Ask a master to explain how they work and you will usually get a story, a theory, or a shrug. Very rarely do you get the procedure. That was the situation in California in the early 1970s, and it is the single most useful fact about how NLP started. The therapists everyone wanted to learn from could not teach what they did, because they did not know what they did.
The brilliance was real. So was the silence.
Fritz Perls, the founder of Gestalt therapy, could walk into a room with a stranger and, inside twenty minutes, strip a lifetime of pretence down to one live conflict. Watch the films and it looks effortless. Watch them with a stopwatch and a notebook and it becomes strange: the same moves, the same rhythm, the same timing, again and again. Perls did not teach those moves. He taught Gestalt theory, which is a different thing entirely.
Virginia Satir, the family therapist, could take a household that had been tearing itself apart for years and change the conversation inside a single session. Families that had rehearsed the same fight a thousand times left her office talking to each other differently. She wrote books about the spirit of the work. The mechanics stayed in the room with her.
Milton Erickson, the hypnotherapist, was the strangest case of the three. He produced changes in people that other psychiatrists called impossible, using sentences that looked like small talk. When asked how, he told stories. The stories were true and useless at the same time, because the knowledge was not stored in a form he could hand over.
Theory is not procedure
Here is the part that matters. Each of these three had a theory about why their work worked. Perls had Gestalt theory. Satir had family systems theory. Erickson had his own idiosyncratic hypnotism. And in every case, the theory did not match the behaviour. What they actually did, move by move, sentence by sentence, was far more specific and far more learnable than what they said they did.
This is a normal fact about human skill. A great athlete usually cannot coach. A natural chef cannot write the recipe, because she never measured anything. Expertise gets compressed into intuition, and intuition does not travel. You cannot download it. You can only observe it in action and code what you see.
What Bandler and Grinder actually added
Richard Bandler, then a student, and John Grinder, then a linguist, did not invent therapy and they said so. What they did was supply the missing step. They sat with recordings and transcripts of these masters and asked a narrow, almost mechanical question: what is actually happening here? Not why, at first. What. Which words, which tense, which direction of attention, which shift in tone, which change in the client's posture came before which change in the client's story.
Out of that came the Meta Model, a set of precision questions built from watching Perls and Satir give directions to language. Out of watching Erickson came the Milton Model, the deliberate use of vague, permissive language to open experience. Neither pattern came from a theory. They came from observation.
The claim NLP made, and still makes, is blunt: if you can capture the structure of what someone does, you can teach the structure to someone else, and they can get a version of the result. The masters themselves were not required to understand it. In several cases they never agreed to the interpretation. Perls died before the first book appeared. Satir was more generous and lent her name to the work. Erickson tolerated the young men who kept showing up with questions.
The honest caveats
Two things should be said plainly. First, modelling always simplifies. What Bandler and Grinder captured was a version of these people, filtered through what they knew how to see. You get a working skeleton, not a whole person. Second, the extraction was contested from the start. Students of Perls and of Satir have argued for decades that the NLP versions miss the heart of the work. They are partly right. The patterns transfer. The genius does not fully.
That is the price of making something teachable. You trade depth for portability.
Why this history is not just history
If you are learning NLP, this story is the first practical lesson. The field is not a collection of techniques handed down from on high. It is the output of a procedure: watch excellence, code the structure, transfer it. Every technique you will ever learn, the swish, the reframe, the anchor, is somebody's finished homework from that procedure.
Which means you can do the same thing. Not to Perls or Erickson. To anyone who is genuinely excellent at anything. That is the actual promise of this field, and it is bigger than any pattern in the catalogue.
We build the whole of the Spirit of NLP programme on this: the attitude and the method, before the techniques. If you want the long version of how the field began, start with the history of NLP, then see where the patterns come from.
The geniuses did their part: they were brilliant. The coding was somebody else's job. Now it can be yours.