
When David Graeber set out to learn Malagasy, the national language of Madagascar, his teachers gave him a surprise. They drilled him on official grammar. Then they turned to one another and spoke a different language altogether.
Asked why they would not teach him the version people actually used, they said it was slang. Asked why almost no one spoke the official version, they said people were lazy. Graeber heard something else. A whole population was treating an old description as a living command.
The first written grammars of Malagasy did not invent the language. Outsiders recorded how people spoke at one moment, so they could study it. Those notes later hardened into the "official" form. Children learned that older grammar in school while speaking something else at home. If the records had been made two centuries later, today's speech would be called correct, and the old forms would be the errors.
Languages do not begin when someone writes the rules. They change as speakers play, borrow, and answer new needs. Official rules make life easier for learners and for people from different places. They also freeze a moving system. People keep changing the language anyway, then blame themselves for not speaking "properly." They have confused a description with a prescription.
Why not just follow the book? A frozen language would help historians and classrooms. It would also make the language less useful. Graeber's answer is that people like to play. Even small, occasional experiments add up. Conventions are still needed, or no one would understand anyone else. The useful balance is enough structure to share meaning, and enough flexibility to let the system update.
"The map is not the territory" names the same mistake. Any representation of a complex system has to simplify. It leaves things out, ages, and carries the maker's choices. The London Underground map ignores real distances because commuters do not need them. A map that copied the city exactly would be useless. Dictionaries and grammars are maps of this kind. They help you enter a language. They are not the language.
We run into trouble when we treat a snapshot as an order: make reality match the page. Graeber called this the grammar-book effect. People do not invent languages by writing grammars. They write the first grammars by watching the unspoken rules speakers already use. Once the book exists, especially in a classroom, the description starts to feel like a prescription.
The same mix-up appears at work. You join a new job and receive step-by-step process documents. The person you replace does not follow them. They explain that the pages were written before anyone had actually done the work. The documents still help a beginner, and they are the safest option if something goes wrong. Once you know the territory, you can change a step without treating the old page as fixed forever.
Descriptions help you navigate and still leave room to revise. Prescriptions help people coordinate, but they go stale. When you meet a representation of the world, ask which parts describe how things were, and which parts quietly tell you how things must stay.