IONI — Institute of NLP India

The Meta Model

The Meta Model was the first product of NLP — the questioning system Bandler and Grinder reverse-engineered from Fritz Perls and Virginia Satir. It remains the field's most defensible tool: a grammar of how language deletes, distorts and generalises experience, and how to ask it back.

The three families

  • Deletions: information is missing — "they rejected the proposal" (who, exactly, rejected what, on what grounds?)
  • Distortions: inference presented as fact — "he resents me" (how do you know?), mind-reading, cause-effect ("she makes me anxious")
  • Generalisations: universals and rules — "I always freeze in interviews" (always? every interview?)

What questioning well looks like

Raw Meta Model questioning feels like interrogation; skilled questioning feels like interest. The difference is pacing, rapport and intent — you challenge a generalisation to recover specificity the speaker needs, not to win. "Always?" asked with genuine curiosity opens a problem; asked with smugness it closes the conversation.

Where it is used

Coaching and interviews (recovering actual criteria), sales and negotiation (what does "too expensive" specifically mean?), self-work (auditing your own absolutist sentences), and management (meetings run on deletions). It pairs naturally with clean language — our journal piece compares the two models directly.

Practice, not study

The Meta Model is a listening reflex, which means it is built through live drilling: hear a sentence, classify the violation, ask the recovery question, repeat until automatic. That drilling — with feedback — is core curriculum in Spirit of NLP.

Free live session

You've read about it. Now watch it work.

Stop Doing NLP is a live, practical session. See the methodology demonstrated, then decide.

Frequently asked questions

What is the difference between the Meta Model and the Milton Model?

They are opposites that pair together: the Meta Model recovers specificity from vague language; the Milton Model uses artfully vague language to invite the listener to fill in meaning. Our journal compares them.

Is the Meta Model scientific?

It borrows from transformational grammar, which has moved on. As a practical questioning toolkit it is defensible; as linguistics it is historical. The evidence pages cover this honestly.

Written by ShriViral Manek, Institute of NLP India · reviewed 30 Sept 2026 · see it live, free →