It is trite that one of key tasks of a mediator is to sort out what Patton, Stone and Heen in their book “Difficult Conversations” refer to as the “What Happened” conversation. Although they write about this and two other conversations in the context of negotiation, their insights are pertinent to mediators as well.

One of the challenges of sorting out the “What Happened” conversation is the assumption that a certain objective reality exists. And this writer is not saying that an objective reality does not exist, but only that we, as humans are unable to remember and represent an objective reality. At the most fundamental level, our memories of what happened are subjective. Of course, there is no problem if both parties’ subjective realities match. Which reminds me of the old joke that the only difference between sane and insane people is that more people share sane people’s realities.

But why should subjective realities differ, in some cases significantly, from objective reality? Information scientists estimate that at any one point, 2 billion bits of information bombard our nervous system. It would be impossible to pay attention to every bit of that information. As such, our nervous system must filter out information that is not relevant to us at that point in time. These filters comprise of our, among other things, memories, experiences and values. These form an overlay through which we view objective reality. These overlay have clear areas that allow us to see parts of objective reality clearly, translucent areas that distort what we see and opaque areas that prevent us from seeing certain things.

Accepting this to be correct, then one’s subjective reality is merely a shadow of objective reality. And there is nothing wrong with this as long as we are aware of it. The problem of course is when we assume that this subjective reality is objective reality. This is when misunderstandings and misperceptions occur. In conflict, mediators know this to manifest as, among other things, attribution bias and blaming behaviours. It is not surprising that narrative mediation proceeds on the assumption that parties have different incompatible stories and part of the mediator’s task is to help them to deconstruct their initial stories and reconstruct a better one.

One of the ways that the authors of “Difficult Conversations” suggest to deal with this problem is to take parties down the ladder of inference. Put another way, parties select different aspects of objective reality and make inferences and conclusions from that selection. The process of selection and making inferences is largely unconscious. Getting parties to be aware of the data they used to make their inferences and conclusions allow them to be open to other possibilities based on other data. Sharing that data and their reasoning with the other party can also help them understand one’s own conclusions. This is oftentimes a far more productive conversation than parties tossing conclusions at one another.

The field of NLP utilizes another method to deal with the “What Happened” conversation. The Meta-Model of questioning was modeled from Virginia Satir’s systemic family work. What the co-creators of NLP noticed that the questions that Virginia asked her clients often led them to readjust their subjective realities. The assumption here is that since our subjective realities shape our behaviours, a readjustment in subjective realities might lead to an adjustment of behaviours, hopefully for the better. This idea, of course, is not new to mediators. Our very act of reframing statements is an adjustment of parties’ subjective realities, albeit at a surface level.

Much has been written about the Meta-Model and its use in various contexts including dealing with various attribution biases in conflict situations. I do not propose to reproduce that here. Instead, I would like to share with readers three Meta-Model linguistic patterns that mediators may find helpful when seeking to adjust parties’ subjective realities.

The first Meta-Model linguistic pattern is known as a mind-read. This is where the party makes a statement that purports to know what the other party is thinking or feeling.

An example would be “He doesn’t care about the company”. Many mediators will generally accept this statement and seek to gather more facts. Some might seek to reframe with a “So the company is important to you?” or acknowledge the emotion beneath with a “You must really feel upset about that”.

NLP prescribes finding out the evidence for that statement eg. “How do you know he doesn’t care?” Making the speaker identify the evidence achieves 2 things. First, it recovers data, much the same way as taking someone down the ladder of inference does. Secondly, it surfaces to the speaker’s attention that this view is an attribution of intention on the part of the speaker and not reality. In an ideal situation, this question will get the speaker to acknowledge his subjective attribution and become more open to other realities.

The second Meta-Model linguistic statement is a connecting statement. Essentially, two concepts X and Y are subjectively connected. For example “He is a bad father because he comes home late everyday.” In this statement, a subjective connection is made between “Coming home late” and “Being a bad father”.

NLP seeks to question the connection between X and Y. The mediator might ask “How does coming home late mean he is a bad father?” The answer to this might get more information about the interests involved eg. spending time with children. The mediator might also ask “In what ways can one come home late and still be a good father?” This question will cause the speaker to reconsider the connection made. There are other ways in which this connection can be delinked via the mathematical construct of Cartesian Coordinates. The writer will reserve this for another entry.

The third and last Meta-Model linguistic pattern is known in linguistics as a nominalization. Abstract nouns (words that look like nouns but do not exist in reality) like success, love, relationship, problem, integrity, happiness, communication and peace are all examples of nominalizations. These words are essentially processes that have been made static and generalized across time.

Examples of statements containing nominalizations would be “Our relationship is not working out” or “We have bad communication”. The problem with these statements is that the nouns seem immovable making the situation seem final.

NLP prescribes denominalizing abstract nouns. For example, in response to the first statement, one could ask “In what ways would you like to be relating instead?” Denominalising “relationship” back into the process of relating adjusts the speaker’s subjective reality so that there is now some movement possible. Things don’t seem so final. The same can be done to the second statement by asking “How would you like to be communicating instead?”

It is useful in closing to make 2 points about using the Meta-Model.

First, the Meta-Model is not a magic bullet. Do not expect a single question to make a complete shift in the speaker’s mind. It is like a reframe. It can sometimes make a significant shift or you may need to string a number of interventions together to achieve the shift in subjective realities.

Secondly, asking these questions can sometimes be threatening as it causes the listener to question their subjective reality. It is therefore vital for a mediator to maintain rapport with the parties and utilize softening frames like “Help me understand how…” Or “Let me play devil’s advocate here and ask you…”

I have only shared 3 patterns in this entry. There is of course much more to the Meta-Model. I hope that through this entry, I have whetted the readers’ appetite to try this out and to find out more. Enjoy!


To make sure you do not miss out on regular updates from the Kluwer Mediation Blog, please subscribe here.

Kluwer Arbitration


  1. Fascinating stuff Joel. I can see the benefits of each move. Your post reaffirms how much mediators can learn from NLP.
    One worry, however: some of the moves you describe seem to position the mediator as therapist, asking the questions (possibly on behalf of the other party, but still doing the speaking). Might there be a subtler form of attending to mind-reads and connecting statements in which the mediator simply notices and paraphrases them? For example, in your first instance, the mediator might say, ‘So, the situation has led you to conclude that he doesn’t care about the company?’ Rather than probing the evidence herself, the mediator could simply hold the space for the other party to do so. Or the mediator might add, ‘She’s sitting right here. Do you mind if I ask her if you have got that right?’

    1. Hi Charlie, my apologies for a late response. Your concern is quite correct of course and I agree with your suggested approach. If one believes in private sessions, then one could ask these questions in that context a bit more directly. The main idea remains the same i.e. attending to the statement made in a way that opens up the space for movement.

      As for the line between mediator and therapist, don’t get me started. At least not here! Perhaps in a future entry! 🙂

Leave a Reply

Your email address will not be published. Required fields are marked *