Thirty‐Five Years of Research on Turning Points: Insights Gained and Gaps to be Filled Open Access

A special issue provides an opportunity to take stock of the research that has been undertaken to date and develop questions for the next steps in the program. It is also an opportunity to reflect on progress since the last Negotiation Journal special issue on critical moments was published in 2004. I begin with a brief historical discussion of how the turning points (TPs) concept¹ was discovered followed by insights into how negotiators monitor the conflict resolution process and make adjustments as well as how they respond to crises in and around the talks. I then review an analytical framework constructed for a comparative study of international and domestic negotiations. More recent extensions of that framework are also presented. These research insights suggest paths for developing a theory of TPs, known also as negotiation departures. Promising ideas take the form of questions to be addressed in future research in order to move closer to a theory that explains how TPs occur and their consequences.

My discovery of the TPs concept came from two sources. One was a collection of transcripts of a 1975 negotiation between Spain and the United States over military base rights. I noticed that the talks alternated between periods of calm and eruptions that threw a wrench into the process. At times the eruptions brought a halt to the discussions, sending the delegates home (either to Washington or Madrid) for further planning. Interestingly, each impasse was resolved and the talks moved forward toward an eventual settlement (Druckman 1986, 2019b).

The second source was a 1972 book chapter by Charles McClelland. The turning‐points discovery made during my transcript reading motivated me to search the literature for relevant research. McClelland’s work was particularly applicable. His time‐series analysis of the Taiwan Straits conflict showed that crises were an impetus for progress in resolving the conflict. His finding comported with my observations of the base‐rights process. His time‐series approach struck me as suitable for analyzing the negotiation. The result was evidence for a relationship between crises and TPs. It was also the impetus for a line of research that included additional case studies and simulation experiments. In the sections to follow I summarize key findings obtained from these studies.² Each section concludes with a summary of the key insights obtained from the research.

The definition of TPs in the early work was related closely to the idea of negotiating stages—“as benchmarks signaling passage from one stage of the negotiation to another” (Druckman 1986: 356). The benchmark events consisted of agreements on an agenda, on a framework or formula, on negotiable issues, on final agreements, and on implementation provisions. These events provided a structure for an analysis of the give‐and‐take during the process of negotiation. It is also the case, however, that the definition of TPs evolved as the research moved to other types of cases and laboratory simulations.

The primary objective of the base‐rights analysis was to explain how the process produced these TPs. Following McClelland (1972), I performed a time series of conversations between the delegations. The conversations were coded according to the categories of a system known as bargaining process analysis (Walcott and Hopmann 1979). The codes consisted of both soft verbal statements (initiation, accommodation, praise, and promise) and hard verbal statements (retraction, commitment, threat, and accusation); all produced acceptable inter‐coder reliability. These analyses revealed a particular kind of process. Negotiators monitored the respective statements or moves (including offers and counteroffers) and made decisions based on these comparative evaluations. A key distinction made by the negotiating delegates was between two types of responding to the other delegation’s moves: synchronous and asynchronous. By synchronous responding, I refer to statements that match in a tit‐for‐tat exchange. Asynchronous responding consists of statements that do not match such as an initiation (new offer) countered with a retraction or accusation. Alternations between matching and mismatching were noted with particular attention paid by the negotiators to moments when a large gap between them occurred. In the base‐rights study, these moments caused impasses when the softer negotiating team adjusted its moves in the direction of the harder team (Druckman 1986). Similar patterns were found in an earlier laboratory study (Druckman and Bonoma 1976) and in a comparative case study of responsiveness (Druckman and Harris 1990). In each of these studies, negotiators were alert to a “hardness gap” that often produced a negotiating crisis that precipitated a TPs.

A more general insight from these findings is the connection between monitoring and TPs. My earlier writing about the monitoring function focused on the kind of diagnosis captured by Coddington’s (1968) expectation‐evaluation‐adjustment cycle (Druckman 1978). Expectations are benchmarks often defined in terms of fairness. Evaluations are appraisals of the other negotiator’s moves in relation to fairness criteria. Adjustments are the changes made based on the evaluations. Of particular interest is the adjustment from offers or statements that are mismatched (which I have referred to as a hardness gap) to those that are matched (reducing the gap). This is the kind of adjustment that produces an impasse with recalcitrance on all sides. Many of these impasses lead to breakdowns in the talks with all sides hunkering down for “battle.” The breakdowns motivate actions to recover from the impasse. These actions are regarded as TPs, positive when they break the impasse, negative when they perpetuate it.

I called this process “threshold‐adjustment” (Druckman 1986). Some studies on arms control talks referred to the process as comparative reciprocity (see Stoll and McAndrew 1986; Druckman and Harris 1990; Patchen and Bogumil 1995). The threshold triggers changes in evaluations that lead to changes in actions. For example, when a large difference in offers is noticed, or a threshold is reached, the “softer” negotiator adjusts her moves in the direction of the “harder” opponent. Levine’s (2014) truth‐default theory captures a similar process with regard to detecting deception. Apparently, people need a trigger to snap out of their tendency to default to their expectations about whether the other is telling the truth. They give the other a pass until something dramatic happens. In his recent book, Gladwell (2019) provides a number of compelling real‐world examples of this pattern: Fidel Castro’s duping of the CIA, Hitler’s deception of Chamberlain, Bernie Madoff’s fraud, and the pedophilia trail of former Penn State coach Jerry Sandusky. Each of these examples evinced a conservative bias: People were willing to suspend judgment until the evidence was too overwhelming to ignore. We recently faced a similar process in the partisan impeachment trial of Donald J. Trump: What was needed to convince Republicans to allow witnesses? The needed evidence is the trigger that alters evaluations leading to changed actions that can be construed as TPs.

These dynamics take place inside negotiations. They describe the way negotiators respond to each other’s moves, counter‐moves, and statements over the course of their interactions. An important feature of this pattern is the difficulty of shifting gears or changing an ongoing exchange trend. There is an apparent reluctance to alter expectations, which often are established over a long period of time. Reputations also come into play; Sandusky and Madoff were regarded as pillars of their respective communities. It takes a jolt to unlock these perceptions. One jolt discussed above is overwhelming evidence presented during a legal trial. Another is a crisis caused by continued recalcitrance on all sides as the stakes for agreement increase. The perceptual changes or departures triggered by these jolts or crises are regarded in our framework as TPs.

These analyses add up to a more general observation about interactive processes and TPs. Negotiators are motivated to close gaps between statements and moves made by both (all) parties. These adjustments are in the direction of synchrony or matching as shown in the study of eight international negotiations by Druckman and Harris (1990). When the adjustment is in the direction of matching hard behavior an impasse occurs. Attempts to resolve the impasse may place the talks back on track, producing a TP. TPs are less likely to occur when negotiators do not respond or adjust to asynchronous behavior. The out‐of‐sync pattern may be regarded as an opportunity for a TP.

A relationship between negotiating crises and TPs was also discovered in the context of the base‐rights case study (Druckman 1986). The parties’ inability to see a way out of their mutual intransigence motivated them to change course. With limited options for extricating themselves from this situation, they turned the talks over to their respective foreign ministers. The ministers were authorized to alter foreign policy prerogatives, make major procedural changes, and craft framework agreements beyond the mandates that guided the negotiating delegations. Their decisions placed the talks on a trajectory toward resolution of the conflict. However, it took a larger crisis to seal the agreements. Franco’s death in December of 1975 threw the Spanish delegation into chaos and desperation, leading them to concede to the US demands.

A similar dynamic occurred in the 1987 talks between the U.S. and the Soviet Union over reductions in intermediate nuclear weapons (Druckman, Husbands, and Johnston 1991). The negotiators were stuck in their inability to find a formula to guide reductions. Presidential leadership was critical in breaking the deadlock. The leaders redefined the scope and focus of the negotiations, including the exclusion of French and British nuclear forces, the unlinking of the negotiations to progress in the parallel START talks, and an agreement to exclude from the negotiations any discussion of strategic and space weapons. By limiting the range of issues, these leaders made the agenda manageable. Their involvement led to an agreement that not only reduced nuclear weapons but also catalyzed Gorbachev’s unilateral force reductions and the end of the Cold War.

Both these cases—the base rights negotiation and the talks to limit intermediate‐range nuclear forces (INF)—provide examples of how issue framing and reframing resolve crises. These cases set the stage for later experimental research on crises and TPs. Focusing on crises that occur outside the negotiation, Druckman and Olekalns (2013) examined decisions made by laboratory negotiators at the moment of a crisis resembling the event of Franco’s death in the base rights case. The negotiators were given three choices: reach an immediate agreement, continue negotiating, or reframe the issues. These choices were made in the context of high or low transaction costs, level of mutual dependence (attractive or unattractive alternatives), and a shared or unshared identity. The results showed that mutual dependence led often to reframing choices. This relationship was amplified by low trust: When distrusting negotiators had no place to go, they opted to think differently about possible outcomes. Similar to the desperation experienced by the base‐rights and INF negotiators, these bargainers sought a way out by reconceptualizing the issues. The laboratory case differed, however. It did not involve leaders with the authority to alter the framing of the issues and the high stakes of foreign policy decisions.

Taken together, the base rights, INF, and laboratory cases suggest that TPs are often motivated by a sense of desperation stemming from either an internal or external crisis. But they may also be motivated by a sense of opportunity that results from a more trusting social climate (Druckman 2019a). More generally, I have shown that the crisis–TP relationship occurs in a wide variety of international negotiations: De‐escalatory moves typically followed escalatory moves in those cases (Druckman 2001).

Although the empirical relationship between crises and TPs may be stated as a general proposition, it may also be contingent on several factors. One is the type of crisis as, for example, a discrete temporal event or an unrelenting impasse. Another factor is whether negotiators are motivated by desperation (no way out) or optimism (perceived opportunity). Higgins’s (1998) distinction between a focus on prevention or promotion is useful in this regard. A third factor is a willingness of negotiators or their principals to reframe the issues in the direction of making them negotiable. Each of these factors can produce TPs in the context of crises.

The decisions made in the earlier cases had several features that helped to move the talks toward an agreement. They departed from previous events, which were often sticking points in the discussion. They occurred with varying degrees of abruptness, and they had consequences for both parties. These features were a basis for a framework that organized a comparative analysis on TPs in thirty‐four cases.

The framework incorporates a slightly different definition of TPs. The key idea is abruptness. Abrupt changes were regarded as sudden departures from a pattern or unexpected transitions from one stage to another. Less abrupt changes “include new proposals … that alter the discussions somewhat or adjust the terms of trade and somewhat predictable stage transitions” (Druckman 2001: 527). In my article in Negotiation Journal’s 2004 special issue on critical moments, I added two more features to the framework. One was that the changes are clear or self‐evident to both the negotiators and the analysts. Thus, a TP requires acceptance of the move in order to be a TP. When one party seeks a departure but is rebuffed, a TP has not occurred. Another feature was that actions taken by one of the parties result in consequences for both (all) (Druckman 2004).

The three parts of the framework are precipitants, departures (TPs), and consequences. This is a linear trace of events, decisions, or moves taken during the process. Each of the thirty‐four cases, divided into security, trade, and political or environmental negotiations, was analyzed in terms of these process traces. A key finding was a difference among the three issue areas: security talks were characterized by external precipitants, notably mediation, and abrupt departures; the other types of cases were characterized by internal precipitants, either procedural or substantive. Security negotiators depended less on their own initiatives than on external interventions. They were more sensitive to possible losses than to potential gains (see Neale and Bazerman 1991). As noted by Druckman (2001), this negotiating pattern hinders the development of the sorts of cooperative regimes prevalent in the trade and political domains. (See also Jervis 1982 on difficulties in establishing security regimes.)

A question arises as to whether there are ways to structure security negotiations to reduce the perceived risks of possible agreements. One idea is to precede the talks with joint fact‐finding tasks. Another is to bolster the verification procedures as part of the implementation. A third is to reframe the issues by reducing their number, as was done by the US and Soviet leaders in the INF talks. Whether reframing can be done by the negotiators themselves or depends on actions taken by their principals remains to be learned.³ More broadly, the relationship between taking risks in negotiation and the formation of international regimes provides a window into the way that micro‐level processes interact with macro‐level institutional level structures. We will return to these levels in the next section.

The TPs framework has been applied in several other case study contexts. Focusing attention on global conferences, Crump and Druckman (2012) analyzed intellectual property rights in two phases of multilateral trade negotiations: the Uruguay round of the Agreement on Tariffs and Trade (GATT) from 1986 to 1993 and the Doha Declaration negotiated within the World Trade Organization (WTO) between 2001 and 2003. Crump and Druckman (2012) analyzed these negotiations in three ways. One analysis focused on the role played by reframing issues in order to encourage bargaining and problem solving. The reframing was regarded as a substantive precipitant in our framework. Another analysis accepted the legitimacy of the new framing, a departure in our framework. A third analysis focused on the conflict management roles played by coalitions and how they led to an eventual agreement at the Uruguay round. These coalitions were regarded as the glue that connected the three parts of the framework, precipitants to departures to consequences. This process worked better in the earlier successful Uruguay round than in the later failed Doha Declaration talks.

The authors then studied the same two multilateral trade cases with a focus on their organizational structures: a Committee‐level structure in GATT and a Ministerial/Council‐level structure in the WTO (Crump and Druckman 2016). The TPs framework was a useful analytical device for understanding the impacts of these structures. The two structures differed chiefly with regard to the type of their precipitating factors, with procedures dominating the Ministerial‐level WTO negotiations and substance dominating the Committee‐level GATT talks. The more centralized authority of a Ministerial structure takes procedural discretion away from delegates. As a result departures are largely determined by changes orchestrated by the Ministers. The less centralized Committee structure provides more discretion to delegates. As a result departures occur when new ideas emanate from the membership. This finding adds a structural dimension to the TPs framework. It also suggests a more general proposition that takes the following form: Conference structures influence the locus of change (TPs) as from the top down or from the bottom up.

Third parties were introduced as an important feature of the framework constructed by Hall (2014). His analysis of twenty‐nine cases of domestic environmental negotiations showed that third parties were instrumental in creating procedures that triggered departures during the early stages when the processes were more flexible or not narrowly prescribed. The negotiating parties themselves had more influence over the substantive decisions that led to agreements. Of particular interest was Hall’s comparison with a set of eleven international environmental cases analyzed earlier by Chasek (1997). In those cases, substantive precipitants occurred early and frequently without the help of third parties. Unlike the domestic cases, international environmental talks are managed with prescribed procedures, allowing the parties more time to focus on substance during the early phases. Procedures come into play during ratification processes that follow the negotiations. These findings add a contextual dimension to the framework. (See also Druckman 2001, for another domestic‐international comparison of the way that TPs occur.)

More recently, an analysis of the 1989 Round Table talks in Warsaw, Poland by Druckman, Bulska, and Jochemczyk (2019) provides another application of the TPs framework. These historic talks ushered in a new era of democratization in Poland. Each of the seven key issues was analyzed in terms of tracing the three‐part TPs process. A summary trace across the issues consisted of substantive + procedural precipitants to abrupt departures to de‐escalatory consequences. More importantly, this case illustrates a strategy for bringing about a departure with implications for the eventual settlement. Rather than deal with the large issue of regime transition, the delegations focused on smaller chunks of societal reforms that would make progress toward democratization. By discussing the key issue of free elections first, the delegates paved the way for agreements on other issues (mass media, role of the president, rehabilitation). By postponing the most difficult issue—referred to as indexation—until the end, the delegations were able to walk away with a near‐comprehensive agreement. The indexation issue was not resolved but three factors made the agreement possible: high stakes, unattractive alternatives to reaching an agreement, and severe time pressure. The delegations’ choice was to let go of this big issue or suffer the substantial consequences of political instability. Under these conditions, the TP is the agreement that was precipitated by the decision to postpone the key sticking point.

These studies contribute to an expansion of the TPs framework by taking into account organizational structures, domestic and international contexts, and the ways that issues are sequenced. They shape the kinds of precipitating factors—either external such as context or internal such as issue orders—that produce departures during the negotiation process. Each of these factors also ramifies the consequences of departures and agreements to a larger policy context. The changes that are wrought have implications both for the way that the negotiation unfolds and for the societies in which the talks take place. A general insight from these studies is that while TPs are observed during the course of negotiating they are shaped by the institutional or organizational settings within which the talks occur.

An important challenge is to chart directions that move us closer to an explanatory theory of TPs. One direction comes from the levels‐of‐analysis approach discussed in my contribution to the first special issue on critical moments (Druckman 2004). In that article, I probed the psychological and social processes that were considered as impetuses for departures.

At the individual level, I concentrated on the way negotiators keep track of developments, how they attend to cues, and how they notice moments (or thresholds) that signal change. The earlier discussion of threshold‐adjustment provides some clues on directions for theory. Three questions are suggested: (1) How do negotiators attend to the exchanges? The tennis metaphor rings true—do they keep their eyes on the ball? (2) How do they detect changes, particularly those that defy expectations or pass a threshold during the give‐and‐take phases? (3) How do they react to changes when they occur? The kind of individual‐level theory suggested by these questions is along the lines of Coddington’s (1968) expectation‐evaluation‐adjustment cycle. The cycle was modeled by Coddington and developed further in my article on boundary roles (Druckman 1977). The way negotiators adjust or react to their evaluations brings us to probes at the next level of analysis.

At the level of interactions, I described the processes of adjusting to moves, of reciprocity or synchronicity, and of coordinating in the face of competitive jockeying for advantage. Several questions are suggested. How do we capture a dyadic (or multiparty) process? Another tennis metaphor is relevant: Watching the American professional doubles players (and twin brothers) Bob and Mike Bryan play a match, I heard their coaching team shout “synchronicity.” This occurred at a rare moment in the match when the players appeared to be confused. The coaching team diagnosed the problem as being “out of sync.” They urged their players to return to a familiar routine of working together in a harmonious way. The theoretical question is: How does a team (including negotiating opponents) switch from one state (out of sync) to another (in sync)? The switch in either direction can be considered as a critical moment during the competition or interaction. The answer to the question of moving in and out will not come from an analysis of individual team members or delegates. It will come from understanding emergent processes as these were analyzed in early work by the social psychologist Muzafer Sherif (1956).

It is also the case that synchronous exchanges can be mutually tough or soft. They can orient a process toward or away from the agreement. With regard to the latter, negotiators mimic each other’s tough postures, similar in some ways to an arms race. A practical question is how they manage to unlock themselves from this deleterious behavior. Perhaps the most prominent proposal was to take a unilateral initiative. Offered a long time ago by Charles Osgood (1962), this initiative has had mixed evidence (e.g., Ramberg 1993). Nonetheless, the initiative may precipitate a TP with important consequences for the future of relationships. It raises a pair of theoretical questions: How do we understand the way negotiators lock into mutually recriminating exchanges? How do we know when opportunities arise for breaking out of these patterns? Relevant theoretical leads are likely to come from work on conflict escalation cycles and on cooperation in the face of competitive wrangling. Going further than the current research, it would be interesting to examine tactical moves made by mediators or negotiators to prevent further escalation, which could precipitate the sort of stage transitions that are regarded as TPs.

At the collective level, I treated negotiating delegations as collectivities with networks, rules, and organizational cultures. When thought about in this way, departures (TPs) are construed as changes in a system rather than in the negotiators or their interactions. A focus on systems is particularly germane to negotiations conducted in institutional contexts such as the World Trade Organization or the United Nations Committee on Disarmament. The research by Crump and Druckman (2012, 2016) discussed earlier illustrates how institutions impact negotiation processes. These studies made evident a rule structure that constrains the behavior of delegates and provides predictability. With regard to the TPs framework, departures during the process are likely to be triggered by procedural rather than substantive events. They also suggest a different kind of theoretical approach guided by the following questions: How do institutional structures shape the way negotiators make decisions? What are the critical differences between institutional and noninstitutional negotiations? Where are the junctures that provide opportunities for changing the negotiation process? Underlying these questions is a larger theoretical issue prominent in the current landscape of international relations: What is the relative importance of human agency and system determinism in negotiations generally and in institutional negotiations in particular? (See also Delanty and O’Mahony 2002, for a treatment of systemic constraints on agency in the context of national collective action and social change.)

Each of the three levels provides a path for theory development. For individual negotiators, a key explanatory process resides in the idea of threshold‐adjustment. Attending to moves made by one’s own and the other delegation, negotiators are sensitive to changes that invoke fairness considerations. The TP occurs when a threshold is realized; the threshold motivates adjustment as captured by the interactional level of analysis. For interacting negotiators, the key concept is synchronicity. Working in concert, negotiators strive to balance their moves in a kind of equilibrium where out‐of‐sync moves are restored. The TP occurs at the moment when synchronization or coordination occurs. For large delegations engaged in multilateral talks, a key theoretical issue is the way structures influence processes. Operating under institutional constraints in the form of rules or norms, delegations have reduced agency, causing them to respond to conference procedural changes. The TPs occur when changes lead to departures during the process.

The levels approach adds complexity to the development of TPs theory. The three types of theories discussed—monitoring, synchronicity, and agency/structures—can be regarded either as separate kinds of explanations or a more general synthesized explanation. The former aligns with disciplinary approaches to negotiation such as psychology (individual), social psychology/sociology (group), and international relations (collective). Much of the literature is segmented into these disciplinary approaches. The latter may be a more substantial challenge where micro‐ and macro‐level processes are linked. The linking can go up or down a latter of complexity. The “going up” perspective focuses attention on the way that negotiating delegations influence policies. The “going down” perspective concentrates on how policies and structures shape negotiating delegations’ decisions. Both directions bring us closer to an integrated theory of negotiation TPs.⁴ We have begun to work on this challenge while also engaging in a pursuit of the research gaps that remain.

The theoretical ideas discussed in the previous section are closely related to research findings. At the level of individual negotiators, there is much to be learned about the monitoring function. The threshold‐adjustment model highlights a comparison of offers and counteroffers. We know that negotiators engage in this process but know less about what specifically they are attending to and the consequences of the adjustments that they make. For example: What are these judgments like in the moment? How do they distinguish small and large discrepancies between what a negotiator gives and receives? What is meant by “moves”—offers, demands, pronouncements, proposals, postures? These questions can be addressed with experiments or case studies. For either methodological approach, the key challenge is how to generate the appropriate perceptual information. Face‐to‐face interviews or survey questions may fit the bill. We can also consider the think‐aloud protocols pioneered by Ericsson and Simon (1980) or the ecological momentary assessments used by some clinical psychologists (Schiffman et al. 2008) supplemented by baseline‐sensitive content analyses of the conversations held during the process of negotiating.

At the level of interactions, the idea of synchrony merits further research (see also the Donohue article in this issue). We noted that being in sync can encourage coordination toward agreement or enduring impasses that stymie progress. We learned in earlier research that disappointed expectations lead to synchronized toughness and reduced concession making (Druckman and Bonoma 1976). We also learned that when impasses are severe, negotiators stop the process and entertain issue reframing to get the process back on track (Druckman and Olekalns 2013). These findings suggest new research questions: What are the precipitating circumstances for synchronized toughness or coordination? What strategies do negotiators use to unravel the deadlock (mutual toughness) or continue on a synchronized path toward agreements? How does a switch occur from toughness to coordination? The last question has particular relevance for TPs research and is discussed also in the Donohue article in this issue. Also relevant is the idea of developing connections outside of formal negotiations (see the Kolb article in this issue). With regard to methods, the old tradition of research on process analysis may prove helpful for documenting patterns and transitions through the course of negotiation (see, e.g., Walcott and Hopmann 1979).

At the collective level, we focus attention on multilateral negotiations that occur within institutional contexts. Here issues of structure and agency come into play. In my research with Larry Crump discussed earlier, a notable finding was how a structural difference (Ministerial Councils‐ versus Committee‐level structures) shaped the way TPs occurred. Delegates to these conferences had limited latitude for influencing key decisions. Several research questions are suggested: At what junctures during the conference do delegates or delegations have more or less influence over decisions? Is the agency influenced by the size of the conference? Is there more or less agency in longer‐term, more institutionalized conferences (e.g., WTO versus bilateral or multilateral free trade talks)? With regard to methods, observing conferences or reading transcripts of proceedings would be advantageous. Once access to conferences or transcripts is obtained, the next challenge is to devise coding systems for depicting how influence occurs. With a coding system in hand, a researcher would benefit from recent developments in social network analysis and comparative case methods.⁵

The overview of turning‐points research provided in this article is intended to look back on what has been accomplished and forward on what remains to be done. With regard to looking back, I described the findings obtained on monitoring and adjustments, on the relationship between crises and TPs, on issue areas in comparative case analyses, and on the role of institutions and domains in shaping negotiation processes. A key insight that emerges from these findings is that TPs can be understood through the lenses of different levels of analysis. They occur because negotiators attend to cues that tip off a change in the flow of offers and counteroffers. But the change is also due to negotiators’ reactions; rather than being passive observers or analysts, they are the agents of change. And yet the change depends on coordinating their own reactions with those of other negotiators. This sort of bilateral/multilateral agency is limited in more complex negotiations where the context intrudes on the process. The multilevel perspective, developed from more than three decades of research, is a step in the direction of synthesizing our knowledge about TPs. It is also a step toward the development of an integrated theory.

Looking forward, I discussed the challenges for theory development and the research gaps that remain to be explored. Two paths to theory development are possible. A more straight and narrow path would remain within the confines of the three levels of analysis. Theories of monitoring and adjusting, of synchrony as an emergent process, and of the way that structures impact on agency can be developed separately. A more inventive approach would seek liberation from the levels concept and work in the direction of an integrated theory. Such a theory would explain the way that monitoring and coordination work together within more and less institutionalized contexts. Both theoretical paths depend on filling research gaps about the meaning of moves, documenting transitions during the negotiation process, and discovering the way that negotiators find opportunities for taking initiatives within larger institutional (or bureaucratic) structures. The challenges ahead include crafting innovative research methods such as momentary assessments, new process coding systems, and network analysis.