Abstract
An exploratory qualitative study explored the effect of attachment styles on disputants' speech during real‐life mediations. Drawing on attachment theory, we classified disputants as secure or insecure individuals using a self‐report attachment‐style questionnaire.
Subsequently, they entered their mediation sessions, where their entire speech was recorded. Qualitative analysis of their speech yielded consistent and sometimes striking differences that portrayed secure speech as remarkably more useful and enhancing toward resolution compared with insecure speech. The findings, presented with many examples, strongly indicate the relevance of attachment to the research of communication during mediation sessions. In this report, we also consider the practical implications of the association between attachment and disputants' behavior, emphasizing the role of mediators.
Introduction
Mediation is a voluntary process that relies on the disputants' desire and ability to participate in the mediation process. Disputants' communication during mediation can materially affect the chances for resolution, and the mediator's ability to approach and involve the participants is thus crucial. Understanding who the disputants are and having knowledge of their dispositions can guide the mediator's approach, and recognizing the disputants' communication styles can enhance that knowledge.
This article presents an initial, innovative study into the effect of personality on disputants' communication style in real‐life court mediations. We chose to begin this inquiry using well‐known theories involving attachment styles. We drew on John Bowlby's (1969, 1973) Attachment Theory not only because it has been empirically well established (Shaver and Hazan 1993), but because it is also relevant to conflict, as we will subsequently show.
Attachment Theory and Conflict Resolution
Bowlby's (1969, 1973) concentrated on the development of the infant–caregiver relationship and its effect on the infant's self‐perception and perception of others. He assumed that babies are born with communication abilities that encourage the mother (or other caregiver) to bond with the baby, thus ensuring him or her a warm, safe environment. If the caregiver is available and responsive to the infant's communication, she or he (usually it is the mother) becomes a secure base for the baby, from which the baby can draw away and explore the world, coming back when she needs to. This secure base allows the baby to develop a sense of security in herself, her environment, and her ability to create relationships. Gradually, based on this first relationship, the child develops internal working models that represent the self and others, with respect to each other, and they become an interpersonal style, or the person's attachment style (Bowlby 1973). A secure base allows for the development of a secure style. If the caregiver were not adequately responsive, however, Bowlby's theory suggests, insecure styles would develop.
Mary D. S. Ainsworth and her colleagues tested Bowlby's theory in a famous series of studies based on “the stranger paradigm.” They observed one‐year‐old babies who were left by their mothers for a few moments in a room with a stranger and defined three different attachment styles displayed by the babies. “Secure” babies expressed only mild discomfort when the mother left, and immediately on her return showed a wish for close contact with her. They used the mother as a secure base and explored the room. “Anxious‐ambivalent” babies strongly protested when their mothers left and, on the mother's return, asked for her closeness on the one hand, but angrily rejected her on the other. They did not explore the room freely and seemed preoccupied with their mother's availability. “Avoidant” babies drew within themselves when left by the mother and pointedly ignored her on her return. They seemed to be suppressing their attachment needs, possibly to avoid the pain of rejection (Ainsworth, Bell, and Stayton 1971; Ainsworth et al. 1978).
The theory of attachment styles was further developed to include adults, and at this phase, four attachment styles were proposed (Hazan and Shaver 1990). The four‐styles model was tested by Kim Bartholomew and Leonard M. Horowitz (1991), who suggested that attachment styles are defined by the combination of a persons' self‐image (positive or negative) and his image of others (positive or negative; see Figure One).
As shown in Figure One, negative self‐image combined with positive other‐image is associated with fear of desertion and with high dependence on a close other, resulting in the anxious style (or anxious‐ambivalent). Negative other‐image combined with positive self‐image is associated with avoidance of closeness with others and high self‐reliance, resulting in the avoidant style. Positive image of both is associated with trust in self and others and with confident behavior, which leads to the secure style. Having a negative image of both oneself and others is associated with a negative experience of both self and environment, and this is the fearful‐avoidant style (or anxious‐avoidant), which is considered the least adaptive (Brennan, Clark, and Shaver 1998).
Bartholomew and Horowitz's (1991) axes of self‐image and other‐image bring to mind M. Afzalur Rahim's (1983) conflict resolution styles model (see Figure Two), which is also defined by the combination of two axes, that of concern for self and that of concern for others.
As shown in Figure Two, Rahim's theory postulated that a person who cares only for himself will be competitive and contending, while a person who has a high level of concern for both himself and others will take a problem‐solving attitude and try to achieve a more integrative solution to a conflict, addressing as many of each side's interests as possible. In between these poles, individuals who have moderate care for themselves and others will be more likely to compromise, according to this theory, while those who care only for the others will be more likely to make full concessions to the wishes of others. Finally, having low concern for both oneself and others, Rahim argued, will lead one to completely avoid coping with the conflict (Rahim 1983).
Much research has established an association between attachment and conflict. Studies have shown that attachment styles affect conflict resolution styles (Dawson 2003) in adult romantic relationships (Pistole 1989; Fischer 1994), as well as in nonromantic friendships, where secure participants displayed more integrating and compromising conflict behavior than insecure ones (Bippus and Rollin 2003). Several recent studies explored the association between attachment styles and communication during conflict, particularly among married or engaged couples and among depressed people (Crowell et al. 2002; Heene, Buysse, and Van Oost 2007; Pearce and Halford 2008). For example, compared with people with other styles, avoidant persons showed a higher frequency of negative behaviors and a lower frequency of positive behaviors during a marital conflict (Gouin et al. 2009).
In school settings, avoidant teachers tended to use less compromise and integration to resolve conflict in class, and anxious‐ambivalent teacherstended to use less integration, compared with secure teachers (Morris‐Rothchild and Brassard 2006). A longitudinal study that followed children over four to five years indicated that secure attachment at two and three years old predicted better social problem‐solving skills in first grade (Raikes and Thompson 2008). Attachment styles were also correlated to college students' attitudes toward conflict. Secure respondents were less threatened by arguing than anxious‐ambivalent or anxious‐avoidant respondents, and avoidant respondents reported avoiding conflict more frequently than secure ones (Pistole and Arricale 2003).
Some workplace research has also found that secure employees tended to seek help when stressed and to concentrate on solving the problem. They could regulate their stress levels and “bounce back” from stress more easily than insecure employees (Mikulincer and Erev 1991; Shaver and Hazan 1993). Another recent study at the workplace showed that secure employees felt more trust in their supervisors, which accounted for their better performance (Simmons et al. 2009). While this last study did not research conflict, it may imply that secure disputants would be more trusting of a mediator than insecure ones.
Mario Mikulincer (1997) has argued that the secure attachment style is a psychological strengthening resource, useful during conflict: secure persons are more optimistic, more trusting of others, and better self‐controlled under stress or conflict. Conversely, according to this category, people with insecure attachment styles are characterized by a “fight or flight” response to conflict or stress (Berant 1998): anxious people use the fight response — they “fight” for (demand) the attention and support of the other person, thus trying to enlarge their “secure base.” Avoidant persons use the flight response — they avoid conflict and count on themselves when stressed, emphasizing their autonomy (Mikulincer and Erev 1991; Shaver and Hazan 1993).
Attachment is also associated with cognitive, emotional, and social abilities and traits that can affect behavior during interpersonal conflict. Secure people show cognitive flexibility (Shaver and Hazan 1993); they can better deal with ambiguity and grasp new information even if it leads to confusion (Mikulincer and Nachshon 1991; Mikulincer 1997). In comparison, insecure people are not open to new information and show relative lack of curiosity and more cognitive rigidity that can lead to bad decisions. Additionally, insecure people tend to be less creative (Mikulincer 1997; Paitan 2003). Among college students, anxious‐ambivalent attachment styles were associated with limited social flexibility, possibly because that style interfered with the students' ability to process information in complex social situations (Miller 1996). Avoidant style, on the other hand, was in fact associated with better cognitive abilities (the ability to switch between tasks and resist distractions), but only in tasks unrelated to attachment (not social or interpersonal). Additionally, when the avoidant participants were asked to remember a time of insecurity just before attending to the cognitive task, they no longer did better than others (Gillath, Giesbrecht, and Shaver 2009).
Secure attachment was positively related to different measures of emotional intelligence, including interpersonal skills, adaptability, and stress management (Hamarta, Deniz, and Saltali 2009), as well as with the ability to facilitate, understand, and manage emotions. Interestingly, in another study, the ability to understand emotions was also positively related to avoidant style, a result that was contrary to hypotheses (Kafetsios 2004). Secure people tend to express their emotions and share them more easily than people with insecure styles (Bar‐ilan 1997). They feel more competent and less afraid of failure than insecure people when dealing with ambiguous and intimidating situations (Mikulincer and Nachshon 1991; Shaver and Hazan 1993; Mikulincer 1997).
Attachment has also been shown to be associated with communication: secure people communicate more directly and openly with others, even during disagreement, and are more likely to expose themselves and share information than insecure people. Anxious and avoidant people seem uncomfortable during communication and do not tend to expose themselves (Shaver and Hazan 1987; Mikulincer and Nachshon 1991; Mikulincer 1998; Paitan 2003).
Attachment and Communication during Mediation
The extensive literature linking attachment to conflict behavior and to cognitive, emotional, and social skills that can be helpful or hindering during conflict suggests that attachment would also play a part in disputants' communication during mediation. On entering mediation, disputants are likely to be stressed. They are already involved in conflict, and they are going to confront the other party in an unfamiliar procedure, guided by an unfamiliar mediator. It is an ambiguous, stressful social situation, and it is likely to catalyze many of the individual differences attributed to attachment style in the review above. The role of attachment in disputants' behavior during mediation has not been studied before, but the literature we have reviewed offers enough background to form expectations.
We may expect secure people to communicate in ways that enhance the mediation and help resolve the conflict. They are likely to display a problem‐solving attitude in mediation and to show more flexibility and creativity in their thinking than insecure disputants. Secure people are also likely to express more trust in the mediator and/or the other party. Secure people would probably express more emotion during conflict compared with avoidant disputants.
Anxious disputants will probably express emotion, but it is more likely to be negative and to be expressed vehemently, compared with secure disputants' expression of emotions. Additionally, because secure disputants feel more comfortable and optimistic during conflict resolution, and tend to expose themselves more, we may expect them to avow more responsibility and blame for the conflict than insecure disputants. Secure disputants would probably also display more care for the other party and better interpersonal skills, such as empathy and leniency.
In comparison with secure disputants, insecure disputants will probably display more communication that hinders the mediation and sabotages agreement. They are likely to express more rigid, dichotomous (black and white) thinking. They may be more negative and aggressive in their speech. Because they tend to be less self‐exposing and less communicative, we may find that their speech is more laconic. Because they feel more threatened during conflict resolution, they may be more defensive, renounce responsibility, and blame the other party.
The current study analyzed the full texts of real‐life court mediations in a small claims court in Israel and found consistent and sometimes striking differences between secure and insecure disputants in many of the above verbal behaviors and in some forms of communication that were not predicted but appeared in the text. The next section describes how the study was conducted. Following that are the description of our results and a discussion of their implications.
Method
Participants
Twenty‐five adults, seven women (28 percent), and eighteen men (72 percent), participated in the study. Thirteen were plaintiffs (52 percent) and twelve were defendants (48 percent). Among the women, four out of seven were plaintiffs and the other three were defendants. Among the men, nine were plaintiffs and nine were defendants. Twenty‐one of the participants were Jewish Israelis (84 percent) and four were Arab Israelis (16 percent). Six participants (24 percent) represented commercial companies. As mediations in Israel are, overall, strictly confidential and not open for observation, it was extremely difficult to reach a larger sample.
Procedure and Materials
We acquired data at two locations: at an Israeli court and at the Campus Mediation Center of the Bar‐ilan University, which also receives cases from an Israeli court. Our research procedure was identical in both locations. To begin, each participant, while waiting for the mediation session to start, was offered the opportunity to answer the relationship questionnaire (RQ; Bartholomew and Horowitz 1991), which is a self‐report of attachment styles.
The questionnaire includes four items, each describing one style. The respondent was required to score the extent to which each item describes himself or herself, on a scale from 1 (completely untrue) to 7 (completely true). After scoring each style, in a fifth question the respondent was asked to choose one of the above items as the style that best fits him or her. The questionnaire result is recorded from the choice in the fifth question; we used the first four scale scores to confirm that the respondent chose a style that he or she previously scored as highly self‐descriptive compared with the other styles (thus ensuring that he or she understood the questionnaire and meant his or her answers).
This questionnaire is short but commonly used and has been validated in attachment research (Backstrom and Holmes 2001). We chose it over longer attachment measures to avoid disrupting the mediation timetable and burdening the parties with additional stress and cognitive load, which risked reducing their effectiveness during the mediation.
We assured disputants that the study was anonymous and that their participation was completely voluntary and would not affect the mediation in any way. Respondents who agreed to answer the questionnaire were further asked to allow a researcher and an assistant to be present in the mediation room and transcribe the mediation.
During the mediation session, a research assistant sat in a corner of the room, unviewed by the disputants, and transcribed disputants' entire speech without any editing. A researcher was also present in the room and took additional notes about the proceedings. (Audio and videotaping is strictly forbidden in mediation sessions in Israel. We used two transcribers to reduce the possibility of missing something.)
To analyze the content of the disputants' speech, we took a qualitative approach, using more than one set of independent judges to create the analytical tools and to perform the analyses. At first, we formulated a list of verbal behaviors that we expected secure and insecure disputants to differ in, based on literature (see the introduction above). Additionally, our research team, which comprised a psychology professor who is also a mediator and three students enrolled in the conflict resolution program at Bar‐ilan University, read several of the mediations and suggested additional verbal behaviors to be added to our original list, marking places in which these behaviors occurred in the transcripts.
We presented the combined list of specific verbal behaviors to ten judges (psychology and conflict resolution professors and students) who independently classified each listed behavior as one that would either enhance or hinder conflict resolution. (Our analysis did not include behaviors that could be considered as neutral.) Only items that all judges agreed were enhancing or hindering were included in the final list of verbal behaviors we used in the study. The final list appears in Table One.
Enhancing and Hindering Verbal Behaviors in Mediation
Enhancing Speech . | Hindering Speech . |
---|---|
Making concessions Using people's first names Expressing empathy Taking responsibility Showing creativity Showing flexibility Participating in problem solving Expressing trust Expressing emotions clearly | Expressing scorn Showing arrogance Expressing negativity Showing aggressiveness Having a low‐reaction threshold Exaggerating Referring to people without using their names Expressing cynicism Blaming Renouncing responsibility Showing defensiveness Using rigid/repetitive speech Curtness Expressing dichotomous thought Demonstrating inconsistency Showing disregard for the future Expressing mistrust |
Enhancing Speech . | Hindering Speech . |
---|---|
Making concessions Using people's first names Expressing empathy Taking responsibility Showing creativity Showing flexibility Participating in problem solving Expressing trust Expressing emotions clearly | Expressing scorn Showing arrogance Expressing negativity Showing aggressiveness Having a low‐reaction threshold Exaggerating Referring to people without using their names Expressing cynicism Blaming Renouncing responsibility Showing defensiveness Using rigid/repetitive speech Curtness Expressing dichotomous thought Demonstrating inconsistency Showing disregard for the future Expressing mistrust |
Finally, we gave this list of verbal behaviors to a new set of four independent judges (psychology and conflict resolution students), who were unaware of the disputants' answers to the attachment styles questionnaire. The judges read the full mediations and divided each disputant's speech into small verbal units. Then they independently classified each unit of speech as one of the behaviors on the list. We included the units of speech that were classified identically by all judges in the result calculations presented in Figures Three and Four.
Rate of Secure and Insecure Participants Who Used Each Hindering Verbal Behavior
Rate of Secure and Insecure Participants Who Used Each Hindering Verbal Behavior
Rate of Secure and Insecure Participants Who Used Enhancing Verbal Behavior
While the mediations were conducted in Hebrew, the verbal behaviors that we noted were not language specific. The examples that we will present subsequently were translated to English by a bilingual mediation researcher.
Results
The RQ (Bartholomew and Horowitz 1991) yielded a striking result: of twenty‐five respondents, no less than twelve (48 percent) described themselves as avoidant, three (12 percent) as anxious, and one (4 percent) as fearful; only nine (36 percent) described themselves as secure. Our sample size was not large enough for us to perform statistical analysis, but we did find an uncommon distribution of styles in our sample compared with previous attachment research, which we will discuss later in the article.
To have some measure of the differences between our secure and insecure participants, we chose to obtain data from the two largest groups — secure and avoidant, and we calculated the percentage of disputants in each of these groups who used each verbal behavior. For example, we calculated the percentage of secure and insecure (avoidant) participants who expressed scorn at least once.
Hindering Communication
Figure Three above displays the percentage of secure and insecure participants who used each hindering verbal behavior.
As shown in Figure Three, insecure disputants used all the hindering verbal behaviors more frequently than did the more secure disputants.
Scorn
Forty‐two percent of insecure participants used scorn in their speech at least once compared with none of the secure disputants. Scornappeared in a mediation concerning a salary payment conflict in which an employer was sued by a worker for failure to pay. At the end of a long explanation by the worker concerning the requirements of the job and its burden, the employer “smiled scornfully” and said: “How about you tell what you actually know, not what you don't.” Later he said, “It's a pity you talk idly, because we [the employers] know what really happened there.” And he later added: “Are you kidding? Were you even there? They're not judges, you see [referring to the mediators]. Bullshit . . . you guys got drunk.”
In this example, the employer undermined the worker's professional knowledge, called his words “bullshit,” and implied that he was lying. Suggesting that the worker had been drunk was intended to challenge the worker's reliability as a witness. And finally, by “explaining” to the worker that the mediators were not judges, he suggested that the other party did not understand the procedure as well as himself.
Arrogance
Thirty‐three percent of the insecure disputants displayed arrogance in their speech; 22 percent of the secure ones did so. One example occurred in a mediation in which a graphic artist sued a customer for stopping payments on completed jobs. In her claim, the graphic designer alluded to the efforts she had made to complete the unpaid work, including a premature return from a maternity leave. The customer called her argument “dramatization.”“I'm a lawyer in my profession,” he said, “and I'm well familiar with the need to dramatize the nursing baby issue. In my eyes, it's unprofessional.” In this example, the speaker suggests that he is a professional while the other party is not.
Negativity
While 83 percent of insecure participants used speech that was judged as negative toward the other party or the procedure, only 33 percent of secure participants did so. For example, in one case the defendant said to the mediator, referring to the plaintiff: “I don't want, don't want to hear him or see him anymore.”
Aggressiveness
Fifty‐eight percent of insecure disputants used speech that was judged as aggressive, while only 22 percent of the secure disputants did that. For example, during one mediation, a renovator had argued that the homeowner had stopped payment before he had terminated the work. At one point, the homeowner screamed at the defendant: “You're a liar . . . I can't stand liars . . . you're a liar.” She also screamed at the mediator: “You don't understand what I'm talking about!” Later, at the end of the session, she angrily stood and called the defendant a “bastard, liar!” During her speech, the plaintiff also waved her hands and inclined her body in an intimidating way toward both the defendant and the mediator.
Low Reaction Threshold
Seventeen percent of the insecure participants and 11 percent of secure ones displayed a low threshold for reacting angrily. The low threshold was expressed by interrupting the speech of the other party or mediator with an agitated, vehement statement that seemed out of proportion to what they were responding to.
Exaggeration
Sixty‐seven percent of insecure participants exaggerated at least once compared with only 33 percent of secure participants. In the mediation in which a customer sued her hairdresser, the defendant said: “She burst, she kept screaming. . .for every little falling of hair she moved heaven and earth” (emphasis added). The hairdresser exaggerated her client's responses while simultaneously minimizing the damage she caused (the “little falling of hair” referred to wounds on her client's head and the skin and hair that fell from it). In another mediation, the defendant responded to the plaintiff's claims for certain proceedings and payments by stating that “it doesn't work that way anywhere and in no business” (emphasis added). In this case, the defendant cannot possibly know how payment works everywhere and in every business, so the “rule” he's presenting must be an exaggeration.
Referring to People without Using Their Names
Ninety‐two percent of the insecure participants referred to the other party using only a pronoun or some kind of descriptor (“he” or “the defendant,” etc.) without using his or her name during the mediation, but only 67 percent of the secure participants did this at least once. In Israel, avoiding one's name and using a title or pronoun instead is notably impersonal and unfriendly. Israeli culture encourages using names, preferably first names, to express goodwill in interpersonal encounters, especially among equals. This will vary, of course, from culture to culture.
Cynicism
Thirty‐three percent of insecure participants expressed cynicism compared with 11 percent of the secure disputants. For example, a graphic designer who was suing a customer described her efforts to complete the job while on maternity leave, such as finding child care. Expressing cynical doubts about her sincerity, the customer replied, “This girl made a big charade of how she's going to leave her kids for our sake.”
Blaming
Eighty‐three percent of the insecure participants used speech that was judged as blaming, while only 22 percent of the secure ones did so. One example comes from a mediation in which a woman sued her renovator: “He simply disappeared, didn't show up. . .each time, he got the money properly and then ran away.” In another mediation, the plaintiff indirectly blamed the defendant for theft by turning to the mediator and, referring to the other party, asked: “[Do] you want to justify the one who stole?”
Renouncing Responsibility
While no less than 92 percent of the insecure participants renounced responsibility at least once, only 44 percent of secure participants did so. For example, the hairdresser who injured her client with hair glue shifted the responsibility for that decision away from herself — even though as a professional she knew it was unwise — by saying, “I explained to her that gluing so much hair is a problem. Her friend insisted on it, I did not recommend it, but if she wants me to do it, I will.”
Defensiveness
Fifty‐eight percent of insecure participants compared with 33 percent of secure disputants expressed defensiveness at least once. We characterized as defensive those instances in which disputants made arguments designed to “save face” but did not directly address the other parties' actual argument or allegation. For example, when he was accused of having been involved in a physical fight, one defendant replied, “I really don't know how to beat anyone, certainly not someone who is even one day older than myself.” Rather than simply denying the allegation of violence, he made a face‐saving statement about his character.
Rigidity/Repetitive Speech
Among insecure disputants, 83 percent displayed repetitive speech compared with only 55 percent of secure participants who did so. For example, in a mediation in which a plaintiff sued a company for patent implementation, he repeated the same words again and again: “implement the idea. . .test for implementation . . . implementation impossible. . .there's no way to implement. . .test for implementation. . .see if the proposal can be implemented. . . impossible to implement. . .and implement the idea. . .test for implementation potential.” We deemed this speech hindering because it repeatedly concentrates on case details that are already known and does not help resolve the conflict (as opposed to expressing new ideas or verbalizing one's interests).
Curtness
Fifty‐eight percent of the insecure disputants spoke curtly at least once compared with 44 percent of the secure ones. This verbal behavior describes minimal descriptions or short answers to questions about personal preferences or wishes, often without giving reasons.
Dichotomous/Absolute Thought
Fifty‐eight percent of insecure disputants showed dichotomous, dualistic (black/white) or absolute thinking in their speech versus 33 percent of secure ones. For example, a tenant suing his landlord because the house lacked a hot water heater said: “I filed this claim because I know I'm right and he's disregarding my claim,” although the landlord claimed there was in fact a water‐heating system in the house, just not of the particular type that the tenant expected. Seen objectively, it is not clear that there is any one right party in this case. In another case, the defendant, when asked by the mediator if there are any other things that can be done, answered, “not a chance, we'll [have to] reach the Supreme Court.” With this statement, he displayed absolute thinking, not considering even the possibility of agreement. (This is also an example of exaggeration, as such an insignificant case would never reach a high court.)
Lack of Consistency
Among insecure participants 42 percent showed lack of consistency, while only 22 percent of secure participants did so. This verbal behavior refers to cases in which disputants contradicted themselves in their own descriptions of the facts, or cases in which they went back and forth on their demands or proposals, not advancing in one direction, whether toward compromise or contention.
Disregard for the Future
While 58 percent of insecure participants displayed disregard for the future in their speech, only 11 percent of secure participants did so. In one case, a worker sued a customer for lack of payment because the customer had paid for none of the job even though the dispute concerned only one‐eighth of the area in which the work was done. Despite the fact that the employer could have been charged with delaying rightful payment (for most of the work at least), he did not hesitate to foresee future court procedures: “Muhamad wants 8,000, so we'll go to the judge.” When the mediator suggested a compromise involving future work, he answered: “Give it a break guys, what do I know of the future?”
Expressing Distrust
Seventy‐five percent of insecure disputants showed distrust in the other party at least once in their speech compared with only 22 percent of secure disputants who did so. The customer who sued the hairdresser expressed distrust in the defendant's professional skills as well as in her truthfulness, saying, “I don't want any more services from her. . .I'll never do any job with her. . .it's not true, she's lying.” In the case in which a patent‐implementation company was sued, the defendant said of the plaintiff, “I think he's not acting in good faith. . .his claim conceals essential things from the court.”
Enhancing Communication
Figure Four shows the percentage of secure and insecure participants who used each enhancing verbal behavior.
As can be seen in Figure Four, with two exceptions — making concessions and expressing emotions — the secure participants were more likely to use the enhancing verbal behaviors than were the insecure participants.
Making Concessions
Contrary to our expectations, a greater percentage of insecure than secure participants made concessions. Only 33 percent of secure participants made concessions to the other party at least once compared with 42 percent of the insecure disputants. For example, one of the plaintiffs initially said he would concede smaller amounts to the other party (“what I was already paid for the screening”) but then increased the amount of his concession as the mediation progressed (“even if you ask me for 100 percent of the investment and 50 percent [of the revenue] is mine, I'll agree”).
Using First Names
Fifty‐six percent of secure participants and 50 percent of insecure participants used the other party's first name at least once. In Israeli culture, first names are considered appropriate and friendly among equals.
Expressing Empathy
Seventy‐eight percent of the secure participants used speech that expressed empathy toward the other party at least once. In comparison, none of the insecure participants did so. In a mediation in which a customer sued an air‐conditioning repair service, the service representative said: “I came here to hear him. . .[I admit that] sometimes our service centers are over‐loaded. . .all the anger in his claim, I'm not judging it. . .his call was attended to, although it could be that he felt unattended to.” By acknowledging the other party's feelings and not judging him, she expressed empathy without necessarily agreeing with all his claims.
Taking Responsibility
Forty‐four percent of secure participants took responsibility at least once compared with 25 percent of insecure participants who did so. In a lawsuit against a door manufacturer, the manufacturer's representative said, “I'll keep this short. . .most of their claims are justified. . .the company's conduct was not what it should have been. . .I'm willing to fix [the door] for them and give them a two‐year warranty.” (See Poitras 2007 for an extensive discussion of the dynamics surrounding the acceptance of responsibility within mediation.)
Creativity
Fifty‐six percent of secure participants used speech that showed creative thinking, while only 17 percent among the insecure participants displayed creativity. In the tenant/landlord dispute the parties argued about the price of a hot water heater. After both made different estimates, the tenant suggested: “I'm ready to take you to Habib [a house goods store] and he'll give his price and that will determine.” The tenant found a way out of a typical “loop” in which arguments can get stuck by suggesting objective criteria (Fisher and Ury 1981). In the same mediation, the landlord responds to the mediator's request to find mutual gain through alternatives to financial compensation by asking “Maybe he [the tenant] would like an olive tree?” He then explains that his family sells olive trees and that a mature tree is worth thousands of shekels. The tree represents a creative idea beyond the issues that are on the table, which could create joint benefit.
Flexibility
Among secure participants, 78 percent used speech that showed flexibility, while only 50 percent of insecure participants did so. Disputants expressed flexibility when they were willing to change a former position on the basis of new information or considerations that were brought up during the mediation. In the case of the hairdresser, the defendant shifted from her original position by offering to fix the damage at her own expense and more: “I'm willing to forego the cost of the work, pay her 500 [shekels] and absorb [the cost of] the new hair.” The defendant was able to give up her original position after consideration and change her attitude in response to new information.
Problem Solving
Eighty‐nine percent of secure participants used speech that was judged as problem‐solving compared with 75 percent of insecure participants who did so. In a mediation in which a father sued a school on behalf of his daughter, the representative of the school said, “I want to help all the students who approach us — I did not know that she was a minor [by age]. . .we'll be happy to help. . .we commit ourselves that after she reaches eighteen she'll be able to take the test.” In response to the father's subsequent speech, which described some of the difficulties associated with this solution, the school representative added, “We are now here to reduce damage. . . there are other branches in the area for students. . .where she could also rehearse, take extra lessons before the test.” The school representative not only offered practical solutions but also framed the mediation as a problem‐solving procedure rather than a dispute (“we are here to reduce damage”).
Expressing Trust
Forty‐four percent of secure disputants used speech that expressed trust compared with only 17 percent of insecure ones. In the case of the tenant against the landlord, the latter said: “The guy spoke truthfully, except for the last part. . .there isn't a device, he's right. . .neither one of us is a liar.” Although he doesn't accept his tenant's version fully, he makes a point of saying (or implying) three times that the other party is to be trusted.
Expressing Emotions Clearly
Forty‐four percent of secure participants clearly expressed an emotion at least once. Contrary to our expectations, more insecure participants, 58 percent, also did so. This also contradicts previous findings, and although the difference between the groups is small, it is notable because the insecure participants reported here were avoidant, and avoidant persons were found in former studies as less emotionally expressive than secure persons.
Summary of Results
The rates of secure versus insecure disputants who used the verbal behaviors documented in our study showed clear and consistent differences. Hindering verbal behaviors composed the majority of speech documented in the study. Insecure participants were more likely than secure participants to use each of the hindering verbal behaviors. They were more likely to express scorn (which was not expressed by any of the secure disputants) as well as arrogance, to renounce responsibility, to blame the other party, and to express distrust, negativity, cynicism, and aggressiveness. They also tended more frequently to avoid the other party's name and use titles or pronouns instead, to be curt, inconsistent, and show disregard for the future. They tended more to exaggerate, and their speech was more rigid (repetitive) and expressed more dichotomous (black/white) or absolute thought.
The differences in enhancing speech were almost as consistent. Secure participants were more likely to use almost all the enhancing verbal behaviors, including expressing empathy (which was not expressed by any of the insecure disputants), expressing trust, demonstrating creativity, showing flexibility, participating in problem solving, and taking responsibility. Secure disputants were also more likely to use the other party's first name, which shows friendliness in Israeli culture. Contrary to our expectations, insecure participants were more likely to make concessions toward the other party and to express their emotions clearly, although the difference between the two groups for this final behavior was very small.
Discussion
Mediation is a delicate process, involving people in stressful conflict situations requiring them to deal face‐to‐face with a person with whom they are in conflict as well as with a mediator whom they have never met before. Moreover, they are asked to clarify their own interests and needs, rather than positions, while acknowledging those of the other party. They need to move from the strategy of attacking people to the strategy of attacking the problem, to move from the past to the future and from “I” to “us” (Fisher and Ury 1981; Kovach 1994). The main tool to achieve all these is communication: they talk until they resolve the conflict (or not).
Attachment theory and research posit that people with different attachment styles have diverse resources and ways of coping with stress and conflict (Bowlby 1973; Shaver and Hazan 1993; Mikulincer 1997; Raikes and Thompson 2008; Gouin et al. 2009). We therefore expected, and found, that communication styles during mediation differed accordingly. Insecure disputants' speech was more scornful and arrogant, more aggressive, negative, and cynical. These disputants tended more often to blame the other party while renouncing their own responsibility. Such language expresses the opposite of seeing the other party as part of an “us” and hinders the possibility of becoming “us” and of separating the problem from the people. Likewise, the insecure participants' tendency to be more curt at times and to express distrust of others can stand in the way of building an effective relationship between the parties.
Other characteristics of insecure speech were that it was more repetitive, seemingly “stuck” on the same issues, which shows cognitive rigidity. It was also more rigid in its tendency to express dichotomous, absolute “black and white” thought (e.g., “I'm right and you're wrong.”). This last sense of being rigid is somewhat reminiscent of the process of “rigidification” described by Terrell Northrup (1989). She referred to enemies' tendency to build a strong psychological wall between themselves and their opponents, seeing themselves as more and more unlike each other, thereby escalating the conflict. In the small claims mediations that we studied, an “I'm perfectly right and you're perfectly wrong” approach might also have escalated rather than resolved the conflict. Combined with insecure participants' stronger tendency to express disregard for the future, which we found, these rigid and reckless approaches would make it harder for a disputant to see the disagreement in its full complexity — rather, they would encourage a one‐dimensional view of the conflict (Mikulincer 1998). Additionally, rigid and absolute cognition likely damages one's ability to come up with creative solutions and “think outside the box” about the dispute.
By hindering effective mediation, such verbal behaviors may protect an insecure participant from the painful recognition that truth is not solely on his or her side and from the tedious, rather intimate, process of resolving differences and recognizing interests and possibilities. Bearing in mind that the insecure participants from whom we eventually obtained data were in fact avoidant, one can imagine that such an intimate course, which involves trusting and opening up to others, is not likely to be comfortable for those who prefer to draw away from people and trust themselves (Shaver and Hazan 1993; Kobak and Duemmler 1994).
As for positive, enhancing instances of speech, these were used more frequently by secure participants than by insecure participants. Secure participants were more likely to use speech that indicated empathy toward the other party, while none of the insecure, avoidant participants expressed empathy. Empathy is the ability to understand and identify with the other through a temporary entrance into his experience, trying to see reality through his eyes and feeling a wish to help him (Leith and Baumeister 1998). A previous study found that avoidant persons were in fact good at understanding emotions (Kafetsios 2004), but empathy during mediation requires more than understanding the other. Recognizing the other's feelings during conflict takes not only care for the other but also trust and confidence in the other — an empathic party has to believe that his or her empathy would not be used as a weapon to weaken his or her own cause. Insecure participants — particularly avoidant ones — might not have felt confident that expressing empathy toward the other party would not hurt his or her own interest. Also, inasmuch as their avoidant style indicates their preference to avoid intimacy with the other party during conflict, they might have felt that expression of empathy would create too much closeness and be out of place (Hazan and Shaver 1990; Bartholomew and Horowitz 1991).
The more resolution‐enhancing speech used more frequently by secure participants seemed to “mirror” the shortcomings of rigid, absolute, and untrusting speech. Secure disputants' speech was more creative, flexible, trusting, and bent toward problem solving while willing to take personal responsibility. These verbal behaviors are pretty much the building blocks of mediation. Two other verbal behaviors that were deemed positive and enhancing by our judges were making concessions toward the other party and clearly expressing emotion. Interestingly, insecure disputants performed these two behaviors at a higher rate than did secure ones. Emotional expression and concession making that results from strong need for the other party's affiliation could characterize anxious style, but as we obtained data from avoidant insecure participants, these findings are intriguing. Disputants might have made concessions in the spirit of resolving the conflict and ending the mediation session more quickly than it is possible to do through problem solving. Inasmuch as avoidant style leads to the wish to avoid conflict, understanding concessions in this light may explain why they were made more often by our avoidant participants compared with our secure ones. It would be interesting to see if these findings are repeated in a quantitative study with a larger sample of avoidant participants.
All in all, our findings suggest that disputants who demonstrate insecure attachment styles are more likely to communicate in ways that hinder mediation effectiveness and that disputants whose styles are more secure are more likely to communicate in ways that can enhance the mediation process. It is worth noting that we identified more instances in which hindering, negative speech occurred than instances of enhancing speech because it could suggest that the effect of these differences in speech on the other party is greater. Research on the “negativity effect” (Kanouse and Hanson 1971) found that negative stimuli affect judgment more strongly than do positive or neutral stimuli that are equivalent in intensity (Peeters and Czapinski 1990), and such stimuli can include speech, which was tested by determining the effects of positive versus negative descriptions of people in stories (Peeters 1989).
It applies, then, to our subject matter, speech, but also more particularly to the situation we studied, as the negativity effect was the basis for the well‐established “prospect theory” (Kahneman and Tversky 1984). According to this theory, a negative result (possible loss) affects our decision‐making process more so than does a positive result (possible gain) even when it leads us to irrational choices. Thus, the negativity effect can be relevant to the process of decision making, and negotiation and mediation are essentially processes of joint decision making. Therefore, the many examples of hindering, negative language that was used more frequently by insecure disputants is likely to not only have affected the emotions of the other parties but might also have affected their decisions. Negative speech could reduce the other party's readiness to cooperate or could even encourage a premature termination of mediation just so that the recipient of such language will no longer have to suffer it.
On top of the consistent differences we found in speech between secure and insecure disputants, our study yielded another intriguing finding, which we did not expect: of twenty‐five respondents, no less than twelve (48 percent) described themselves as avoidant; only nine (36 percent) described themselves as secure, three (12 percent) as anxious, and one (4 percent) as fearful. This is quite uncommon; the usual distribution of attachment styles among the many populations that have been studied by attachment researchers was roughly 60 percent secure, 25 percent avoidant, and 15 percent anxious (Mikulincer, Florian, and Weller 1993). Our remarkable rate of avoidant participants may indicate that insecure people are more likely than secure ones to resolve disputes at court, which sent them to mediation, while secure individuals who are involved in a disagreement are possibly more likely to resolve it on their own. But because our sample may just have been too small to be representative, drawing any conclusions about the distribution of attachment styles among court disputants would require further study with a larger sample.
Theoretical and Practical Implications
This research is the first to tie together the study of attachment styles to the practice of mediation. It suggests avenues for deeper exploration of the impact of disputants' personality factors while offering a new real‐life setting in which to study the premises of attachment theory. Mediation studies in particular, as a young academic field, can benefit from researching the relevance of well‐established psychological knowledge.
We also believe this study has practical implications as well. To begin, it reinforces the idea that mediators should consider the individual characteristics and differences of their mediation clients. Mediators' understanding of personality differences, in this case those differences involving attachment styles, can become another rich tool in their kits. A mediator cannot reasonably be expected to diagnose his or her disputants' personalities. As more research linking attachment and mediation emerges, however, mediators will be offered more and more useful knowledge about the different styles, their typical behaviors, and their different needs.
Second, the findings of this initial study of attachment and speech patterns during mediation may already be used practically. An aware mediator who recognizes the speech patterns associated with insecurity should take steps to address that underlying insecurity. The mediator should take steps to minimize the incidence of scornful, aggressive, cynical, or rigid speech that prevents the speaker from participating effectively in problem solving and also naturally arouses the other side to respond similarly (see Poitras 2005.) One way is to halt the mediation and caucus privately with the parties. In caucus, the mediator could seek to calm the insecure party and attempt to make him feel more secure about the process as a whole, how it works, and everyone's role in it. In terms of attachment theory, the mediator can serve as a “secure base” for an insecure party, using mostly verbal behavior to provide reassurance (see Crowell et al. 2002). Inasmuch as the insecure speech pattern is induced by the activation of an insecure attachment internal model (Bowlby 1973), the mediator can use her or his position to offer support that would calm insecurity. She can ask questions and reframe the party's statements to gradually find out what it is that intimidated the insecure disputant and what might make him feel more confident. This behavior would show care on the part of the mediator and could also gradually guide the disputant away from his or her hindering position. Additionally, the mediator may turn to the other party and find out what might be done to answer the insecure party's needs without hurting the other party's interests. As for counteracting insecure cognitive tendencies, the mediator can broaden the rigid and less creative thinking of insecure participants by offering scenarios and solutions.
Limitations and Further Research
This was an exploratory, mostly qualitative study. Its most evident limitation was its small sample. The strict confidentiality that governs mediation procedures in Israel — and elsewhere — makes direct study challenging. Research on mediation would benefit from more opportunities to study it directly. Despite our small sample, we were able to show consistent differences in speech between secure and insecure disputants, but it would take further research with larger samples to statistically validate our findings.
Another limitation of our study is that it did not include enough anxious insecure participants. While the high rate of avoidant participants presented an interesting finding in its own right, further study is required to test anxious speech patterns in mediation, which, because of significant differences between the avoidant and anxious attachment styles, can be expected to differ.
A third and important limitation is that we studied only small claims mediations. It is possible and even likely that disputants' behaviors vary according to the weight of the issues discussed and the relationship of the participants to each other. Parties in conflict can be expected to engage in more integrative negotiation behaviors when they have a relationship to sustain; parties in small claims mediation are less likely to have ongoing relationships. However, when parties try to resolve a conflict in an important relationship, their ability to integrate may be handicapped by intense negative feelings. So, while our findings demonstrated that small claims disputes are enough to encourage many instances of aggressive and hindering speech, some of the differences we found may diminish when more is at stake. For example, in divorce mediation, where both parties are experiencing substantial distress, a secure attachment style may not necessarily preclude insecure behavior.
Our study was innovative in that it connects attachment style and speech using data from real‐life mediations. Additional research would enable a greater understanding of the role of personality — attachment styles in particular — in real‐life mediation. No less importantly, the practical implications of our findings concerning the mediator's role in making his or her clients feel more secure would also benefit from further direct study.