REFLECTIONS ON THE MIRROR-MODEL
by Philip Harland
An example of conversational clean language and
"If you can reflect the client's problem
the client is relieved of the responsibility for holding
The problem shifts and the system will spontaneously
In the previous issue of Rapport (Winter 2001), I
summarised the 'Mirror-model' as a self-reflective model of
questioning that could be used to facilitate pretty much anyone
anywhere. (See Part 1.) And I drew
distinctions between conversational clean language and Clean Language
as a therapeutic modality.
Conversational clean language is for the most part semantic, as
you would expect: the information it gleans for the client is sourced
in what is said, and in the main engages the client's ego-state.
Grovian Clean Language, on the other hand, is geared to identifying
and developing embodied perceptions, which are sourced not only
semantically (in the client's language), but also somatically (in the
client's body), spatially (in the client's metaphor landscape), and
temporally (in the client's coding of biographical, ancestral and
cosmological time). David Grove's therapeutic model stems from his
clinical work with survivors of abuse, treating negative symptoms as
coded solutions from the unconscious containing positive resources
for healing. Hardly a casual procedure. Yet even a casual
conversation which honours the underlying principles of Clean
Language can have a far-reaching systemic effect.
Applied to conversational facilitation these principles are:
- respect and reflect client's own words
- avoid obvious metaphors in your words
- no assumptions, judgments or interpretations
- no suggestions, reframes or linguistic challenges
- follow client information
- keep questions clean
I don't propose to elaborate here on the philosophy of clean, as
opposed to 'contaminated'or assumptive language, or about the virtues
of inner-directed, rather than facilitator-directed, change. If you
want to catch up, or have yet to be convinced, you can find more on
the subject elsewhere. 1
This second and last part of the article is a practical guide to
using the questions and frames of the Mirror-model. The outcome, if I
may remind you, is not to try and understand the other person, but
to facilitate them; it is not to suppose what they mean by what
they say, but to help them know themselves; and it is not to
suggest solutions, but to help them generate solutions for
themselves - if that is what they wish to do. Honouring
Take the statement,
'It's all too much, I don't know which way to
Your colleague or client has said this kind of thing before about
their work / lover / drinking / life, and again you're tempted to
commiserate, or to suggest which way they should turn / how to deal
with the the other guy / why they should come off the booze / what
they must do to pull themselves together. Has this made any
difference in the past? Not much. Today before you attempt to
empathize, or make intelligent suggestions, or offer linguistic
challenge, or put them into hypnotic trance, or launch into an NLP
process, you take the opportunity to honour the expertise they have
in themselves. You reach for your Mirror-model crib, kept handy in
your purse or wallet ...
X = client statement or part of statement
from any frame
© 2001 Philip Harland
* What happens/ed
just before X?
* Before that?
* Importance of
* Purpose of X?
* Meaning of X?
* Enables X?
do you know X?
* What X
* What kind of X?
* What part/aspect of
* Anything else about
When X, what do you
* When X,
what about (other) X?
*When X, what
* Then what
* What symbolises X for
Or That's an X like what?
* What kind of [part of metaphor]?
* Anything else about [part of
What exactly is your client's Present state? What is the wider
Context for them of this? What in the immediate or distant Past may
have prompted it? How will it carry over into the Future? Are there
Higher considerations that might help them deal with it? Or Metaphor
correspondences that will allow them a different
If you take the original statement as a starting point you can
question it from any frame and move to any frame. The itinerary is
normally dictated by the logic of the information. In this instance
it will take us from Present --> to Context --> to Past -->
to Future --> to Higher --> to Metaphor.
An immediate awareness, or foreground, frame. It's where the
client's attention is right now.
Client: It's all too much, I don't know which way
This client is being very general. We can ask clean specifying and
clarifying and questions:
You: What specifically is too much?
Client: I told you - everything.
You: What kind of everything?
Client: Oh, my marriage, and my new boss, and my own
stupidity, I suppose.
The problem areas are defining themselves. They may now be
considered one at a time.
You: What kind of marriage is that?
You: What aspect of your new boss?
You: Is there anything else about your own stupidity?
'Is there anything else about [anything the client has
said]?' is a generative question that can be asked in any
frame, because there always is. Even if the answer is, 'No, there
isn't'. This will mean either that the client knows enough for
now - when without your question they may not have known they knew,
so it was still worth asking - or that the question has evoked
something the client prefers not to consider there and then, so the
question will have encouraged them to make a choice. Information
either way: for them, not you. You have no need to know.
A wider present, or background awareness, frame. No present
behaviour, thought or feeling takes place in a vacuum. In the diagram
the Context frame surrounds the Present frame, suggesting a state
that includes and transcends immediacy. A shift in the client's
attention from partial to fuller awareness may itself be enough to
unstick a stuckness.
In fact this client has already answered a 'What kind of?'
question in the Present frame with further information, and now knows
that their 'all too much' has marriage, work and self
Another clean Contextual question:
You: How do you know it's all too much?
has the potential for opening things up by inviting the client to
check for sensory evidence, which may offer a new kind of
Client: How do I know? I have butterflies in the
You: And what might butterflies in the stomach be related
Client: Hm, it reminds me of the feeling I used to get before
You might equally have asked 'What kind of 'all too much'
[statement #1] could that be when you 'don't know which way to
turn' [statement #2]?' As a facilitator you can speculate
about two statements like these until the cows come home, but only
the client can decide if they relate in any particular way. A clean
'related to' or 'what kind of X when [other] X?' question enables the
client to consider a connection between any two states or statements
however far apart.
You: Is there anything else about that feeling?
Client: No. It wasn't very nice.
Let's go back a moment. The client discovered a clue ['the
feeling I used to get before exams'] that you might want to
examine in the Past frame. First, however, you can ask the ultimate
Contextual question. I omitted 'What do you want?' from the
original model. The Grovian equivalent of this key NLP question is
'What would you like to have happen?', and I suppose I was
saving that select starter for the special occasion, the full
five-course therapeutic process. In the meantime all my experience as
a psychotherapist, trainer and citizen has convinced me that the
outcome question is indispensable for any kind of facilitation, and
may be particularly relevant in those everyday situations where we
are more used to hearing what the boss, or the partner, or the
average traffic warden thinks is the purpose of the conversation, and
haven't normally had the privilege of being asked for ourselves.
Although most of us will revert to our own motivation whatever
others may want for us, I believe it helps to answer outcome
questions consciously, and if possible out loud. I was a film-maker
before I became a psychotherapist, and the best tip I learnt for
working with actors came from watching theatre director Mike Alfreds
at work. He would always say to his actors: "When you come on
stage, when you enter this scene - whatever the situation, whatever
your lines - what does your character want?"
You: And when you have these feelings of all too
much and butterflies in the stomach, what do you want?
Client: [Pause] To understand them and be able to deal with
Your questions can now be directed to the outcome or to the
original state. I don't think it matters much. All your questions
will be clean, and it is the client who makes the real decisions.
Their response prompts the next question, so the process is always
client information-led. The important thing is that the client now
has a self-defined template - in this case 'understanding' and
'being able to deal with' (though you might like to help them
define these more precisely) - against which to measure all new
As you ask questions that help relate the problem to what may have
prompted it, your client will begin to get a sense of sequence. This
can be a powerful means of escape from the confines of a present
state. Especially for a client who wants 'to understand'.
You: When you want to understand these feelings of
'all too much' and butterflies in the stomach and you want be able to
deal with them, what could have prompted the feeling you used to get
Client: I guess the idea that I am being tested.
You: And what happens just before the idea of being tested?
Client: Just before? It's the thought that I have to solve
someone else's problems.
You: And before that?
Client: My partner or my boss is telling me about their
You: I wonder is there anything else about problems like that
you think you have to solve?
Client: Yes. I feel incompetent if I can't.
The client has indicated a consequence ('I feel
incompetent') of a present belief ('I have to solve
something'), and this can be explored in a Future frame.
You: And when you feel incompetent, what
Client: I get defensive.
You: And when you get defensive then what happens?
Client: Ha, I get angry.
You: And then what happens?
Client: I sulk.
You: And then?
Client: My partner and my boss get pissed off at me.
More links in the chain. The client now has a sequence. Given a
sequence, a familiar pattern may be recognized - or a little-known
pattern may be revealed - or past and present, memory and immediacy,
may seem less disconnected and become more like changes in the
seasons: temporary transitions in a continuously unfolding.
You: Is there anything else about feeling
Client [Pauses]: Well, I don't feel loved. And then it all
becomes too much.
In pursuit of a consequence the client has arrived back at a
'cause' ('I don't feel loved'). In the context of their
relationship with the boss this may be new and surprising
information. The sequence now goes something like:
all too much --> other people's problems -->
have to solve -->
feel tested --> can't solve --> feel incompetent --> get
angry --> sulk --> [??] --> don't feel loved --> all too
This may well be a repeating pattern.2 There seem to be
many ways of getting into such a pattern, and maybe as many ways of
getting out. What else is there here? A modal operator of necessity
('have to') that could be examined. A possible gap in the sequence
(between 'sulk' and 'don't feel loved'?) to be filled. And a
likely bind (things all too much <--> don't feel loved
<--> things all too much) for unravelling. Interesting,
speculative, and none of it need concern us directly. We are simply
facilitators of conversational change here, and the client will work
through these and other matters as they choose.
The client now has information from Present, Context, Past and
Future frames. What next? We could reflect and reiterate the
recursive sequence 'I don't feel loved and then it all becomes too
much.' by returning to the Past--Present--Future axis until the
client really gets to know it. Or take time out and invite the client
into a 'higher' frame.
Invitations to consider higher purpose or meaning are not always
welcomed with open arms, so do check your level of rapport and don't
expect to start off in this frame.
Questions here relate to the 'higher' logical levels of NLP -
beliefs & values, identity, spirituality. They can help a client
return to the Present having evolved beyond their original state, or
at the very least having a broader view of it. In the diagram the
Higher frame includes both Present and Context, representing a state
that both embraces and transcends them.
You: What might be the importance of not feeling
loved for you?
Client: The importance? Well, when I don't feel loved I can
blame someone else for my feelings.
You: Is any purpose served for you by blaming someone else?
Client: Yes, I can escape responsibility.
You: What could be the meaning for you of escaping
Client: I guess it means I don't grow - mentally or
You What may enable [helps/supports] you not to
Client: Oh, not taking time to sort out the confusion, thinking
there's only one way to turn and that someone else knows what it is,
You: Is there anything else about all this?
Client: [Pause] Yes, I think I understand more about the
These importance/purpose/meaning/enabling questions are clean in
that they hold no assumptions, other than perhaps that there
is an importance, a purpose, a meaning and an enabling. They
do not appear in the strict Clean Language modalities, Grovian
Metaphor and Symbolic Modelling. For three inter-related reasons, I
believe. One, they are at a higher level of abstraction than
other questions, and this makes it difficult for a therapist to put
them directly to the form, location and temporal coding of a client's
embodied perceptions. Two, they require cognitive processing,
which means the client has to come out of inner-directed trance into
everyday narrative. And three, they contain metaphors that arguably
presuppose a therapist world view. 'Importance' has associations with
external authority - it derives from the Latin importare, 'to
introduce from abroad' - and in that strict extrinsic sense may not
be considered self-generatedly 'Clean'.
A colloquial model, however, can cast its net wider. 'Importance'
also has a sense of 'personal significance' nowadays, so it gets my
vote as kind of clean. Not perfect, but serviceable, like my
almost-white shorts. And there are notional modifiers for these
'importance' etc questions. 'What might be ..?', 'Is there
any ...?' and '... for you?' can be employed as mild
Asking these questions of ostensibly negative states ('not feeling
loved', 'escaping responsibility', 'not growing'), by the way, does
not imply that the client should want to feel loved, or to
accept responsibility, or to grow. This may be an awkward concept for
a facilitator to grasp. Of course the client wants to feel loved, to
grow, to know which way to turn! Well you could be right, but ask
yourself if having someone else identify with their not feeling loved
('That must be awful for you...') is really helpful,
especially if it's in a metaphor ('awful') that may be far
removed from their own.
If you can embrace this fundamental premise of Clean Language -
MAKE NO ASSUMPTIONS! - it could change the way you do therapy,
counselling and facilitation forever. 3
I have introduced a 'Learnings' question to this Higher frame.
Again, omitted from the original model. Although at the time I was
cottoning on to the importance of monitoring learning in trainings, I
wasn't doing it much during therapy. 'What have you learnt?'
is a truly enabling question. The client is unlikely to move on until
they know something new, and may not move on until they know they
know. Articulating knowledge brings the unnamed into awareness, and
is a way of combining the strengths of both unconscious and conscious
Timing the question should allow time for significant learning.
I've placed it apart from the other questions in the frame as a
reminder that it can be left for a while.
You: Now what have you learnt about all this?
Client: [Thinks] Well, I've learnt that I can deal with the
feelings as long as I acknowledge them rather than try and deny them.
And I still want to know which way to turn!
Not knowing 'which way to turn' is of course a metaphor
that has been around from the start of the session. You may also
recall the client having 'butterflies in the stomach' in the
Context frame, or their experience of partner and boss getting
'pissed off at me' in the Future frame.
You have a choice here. You can mosey along at the momentum of the
other frames or move up the gears into full Metaphor process using
Clean Language proper. 4
Even without Clean Language training you can utilise
client-generated metaphor - which may appear at any time, in any
frame - or you can elicit one specifically:
You: What for you symbolises wanting to know which
way to turn?
When you want to know which way to turn that's like what?
Client: It's like I'm at a crossroads.
There are upwards of 30 Clean Language questions, but to explore
this metaphor we need only two, and you are familiar with these
already. 'What kind of ...?' and 'Is there anything else
about ...?' are key clean questions that can unlock a phrase in
any frame and open up a treasure chest of new information for the
client. Once you identify and engage client metaphor, you are
accessing a container for very precious, deeply coded information
that comes directly from the unconscious.
You: And what kind of crossroads?
Client: Well, the kind where there's a main road that's
obvious but goes nowhere and a narrow side road that looks
interesting but may be a long way round.
You: And is there anything else about a side road like
Client: It could be more interesting and useful and certain
than I first thought.
It is my belief that inner-directed change will always be more
interesting, useful and certain than change that comes about as a
result of external direction, suggestion or challenge. And will also
This self-reflective model asks questions which do not presuppose
answers, which prompt for self-knowledge and encourage people to take
charge of their own minds. And if that, in James Lawley's words, has
"the ripple effect of influencing deeper organizing metaphors and
neuro-chemical processes", so be it. We're doing our job as
facilitators of change.
1 The website: www.cleanlanguage.co.uk and the book: James Lawley and
Penny Tompkins, Metaphors in Mind: Transformation
through Symbolic Modelling, The Developing Company Press 2000.
1 of this article: Rapport 54, Winter 2001.
2 More on Patterns in Chapter 7 of Lawley and Tompkins,
Metaphors in Mind; and in Philip Harland, Resolving Problem Patterns, Part 1 and
2, Rapport 49
& 50, Autumn & Winter 2000, and www.cleanlanguage.co.uk.
3 Remember the therapist working in a prison who was told by
an inmate, "I broke the jaw of the last person who told me 'time
4 The French version of Clean Language is Langage
which the word propre nicely combines the sense of 'uncontaminated' (the language
of the therapist) with the sense of 'one's own' (the language of the
client). If you're interested in the therapeutic possibilities of
client-generated metaphor and symbol in any language I strongly
recommend you train in the process. See below.
David Grove, Penny Tompkins & James Lawley, and the London Clean
Language Practise Group.
© 2001 Philip Harland
First published on this site 27 December 2001.