The Three-Chair Model

By Ike Lasater & John Kinyon
With Julie Stiles

Over the past five years, in approximately 40 weekly classes per year and periodic weekend workshops, we (John and Ike) have engaged in a process of experimentation and discovery in creating a context for learning NVC mediation. This context, which we call the “three-chair model,” provides in-depth experiential education and is the most powerful framework we have come across for not only learning a Nonviolent Communication (NVC) approach to mediation, but also for accelerated development of core NVC skills, centered presence, and the capacity to heal and reconcile with self and others. In this article we offer our experience on creating a learning context, specifics on the approach we have developed, and the benefits people report from our trainings.

Creating a Learning Context

When we first began teaching this course, we thought our objective was to “teach” mediation. Early on, however, we realized that trying to teach did not feel in integrity with what we so valued about Nonviolent Communication, namely the power we had experienced of people choosing how they want to meet their needs. We were also influenced greatly in our views by Carl Rogers, and in particular an article he wrote entitled “Personal Thoughts on Teaching and Learning.” Rogers' thesis, in essence, was that "teaching" in the traditional sense (students learning what an authority tells them to learn) was at best ineffectual, and at worst detrimental; self-directed, self-discovered learning was where real growth and meaningful behavioral change occurred. We began to view ourselves as creating a context for learning to happen rather than teaching people what to do.

As part of our learning context, we do provide a structure for people to learn mediation, utilizing a model drawn from Marshall Rosenberg’s five-step mediation model. Besides presenting this model, we also present nine skills that we have found important in NVC mediation. We do coach to this model and we share what we value and have found effective based on our learning and experience. However, when we are coaching people in the mediator chair, more and more we offer possible options and choices (including to come up with their own mediation choice) at various points and phases of the mediation process, rather than directing people what to do. It is for the person sitting in the mediator's chair to try out for themselves what works and does not work in their own experience. Coaching in this way, we as facilitators also continue learning and growing by seeing over and over again what choices lead to what results and by being open to the emergence of new, more effective choices.

In addition to students receiving in-the-moment coaching while mediating, a tremendous learning for people in the mediator chair has also been to receive feedback from people playing the roles of the "disputants" in the conflict, as well as from those observing. Another aspect of this learning context is that when students ask questions, we suggest "putting the question into the chairs." We set up the situation in the three-chair model to address the questions that are asked. For example, when participants ask, "What do you do if such and such happens?" we set up a role-play to simulate "such and such" happening, and people learn by doing. In this way we attempt to support people discovering for themselves what is meaningful and effective.

The Three-Chair Model

At the heart of our NVC mediation training is the three-chair learning model. Participants in our trainings have the chance to play mediator and disputant in conflict situation role-plays. While sitting in the mediator’s chair, students practice staying present and connected with themselves and others moment by moment, get real time coaching, and receive invaluable feedback from those in the disputant roles or observing. From being in the role of disputant or observer, people then have the opportunity to move into the mediator’s chair and put into practice what they have just learned. There is a back and forth flow between experiential doing and stopping to “harvest” the learning from the role-plays, and between being in the whole group and breaking into smaller groups. As trainers, we actively coach, facilitate, and share our perspectives and experiences throughout.

In this model, we create a role-play of a conflict situation, either one that participants make up or a situation they bring in from their experience. One person decides to sit in the mediator’s chair, and the other two people take on the role of the disputants. As coaches, we ask people how much coaching they want from us, and encourage people to let us know if they do not want coaching at any time during the role-play. As long as people are OK with coaching, we tend to interrupt a lot, particularly when people are new in the mediator’s chair. We interrupt mostly to remind people of the choices we see at various decision points; the model provides a skeletal structure of choice points at transitions between the steps and stages, and we coach the mediator to a basic set of choices that she can try out.

Using this model supports people in having an experience of being in the mediator’s chair with some way to find their bearings. They start with the basic structure, and through our coaching, develop at first a limited set of choices that they can try out, building upon those choices with their own experience in the mediator’s chair and through observing others. Over time and with repeated experience, participants develop a larger set of options and those options become more readily recallable. We create the basic platform and support them with a basic structure that they then build upon themselves. An analogy might be learning to dance; when someone learns to dance they learn the basic dance steps, but at some point, with enough experience, they stop thinking about the dance steps and begin to simply dance. In our learning context we offer a conceptual structure for what a person is doing as mediator, but we remind them that it is simply a map, and we make it clear that we are not teaching this model because it’s the “right” way to mediate; in fact, we sometimes diverge significantly from the form of the model when we are mediating real conflicts. Once someone learns the model and has enough practice with it, they can begin to embody the map and the choices and do it in their own way, following the flow of the conversation much as someone might follow the flow of the music, building upon the basic dance steps to create something unique.

Utilizing this model has yielded rich and unexpected learnings for us, from our own observations and from what participants have reported. In the remainder of this article we’d like to share the benefits we’ve experienced using this method.


Accelerated Learning of Core NVC Skills

Early on, one of the first pieces of feedback people gave us in our classes was how quickly the three-chair model helped them develop the core NVC skills and integrate them into their daily life. People would basically tell us, “I have tried all sorts of ways to learn NVC, including practice groups, books, CD's and empathy circles, and I have integrated more NVC into my day to day life as a result of doing this training than anything else I’ve tried.” As we began hearing this from various participants, we began to pay attention to what was happening. The two areas we really saw this taking place were in the skill of offering empathy and the core distinctions of NVC.

The first skill we work on with participants is the process of offering empathy. Every time someone takes on the mediator role, she is practicing being present and keeping her focus lightly resting on the person speaking. She is also learning how to language guesses that inquire with curiosity what is going on for a disputant, and remembering to hazard a guess as to what the disputant is needing instead of asking them what they are needing. Since the mediator is in the midst of a role-play making these guesses and practicing being present, she gets to see immediately what works (meaning whether connection was created) and what does not as the other participants react to what she has said. Thus, the skill of empathy is practiced and reinforced each time a person takes on the role of mediator.

The core distinctions of NVC—observation versus judgments, feelings versus evaluations masquerading as feelings, needs versus strategies and requests versus demands—also get reinforced in the role-plays. As a mediator hears the story of each disputant, he is practicing translating the judgments into observations and listening behind the story for what feelings, needs, and requests the disputant might have. In a round about way, disputants also can have these distinctions reinforced. Often the role-play is a basic situation that is made up, for example, a landlord and tenant are arguing over who will pay for repairs to the house. In order to take on their respective roles, the two disputants to a certain extent must put themselves in the role, and imagine what they might be thinking and feeling in that situation. Though they are likely to play the role in a judgmental and reactive fashion, thinking about what that person might be observing, feeling, and needing can assist in making the role-play as real as an imagined situation can be.

The skill of making requests is additionally reinforced through the fact that the entire mediation process is facilitated by the mediator making a series of doable, present tense, action language requests. For example, asking a disputant “are you feeling irritated because you have a need for respect?” is a request that follows the guidelines of NVC. Asking a person to reflect needs is as well; “I heard Person A say that they have a need for respect, would you be willing to tell them that you also heard them say they have a need for respect?” As the mediator gets coached on and practices these requests, the skill is reinforced for all present. The mediator may also help the conflictants make doable, present tense, action language requests, further enhancing this learning.

Finally, as mediator, the entire process of mediating a conflict revolves around being aware of where your attention is and what’s happening in the process. Are you focused on one disputant, offering empathy, or are you doing self-empathy or revealing what’s going on for you? Practicing this kind awareness in class situations increases participant's ability to carry it with them throughout their daily lives.

Enhanced Learning Through Immediate Feedback

A crucial factor that turns the three-chair model into a powerful learning context is immediate feedback. The mediator gets three types of feedback, each important. First, he gets the immediate feedback during the role-play of how each disputant responds to what he says; perhaps the disputant pushes back, or evidences resistance or frustration, which can be a message to the mediator that something isn’t working. Or, the disputant may show signs, through body language or tone of voice, that he has shifted into a more open place, which is immediate evidence that he is feeling more connection.

Second, during the harvesting after the role-play, the mediator gets feedback from the disputants, the coach, and from any additional observers who are present. For example, out of role, a disputant may review and say what was happening internally for her at various points in the role-play. This feedback is still close in time to what the person in the mediator’s chair experienced. This tends to be very powerful learning for the mediator, to hear first-hand the needs met or not met in others in relation to his choices. When we coach people in the debrief session, we request the feedback be stated using the distinctions of NVC—giving the observations of what the mediator said or didn’t say and how the disputant felt about it before and after.

The third kind of feedback comes in the form of coaching. One of the reasons we think people report such value from using this model is the implicit and explicit permission for us to coach as closely as we do. The closer in time the feedback is to a person’s action, the faster they will be able to learn. For example, if you are learning a foreign language and you speak with someone who will interrupt you right after you’ve used an incorrectly conjugated verb and drop in the correct conjugation for you to pick up and keep going in your sentence, you will probably learn more quickly. This is the kind of thing we try to do in coaching. While interfering with the flow as little as possible, we drop in suggestions such as, "I might try this way of saying it...," or "A few choices I see right now are...." We might offer another way of saying something, and as the mediator immediately uses it in the context of his role, that suggestion gets stored in the brain in a way that is more readily accessible next time he is in a similar situation. We don’t offer the suggestion as if it's the correct one, but rather as something we might do that in our experience is more consistent with the intention of creating connection.

The combination of the immediacy of the feedback and the permission for it provides a powerful learning context. What we are engaged in, after all, is adult learning—as children, our grid of discernment is not very strong, but as adults it can be. When people come in actively wanting to learn, asking for feedback, and willing to be coached, it reduces that grid of discernment and makes learning more likely to occur. We find it crucial in adult learning contexts that the intention to be open to feedback is present, and that feedback is immediate.

Returning to Presence

Repeated experience with the three-chair model also helps people embody the reality that they might be the stimulus for another person's reaction, but they are not the cause. Since you can never be inside another person’s experience, you never really know what is true for them, and thus, when trying something that you hope will connect, you never really know whether it will work. Even if something similar has created connection with a lot of people prior to this situation, with this person, you don’t know. When you make a guess that has been connecting in the past, and in this situation the disputant evidences irritation, it can really sink in that their reaction is about their internal makeup and not your guess.

In practical terms, this realization gives you more freedom; if you are not the cause for the other person’s reaction, then you don’t need to go into your own reaction about their reaction. You can skip the whole stage of being in reaction, of judging yourself or the other person. You don’t get blocked in your own learning by reacting, going into explanation, or shifting the focus. You can simply stay present with that person, essentially saying to yourself, “oh, that didn’t land the way I wanted it to land, but I have these other things I can try, let me try this one.” This helps you stay present with the intentions of wanting to connect and learn.

This point is worth a little more explanation. We can never predict with certainty what is going to connect with another person, though over time our guesses are likely to begin to land more often; however, over time we develop more confidence that if our guess does not land the way we wanted, that we will be able to come up with something that has a likelihood of contributing to connection. We can use this phenomenon—not knowing what will connect because we don’t know what is going on with the other person—as a way to support returning to presence. At every moment in the conversation, we have new information based on what the other person has said or done, and as long as we are not stuck in the past through judging what has already happened—that we didn’t language a guess that was connecting for the other person—we have an opportunity in this moment to make another attempt at connecting. Being in the mediator’s chair is a real practice in returning to presence.

Shifting out of Reactive States 

We have found that even when the role-plays are based on imaginary situations, the disputants become emotionally involved in their roles. For many generic conflict situations, most of us can imagine the thoughts and feelings we would have being in that situation, and as a result, role-plays can take on a certain reality. Of course, there are not as many layers as a real dispute between people, particularly people in long-term employment and personal relationships, but nonetheless, disputants get into their roles, and often take on a kind of resistance to shifting out of their position. It is particularly powerful, then, when a disputant experiences a shift, and this provides rich learning for everyone.

People in the disputant role often report after a role-play something like the following; “I had decided that I wasn’t going to shift; I was not going to look at the other person or be moved by anything the mediator said. When the mediator asked me to reflect, I refused, and he then gave me empathy so I was willing to reflect, but I still wasn’t going to soften. And then when I said the words back, even though my tone was grudging, I did feel the beginning of an opening.” We have experienced the same phenomenon when we have role-played as disputants.

Hearing about these shifts is helpful to the mediator, though they don’t happen every time. In fact, while in the middle of offering empathy to one disputant, it is never clear whether the other person is getting connected through hearing the conversation or whether they’re in reaction because they are hearing judgments instead of the translation to needs. As facilitators and mediators, we have officially given up on trying to predict whether one conflictant is in reaction or whether their heart is opening based on what is taking place. There are certain signs that we can read, particularly body language and facial expression, but we are so well trained in our culture to restrain our movement and keep a poker face that it can be difficult to read what is going on for people. Sitting in the mediator’s chair over and over again reinforces the fact that while the mediator can try to be aware of how each disputant is doing, prediction is imperfect, and rather than trying to manage each person’s reactions, the point is to develop the trust that you can handle whatever reactions show up.

Experiencing the Effect of Interior States on Mediation

Another rich source of learning in using the three-chair model is how the internal state of the mediator affects the disputants and the course of the mediation. For example, if there is any sense of demand energy from the mediator, the disputants will react to it in the role-play or report it in debriefing. It is incredibly valuable to people to experience first-hand how quickly they and others pick up any sense of judgment, anxiety, or demand. Similarly, if the mediator holds any enemy images of themselves or the disputants, this will typically come out during harvesting, with disputants noting that they felt the judgment. It is natural for a mediator, as a human being, to sometimes find he begins to side with or have an affinity for one disputant’s story, perhaps because he relates to it more than to the other side; typically, however, the other disputant will tend to pick this up—even if they don’t know exactly what was going on—and report that it affected how they felt.

The three-chair model also provides the opportunity for people to experience the power of self-empathy. Particularly when a person is new, we offer self-empathy when he first sits in the mediator’s chair, using the reactions that are coming up and the anxiety, distress, or discomfort he may be experiencing about his choice to mediate. At times, we also push the pause button on the role-play and ask the mediator to do self-empathy out loud, often supported by the coach, then return to the role-play. Quite frequently, one or both disputants will report in the debrief that they noticed a distinct shift in the mediator’s presence from before self-empathy to after; they both felt more connected to the mediator and liked what he did more after self-empathy. This is a particularly rich learning experience for two reasons. The first is that it actively coaches people as to what self-empathy actually is. Often as facilitators, we hear people report that they have done self-empathy, but when we ask them what language they used with themselves, it was not what we would call self-empathy. In these self-empathy moments before or during a role-play, the mediator and the disputants get coaching on how to do self-empathy. Second, all people present get the positive feedback of experiencing the tangible benefit of doing self-empathy; in real time they get to experience the shift in a person’s being that happens when they get connected with their needs.

We also encourage out loud self-empathy because it highlights the model we suggest people use after they have participated in mediation, have judgments about that mediation and want to shift out of those judgments and into learning. Self-empathy can be an essential step in this shift.

Skill and Confidence Mediating Conflicts

Something we have seen and heard from participants since the beginning of offering our trainings was how relatively quickly people gained skill in the mediator role. It has not been uncommon for people to begin our classes with anxiety and great difficulty moving through the steps and skills of the mediation process. Within weeks, however, and particularly by the end of an eight to twelve week series of three hour classes, most of these participants demonstrate significantly increased ease and confidence when in the mediator's chair.

Healing and Reconciliation in Personal Relationships

Sometimes participants in our trainings choose to work with real conflicts in their personal, community, and work lives. The person whose conflict it is can elect to play herself in the role play or be the mediator of her own conflict while others play her role and the role of the other person. Additionally, she could choose to be an observer and watch her conflict being mediated and played out by others. The benefit we have seen (and occasionally experienced for ourselves) is that the person with the real conflict often experiences a significant shift towards greater inner peace, as well as clarity and confidence in how they would like to respond to the actual conflict. It is also not uncommon for us to later hear from people who do this work on their conflicts in training that they subsequently communicated with the others in the conflict in a way that led to resolution, as well as healing and reconciliation.

Awareness of Interpretations

One of the benefits that indirectly comes from participating in the three chair model particularly shows up when people are in the role of observer. In the feedback portion, I have many times see people new to the observer role start their feedback with a description of the internal state or reaction of one of the disputants or the mediator as if their description is a reality, then proceed to comment about what happened subsequently. For example, John has been observing Joe mediate a role play between Mary and Diane, and says during feedback, “Well, after Joe said… then Diane got resistant and reactive; her feelings were clearly hurt, which was understandable in that situation, and Joe should have anticipated that she was going to react that way.”

Invariably, what I will do in this type of situation is to say, “Excuse me a second, before you go any further I’d like to check to check that with Diane. Diane, was that what was happening for you at that point?” Before John continues any further with his feedback, I want to check to see if that was really Diane’s internal state. Quite often it turns out that John’s interpretation is incorrect. Diane might respond by saying something like, “Well, no, right then I was just feeling confused and didn’t really understand what was going on. I just wanted the mediator to tell me what was happening. I wasn’t hurt, I just didn’t understand.”

Of course, we all interpret what is going on for other people all the time; this isn’t a problem in and of itself. The problem is twofold: first, we don’t check our interpretations; we simply believe they are true, and second, our belief leads us to respond in various ways that do not necessarily lead to what we want. In some cases, our certainty that we are right gives us permission to tell the other person that they shouldn’t be experiencing what we believe they are experiencing. In others, we limit ourselves; we edit our own conduct or speech because we are so certain of our prediction of what someone else will say or do in a situation, based on our belief of what they think or feel. When we believe that we know someone else’s internal state without checking with that person, it can lead us into a dead end; we put ourselves in a box and get either depressed or angry because we don’t see any options in how to proceed. In my view, believing our interpretations of what is going on for someone else is a core source of distress in the human condition.

As the above example implies, what we find out when we do begin to check with people whether our interpretations are correct is that, often, they are incorrect. What we think is going on internally for someone is, in fact, not what is happening. Finding this out puts a crack in the rigidity of our belief that we know what other people’s feelings, motivations, or intentions are. This crack begins to provide more freedom; if we can get some sense that the other person may in fact actually be different than the way we are judging them to be, we begin to see more options in front of us. When I began checking my interpretations with people and found out how often I was completely incorrect, I became much more open to trying things out with people since I had loosened my certainty that I could predict what they would do.

The three-chair model, and particularly when people are in the observer role, provides a training ground for people to begin to gain awareness of their interpretations of other people’s internal state and the cascade of responses that emerge from their belief in their interpretation. People report that they gain awareness that they don’t necessarily know what’s going on for people even when they think they do, and this awareness then gives them more options to choose from in how to respond.

Variations on the Three-Chair Model

Though in its simplest form, the three-chair model includes only a mediator and two disputants, at times variations on this model can yield interesting results. Sometimes the numbers in the room work out such that a group also has an observer, and we have found this to be valuable for everyone. People report another level of learning that takes place from being an observer, and in this role participants contribute to the learning for the mediator, the disputants and us through the feedback they offer. Another benefit to having an observer is simply in the additional person; it can be difficult for some people sitting in the mediator’s chair to be able to maintain presence with additional observers, and this becomes part of the learning process in these cases.

Another variation is to add an additional disputant. This adds to the complexity the mediator holds, and thus adds another layer of learning. This scenario is particularly interesting when externalizing an inner mediation, that is, when an inner conflict of a participant is being role-played with different people in the roles of different internal voices. Since we all have many internal voices that can be in conflict with each other, having another person available to embody an additional voice that shows up helps maintain a clarity and focus to the mediation.

Often with the way we work, the coach is in effect in the role of an observer; however, this is not always the case. In another variation, the coach can play one of the disputants, or may even hop into the mediator’s chair if participants request it, to model working through the situation at hand. 
Finally, another variation adds people acting in the role of agent—someone who is acting on behalf of someone else. An agent, depending on the situation, might be a lawyer, a real estate agent, or a health care advocate. The agent is tasked with representing their principal, who might be present in the mediation or not. A role-play situation can be set up with just the two agents participating in the mediation or with both agents and principals participating, in which case you have four disputants and might then also choose to have more than one mediator.

This variation adds another level of complexity to the mediation; besides the conflict that brought all the parties together, the agents are also parties to a potential conflict since they have their own internal motivations for completing a deal. A person acting on their own behalf has much greater flexibility in the positions they take in a dispute because they are doing their own internal calculus of whether a proposal will meet their needs. When you have agents involved, however, the agents are not only doing their internal calculus of what will meet their needs as an individual in the situation, but they are also thinking about their duties and obligations to their principle. An agent will advocate for the position that is most favorable to their principal unless they have been specifically instructed that they can accept something less than the optimum position with regard to each point. This creates difficulties in mediation and gives mediators and disputants in a role-play rich material to practice with.


We find the three-chair model to be a powerful pedagogical tool that goes beyond people simply learning conflict resolution skills. It provides a unique way to practice with close in time feedback from a coach and others, it offers ample opportunity for personal growth, helps people integrate core NVC skills into their daily life, and supports the development of concrete mediation skills. As we have used this model over the course of years, we see the small shifts people make accumulate over time, fundamentally shifting their ability to be present and connect with themselves and others even in difficult conflictual situations. This gives us hope. With the looming crisis we see ahead—exponential population growth and climate change resulting in huge dislocations of people, all greater stimulus for conflict—we would like people to be able to cope more effectively with whatever lies in the future, collectively and individually. For us, NVC and the three-chair model provide a strategy for people to integrate skills into their lives that allow them to learn a whole new way of being and be agents for the creation of the kind of world that they want to live in.

© 2009 Words That Work in Business