Symbolic Modelling Lite

Emergent change though metaphor and Clean Language
Print Friendly, PDF & Email

First published as Chapter 4 of Innovations in NLP: Innovations for Challenging Times (Eds. L.Michael Hall & Shelle Rose Charvet) Download chapter here

Why Is This Model Important?

Before we introduced David Grove’s work into NLP, metaphors were mostly used to tell Milton Erickson-style stories. There was little use of autogenic metaphor—metaphors generated by the client.1 When we stumbled upon David Grove we realized he had devised a new way to “study the structure of subjective experience”—the raison d’être of NLP.2

David Grove is best known for Clean Language—a questioning model designed for working with the metaphoric and symbolic domain of experience.3 In 1995 we decided to model his innovations which led us to write Metaphors in Mind: Transformation through Symbolic Modelling. While we incorporated many of Grove’s ideas we also drew upon recent findings in cognitive linguistics, self-organizing systems theory, and evolutionary dynamics. Our aim was to create a model that could be applied to a range of contexts in addition to psychotherapy — in education, management, research, and so on.

When we recently reviewed our model we saw that the process has a central core — Symbolic Modelling Lite — which is presented here for the first time.4 Mastering the Lite version means you will be able to facilitate people to identify, develop, explore, and evolve their metaphors using the basic Clean Language questions. If you want to go further you will need to acquire the skills of a symbolic modeller; which will take somewhat longer.

Using Symbolic Modelling in a Troubled World

Just about everyone uses metaphor all the time — often six times a minute! We are surprised by this figure because most metaphors are constructed, spoken, and received out of awareness. Research over thirty years has shown that we not only speak and gesture in metaphor; we think and act on the basis of our metaphors. Your clients, colleagues, customers, friends, and enemies will all use metaphor — and not just occasionally, but much of the time.5

Metaphors can be a source of creativity but they can also specify and constrain ways of thinking, thereby maintaining unproductive and self-destructive patterns of behaviour. Knowing how to listen and observe a person’s metaphorical words and gestures gives great insight into how their inner world works, and Clean Language is tailor-made for modelling the process and structure of that world.

Symbolic Modelling has been used successfully as a change, creative, and educational process in settings as diverse as a maximum security prison, the NASA Space Center, and an elementary school. There is a growing recognition of its value in business, life, and sports coaching. While it is highly effective for working with “everyday” issues, it is particularly suited to:

  • The big issues of life, e.g., finding a sense of purpose.
  • Ill-defined feelings, e.g., something is wrong, fearful, unsafe, or missing.
  • Identity and spiritual levels.
  • Internal conflicts.
  • Intractable and double-binding patterns.
  • Trauma work.

While Symbolic Modelling has traditionally been used to facilitate individuals to develop themselves, it has also been applied in large-scale environmental projects, such as creating a strategy for European-wide sustainable land management; using hydro-thermal water to heat ex-mining communities in Holland and Scotland; and responding to rising sea-levels by re- visioning the Dutch attitude from “holding the ocean back” to “living with water” and “water cities.” The Modelling Shared Reality process used in these projects provides a snapshot of the current collective experience of those involved and a communal voice for those not usually engaged in decision-making—thereby turning metaphors into action.6

What is Symbolic Modelling?

All change processes require a medium, a method, and a means. In Symbolic Modelling these are: metaphor, modelling, and Clean Language. Together they can be used in three ways: to model successful strategies and states of excellence; to facilitate change; and to facilitate individuals and groups to create new metaphors (see Figure 1).7

Figure 1. Three ways of applying the components of Symbolic Modelling

Unlike the majority of psychology which aims to discover generalities about humans, Symbolic Modelling seeks out the distinctive and idiosyncratic organization of each individual’s map of the world. For example, a client says they are “at a crossroads” in life. This is quite a common metaphor. However, after a few Clean Language questions it becomes clear that this crossroads is unique. It involves train tracks, an approaching train, barren land on one side of the tracks and lush nature on the other. Also, while the desire in the client’s chest wants to take the small leap across the tracks, their legs are stuck because of the hands of responsibility in their stomach holding them back. The idiosyncratic is important because as Aristotle said, “There is no science of the individual.”

Metaphor—The Medium

As George Lakoff and Mark Johnson observe, “The essence of metaphor is understanding and experiencing one kind of thing in terms of another.”8 Research shows that the majority of our metaphors and similes are derived from our understanding of the body and the workings of the physical world.9 When a client says they feel like a fish out of water we instantly grasp how they perceive their situation. Because we already know the nature of “fish,” “out of,” and “water” we can carry across our experience of the physical into the abstract mental realm.

Metaphor makes the intangible tangible, it embodies relationships and patterns, and it captures the essential nature of an experience. Metaphor provides a number of advantages:

  • It allows a client to work content-free.
  • It makes it easier to work at the process and structural level.
  • It operates at multiple levels simultaneously.
  • It is like a gift that keeps on giving. In the session a process unfolds which can continue for days, weeks, and sometimes years. A good resource metaphor can last a lifetime.

Modelling—The Method

Symbolic Modelling differs from other forms of modelling in three ways. The first two are obvious: we model the organization of people’s metaphors, and we use Clean Language to do it. The third is more subtle, our primary aim is for the client to self-model.

The entire focus of Symbolic Modelling is an exploration of the client’s metaphoric model of the world from their perspective, within their perceptual time and space, using their words and non-verbals. Instead of a conventional dialogue there is what David Grove called “a trialogue” between facilitator, client, and their “metaphor landscape” — the four-dimensional, psychoactive world that emerges within and around the client. The facilitator sets aside their own perceptual space so that only one metaphor landscape occupies the physical space — the client’s.

Four Fundamental Modelling Processes

Having studied hundreds of Clean Language sessions we have concluded that experienced facilitators make maximal use of just four modelling processes: Identify, Develop Form, Relate over Time, and Relate across Space:

  1. Identify: To establish, recognize, or distinguish what something is; to name and give something an identity; to individuate an element or characteristic. At each level a different kind of something can be identified: an attribute, a symbol, a relationship, a pattern, a context.
  2. Develop Form: To elaborate what has been identified; to identify enough attributes of something that its nature becomes apparent; to bring a (symbolic) perception to life — like a pre-digital photograph emerging from developing solution.
  3. Relate over Time: To identify a sequence of events (Before–During–After); to identify temporal relationships such as cause, effect, contingency, precondition, provenance, and expectancy.
  4. Relate across Space: To identify relationships between separate things, places, perceptions, frames, contexts, and so on.

The four modelling processes are fundamental because they are so widely applicable. We have used them to model resources, desired outcomes, problematic situations, changes, the structure of excellence, conflict, corporate metaphors, and so on. Figure 2 shows how the four processes relate to each other.

Figure 2. Four fundamental modelling processes

At first it may seem strange that a process based on modelling—with no intention by the facilitator for the client to change — can produce significant and long-lasting development.10 That’s the mystery of self-organizing systems. As a client’s metaphors are identified, developed and explored, their system learns from itself. As the landscape evolves they discover new ways of perceiving themselves and the larger systems of which they are a part. In doing so, their everyday thinking, feeling, and behaviour correspondingly change.

Clean Language—The Means

The three functions of Clean Language (to acknowledge, orientate, and send on a quest) and its four components (the syntax, vocal qualities, gestures, and clean questions) have been well documented and are freely available on the web, so we have not repeated them here.11

To be clear, Clean Language influences and directs attention—it wouldn’t be useful if it didn’t. However, unlike other uses of language, Clean Language is “clean” because it is sourced in the client’s exact vocabulary, it is consistent with the logic of their metaphors, and it only introduces universal metaphors of time, space, form, and perceiver.12

There are nine basic questions which form the beating heart of Symbolic Modelling because they are asked so often.13 Over the years we have devised a number of ways of organizing the questions.14 Below they are arranged according to the four fundamental modelling processes:


And what would [you/X] like to have happen?

And that’s [ ] like what?

Develop Form:

And what kind of [ ] is that [ ]?

And is there anything else about (that) [ ]?

And where/whereabouts is [ ]?

Relate over Time (within and between events):

And then what happens? or And what happens next?

And what happens just before [event]?

Relate across Space (within and between perceptions):

And when/as [X], what happens to [Y]?

And is there a relationship between [X] and [Y]?

[ ] = A client’s exact word or phrase.

How Does Symbolic Modelling Lite Work?

Symbolic Modelling is an outcome-orientated approach. By making a client’s desired outcome the focus of the four fundamental modelling processes, a simple framework for change is created. Figure 3 illustrates the six phases and the iterative loops involved.15

Figure 3. Symbolic Modelling Lite

Phase 1: Set Up16

Setting up a clean process itself needs to be clean. We want to offer the client a chance to align their inner perceptual world with the outer physical world where the session takes place. Thus the client decides where they want to be, and where they want the facilitator to be:

With both client and facilitator standing:

And where would you like to be?

When the client has positioned him or herself:

And where would you like me to be?

Phase 2: Identify a Desired Outcome—The Problem-Remedy-Outcome (PRO) Model

Once the client has arranged the seating just as they would like, they are asked the opening question:

And what would you like to have happen?

A person will invariably respond to this question in one of three ways—with a statement of: (1) a problem, (2) a proposed remedy, or (3) a desired outcome. To cleanly facilitate a client to identify a desired outcome we created the Problem-Remedy-Outcome (PRO) Model.17 The PRO model has two stages. First we use the client’s precise language to determine whether they are attending to:

A Problem—a current difficulty they do not like; or
A proposed Remedy—a desire for a problem to not exist, be reduced, solved, or avoided; or
A desired Outcome—a desire for something new to exist.

When the client responds with a problem or a remedy (which will also refer to a problem), the problematic aspect is acknowledged and noted for later use. In the second stage we respond with a question that invites the client to shift their attention to a desired outcome. Depending on how committed the client is to problem-thinking or to problem-solving, he or she may need to iterate round the loop a few times before they settle on an initial desired outcome.

When the client gives a desired outcome statement we ask a question that keeps their attention on that aspect of their experience. Figure 4 shows the PRO model in its entirety.

Figure 4. The Problem-Remedy-Outcome model

The following transcript shows how the nine basic questions are used during the developing, exploring, and maturing phases of Symbolic Modelling Lite. All words introduced by the facilitator are highlighted in italics. This makes it easy to see the syntax of the questions and to distinguish between client- and facilitator-introduced words:

F: And what would you like to have happen?

C: I feel like a fish out of water. [Problem]

F: And when you feel like a fish out of water, what would you like to have happen?

C: I want to stop suffocating. [Remedy]

F: And you want to stop suffocating. And when you stop suffocating, then what happens?

C: I can be at home with myself. [Desired Outcome]

Since a client can express a problem, propose a remedy, or desire an outcome at any time, we run PRO continuously in the back of our mind so that we are always paying attention to what the client is paying attention to, and are ready to respond with the relevant P, R, or O question accordingly.

Phase 3: Develop a Desired Outcome Landscape

Once a client has identified a desired outcome we facilitate them to develop that statement into an embodied metaphor landscape. As Grove put it, “to make words physical.” We do this by repeatedly asking the three classic developing questions:

F: And when you are at home with yourself, what kind of home is that?

C: It’scomfortable.

F: And when it’s comfortable, where is it comfortable?

C: In my heart [touches chest].

F: And whereabouts in your heart?

C: Right at the core.

F: And it’s comfortable right at the core, in your heart. And is there anything else about that comfortable there [gesture to client’s chest]?

C: It’s flowing.

F: And when it’s comfortable and it’s flowing, it’s comfortable and flowing like what?

C: Like a river.

Metaphors comprise a number of interacting symbols (in this case, “fish,” “water,” “home,” “heart,” “river”). Symbols exist in a client’s perceptual space when they have a location and their attributes can be described. This usually means the client can point to the symbol and could, if asked, draw or enact it. While symbols are being identified, located, and described the client will usually mention the relationships between them. These too can be developed with the same clean questions.

Symbolic Modelling is entirely additive.18 We aim to facilitate the client to retain everything relevant to their desired outcome in one perception. Developing a desired outcome landscape is not something to be “got though.” It is central to the whole process and encourages the conditions under which change is the natural response. In other words, the client’s self- modelling is preparing them to evolve in ways they are not yet aware of.

Phase 4: Explore Effects of Desired Outcome Landscape

Once the ground is prepared, the client can explore the effects of their desired outcome happening (the “ecology”). This is done in two ways. We invite them to attend both to what happens after their desired outcome occurs, and how the desired outcome handles the problematic situations previously described (this is why we noted the exact words for the client’s problems in Phases 2 and 3):

F: And as a river is flowing at the core of your heart, then what happens?

C: It’s my destiny.

F: And when it’s your destiny, what happens to a fish out of water?

C: It’s trying to get to the spawning grounds but it’s been frozen in mid-jump.

In Symbolic Modelling we are always on the lookout for what surprises the client about their inner world. Rather than trying to solve a problem or to make something happen we remain vigilant to the out-of-the-ordinary, and make that the focus of attention. Then we “follow the white rabbit,” just like Neo did in the film The Matrix.

Should a new problem arise at any phase (such as, “frozen in mid-jump”) we apply the PRO model and return to developing the enhanced outcome landscape:

F: And when it has been frozen in mid-jump, what would that fish like to have happen?

C: To get back in the water where I belong.

F: And what happens just before that fish gets back in the water where you belong?

C: I trust myself—it always comes back to trust.

F: And when you trust yourself, where is that trust?

C: In my heart again.

F: And is there a relationship between trust and flowing river at the core of your heart?

C: Yes, I can trust myself when the river flows.

We facilitate the client to keep going round their iterative loops until one of two things happen: they notice a pattern or a change occurs.

When a pattern is indicated (in this case by the client saying “it always comes back to”) we continue as before but now use the client’s metaphors to address the pattern in its entirety. Then, when the client makes a change, they are not just resolving the presenting problem but the class of experience of which the problem is just one example. In the future, when life presents similar but different problematic situations they will handle them in new ways.

F: And you trust yourself when the river flows at the core of your heart, and you’d like to be at home with yourself, and that fish would like to get back in the water where you belong, and it’s your destiny … and then what happens?

C: I realize I’ve been fighting who I am. I have to accept it’s a struggle but I am meant to be in the water. That’s the only way I am going to get upstream.

F: And when you are meant to be in the water, and that’s the only way you’re going to get upstream, what happens to that fish out of water?

C: She completes the jump.

A change is indicated by the client’s multiple realizations and because the fish “completes” the jump. When a change occurs at any time we immediately move to Phase 5 — maturing the change.

Phase 5: Mature Changes As They Occur

Figure 5 depicts a beautiful symmetry; the same four processes used before a change occurs are used after a change—but for a different purpose.

Figure 5. Four modelling processes in change work

Our aim during the maturing phase is for the client to find out whether the change starts a contagion which creates a new or reorganized metaphor landscape, or whether it invokes doubts, concerns, or fears. A problem reaction is not a sign of failure — quite the opposite. It indicates the client’s system is revealing more of its complexity and is acknowledging its “current reality.”19 This increases the chance that the client’s system will reorganize in a way that is more than relief of a symptom, more than a remedy. Rather it will be a generative change — a robust, resilient change that keeps on giving.

Maturing is also not a phase to “get through” as quickly as possible. We were amazed to see David Grove devote up to a third of a session to what happens after the first change occurs. Just as a desired outcome landscape was developed and explored in Phases 3 and 4, so each change is matured in Phase 5 using exactly the same questions:

F: And as she completes that jump … then what happens?

C: I can continue.

F: And as you continue what kind of continue is that?

C: I’m comfortable in my own skin.

F: And when you’re comfortable in your own skin, what happens to a flowing river at the core of your heart?

C: I can trust myself even when it’s difficult.

F: And that fish completes that jump, and you’re comfortable in your own skin, and that river flows, and you trust yourself, and then what happens?

[And so on]

The maturing process helps consolidate the evolving landscape and gives the client ample opportunity to model how they now respond to the previous problems (“suffocating,” “frozen,” “fighting who I am”), and to find out what happens next.

Phase 6: Set Down

We know the process is finished when:

  • The changes come to a natural conclusion or resting place.
  • The logic of the new landscape is complete, coherent, and consistent.
  • No new problems surface and the new metaphors can handle the previously problematic
  • The client says “no” to the question: “And is there anything else you need right now in relation to [their original desired outcome statement]?”

If there is limited time left in the session we invite the client to continue the process between sessions by drawing their metaphor landscape and saying, “And get to know more about … [list their key metaphors], and to find out, what happens next …” To finish we ask something like, “And is it OK to leave it here?”

Caveats about the Use of Symbolic Modelling

Symbolic Modelling is inherently rapport-ful, forgiving, and robust — as long as you stay clean and connected to the client’s desired outcome. Even so, we recommend people only use it within their sphere of expertise. If you are a coach, use it in coaching; not with someone who suffers from a severe mental disturbance.

If a client gets into difficulties the first thing to remember is that most people have experienced their problematic patterns for many years, and while the pattern might be painful they know how to cope with it. Many facilitators step in too quickly and disrupt the client’s process. Your first response should be to stay within the metaphor and do as little as possible. In the rare case that a client does not naturally transition to a different state you can ask clean questions to invite them to attend to a resource symbol, their desired outcome, or a meta- perspective.

Often a facilitator has said to us that their client was “stuck”, when in fact it was they who were stuck. It is important not to confuse your state with what is happening for the client. When you don’t know what to do (and it will happen, it’s all part of bottom-up modelling 20), it is best to let the client set the next direction. You can:

  • Just wait.
  • Ask: “And is there anything else?”
  • Ask: “And what would you like to have happen now?”
  • Go back to the exact wording of their desired outcome, and ask about that.
  • Invite them to draw their metaphors.


Even after fifteen years of facilitating people with Symbolic Modelling we are still in awe at the unexpected and creative ways people find to change themselves. One client found that not only was her anxiety like butterflies in the stomach but that one particular butterfly had to open its wings and fly out of her mouth — but it couldn’t. You can probably think of several ways to help that butterfly, but we doubt you would ever come up with what worked for this client: all the other butterflies had to stand in a line and on the signal “go” flap their wings in unison to create sufficient updraft to propel the particular butterfly up and out of her mouth!

A clean approach encourages conditions whereby what changes fits the uniqueness of each client’s system. These changes are ecological because they are self-generated. We have found that these conditions are encouraged less by an expert magician and more by a facilitator who is an expert at getting out of the way.


  1. Notable exceptions were Faulkner (1991); Hejmadi and Lyall (1991).
  2. Dilts et al. (1980).
  3. Less well known are his other innovations such as Clean Space and Emergent Knowledge. What unites all Grovian processes is the notion of working “cleanly.” See Tompkins and Lawley (2003), Grove and Wilson (2005) and Harland (2009).
  4. “Lite” is less about food with fewer calories and more about fully functioning computer software aimed at entry-level users. Extra features require an upgrade and a more skillful operator.
  5. For a very readable and up-to-date account of the importance of metaphor see Geary (2011).
  6. The late Dutch environmentalist Stefan Ouboter and his colleagues devised “Modelling Shared Reality” which incorporates a clean approach in a number of innovative ways. See Tompkins and Lawley (2006c).
  7. Examples of Symbolic Modelling and Clean Language used to model excellence (Tompkins and Lawley, 2010); to create new metaphors (Lawley, 2001); and as an academic research methodology (Lawley et al., 2010).
  8. Lakoff and Johnson (1980), p. 5.
  9. Kovecses (2002).
  10. We coined the phrase “therapeutic modelling” to distinguish between modelling for change, and “exemplar modelling,” which aims to replicate and codify excellence. See Lawley and Tompkins (2006).
  11. Tompkins and Lawley (1997, 1999).
  12. Philip Harland elegantly presents “the case for achieving healing, change, and self-knowledge with minimal outside intervention” (Harland, 2011, p. 10).
  13. Other more specialized questions, which are only used when the logic of the client’s metaphors permit, are available in Lawley and Tompkins (2000) and Sullivan and Rees (2008).
  14. For our other models see Lawley and Tompkins (2004).
  15. Iteration is a process that repeatedly applies a rule, computation, or procedure to the result of the previous application of the rule, computation, or procedure. See Tompkins and Lawley (2007).
  16. “Set up” and “set down” are terms borrowed from McWhirter (2000).
  17. Tompkins and Lawley (2006a).
  18. Thanks to Steve Andreas for making this point explicit. See Andreas (2006, pp. 130-135), with prepublication comments in Lawley (2005).
  19. “Current Reality” means the entirety of what is true for a client at that moment in time. The term is borrowed from Fritz (1989).
  20. In top-down approaches, such as a step-by-step procedure, the facilitator has an idea of where the process is going and their job is to guide the client toward that end. In a bottom-up systemic outcome-orientated approach, like Symbolic Modelling, the end result is not known until the client gets there. See Tompkins and Lawley (2006b).


Andreas, S. (2006). Six Blind Elephants. Vol. 1: Understanding Ourselves and Each Other. Boulder, CO: Real People Press.

Dilts, R., Grinder, J., Bandler, R., and DeLozier, J. (1980). Neuro-Linguistic Programming. Vol. 1: The Study of the Structure of Subjective Experience. Capitola, CA: Meta Publications.

Faulkner, C. (1991). Metaphors of Identity: Operating Metaphors and Iconic Change. Audiotape and booklet. Lyons, CO: Genesis II.

Fritz, R. (1989). The Path of Least Resistance. New York: Ballantine.

Geary, J. (2011). I Is An Other: The Secret of Metaphor and How It Shapes the Way We See the World. New York: HarperCollins.

Grove, D. J. and Panzer, B. I (1989). Resolving Traumatic Memories: Metaphors and Symbols in Psychotherapy. New York, Irvington.

Grove, D. with Wilson, C. (2005). Six degrees of freedom: intuitive problem solving with emergent knowledge. ReSource Magazine 5 (Summer).

Harland, P. (2009). The Power of Six: A Six Step Guide to Self Knowledge. London, Wayfinder Press.

Harland, P. (2011). Trust Me, I’m the Patient: Clean Language, Metaphor, and the New Psychology of Change. London, Wayfinder Press.

Hejmadi, A. V. and Lyall, P. J. (1991). Autogenic metaphor resolution. In C. Bretto, J. DeLozier, J. Grinder, and S. Topel (eds), Leaves Before the Wind. Bonny Doon, CA: Grinder, DeLozier & Associates, pp. 99–122.

Kovecses, Z. (2002). Metaphor: A Practical Introduction. Oxford: Oxford University Press.

Lakoff, G. and Johnson, M. (1980). Metaphors We Live By. Chicago, IL: University of Chicago Press.

Lawley, J. (2001). Metaphors of organisation, Part 2. Effective Consulting 1(5).

Lawley, J. (2005). Steve Andreas on Symbolic Modelling. [Link available soon]

Lawley, J. and Tompkins, P. (2000). Metaphors in Mind: Transformation through Symbolic Modelling. London: Developing Company Press.

Lawley, J. and Tompkins, P. (2004). Clean Language revisited: the evolution of a model. Rapport 64.

Lawley, J. and Tompkins, P. (2006). What is therapeutic modelling? ReSource Magazine 8. [Link available soon]

Lawley J., Meyer, M., Meese, R., Sullivan, W., and Tosey, P. (2010). More than a balancing act? Clean Language as an innovative method for exploring work–life balance. Available at

McWhirter, J. (2000). Re-modelling NLP. Pt 6: Understanding change. Rapport 48

Sullivan, W. and Rees, J. (2008). Clean Language: Revealing Metaphors and Opening Minds. Carmarthen, UK: Crown House Publishing.

Tompkins. P. and Lawley, J. (1997). Less is More … The Art of Clean Language. Rapport 35.

Tompkins. P. and Lawley, J. (1999). Clean Language without words. Rapport 43.

Tompkins. P. and Lawley, J. (2003). Clean space: modelling human perception through emergence. Anchor Point 17(8).

Tompkins. P. and Lawley, J. (2006a). Coaching for P.R.O.s. Coach the Coach (Feb.).

Tompkins. P. and Lawley, J. (2006b). Modelling: top-down and bottom-up. [Link available soon]

Tompkins. P. and Lawley, J. (2006c). Using Symbolic Modelling as a research and interview tool. [Link available soon]

Tompkins. P. and Lawley, J. (2007). Iteration, iteration, iteration. [Link available soon]

Tompkins. P. and Lawley, J. (2010). Modelling Robert Dilts modelling. [Link available soon]

Print Friendly, PDF & Email
body * { color: inherit !important; }