Negotiation Processes As Sources of (And Solutions To) Interorganizational Conflict

We investigate how structural features of negotiations can affect interaction processes and how negotiations can be not only a solution to, but also a source of, inter-organizational conflict. Principals, agents, and teams face different sets of constraints and opportunities in negotiations. We develop grounded theory detailing how the micro-interactions comprising a negotiation are shaped by the representation structure (principals, agents, or teams) of the parties. In qualitative and quantitative analyses of negotiations carried out by principals, agents, and teams in a laboratory experiment, we find that negotiators’ efforts to manage the constraints and opportunities of their representation structure are reflected in the micro-interactions, the broad improvisations, and the resulting substantive and relational outcomes.

In organizations, negotiations are used to settle disputes and to determine the terms of agreement in transactions and cooperative endeavors. As a result, negotiations are often conceptualized as a means of managing or resolving conflict. But, while the process of negotiation may be a solution to conflict in some cases, it may be a source of conflict in others.
The constraints and opportunities experienced by parties negotiating on behalf of their organizations and the need to simultaneously manage relational and task dynamics in inter- Inter-organizational negotiations present choices regarding who will negotiate on behalf of the organization. Owners or principals might engage in negotiation, but individual agents or teams may also represent the organization. While research comparing team versus individual negotiations suggests that different party representations affect the process and outcomes of negotiations (Polzer 1996;Thompson et al. 1996), no consensus has evolved regarding the advantages and constraints of the various negotiation party structures. Despite the critical resources at stake, little is known about the relative pros and cons of negotiating alone for one's own interests, sending an agent, or relying on a team.
In this chapter, we explore how party representation affects negotiation processes and how those processes may affect conflict during and after the interaction. We build on the extant literature on negotiation processes, which emphasizes the interactive, dynamic, and improvisational nature of negotiations (e.g., McGinn and Keros 2002;Weingart et al. 1999) but does not provide a theory for how negotiation interaction and outcomes might be shaped by structural variables, such as party representation. Our perspective is inspired by a grammatical approach to examining organizational processes (Pentland 1995). This enables us to bridge micro and macro understandings of negotiation process and provide theoretical and analytical traction on the question of how structural features play out in negotiations.

A Process Perspective on Negotiations
A process approach to negotiations concentrates interest and analyses on the interaction itself. Outcomes are seen as necessarily dependent on the emergent interaction, rather than directly attributable to exogenous independent variables. Process and outcomes are inextricably linked as perceptions and expectations of outcomes affect the unfolding process, which in turn affects the outcomes realized.
Research on negotiation processes considers task-and relationship-oriented behaviors (Gelfand et al. 2006;Wilson and Putnam 1990). Task-related behaviors include both substantive and procedural acts. Substantive acts-such as exchanges of information, questions, and offersare the heart of most negotiation process analyses (e.g., Thompson 1991;Weingart et al. 1999).
Procedural acts, such as discussions of how to structure the interaction (Lytle et al. 1999) and assessments of progress against time (Lau & Murnighan 2005), help define a structure for substantive exchanges. Relational acts can serve task-related functions as well (Kolb andWilliams 2000, 2003), but they are distinct in that they reveal or affect the relationship between the parties and contribute to the relational positioning between actors (e.g., Adair and Brett 2005;Gelfand et al. 2006). Relational acts may be central to understanding the effects of party representation on negotiation process and outcome.
Process research has ranged from the micro level, counting the frequency of different words or phrases (e.g., Weingart et al. 1999;Weingart et al. 1990), to a higher level of analysis examining the patterns and stages of interaction (e.g., Adair and Brett 2005;Olekalns et al. 1996), to a more macro approach exploring how emergent logics shape the repertoires of action within a negotiation (e.g., Kolb 1985;McGinn and Keros 2002). Each level of analysis has offered new insights into the paths toward different negotiation outcomes.
Frequency counts of specific types of behaviors, information sharing for example, led to a more dynamic, interactive theory of negotiation (Weingart et al.1993;Weingart et al. 1999).
Research exploring the dynamic aspects of bargaining revealed how meaning is created through conversational structure and language and identified behavioral sequences and strategies that delimit the range of actions available to negotiators as bargaining progresses Valley et al. 1992). Studies examining interaction at the level of the negotiation exposed how negotiator roles lead to different behavioral repertoires (Kolb 1985;Putnam and Jones 1982a). Collectively, these results imply that parties are not simply choosing among all available task-and relationship-oriented behaviors, nor are they following pre-set scripts.
Instead, negotiations exhibit -a coherent sequence of actions and responses created, chosen, and carried out by the parties during the social interaction‖ (McGinn and Keros 2002, 445). Just as jazz musicians' improvisations are guided by the melody, harmony, and rhythm of a song, negotiators co-create the interaction through an interdependent process of active response to the emergent constraints and opportunities of the unfolding interaction (Balachandra et al. 2005). To do so, negotiators must simultaneously structure their interaction and then improvise within the structure that they co-create.
While the improvisational approach suggests that emergent logics guide negotiation interaction, it fails to develop how these logics might act as a theoretical bridge between microinteraction and structural variables, such as party representation. Pentland's (1992; work on grammars of organizational processes presents a useful theoretical and methodological approach to linking structure and process. In their examination of organizational routines, Pentland and Reuter (1994) focus on both micro-interaction and macro-logics of interaction, which they call grammars. Grammars, like improvisations, are logics for interaction that guide, but do not determine, how individuals combine acts in practice. Social actors choose from a repertoire of available actions within a logic of interaction (Giddens 1976(Giddens , 1984Pentland 1995).
Grammars reflect the constraints and opportunities that technological, institutional, cultural, and coordination structures place on the actions available (Pentland 1995).
In this chapter, we present a typology of improvisations that guide and emerge out of micro-interactions between negotiating parties. Just as grammars of organizational routines reflect features of the organizational context, improvisations reflect negotiators' responses to the negotiation context (Lingo and O'Mahony 2010). More specifically, we suggest that improvisations reflect negotiators' efforts to manage the constraints and opportunities presented by the presence of teams, agents, or principals at the negotiating table.

Party Representation: Constraints and Opportunities
Prior work focusing on the effect of party representation in negotiations suggests that principals, teams, and agents face differing constraints and opportunities at the bargaining table, which may differentially affect the extent to which conflict emerges. We conceptualize these constraints and opportunities around three domains: (a) the knowledge and skills that parties bring to bear on the negotiation; (b) the potential development of cross-party identification, trust, or relational conflict; and (c) coordination and communication. In this section, we summarize the challenges and opportunities imposed by each representation structure and its likely influence on the emergence of conflict and negotiated outcomes.
The primary constraint for individuals negotiating as principals is that of knowledge and skill: information processing is limited by the capacity constraints of a single person. However, principals negotiating with other principals need only to interpret the actions of one person, reducing the possibility for confusion due to multiple parties' potentially conflicting statements and actions. Because of the more intimate setting, principals negotiating one-on-one have the opportunity to experience relational connection with the other party, especially in face-to-face bargaining (McGinn and Crosen 2004), increasing the likelihood of cross-party identification and trust. As a result, two principals negotiating with one another may find it straightforward to develop a shared understanding of how interaction should unfold, lessening the risk of unproductive relationship conflict (Jehn 1997).
Agents, like individuals, must deal on their own with the complexity of information and interpretation, but they also face their own portfolio of constraints and opportunities. Agents must struggle with attempting to build working relationships with their negotiating counterparts while simultaneously maintaining credibility and integrity with their constituents not present at the table (Walton and McKersie 1965). As a result, agents may be less likely than principals to develop cross-party identification and trust (Kramer et al. 1993), which may heighten the risk of conflict. For instance, agents have been shown to use threats, pressure tactics, and noncooperative behaviors in negotiations to signal concern for their constituents (Bacharach and Lawler 1981;Enzle et al. 1992) and increased accountability for the outcome relative to teams (O'Connor 1997). Thus, research has shown that while agents may improve outcomes when they reach agreement, they also have been shown to increase the likelihood of impasse (Bazerman et al. 1992) and to negotiate more contentiously (O'Connor 1997).
Negotiations between teams offer the potential for greater information processing capacity and multiple perspectives. Teams' greater breadth in expertise, working styles, and approaches to problems can lead to creativity and robust decision-making (Jackson 1992).
Multiple negotiators allow teams to allocate socio-emotional and task-related roles to different people, whereas individuals may find it difficult to fulfill these roles simultaneously (Friedman and Podolny 1992). But negotiation teams may be constrained by information processing errors unique to groups (Brodbeck et al. 2007;Mannix et al. 1989). Problems such as conformity pressure (Janis 1982), premature arrival at the first solution recommended, and the domination of a particular person over minority or diverging opinions present constraints for teams (Maier 1983). Coordination constraints and in-group bias may inhibit the development of trust and cooperation across negotiating parties (Polzer 1996) and increase the likelihood of dysfunctional forms of conflict (Jehn, 1997;Jehn and Mannix 2001;Bendersky and Hayes 2010). A shortage of trust between parties is a precursor to defensive behavior, including withholding information and aggressive statements attacking the others' position (Lewicki and Bunker 1996). Table 1  The constraints and opportunities of each representation structure are summarized in Table 1. In sum, extant research consistently demonstrates that teams, agents, and individuals face different constraints and opportunities in their negotiations but provides mixed information that limits accurate predictions regarding how party representation might manifest in a given negotiation.

Participants and Overview
One hundred-eight participants, recruited through advertising at universities in a major metropolitan area, participated in one of ten sessions of the experiment. 1 Participants were assigned one of two negotiation roles and matched with strangers in one of three party representation conditions: individual (one principal negotiates with another principal), agent (one individual representing a 3-person team negotiates with one other individual representing a 3person team), or team (three individuals assigned to one team negotiate with three individuals assigned to another team). This resulted in thirty-four face-to-face negotiations-nine individual negotiations, 15 agent negotiations, and ten team negotiations.
The negotiation scenario simulated a transaction between representatives of a town council and representatives of a logging company operating in the area. 2 The role materials described five issues: one compatible issue (both parties received high payoffs with the same option); two trade-off issues (one party placed high value on the first issue and low value on the second, while the reverse was true for the other party); one distributive issue (the parties' payoffs were opposed); and one integrative issue (maximum joint gain was achieved with an alternative that neither party preferred individually).

Experimental Procedure
Upon arrival, participants were randomly assigned to one of two roles (the Town of Silva or the Shaded Glen Logging Company) and a representation structure condition (individual, agent, or team). To manipulate representation structure, we varied the role instructions and number of people negotiating on each side, the priming during negotiation preparation, and payoffs. The details are provided in Table 2. To prime for representation structure while keeping the potential for information processing during preparation constant, all participants prepared for 15 minutes in face-to-face groups of three. However, instructions for the preparation phase varied across roles. All participants were paid a show-up fee, plus their earnings in the negotiation, but the method of calculating earnings varied by treatment.
After the fifteen-minute preparation period, participants joined their negotiation counterparts in private negotiation rooms for up to 30 minutes. All interactions were audio recorded. Following the negotiations, participants turned in agreement forms listing the details of their agreements and individually completed questionnaires regarding their perceptions of the negotiation process and outcome and their perceptions of the other party. Participants were privately paid and given a debrief sheet explaining the study's purpose.

Coding Transcripts for Negotiation Acts
Thirty-three negotiations 3 were transcribed and analyzed. Our coding proceeded through three general steps: inductively generating a coding scheme; coding speaking turns; and identifying patterns or regularities of underlying improvisational logics. Following existing negotiation coding schemes (Adair and Brett 2005;Weingart et al. 1999;Weingart et al. 1990), we coded each transcript at the level of the speaking turn. We inductively developed micro-level coding categories based on an iterative process in the spirit of grounded theory development (Glaser and Strauss 1967). Two of the authors read through the transcripts to get a sense of behaviors arising during the negotiations. In keeping with the extant literature, we distinguished between substantive acts and procedural acts, which we coded as sub-categories of task-related acts, and relational acts focused on the parties themselves. Relational acts worked in two directions: acts of connection reflected or drove a positive relationship; acts of separation reflected or drove a negative relationship. When we found that what we had conceived of as a single code was too broad, we broke the code down further. This resulted in 32 separate codes-17 relational acts; 11 substantive acts; 3 procedural acts; and a code for filler-talk (no content). Descriptions of each code are provided in Table 3. Table 3 about here We attempted to interpret behaviors or messages as they were received, rather than trying to infer the speaker's interpretation. Because messages can serve more than one function (Weingart et al. 1990), our coding scheme allowed each speaking turn to receive up to two of the 32 codes-one of the task codes, either substantive or procedural, and/or one relational code. If there was no identifiable task or relational function for a speaking turn, it was coded as filler.
Each transcript was coded independently by two of three coders (two of the authors and one coder blind to the hypotheses and conditions). After coding, reliability was calculated using the ratio of the number of exact agreements to the number of potential coding units (which varied by transcript due to differences in the length of the negotiations). Average agreement rate across the two coders on each transcript was 71%, significantly higher than chance (< 5%). Disagreements were resolved through discussion among the coders.

Inductively Deriving Improvisations
We inductively derived the macro-logics of interaction, or improvisations, by identifying and analyzing patterns of micro-interaction. We began our analysis of the detailed micro-coding by creating a visual behavioral pattern map for each negotiation, assigning each major code a different color and collapsing to the codes presented in bold in Table 3. Similar codes were assigned shades in the same color range. The pattern map for each negotiation comprised task codes in a top row, over a bottom row of relational codes. Each speaking turn was given one column, resulting in pattern maps that were made up of two rows and dozens to hundreds of columns (depending on the number of speaking turns), as in the illustration, below: By allowing statements to have a separate relational code, we elevated the relational component beyond positive affect or negative affect to more directly address the cross-party identification dynamics that serve as constraints or opportunities across party representation conditions. Relational acts may serve to define, build, and reinforce the emerging relationship between parties, or they may create separation or conflict between parties.
Using the pattern maps, we identified dynamic regularities that might be missed by frequency counts or markov chain analysis. The pattern maps helped us recognize improvisations by the visual similarities within and differences across their behavioral mappings. After transcripts had been coded and mapped, two authors read the transcripts and independently wrote a story of each negotiation, detailing the tone, content, and rhythm of what unfolded, as if they were describing the negotiation to someone else. Combined with the pattern mapping, these stories helped us identify improvisations and the underlying logics that distinguished them. We first grouped together negotiations that were nearly identical in their maps and stories, then pulled in those that were somewhat more distinct, working in rounds until all negotiations fit into Task codes (Substantive and Procedural) Relational codes one of the sets. We then reviewed the negotiations within each set to assure that those included within a set had more in common with one another than they did with negotiations in any of the other sets.

Improvisations
We identified five improvisations varying in the type, pattern, and complexity of taskrelated and relational acts: Building Relationships; Working Together; Haggling; Asymmetry; and Changing Logics. Summary descriptions of each improvisation are presented in Table 4. Table 4 about here Negotiations characterized by a Building Relationships improvisation focused primarily on the parties' interpersonal relationship. In these negotiations, the parties acted with little apparent concern over the negotiation's economic outcome. Relational acts emphasized bringing the parties together and included few, if any, acts of separation. Information exchange was superficial. Questions, primarily inquiring about the other party's preferences, were couched in language of concern and understanding. Both parties tended toward easy acceptance of the other's offers and often invoked the term -compromise‖ to describe the interaction. Procedurally, there was a notable absence of cycling through issues. Instead, parties relying on a logic of building a relationship dealt with one issue at a time, came to a resolution on that issue, and then moved on to the next. Overall, little effort was extended toward the task, but substantial effort was directed toward the relationship.
Working Together improvisations reflected a logic of honesty, cross-party identification, and problem solving toward a mutually beneficial conclusion. Participants exchanged information couched in empathetic and flexible language. When discussions bogged down, parties sometimes used pressure tactics to push toward an optimal solution, or one of the parties would offer a suggestion for an alternative approach and the other party would accept it.
Openness to discussion and shared understanding were evident throughout. Congratulatory statements punctuated the interactions, especially after a mutually satisfactory tentative agreement was reached and at the negotiation's end. The few separation statements used in working together negotiations were not reciprocated by the other party. . For example, in one negotiation, the logging company representative tried to work with the other party toward a mutually advantageous agreement-providing information about underlying preferences and attempting to bring the parties together-while the town negotiator haggled-simply making offers or counteroffers. Negotiations were asymmetric along relational, procedural, or substantive lines, or combinations of the three. In all of these cases, asymmetric negotiations reflected a lack of shared understanding and unmatched approaches to the negotiation.
Changing Logics improvisations were characterized by multiple shared logics, shifting over time. Early task-related acts and bringing together statements celebrated successes, but these hard-earned first agreements were often followed by mixed relational or substantive messages as negotiators attempted to achieve better outcomes. These shifts often stemmed from and, in turn, added to confusion, frequently resulting in a spiral of acts of separation. Eventually, one or both parties realized that the interaction was being unproductive, and someone would begin the process of bringing the parties back together through positive relational and substantive acts, often after explicit procedural suggestions for a different approach. Any one excerpt from a negotiation exhibiting a changing logics improvisation would be insufficient to identify the underlying flexibility of this improvisation; it was through twists and turns involving a full spectrum of relational, substantive, and procedural acts that the parties reached mutually beneficial agreements.

Party Representation, Improvisations and Outcomes
In addition to inducing and describing the improvisational logics above, we also conducted quantitative analyses of the relationships among party representation, improvisations, and outcomes. First, we examined the effects of party representation on three types of outcomes typically studied in the negotiations literature: 1) the joint value of the outcome achieved across the two parties in the negotiation, 2) the difference between or distribution of the payoff across We also analyzed the association between party representation and the type of improvisation. Because the data are categorical and our number of observations is low, we used Fisher's exact test. As shown in Table 5, the distribution of improvisations was not randomly distributed across the party representation treatments (Fisher's exact = .038).

-----------------------------------
Insert Table 5  Individuals were frequently able to jointly coordinate on a shared logic, working together, despite the fact that participants were strangers entering the negotiation. The modal improvisation for agents, however, was asymmetric, reflecting agents' dual constraints of working within their teams' expectations while simultaneously attempting to develop the relational connection that one-on-one interaction calls for. Team negotiations were less locked into one logic, changing logics as needed to move the negotiation forward or as directed by the social interaction.
To examine how improvisations influence economic and relational outcomes, we conducted one-way ANOVAs. These omnibus tests, presented in Table 6, showed that improvisations significantly affected difference in payoffs, trust, and perceived competitiveness of the interaction. Pairwise comparisons revealed that haggling led to significantly more unequal payoffs than did building relationships, working together, or changing logics. Tests of effects on attitudes revealed that those who engaged in haggling trusted the other party less than those who were working together. Parties who engaged in building relationships perceived the interaction as less competitive than those involved in haggling, changing logics, or asymmetry.

DISCUSSION
How do contextual features within organizations affect negotiation process and outcomes, and how do these processes in turn become a source of or solution to interorganizational conflict? A theory of improvisations as logics guiding micro-interactions offers a framework for understanding the link between the unfolding interaction that has been the subject of considerable negotiation process research and more macro features of organizations, such as formalization of roles, culture, or party representation. We have shown that emergent improvisations in negotiations reflect the efforts of negotiators to manage the constraints and opportunities of their party representation. The resulting negotiations can be not only a solution to, but also a source of, inter-organizational conflict. This insight may guide future studies examining how other organizational features affect negotiation processes and outcomes, including conflict between and within organizations.
The modal improvisations across individual, team, and agent negotiations reveal the parties' attempts to manage the constraints and opportunities inherent in each representation condition. In one-on-one negotiations among principals, the parties were able to jointly coordinate on a working together improvisation, despite the fact that they were strangers entering the negotiation. For individuals negotiating for themselves, the working together improvisation reflects the immediate relational imperatives of one-on-one interaction. Individuals privileged the relationship while simultaneously attempting to communicate their own interests. This allowed them to push for better economic outcomes without falling into a negative relational spiral. While individuals were constrained by their limited information processing capability, they faced neither the confusion and information complexity facing teams, nor the conflicting allegiances facing agents. As a result, principals negotiating for themselves tended to engage in working together improvisations, resulting in more equal payoffs without sacrificing joint gain and relatively high levels of trust.
In contrast, negotiations involving agents were most frequently characterized by asymmetric improvisations, reflecting the dual constraints of agents attempting to work within their teams' expectations, while also attempting to develop the relational connection that one-onone, face-to-face interaction elicits (Sally 1995). Asymmetric negotiations often involved one party exploiting the other with a haggling improvisation while the other utilized a building a relationship or working together improvisation throughout. Agents seemed constrained in their ability to manage the trust-testing process needed to develop cross-party identification and trust, and conflict emerged during interaction. Further, agents lacked the information processing capability that enabled teams to develop a shared understanding of the negotiation despite the complexity of the substantive and relational exchanges and potential constraints of within-team bias.
The modal improvisation for teams, changing logics, responds to the relational, substantive, and procedural constraints of this representation structure. Team negotiations changed logics as needed to move the negotiation forward or as directed by the social interaction.
Team interactions reflect the potential constraints on developing cross-party trust and the information complexity inherent to their party representation structure. Negotiations between teams were likely to draw on multiple perceptions, working styles, and approaches. In some negotiations, team members adopted differentiated social or task-related roles as needed in the negotiation. For example, as one team member pushed hard for a certain outcome, another team member offered group identification statements that helped bring the parties back to the table to offset a separation cycle. The changing logics improvisation reflected teams' ability to develop a shared understanding, while also remaining flexible enough to adapt if the negotiation devolved into conflict. Although changing logics allowed teams to achieve relatively high joint gains and balanced distributions, changing logics did not result in the positive relational outcomes of those engaging in the working together improvisation.
We suggest that a focus on both micro-interaction and macro-improvisational logics provides a more robust and dynamic theoretical frame than those that privilege the frequency of certain actions (Pruitt and Lewis 1975;Taylor 2002;Thompson and Hrebec 1996;Thompson 1991), the response-cue relationships among acts and interacts (Bazerman et al. 1992;Taylor and Donald 2003;Weingart et al. 1999), or gestalt improvisations (McGinn andKeros 2002). Multilevel analyses also draw attention to critical transition points in the negotiation (Druckman 2001;, such as when actors diffuse or exacerbate a defensive exchange through closure or process statements . Such transition points were a hallmark of the changing logic negotiations, which involved movement through logics over time.
The approach proposed here emphasizes the relational aspects of negotiation processes and outcomes and offers a coding scheme for capturing relational aspects of micro-interaction.
We were able to tease out improvisations focused solely on the relational aspect of the interaction, others that balanced the relational and substantive aspects, and those that concentrated on the substantive aspects of the negotiation to the exclusion of relational issues.
Further research could investigate how a lack of, or focus on, the relational dimension of negotiations affects negotiated outcomes, conflict, and ongoing working relationships embedded in organizational contexts.
A theory of improvisations as guiding logics for micro-interaction offers a useful methodological and theoretical framework for future studies seeking to examine how organizational features may affect negotiation processes and outcomes, and in turn, how negotiation processes may be a source of, or solution to, inter-organizational conflict.  Kurtzberg et al. 1999), breadth of expertise, working styles, and approaches (Jackson 1992) are particularly relevant to our understanding of negotiations across and within organizations. Considering these features when designing and carrying out negotiations may enable organizations to guide the interactions so that they are a solution to, rather than a source of, conflict. Wilson, S.R., L.L. Putnam. 1990. Interaction goals in negotiation. Communication Yearbook 13 374-406.  Payoff -You will be paid $1 for every $75,000 you earn over your minimum.‖ -Your group outcome will be the outcome of the team member who has the lowest outcome. Each member of your team will receive the same outcomeyou will be paid $1 for every $75,000 the lowest payoff agreement earns over your minimum.‖ -Each member of your team will receive the same outcomeyou will be paid $1 for every $75,000 your team earns over your minimum.‖