Escalation

In its original sense, “escalation” means movement from less extreme to more extreme contentious tactics by one or both parties to a conflict. We can also speak of “level of escalation,” which means the severity of contentious tactics that are being used in a conflict.

Movement from less escalated to more escalated tactics often goes through a series of intermediate stages, in what can be called an “escalation sequence.” There are two types of escalation sequence, unilateral and bilateral (Pruitt, 2005a). In unilateral sequences, only one party escalates, whereas in bilateral sequences the parties escalate in tandem.

Unilateral Escalation

In a laboratory experiment on unilateral escalation sequences, my students and I (Mikolic, Parker, and Pruitt, 1997) studied reactions to persistent annoyance by having confederates of the experimenter withhold supplies needed by the participants to complete a project. The data were based on a content analysis of telephone messages to the confederates. Most participants tried to get the supplies by means of the following orderly progression of tactics: requests, demands, angry statements, threats, harassment, and abuse. Escalation stopped at different points in this progression. Some participants made only requests; others, requests and then demands; still others, requests, then demands, then angry statements, and so on. Groups, on the whole, escalated farther than individuals—following the same orderly progression of tactics.

This study raises several questions requiring further research. The most basic question concerns why persistent annoyance so often produces escalation. There are at least three possible answers to this question, which need empirical testing.

A second question concerns how to account for the orderly sequence of tac­tics seen in this study. Perhaps it reflects a widely shared escalation script. Or there may be increasing levels of inhibition associated with the tactics in this sequence, so that as moral exclusion develops, they are accessed in order. A third question concerns the generality of these findings. We know that severely escalated tactics are sometimes adopted without going through intermediate lev­els. At one moment, the Japanese are making demands in a negotiation; at the next moment, they are bombing Pearl Harbor. The challenge is to identify the conditions that promote a precipitous escalatory jump as opposed to an orderly escalative progression. A fourth question concerns the level at which parties stop escalating. Why did groups escalate farther than individuals in our study? And what conditions or personality traits caused some of individuals to stop escalating at the level of demands while others went all the way up to threats?

These questions are worth answering because unilateral escalation is a com­mon phenomenon in social conflict. Research to answer them can probably use our laboratory paradigm for a while longer, but generality studies in other set­tings will eventually be needed.

Processes Involved in Bilateral Escalation

Bilateral escalation sequences usually develop through conflict spirals entailing repeated retaliation and counterretaliation, or defense and counterdefense. Thus, a frustrating management policy might produce criticism by an employee, pro­voking disciplinary action such as failure to grant a raise.

It is possible for conflict spirals to go round and round without advancing in level of escalation. For example, I yell at you... you yell at me... I yell at you... you yell at me. Contentious tactics are being used, but they are not getting any heavier. This raises an important theoretical question: under what conditions and through what processes do conflict spirals produce bilateral esca­lation rather than simply going round and round?

One possible answer to this question is that bilateral escalation occurs if the participants in a conflict spiral see the situation as one of persistent annoy­ance. Escalation will then occur for one or more of the reasons suggested in the last section. For that to happen, each party must see the other as respon­sible for every new round of the conflict spiral, a phenomenon called biased punctuation. An objective observer would see the conflict spiral as A→B→A→B→A→B→A, where A is party A’s contentious behavior, B is party B’s contentious behavior, and the arrows denote perceived causation. But the parties punctuate the sequence differently, with each of them seeing the other’s input as the driving force in every round. Party A sees the sequence as A, B→A, B→A, B→A; and Party B sees it as A→B, A→B, A→B, A. In addition, bilateral escalation may sometimes occur if only one party engages in biased punctuation, provided that the other party matches the first party’s level of escalation at each new round of the spiral.

Biased punctuation has often been discussed by conflict theorists (for exam­ple, Kramer, 2004), but research on this topic appears to be nonexistent. If this is indeed a source of bilateral escalation, we need to know the conditions under which it occurs and the processes by which it develops.

Another possible answer is that bilateral escalation occurs when each round of the spiral produces structural changes in one or both parties or in the com­munity surrounding them.

Some structural changes are in the psychological realm. Hostile attitudes and perceptions set in, trust takes a nose dive, and new, more competitive goals develop. It becomes important “to look better than, punish, discredit, defeat, or even destroy” the other party (Pruitt and Kim, 2004, p. 109).

If groups are involved in a conflict spiral, changes may also occur in nor­mative and social structures within the groups. Hostile attitudes, perceptions, and goals are accentuated by group discussion, through the process of group polarization. (See Isenberg, 1986.) And these psychological states often become group norms, which are perpetuated by the processes of norm enforcement. Furthermore, groups may become more mobilized as a reaction to continued conflict; and it is often hard to put the genie back into the bot­tle once strong group identities are formed, group grievances are crystallized, group leaders emerge, or activist subgroups are formed. Group mobilization is, first and foremost, a contributor to the development of intergroup conflict and, hence, may at times lead to problem solving rather than escalation. But it often leads to escalation, especially when militant leaders and subgroups emerge.

Structural changes may also occur in the community surrounding the par­ties. In continuing conflict, community members are often pulled into one side or the other of the controversy, a phenomenon known as “community polar­ization.” The resulting social support reinforces both sides’ sense of grievance and hostility toward the other side and can strengthen their capacity for aggres­sive action, encouraging escalation. Community polarization also erodes the ranks of neutrals who might otherwise try to dampen the two sides’ tactics and mediate the conflict.

Conditions That Encourage Bilateral Escalation

An important question remains: under what conditions will the processes just described go forward and produce bilateral escalation? Other related ques­tions concern determinants of the rate of escalation and how severe it will become.

Some testable answers to these questions can be derived from social psy­chological research on the antecedents of aggression. (See Berkowitz, 1993.) For example, people who have been annoyed become more aggressive when they have been recently angered, are autonomically aroused, are under the influ­ence of alcohol, or are so rushed that they don’t have time to think. They become less aggressive when they are in a good mood or are engaged in com­peting activities. Impulsive individuals and those with high, unstable self-esteem are easily provoked, while people with strong needs for social approval are not. All of these conditions are potential contributors to, or detractors from, bilat­eral escalation, especially if found on both sides of the conflict.

Relationships between individuals are also important. Several studies (for example, Bradbury and Fincham, 1992) have shown that people in distressed marriages are prone to retaliate when annoyed. Since retaliation is at the base of conflict spirals, this suggests that bilateral escalation should be greater in dis­tressed marriages and other hostile relationships, a testable hypothesis.

Positive relationships presumably have the opposite effect. An old study by Back (1951) adds an intriguing twist to this speculation. He found some evi­dence that more cohesive groups (involving more positive relations between the members) engage in more conflict but are less prone to escalation. This suggests that strong, positive interpersonal relationships provide enough security that people feel free to raise issues with one another but they inhibit the use of heavy contentious tactics.

Although our study (Mikolic, Parker, and Pruitt, 1997) showed that groups move to higher levels of unilateral escalation than do individuals, it is not clear that they are more prone to bilateral escalation. Groups tend to be more insight­ful about solving the problems they face, which suggests that they may be less prone than individuals to fall into the trap of biased punctuation. Clearly, research is needed on whether groups become more involved or less involved in bilateral escalation.

Pre-existing community structure may also be important in determining whether communities hold together or polarize in the face of conflict between subgroups. The issue is whether the bonds between community members are predominantly crosscutting or overlapping (Kriesberg, 1973). In a heavily cross­cutting structure, important members of most subgroups are bonded (for exam­ple, are members of the same clubs or political parties) to members of most other subgroups. In a heavily overlapping structure, subgroup members are bonded only to other members of their own subgroup.

There has been some research on this issue. Comparative case studies by Coleman (1957) and Varshney (2002) suggest that communities with crosscut­ting structures are much less prone to escalation than those with overlapping structures. Laboratory research by Vanbeselaere (1991) shows that the effects of crosscutting are partly psychological—members of one group who are bonded to members of another group are reluctant to use harsh tactics in dealing with that other group. But Varshney’s study of conflict in Indian cities suggests a more dynamic account, that people with crosscutting ties often take actions designed to head off impending escalation, such as curbing the rise of militant leaders.

Again, there is room for research on this topic. Varshney’s discussion of these counterescalative actions is quite brief, suggesting the need for more case mate­rial on the processes by which crosscutting structures have their effect. Labo­ratory research on the conditions and processes underlying bilateral escalation should also be done. What is needed is a laboratory simulation where we can examine the complex interaction of variables and make careful real-time mea­surements of escalation as it unfolds. Simple laboratory games like the pris­oner’s dilemma were once thought to provide such a simulation, and some things were learned in these studies (Pruitt, 1998). But these games are too sim­plified to allow most of the complex processes just described to go forward. Clearly, the field is in need of a new laboratory paradigm.