Reason, Ideology and Politics. Princeton, NJ: Princeton Univ. Press.Schuster, H. G. (1984). Deterministic Chaos, An Introduction. Weiheim: Physik-Verlag.Senge, P. (1990). The Fifth Discipline, The Art & Practice of The Learning Organization.London: Random House. Suedfeld, P. & Coren, S. (1992) Cognitive correlates of conceptual complexity. IndividualDifferences, 13, 1193-1199. Suedfeld, P., & Leighton, D. C. (2002). Early communications in the war against terrorism:An integrative complexity analysis. Political Psychology, 23, 585-599.Streufert, S., & Streufert, S. (1978). Behavior in the complex environment. Washington, DC:Winston.Suedfeld, P., Tetlock, P. E., & Streufert, S. (1992). Conceptual/integrative complexity. In C.P. Smith, J. W. Atkinson, D. C. McClelland, & J. Veroff (Eds.), Motivation andpersonality: Handbook of thematic content analysis (pp. 393-400). New York:Cambridge University Press.Sukhodolsky, D. G., Golub, A. & Cromwell, E. N. (2001). Development and validation of theanger rumination scale. Personality and Individual Differences, 31, 689-700.Swim, J.K. & Miller, D.L. (1999). White guilt: its antecedents and consequences for attitudestoward affirmative action. Personality and Social Psychology Bulletin, 25, 500-514.Tetlock, P. E. (1991). Integrative complexity of political discourse in American-Sovietrelations. In W. Gamson & C. Tilly (Eds.), Working Papers on Security Discourse inthe Cold War. New York, NY: Center for Studies of Social Change, New School forSocial Research.Vallacher, R. R., & Nowak, A. (1994). Dynamical systems in social psychology. San Diego:Academic Press.Vallacher, R., Nowak, A., Bui-Wrzosinska, L., & Coleman, P. T. (2006). Attracted toconflict: A dynamical perspective on malignant social relations. In A. Golec & K. Moral Conflict and Complexity 29 Skarzynska (Eds.), Understanding social change: Political psychology in Poland (pp.167-182). Haauppague NY: Nova Science Publishers Ltd.Vallacher, R. R., Nowak, A., & Kaufman, J. (1994). Intrinsic dynamics of social judgment.Journal of Personality and Social Psychology, 67, 20-34. Winter, D. G. (2007). The Role of Motivation, Responsibility, and Integrative Complexity inCrisis Escalation: Comparative Studies of War and Peace Crises. Journal ofPersonality and Social Psychology, 92, 920–937. Woike, B.A., & Matic, D. (2004). The role of motives in the organization of traumatic eventsin memory. Journal of Personality, 72, 633-657.Yang, J., and Mossholder, K. W. (2004). Decoupling task and relationship conflict: The roleof intragroup emotional processing, Journal of Organizational Behavior, 25, 589-606.Yuille, J. C., & Cutshall, J. L. (1986). A case study of eyewitness memory of a crime. Journalof Applied Psychology, 71, 291-301.Yuille, J. C., & Cutshall, J. (1989). Analysis of the statements of victims, witnesses andsuspects. In J. C. Yuille, (Ed.), Credibility Assessment. Dordrecht, The Netherlands:Kluwer Academic Publishers. Moral Conflict and Complexity 30 Figure 1: The different patterns of constructive versusdestructive conversations. For the parameters without a * amedian-split was used to divide between more constructive andmore destructive conversations for the entire sample; for theparameters with a * only the 5 most constructive and 5 mostdestructive conversations (N=20) were used for the analysis.4.85.6