Published in Acuity Vol.3, Nov 2012
A previous version first presented to The Developing Group, 14 May 2011
REPROCess: Modelling Attention
James Lawley and Penny Tompkins
This paper presents a new way to model attention. The material is introduced in a coaching/therapy context, however it can easily be generalised for use in many other areas and applications.
The development of REPROCess (Resource, Explanation, Problem, Remedy, Outcome, Change) began in 2002 when we noticed that even though facilitators started by asking a client for a desired outcome or goal, in a short time many of their clients were deep into a problem. Since this rarely happened to our clients we decided to model out these interactions. The result was our PRO, Problem-Remedy-Outcome, model (Tompkins & Lawley, 2006a). Over the years we added three more categories until in 2008 REPROCess emerged in its current form.
This project stimulated a keen interest in what happens to people’s attention when they are asked a question, and what the pattern of their answers reveals about the structure of their subjective experience. On the other side of the coin, we wondered how facilitators knew what their clients were paying attention to, and how they invited them to attend in ways that encouraged learning and development. We concluded that expert facilitators used three guiding principles. Moment-by-moment they:
1. Know what they are modelling, i.e. what kind of experience the client is having.
2. Know their purpose and where they are heading in relation to the overall frame/context and the client’s desired outcome..
3. Calibrate whether they are encouraging conditions for beneficial change – or not.
This article will concentrate on a way to enact the first principle. The method will also supply part of the second principle, and provide a basis for the third. We start by describing the REPROCess categories and how to recognise indicators in a client’s language. This is followed by a comparison of REPROCess with Robert Dilts and Todd Epstein’s SCORE model. The second half of the article is devoted to practical examples of how to use REPROCess to model a client’s words using the categories, and to navigate through a therapy/coaching session.
Because we specialise in using Symbolic Modelling and the Clean Language of David Grove in a personal change context, the examples we use draw from this source – but many other processes or methodologies can be enhanced by using REPROCess.
REPROCess is a linguistic-based model for distinguishing between six classes of experience:
Problem, Remedy and Outcome are highlighted since together they form the PRO Model, a model in its own right (see Tompkins & Lawley, 2006a).
any aspect that is valued or proves to be useful.
a description of the relationships between ideas, actions and events –
especially those giving a cause, reason or justification (also called
a current difficulty the client does not like.
a desire for a problem to not exist, to be reduced or avoided.
a desire for something new to exist (that currently cannot be enacted)
a difference that makes a difference. It will occur over time – i.e.
between an event ‘after’ and an event ‘before’ — and can therefore only
be detected retrospectively.
REPROCess identifies six categories that appear to be more or less universal ways of punctuating experience. Everyone encounters problems and is at times resourceful; everyone has desires to remove or reduce their problems and to create something new; everyone experiences changes and has beliefs and theories about why the world is the way it is, or why it should be different. Definitions
We use the client’s precise language to indicate which of the REPROCess category they are attending to. Of course there is always a degree of interpretation but as we show below in our modelling of a transcript, any mind-reading required should be grounded in the client’s actual words.
REPROCess can be used as an overt framework where the client is made aware of the categories. More commonly it remains in the background helping the facilitator to model the client's information and attention, and to track and guide the process. REPROCess has been designed primarily as an aid to therapeutic modelling, not a substitute for it (Lawley & Tompkins, 2006).
are aspects of a person's experience that they value. A resource can be anything that a person already has or has access to. It can be a general quality, skill or attribute. It can be physical, imaginary or symbolic. It does not matter what form it takes as long as the person regards it as having value, use or goodness in its own right, or in relation to something else or the context. People experience resources as, well, resourceful; or empowering, uplifting, redeeming, problem-solving, mystical, balancing, grounding, protective, enlightening, etc. — depending on their preferred metaphor. When a resource is present and its function fulfilled, it will have a beneficial influence on other parts of the system. A resource symbol such as a key may simply unlock a door, or it may resolve a double bind which transforms the whole metaphor landscape (Lawley & Tompkins, 2000).Explanations
are descriptions or accounts of the relationships between ideas, actions and events. They often involve a cause, reason or justification. Every explanation involves beliefs (either overt or presupposed) about ourself and how we envisage the world working. Explanations account for intangible relationships. They are one of the principal ways we glue our REPROC’s together into a coherent and consistent unity. And therein lies the rub. Sometimes the only thing more frightening than changing our metaphor landscape – a translation – is changing the glue that holds the landscape together – a transformation.
Because client explanations often, either explicitly or implicitly, involve the word ‘because’ we have dubbed the process of explaining 'Becausation
' (Tompkins & Lawley, 2006b).
A Problem is a difficulty that a person does not like. A Problem statement can be recognised because it contains:
- A difficulty in the present even if the situation occurred in the past or will occur in the future
- A word or phrase indicating or presupposing the client does not like what is happening
- No stated desire for anything to be different.
A proposed Remedy is a desire for a problem to not exist; to be reduced or avoided. A proposed Remedy can be detected linguistically because it:
- Has yet to happen
- References a problem
- Contains a desire, want, need or a would like for a Problem to change
- Is usually some variation of ‘I want to not have [a Problem].’
- Comprises a mixture of Problem, Remedy and desire
- Often uses an ‘away from’ metaphor (e.g. stop, remove, lose, get rid of, solve)
- Is usually a desire for less of something.
Remedies are fine for solving one-off problems. But they can end up creating more problems when applied to ongoing or systemic problems. If you have an occasional headache, take a headache tablet. But if you keep getting headaches just taking pain-killers is probably doing you no good at all.
Clients often start sessions by proposing commonsense Remedies but these rarely turn out to be sustainable solutions to their problems. If they were viable solutions the clients would have acted on them, they wouldn’t have the problem and they wouldn’t be coming to see you. In the context of coaching or therapy most proposed Remedies either can’t be implemented, won’t work or will end up making the problem worse. They often lack a key ingredient: the kind of life the client would like to have after they have remedied the problem – a desired outcome.
A desired Outcome is a desire for something new (that currently cannot be enacted), i.e. a desire to create or add something in yourself or the world that does not yet exist. It can be recognized because:
- It contains a word for a would like, desire, want or need for a new situation, state or behaviour
- The outcome has not yet happened
- It does not directly reference a problem.
Very often a desired Outcome statement will contain a ‘towards’ metaphor while a proposed Remedy will use an ‘away from’ metaphor. However, not always. For example, “I want to find my lost soul” uses a towards metaphor “to find”, but by our definition the client is proposing a Remedy to the problem of a “lost soul”. Similarly, “I want to get my marriage back in order” or “I’d like to sort out my life” might be thought of as towards goals but in each case, the PRO model regards them as Remedies.
The distinction between Problems, Remedies and desired Outcomes form the basis of our PRO model. The evidence for whether the client perceives a part of their experience as a Problem, a proposed Remedy or a desired Outcome must come from their exact words and what can be presupposed from those words. It is irrelevant whether the facilitator considers something to be problematic, or a solution, or whether it conforms to SMART goals (Doran, 1981) or ‘Well-formed Outcome’ conditions (Dilts & DeLozier, 2000, pp. 1548-1550).
PRO was designed to help cleanly facilitate a client to identify a desired outcome. The PRO model has two stages. First we use the client’s precise language to determine whether they are attending to a problem, a proposed remedy, or a desired outcome. Commonly clients’ statements take the form of:
I hate X.
X will upset me.
I don’t like X.
(or, in the context, X can be
presupposed to be a problem,
e.g. “I am depressed/in pain.”)
I need to stop X.
I want X to disappear.
I want less of X.
I wish I could avoid X.
I don't want X.
Please take away X.
I want Y.
I want to Y.
I’d like Y to ....
I need more Y.
I would like to Y.
I wish I could Y.
In the second stage we respond with a Clean Language question depending on the category the client is attending to. When a client is attending to a problem or a remedy the problematic aspect is first acknowledged and noted for later use. Then the client is invited to shift their attention to a desired outcome. When the client describes a desired outcome we ask a question that keeps their attention on the outcome:
NOTE * do not include any desire word(s) here.
And when [PROBLEM],
what would you like to have happen?
And when [REMEDY*],
then what happens?
And [OUTCOME Y]
And when [OUTCOME Y *], ...
... what kind of [(part of) Y*] is that?
... is there anything else about [(part of) Y*]?
The first six interchanges in the transcript below illustrate PRO in action.
Change is an interesting class of human experience. It is, as Gregory Bateson made clear, a “difference which occurs across time” (1972, p. 452). He might have added that the difference has to be detectable. That is, one or more attributes have to be seen, heard, felt or in some other way sensed by the client as different compared to how they were before. In other words, the form of the client’s system has to change. Even changes to higher-level patterns will be embodied in changes to lower levels of form—but not necessarily vice versa. When a person moves, all their cells move with them, but when a cell changes, it rarely changes the person. SCORE
To qualify as a Change, there needs to be linguistic or behavioural evidence that the client is either experiencing some aspect of their inner world changing there and then, or something has changed since the last session. Change is not a stage in a technique. It is not something that the facilitator or client does, it is what happens as a result of appropriate conditions being present. By ‘maturing’ each change the client gets to find out whether the difference actually makes a difference.
We aim to model a client’s model from their perspective. Their words, in the context of their nonverbals and what they have said before, will usually enable us to distinguish which of the six kinds of experience the client is ‘having’. If there is ambiguity, a single clean question such as ‘And what kind of ...?’ or ‘And is there anything else about that ...?’ will usually provide enough information to settle the matter. Having said that, we always hold our model of the client’s inner landscape lightly since (a) we may have mis-modelled, and (b) the client’s interior world can change in an instant – a resource taken to an extreme can become a problem; and the most horrid symbol can transform right before the client’s mind’s eye.
Robert Dilts and Todd Epstein introduced us to SCORE (Symptom, Cause, Outcome, Resource, Effects) in Santa Cruz, California in 1992. We have made use of their model many times and it laid the foundation for our thinking about REPROCess. The NLP Encyclopedia
(Dilts & DeLozier, 2000, pp.1155-1178) defines the SCORE categories as:
The most noticeable and conscious aspects of a present problem.
The underlying elements responsible for creating and maintaining the symptoms.
The particular state or behaviors that will take the place of the symptoms.
The underlying elements [including techniques] responsible for removing
the causes of the symptoms and for reaching and maintaining the desired
The longer term [positive or negative] results of achieving a particular outcome.
SCORE is often represented on a timeline:
[Words introduced by the facilitator are emboldened to highlight the syntax of Clean Language questions.]
Figure 1: SCORE represented on a Timeline
Comparing SCORE and REPROCess
When we compare and contrast SCORE with REPROC we find many similarities and some key differences.
Symptoms in SCORE more or less correspond to Problems in REPROCess.
Causes are contained within REPROC’s Explanation category. However in our model, causes do not need to be causes of symptoms. People can, and frequently do, see causal relationships in all sorts of aspects of their experience. Resources can have causes, so can Remedies and Outcomes, and so can Changes. Even Explanations can have causes. Also, there are plenty of explanations that are not obviously causal. In the transcript below for instance, the client explains “When I say I want to see more of the day, I’m not sure that’s true. I suspect I’m trying to shorten the day” (line 3). He or she is explaining that they are “not sure”. It is not a statement about cause and effect. Rather it is an important step in the client getting clear about the problem they don’t want and the desired outcome they do want.
Outcomes are similar in both models except, as described above, the PRO model has a wider and simpler definition than the Well-formed Outcome criteria usually used in NLP. We think there is an important distinction to be made between desired and actual Outcomes. Before anything changes there are only desired outcomes in the here-and-now. Actual outcomes most often happen after the session, often long after. Lastly, we don’t think Outcomes need to “take the place of the symptoms”, rather they can create something new without reference to a symptom or a problem.
Resources refer to the same kind of experience in both models. The only difference is that in REPROCess whether an experience is resourceful or not has to be determined from the client’s perspective. In SCORE, resources can include techniques which are decided and managed by the facilitator, which is why they are not included in the first ‘R’ of REPROCess.
Effects do not appear in REPROCess as a separate category. We considered using the second ‘e’ for ‘effects’ but chose not to. First because there are effects to all of the categories, not just outcomes. In particular, not changing the current problematic state will also have consequences which are likely to be one of the main motivators for the person seeking to change. Secondly in some ways everything is an effect. Even in SCORE, symptoms are the effects of causes and outcomes happen because of the application of resources. Rather than make effects a category of experience we decided the exploration of effects was better placed as a phase in the Symbolic Modelling process (see Figure 2).
That leaves Remedies as one of the two categories in REPROCess that do not appear in SCORE. In our definition Remedies are proposed solutions to a problem. Strict adherence to the criteria of the Well-formed Outcome model effectively banishes proposed Remedies. This is strange seeing as they are the most common starting point for most therapy and change work. Instead of banishing them, in Symbolic Modelling we acknowledge them and notice the metaphors involved (they nearly always contain metaphors) and what these tell us about the way the client’s inner world is organised. Furthermore, Remedies gives clues about what is likely to not work and what to be on the lookout for when, to paraphrase Paul Watzlawick, ‘the Remedy is the Problem’ (Watzlawick, et. al., 1974).
The other category missing from SCORE is Change. Unless the facilitator is on the lookout for spontaneous changes, it is possible for small but potentially significant changes to pass them, and the client, by. We added this category to encourage facilitators to listen for the often subtle clues that something is changing for the client right now.
In REPROCess, every category except Change, together make up a client’s “current reality” (Fritz, 1989) – the sum total of what is true for the client in the here and now. Change is the threshold between different realities. As something changes for the client their REPRO’s change and the result is a new current reality.
As Figure 1 shows, the SCORE categories are often explicitly organised in relation to time. This makes intuitive sense but it is prescribing a particular, albeit common, model of the world for the client. We have found that if a time ‘line’ isn’t presupposed some people have exceedingly creative ways of organising time in their inner worlds. For example, two of Aristotle’s Four Causes can be seen as operating in the present and one, Final Causes, in the future (Tompkins & Lawley, 2006b). Rather than presupposing classical time structures, we have found we are better able to model the idiosyncrasies of the way clients think about time with our straightforward Before-During-After model (Lawley & Tompkins, 2000).
REPROCess can be applied in SCORE-like techniques including spatialising the categories and having the client physically move between states. This is particularly effective when REPROCess is combined with another of David Grove’s innovations, Clean Space (Tompkins & Lawley, 2003, 2009). REPROCess was originally designed to encourage the facilitator to pay attention to what the client is paying attention to.
Modelling a transcript using REPROCess
The remainder of this article will give examples of how REPROCess can be used to aid facilitators modelling clients’ maps of the world, and as a guide to facilitating a client to navigate a beneficial pathway through their inner landscape.
The following transcript is an edited extract from a demonstration of the beginning of a Symbolic Modelling session.
Before you begin, you might like to:
- Read the transcript through.
- Have a go at allocating each sentence, or part of a sentence, to the relevant REPROC category.
- Review the client’s words allocated to each category and notice what you now understand about his or her model of the world (compared to your first read-through).
- Compare your allocation with ours.
- Consider what you have learned about modelling with REPROCess.
[FACILITATOR] And what would you like to have happen?
1. [CLIENT] I feel scared going out and because of that I avoid going out.And when you feel scared going out, what would you like to have happen?
2. I would like to feel comfortable going out.And when you feel comfortable going out, is there anything else about that feel comfortable going out?
3. I want to see more of the day. Get up earlier in the morning. When I say I want to see more of the day, I’m not sure that it’s true. I suspect I am trying to shorten the day. Actually I don’t want to see early morning. I want to have the whole day. I fear I don’t have enough resource to actually face an entire day.And when fear you don’t have enough resource to face an entire day, what would you like to have happen?
4. Obviously I’d like to develop a resource. I’d like to bring something new to facing the whole day, other than fear. I’d like to focus on the positive, what is great about the whole day.And you would like develop a resource. And when you have enough resource to face an entire day, where is that day?
5. It’s here. It is like a band around my middle. Hot – really hot – and heavy. And a bit nauseating. And actually I feel a bit sick.And when a hot and heavy band round your middle, what would you like to have happen?
6. I want to be willing to accept that every moment of that day may sort of bring some sudden sadness and if I’m out I might be feeling everything I might be feeling. My mind is saying it’s too much and I’m just not big enough or strong enough to keep on doing it.And your mind is saying you’re not big enough or strong enough. And when willing to accept every moment of the day, is there anything else about being willing to accept?
7. I think I underestimate my willingness because [pause] I am stronger than I think I am.And when you’re stronger than you think you are, is there anything else about that willingness?
8. A willingness to [pause] allow myself to feel.And a willingness to allow yourself to feel. And where is that willingness to allow?
9. It’s my middle again. It is down my legs and into my feet and across my shoulders.And does that willingness in your middle have a size or a shape?
10. Something fruity. The shape of an avocado. It’s like all over it’s soft and you squeeze it like this [gesture]. You can squeeze it and it springs back. Lots of little projections.And that’s soft and springs back with lots of little projections like what?
11. [Pause] A sea anemone with loads of surface area. It’s quite pliable and it’s also solid. It’s not spongy. And when sea anemone with loads of surface area that’s quite pliable and solid, what happens to band around your middle?
12. That band is still there and still feels the same but not sickening anymore. And when band is not sickening anymore, then what happens?
13. I just realised something. That fear could also be excitement. It could be that just because I don’t know what will come, it doesn’t mean that what will come will be awful. It could be great.
Our allocation of the client’s words to the REPROC categories is as follows:
Symbolic Modelling Lite
7b. I am stronger than I think I am.
1b. because of that I avoid going out.
3b. When I say I want to see more of the day, I'm not sure that it’s true. I suspect I am trying to shorten the day.
7a. I think I underestimate my willingness
1a. I feel scared going out
3e. I fear I don’t have enough resource to actually face an entire day.
4b. other than fear.
5. It is like a band around my middle. Hot – really hot – and heavy. And a bit nauseating. And actually I feel a bit sick.
6b. My mind is saying it’s too much and I’m just not big enough or strong enough to keep on doing it.
3c. I don’t want to see early morning.
2. I would like to feel comfortable going out.
3a. I want to see more of the day. Get up earlier in the morning.
3d. I want to have the whole day
4a. I’d like to develop a resource. I’d like to bring something new to facing the whole day,
4c. I’d like to focus on the positive, what is great about the whole day.
6a. I want to be willing to accept that every moment of that day may sort of bring some sudden sadness and if I’m out I might be feeling everything I might be feeling.
8. A willingness to [pause] allow myself to feel.
9. It’s my middle again. It is down my legs and into my feet and across my shoulders.
10. Something fruity. The shape of an avocado. It’s like all over it’s soft and you squeeze it like this [gesture]. You can squeeze it and it springs back. Lots of little projections.
11. A sea anemone with loads of surface area. It’s quite pliable and it’s also solid. It’s not spongy.
12. That band is still there and still feels the same but not sickening any more.
13. I just realised something. That fear could also be excitement. It could be that just because I don’t know what will come, it doesn’t mean that what will come will be awful. It could be great.
Figure 2: Symbolic Modelling Lite
first it may seem strange that a process based on modelling, with no
intention by the facilitator for the client to change – what
we call therapeutic modelling
– can produce significant and long-lasting development.
That’s the mystery of self-organizing systems. When a
client’s metaphors are identified, developed and explored,
their system learns from itself; and as their inner landscape evolves,
new ways of perceiving themselves and the larger system of which they
are a part emerge. As that happens their everyday thinking, feeling, and
behaviour correspondingly change.
will now turn from using REPROCess for modelling the client’s
words, to tracking the client’s attention as it shifts
between categories. This will show how REPROCess can be used to both
model the client in another way, and to guide the process.
Figure 3 the overall ‘shape’ of the beginning of
the Symbolic Modelling session transcribed above is illustrated by the
combination of dotted lines which indicate when the client’s
attention naturally shifts, and solid lines which indicate a shift in response to a question.
The numbers in the diagram refer to the previous allocation of the
client’s statements to the REPROCess categories. The
approximate transition between phases of the Symbolic Modelling Lite
process are shown in the right-hand column.
Figure 3: Modelling Attention with REPROCes
initial wanderings of the client’s attention are clearly
shown by the dotted lines. The solid lines demonstrate how the
facilitator continually invited the client to attend to their desired
outcome. From response 8 onwards the client’s attention
remains in the Outcome category – allowing their desired
outcome landscape to develop a form (Phase 3).
client identifies an embodied metaphor – a “sea
anemone” – for their desired outcome
“willing to accept/willing to allow”. Then, at
question 12 the facilitator enters Phase 4 by inviting the client to
find out what effect this might have on the metaphor for the problem
“hot, heavy band”. Unexpectedly (for both the client
and facilitator) there is a slight change in the landscape. While
“the band is still there and still feels the same”
it is “not sickening anymore”. Those last three
words are the clue that a change is happening in real time. The
facilitator instantly switches tack from developing the desired outcome
to maturing the change (Phase 5).
that although the client is asked about their desired Outcome in
question 5, they respond with a (metaphor for a) problem. This is a good
example of why facilitators need to model what a client is actually
attending to, rather than what they were asked to attend to.
After line 13 the ‘maturing the change’ process continues. Interestingly the client discovers that
fear is the band. The awareness of life being out there. Thinking that
life is too close and too much and too fast. And too overwhelming. The
band is holding me in, holding me together.”
What was a Problem symbol is now seen as a Resource.Conclusion
is a way to model subjective experience. The six categories correspond
to near-universal aspects that, while easily understood, can take a
little teasing apart. REPROCess gives facilitators a means of keeping
out of the detail of a client’s content, a method for tracking
what the client is paying attention to, and a guide for inviting
clients to attend in ways that are beneficial.
addition to modelling a client’s language into categories and
guiding the facilitator through a client’s change process,
REPROCess can be used to consider how a client’s patterns of
internal and external behaviour contribute to what happens in their life
outcomes), and the likelihood of them achieving their desired
outcomes. As Fritjof Capra notes, patterns “cannot be
measured or weighed; they must be mapped” (1996, p. 98). To
notice a pattern – especially a pattern of attention
– facilitators need to be proficient at working with the
structure, process and organisation of a client’s content, rather than the content itself
. REPROCess is a simple way of keeping the facilitator from getting lost in the wealth of client information.
we have concentrated on personal development contexts, it takes very
little to extend REPROCess into other areas such as couples and teams,
education and learning, consulting and organisational development.NOTES
 Definitions of attention
is the behavioral and cognitive process of selectively concentrating on
a discrete aspect of information, whether deemed subjective or
objective, while ignoring other perceivable information." Anderson, John
R. (2004). Cognitive Psychology and its Implications (6th ed.). Worth Publishers. p. 519.
"Everyone knows what attention is. It is the taking possession by the
mind in clear and vivid form, of one out of what seem several
simultaneously possible objects or trains of thought ... It implies
withdrawal from some things in order to deal effectively with others,
and is a condition which has a real opposite in the confused, dazed,
scatterbrained state." James, William (1890). Principles of
Psychology. New York: Holt & Co. p. 403
We are grateful to Wendy Sullivan and Phil Swallow for helping us to
clarify our model and coming up with the name REPROCess.
G. (1972) Steps to An Ecology of Mind: Collected Essays in
Anthropology, Psychiatry, Evolution and Epistemology, Ballantine.
Capra, F. (1996). The Web of Life: A new synthesis of mind and matter, Harper Collins.
R.B. and DeLozier, J.A. (2000). Encyclopedia of Systemic
Neuro-Linguistic Programming and NLP New Coding. Scotts Valley, CA: NLP
University Press. Available at www.nlpuniversitypress.com
G. T. (1981). There's a S.M.A.R.T. way to write management's goals and
objectives. Management Review, Volume 70, Issue 11.
Fritz, R. (1989). The Path of Least Resistance. Ballantine.
J (2000). Modelling the Structure of Binds and Double Binds, Rapport
47, Spring 2000. Available at cleanlanguage.co.uk/articles/articles/10/
Lawley, J. and Tompkins, P. (2000). Metaphors in Mind: Transformation through Symbolic Modelling. Developing Company Press.
J. and Tompkins, P. (2004). Clean Language revisited: the evolution of a
model. Rapport 64. Available at
J. and Tompkins, P. (2006). What is therapeutic modelling? ReSource
Magazine 8, April 2006. Available at
& Tompkins (2011) Symbolic Modelling: Emergent Change though
Metaphor and Clean Language. Chapter 4 in Hall, L.M. and Charvet, S.R.
(Editors). Innovations in NLP: Innovations for Challenging Times, Crown
House. Chapter available at: cleanlanguage.co.uk/articles/articles/346/
P. and Lawley, J. (2003). Clean Space: modelling human perception
through emergence. Anchor Point 17(8). Available at
P. and Lawley, J. (2006a). Coaching for P.R.O.s., Coach the Coach
(Feb.). Available at cleanlanguage.co.uk/articles/articles/31/
Tompkins, P. and Lawley, J. (2006b). Becausation. Available at cleanlanguage.co.uk/articles/articles/246/.
Tompkins, P. and Lawley, J. (2009). Clean Space Revisited. Available at cleanlanguage.co.uk/articles/articles/255/.
Watzlawick, P., Weakland, J. and Fisch, R. (1974). Change: Principles of Problem Formation and Problem Resolution. Norton.