“The map is not the territory” means that your mental model of something is not the thing itself, only a partial and edited representation of it. You never respond to reality directly. You respond to the version of it you have built, and that version leaves things out. NLP treats this as its first working assumption: change the map and behaviour changes, because the map is what people were actually reacting to.
The phrase belongs to Alfred Korzybski, a Polish American engineer and the founder of general semantics, who used it in a paper presented in 1931 and developed it in Science and Sanity (1933). NLP borrowed it about forty years later.
| Term | What it refers to | Everyday example |
|---|---|---|
| The territory | What is actually there, independent of anyone’s description of it | A two hour meeting, exactly as it happened |
| The map | Your internal representation of it, filtered by attention, language and memory | “That meeting was a disaster” |
| The confusion | Treating the map as the territory and defending it as fact | Arguing about whether it was a disaster instead of what was said |
| The use | Changing the map when it stops being useful, without denying the territory | “Two people disagreed with me and one agreed. What now?” |
Who said “the map is not the territory”?
Alfred Korzybski, in 1931. He presented it to a meeting of the American Association for the Advancement of Science, and it became one of the central premises of general semantics, the discipline he founded to study how language shapes thinking. His book Science and Sanity is where the full argument sits.
Attribution matters here, because the phrase circulates as a piece of loose NLP folk wisdom and is sometimes credited to Bandler, Grinder or Gregory Bateson. Bateson did carry the idea forward, notably in Steps to an Ecology of Mind (1972), and his reading is how it reached NLP. He did not originate it.
Korzybski’s own analogy is worth keeping. A useful map has a structure similar to the territory, otherwise it misleads you. It is smaller than the territory, otherwise it is useless. And it is not the territory, so no amount of arguing about the map moves a single road.
What does the map is not the territory mean in NLP?
In NLP it is the first presupposition, and the one the others depend on. If people respond to their map rather than to events, then changing someone’s experience does not require changing their circumstances. It requires changing how those circumstances are represented internally, which is what most NLP techniques are for.
That is why an NLP session asks how rather than why. Not “why do you find presentations frightening”, which produces a story, but “what do you see and hear just before you walk in”, which produces the map. Once the map is visible it can be edited. The rest of the working assumptions build outward from this one, and our entry on the presuppositions of NLP lays out how they connect.
It also sets a limit that is easy to miss. If everyone runs on a different map, then your interpretation of another person’s behaviour is itself just a map. This is where perceptual positions come in, as a structured way of borrowing someone else’s view before deciding what their behaviour meant.
How does a map get built in the first place?
Through three filters that operate constantly and below awareness: deletion, distortion and generalisation. They are not faults. Without them you would be overwhelmed by detail within seconds. They become a problem only when the edited version is mistaken for the whole.
- Deletion. You leave things out. Of a two hour meeting you keep one sentence, usually the one that stung, and the other 119 minutes vanish.
- Distortion. You reshape what remains. A colleague’s short reply becomes evidence of disapproval, complete with a motive you invented for them.
- Generalisation. You turn one instance into a rule. One difficult presentation becomes “I am bad at speaking in public”, which then applies to every future room.
The Meta Model is the toolkit built directly on this. It is a set of questions that reverses the three filters in conversation: which meeting, disapproval according to whom, bad compared with what. Our entry on the Meta Model and its question patterns covers the full set, and it is the most practical thing in NLP for anyone who works with other people’s language for a living.
Does it mean reality is subjective?
No, and this is the most common misreading. Korzybski’s point assumes there is a territory. The whole value of a map lies in how well its structure matches something real. Saying maps differ is not saying all maps are equally good.
Some maps predict what happens next. Others do not. A map that says “nobody at work respects me” and a map that says “two people interrupt me in meetings and I have not raised it” describe the same territory, but only one of them tells you what to do on Monday. The test of a map is not sincerity, it is usefulness against the territory it claims to describe.
This distinction is what separates the idea from relativism. It is also what keeps it honest. Used properly, “the map is not the territory” is an invitation to go and check.
How do you use it in a real conversation?
By treating the other person’s statement as a description of their map and asking what it was built from, rather than agreeing or disagreeing with it. Four moves cover most situations.
- Notice the claim is a map. “This project is going nowhere” is a summary, not a report. Something got deleted on the way to that sentence.
- Ask for the territory. What happened, specifically? Who said what, and when? Keep asking until you get something you could film.
- Check your own map at the same time. Your read on their tone is also a map. Say what you observed and ask whether you got it right.
- Redraw only if the territory calls for it. Sometimes the detail confirms the summary. Then you have a real problem instead of a vague feeling, which is progress.
This is slower than arguing and it works better, particularly when the other person is upset. Most of what looks like disagreement about facts is two people comparing maps and assuming the other one is lying. Working through that pattern is a large part of what one to one NLP coaching, in Lisbon or online, spends its time on.
Where the idea gets misused
It gets used to end conversations rather than open them. “That is just your map” is a way of dismissing someone while sounding thoughtful, and it appears in coaching rooms more often than anyone in the field likes to admit. The phrase was meant to make people more careful about their own descriptions, not to give them a tool for waving away other people’s.
The second misuse is heavier. Real constraints get reframed as map problems. Poor treatment at work, discrimination, illness and financial pressure are territory, and telling someone their distress is a representational issue is both wrong and unkind. NLP works on the part of a difficulty that lives in how it is represented. That part is often larger than people expect and it is never the whole thing.
In a training room the concept lands fastest when it is applied to the trainer rather than the participant. I would rather have people question my description of what just happened in an exercise than accept it. That is the same skill, practised where the stakes are low. It is also the reason the four things a practitioner is actually building, set out in the four pillars of NLP, include accurate noticing as a separate discipline from technique.
NLP is a set of communication and change tools. It complements professional medical or psychological care and does not replace it.
Carolin Mallmann, Licensed Trainer of The Society of NLP, Lisbon
Frequently asked questions
What does “the map is not the territory” mean?
It means your mental model of something is not the thing itself, only a filtered representation of it. People act on their internal description of a situation rather than on the situation directly. The phrase is a reminder to keep checking the description against what is actually there.
Who came up with the phrase?
Alfred Korzybski, a Polish American engineer and the founder of general semantics. He used it in a paper presented in 1931 and set out the full argument in Science and Sanity in 1933. Gregory Bateson later carried the idea into systems thinking, which is how NLP picked it up in the 1970s.
Why is it the first presupposition of NLP?
Because every NLP technique depends on it. If people respond to their internal map rather than to events, then changing the map changes the response without needing to change the circumstances. Remove that assumption and reframing, anchoring and the rest of the toolkit have nothing to act on.
Does it mean everyone’s version of events is equally valid?
No. Korzybski’s point assumes a real territory exists, and the value of a map lies in how well it matches it. Maps differ in accuracy and in usefulness. A description that predicts what happens next is better than one that does not, even when both are sincerely held.
What are deletion, distortion and generalisation?
The three filters that turn experience into a map. Deletion leaves information out, distortion reshapes what remains, and generalisation turns single instances into rules. All three are necessary for functioning. They cause trouble when the edited result is treated as a complete account.
How is this different from positive thinking?
Positive thinking replaces a description with a nicer one. Working with maps means going back to the territory and finding out what is actually there, which sometimes produces a less comfortable picture. The aim is accuracy and usefulness, not cheerfulness.
Keep exploring
- The four pillars of NLP, for the four conditions every technique quietly depends on, including accurate noticing.
- Is NLP scientifically proven?, for where the research stands on the claims built on top of this assumption.
Carolin Mallmann is a Licensed Trainer of The Society of NLP, trained by Richard Bandler. She has trained more than 300 people and runs English language NLP trainings in Lisbon in groups of no more than 10.
