Most questions a practitioner asks are dirty questions. Not in the moral sense. In the sense that they carry the practitioner's assumptions inside them, the way a paintbrush carries the previous colour.
Ask a client "how does that make you feel?" and you have already decided that feelings are the relevant category. Ask "what do you want?" and you have already decided that wanting is where the answer lives. The question steers. The client follows, because clients are polite.
Clean Language is the discipline of asking questions that steer as little as possible. This post explains where it came from, how it works, and why it matters more than most technique lists suggest.
Origins: David Grove
Clean Language was developed in the 1980s by David Grove, a New Zealand therapist working with clients recovering from trauma. Grove noticed something about his own questions: the more his clients used metaphor to describe their experience, the more powerful the work became. And the more he could stay inside their metaphor, using their words back to them, the faster they moved.
He built a small set of questions designed to do two things only: direct attention to what the client just said, and nothing else. No interpretation. No reframe. No "and how does that make you feel?" The questions are content-free: they carry no assumptions about what the client's experience means.
The core questions
The full Clean Language set is small. The most famous ones are these:
- "And what would you like to have happen?"
- "And what kind of X is that X?"
- "And where is X?"
- "And is there anything else about X?"
- "And what happens just before X?"
- "And what happens next?"
- "And where does X come from?"
Notice the pattern. Every question takes the client's own word, X, and asks about its structure: its location, its sequence, its history, its kind. The practitioner never supplies content. They only point the client's attention back at what the client just produced.
That is the entire trick, and it is harder than it looks. The temptation to add, interpret, or fix is nearly irresistible, because every practitioner has been trained to add value. Clean Language is the training in not adding value, so that the client's own system can reorganise itself.
How it works in a session
A Clean Language session sounds almost boring from the outside. The client says "I feel like I'm carrying a stone." The practitioner asks "And what kind of stone is that stone?" The client says "It's heavy, grey, it was given to me." The practitioner asks "And where was it given to you?" And so on, for twenty or thirty minutes.
What is happening underneath is precise. The stone is a metaphor, and metaphors are compressed models of experience. Every question expands the compression a little, in the client's own terms, at the client's own pace. The practitioner's map never gets imposed, so the client's map does the work. When the metaphor completes its own logic, the felt sense shifts with it. That shift is the change.
This is why Clean Language pairs so naturally with Symbolic Modelling, the method David Grove and Penny Tompkins and James Lawley built on top of it. Symbolic Modelling is the larger framework; Clean Language is the questioning engine inside it.
Why we teach it
At the Institute, Clean Language sits in Month 2 and Month 3 of the Spirit of NLP programme, as part of the content-free work we teach alongside the Meta Model and the Milton Model.
It matters for three reasons.
It protects the client. A practitioner with a favourite technique will find a way to use it. Clean Language makes the practitioner's map visible to themselves, which is the first step to keeping it out of the room.
It is the practical expression of a core value. We say we train process, not content, and that your beliefs stay yours. Clean Language is that sentence turned into a set of questions. It is not a slogan; it is a skill.
It is evidence-friendly. The claims made for many NLP techniques have not survived testing. Clean Language is modest: it makes no claim to read minds or detect lies. It only asks questions. Its results are observable in the client's own words, and its failures are visible in the same place.
Learning it properly
Clean Language cannot be learned from a list of questions. The questions are easy. The hard part is the posture: staying empty, tolerating silence, resisting the urge to help. That is a skill, and skills need reps, feedback and time.
In the Spirit of NLP programme you do not just read about Clean Language. You run it on real material, in front of people who are watching whether you contaminate the client's model with your own. By the end you can hold a Clean session that a client cannot tell apart from your own map, because your map never entered the room.
If you want to feel the difference before committing, the free webinar is a live sample of the work. One hundred and twenty minutes, no replay, no fee.