Definition
A category error is a reasoning mistake that treats things in different categories as if they belonged to the same one, or that imposes clean categorical boundaries on what is actually a continuum.
In Rosling's framework it is the engine underneath the Generalization Instinct. The labels "developed" and "developing," for example, group hundreds of countries into two buckets when income, health, and education vary smoothly across four income levels — the error is not the labels themselves, but the assumption that everything inside each label is alike.
Why it matters
How it works
Two sub-mistakes hide under one name. The first is treating different things as one — assuming, say, that "African countries" share roads, schools, fertility, and politics, when within Africa the variation is enormous. The second is treating one continuum as two — splitting humanity into "developed" and "developing" when both labels span income ranges that differ by an order of magnitude.
Both errors persist because categories are useful — they compress, they speed conversation, they make policy debates tractable. The trouble is that the compression is lossy in a specific direction: differences inside the bucket disappear from view, and the bucket label becomes a stereotype for everything inside it.
The corrective is to keep questioning your buckets. Look inside each one: how wide is the spread? Is the bottom of bucket A actually closer to the top of bucket B than to its own peers? When the answer is yes, the categorical framing is doing more harm than good, and a finer-grained scale — Rosling's four income levels, or simple quartiles — is the cleaner tool.