28 July 2026

The map is not the territory (and why most arguments are two maps colliding)

There’s a sentence that comes up in the first week of almost every NLP course ever taught: the map is not the territory. It’s borrowed from Alfred Korzybski, a Polish-American philosopher writing in the 1930s, and it might be the single most useful idea in the whole field.

Here’s the problem it points at. You never experience the world directly. Everything comes in through your senses, gets filtered by what you’re paying attention to, gets compared against everything that’s ever happened to you, and gets wrapped in language before you’re even aware of it. What you end up with isn’t reality. It’s a map of reality, drawn by a cartographer with strong opinions and a patchy memory.

Maps are useful precisely because they leave things out. The London Underground map is a design classic because it ignores geography almost entirely; it tells you how to get from Oval to Angel and nothing else. Your mental maps work the same way. The map you hold of your boss, your commute, or your own abilities keeps what once seemed relevant and quietly bins the rest.

The trouble starts when you forget you’re holding a map. Then “he ignored my email” stops being one reading of events and becomes the event itself. You respond to your drawing of the situation, coloured in with motive and history, rather than to what actually happened, which is that no reply has arrived yet.

Two maps, one argument

Watch any long-running disagreement, at work or at home, and you’ll usually find two people defending their maps while believing they’re defending the territory. Neither is lying. Each map is honestly drawn from that person’s filters, history, and attention. They just left different things out.

This is why “but look at the facts” so rarely ends an argument. Each side’s facts have already been through the filtering. The argument isn’t about the territory at all; it’s about which map gets to be called accurate.

Knowing this doesn’t make you right, which is rather the point. It makes you curious. The question shifts from “how can they possibly think that?” to “what would have to be on their map for that to make sense?” That second question has an answer you can actually find.

Something to try

Next time someone’s behaviour baffles or annoys you, try this before responding. Take the sentence you’ve got in your head about them, something like “she doesn’t respect my time”, and label it honestly: that’s my map. Then ask what a different map of the same events might look like. Not a kinder one, necessarily. Just a different one.

You’re not obliged to adopt the alternative. Sometimes your first map was fair. But you’ll notice how much of what felt like observation was actually inference, and that gap, once you can see it, is where most of the useful work in NLP happens.

Anecdotally, this one idea defuses more friction than any technique with a fancier name. But that’s an anecdote, and you should test it on your own week rather than take my word for it.