The Meta Model in NLP is a set of questions that recovers the specific information people delete, distort and generalise when they speak, bringing vague language back to concrete experience. When someone says “everyone ignores me” or “I just can’t,” the Meta Model asks the precise questions that reopen what got lost. This guide covers the three language patterns it targets, the questions that answer them, and how it differs from the Milton Model.
What is the Meta Model in NLP?
We never put our full experience into words — we shrink it to fit. The Meta Model is a map of the three ways language loses information and a matching set of questions to recover it. It was the very first NLP model, published by Richard Bandler and John Grinder in The Structure of Magic (1975) after they studied the language of therapists Virginia Satir and Fritz Perls.
The Meta Model at a glance
| What it is | Questions that recover deleted, distorted and generalised detail |
| Created by | Bandler & Grinder, The Structure of Magic (1975) |
| Targets | Three patterns: deletion, generalisation, distortion |
| Good for | Clear thinking, coaching, problem-solving, precise communication |
| Opposite of | The Milton Model (deliberately vague) |
The three patterns it targets
| Pattern | Example | Recovering question |
|---|---|---|
| Deletion | “I’m anxious.” | Anxious about what, specifically? |
| Generalisation | “Nobody ever listens to me.” | Nobody? Ever? Who, specifically? |
| Distortion | “She makes me feel small.” | How, specifically, does what she does lead to that? |
How does the Meta Model work?
The Meta Model works by listening for a vague or absolute phrase, then asking the one question that recovers the missing detail. Each question gently challenges the limit hidden in the language — an unspecified “it,” an absolute “always,” an unexamined cause-and-effect — and returns the speaker to their actual experience, where change becomes possible.
How to use the Meta Model: 3 steps
- Listen for the pattern. Notice deletions (“I’m stuck” — with what?), generalisations (“everyone”, “never”), and distortions (mind-reading, “you make me…”).
- Ask the recovering question. Aim precisely — “who specifically?”, “compared to what?”, “how do you know?” Common mistake: firing questions like an interrogation. Keep rapport first, or it feels like an attack.
- Get specific — then act. Once the vague statement becomes concrete, the real issue and the next step usually appear on their own.
Meta Model vs. Milton Model
The Meta Model sharpens language toward precision; the Milton Model softens it toward artful vagueness. They’re mirror images — one recovers detail, the other invites the listener to supply their own.
| Meta Model | Milton Model | |
|---|---|---|
| Direction | Toward specifics | Toward vagueness |
| Use | Clarify, problem-solve | Guide, relax, influence |
| Modelled on | Satir & Perls | Milton Erickson |
How we teach the Meta Model in Lisbon
The word I listen hardest for in coaching is “just” — as in “I just can’t do it.” That one small word deletes an entire story. Ask “what stops you, specifically?” and the wall someone thought was solid turns out to have a door in it they’d simply never described out loud.
Related terms: the Milton Model, presuppositions and reframing. Back to the full NLP glossary. See also: what NLP is and all NLP techniques.
Sources: Richard Bandler & John Grinder, The Structure of Magic, Vol. I (1975); modelled on Virginia Satir and Fritz Perls.
This glossary is educational and reflects a coaching perspective. NLP complements but does not replace professional advice.
Frequently asked questions
What is the Meta Model used for?
It’s used to turn vague, stuck or absolute statements into precise ones — for clearer thinking, better problem-solving and sharper communication. In coaching it helps a client discover what they actually mean and want.
What are deletion, distortion and generalisation?
They are the three ways language loses information. Deletion drops detail (“I’m upset” — about what?), generalisation turns one case into a rule (“this always happens”), and distortion warps meaning (mind-reading, false cause-and-effect).
Can you give a Meta Model example?
“They don’t respect me.” Meta Model responses: “Who, specifically?” and “How do you know they don’t respect you?” Both recover the missing detail behind the generalisation.
What’s the difference between the Meta Model and the Milton Model?
The Meta Model moves language toward precision to clarify; the Milton Model moves it toward artful vagueness to guide and influence. They’re deliberate opposites.
Is the Meta Model just asking questions?
The questions are ordinary, but the skill is knowing which pattern you’ve heard and which question recovers it — while keeping rapport, so it feels like curiosity rather than interrogation.
