“Intelligence Without Representation” is a 1991 paper by Rodney Brooks, published in the journal Artificial Intelligence (volume 47, pages 139 to 159) and hosted on his MIT site. With his related 1990 paper “Elephants Don’t Play Chess,” it is the manifesto of behavior-based, or “nouvelle,” AI - a direct challenge to the symbol-processing tradition that had defined the field.
Brooks attacked the standard architecture in which a robot builds a detailed internal model of its surroundings, reasons over that model to make a plan, and then acts. He argued the modeling step is both the hardest part and largely unnecessary. His alternative, the subsumption architecture, decomposes a system not into sense-model-plan-act stages but into layers of simple behaviors that each connect perception directly to action and run in parallel. A lower layer might keep the robot from hitting walls; a higher layer might make it wander; another might send it toward a goal. There is no central place where a model of the world is stored. His slogan captured it: “the world is its own best model” - rather than remember where things are, just look again.
To demonstrate the point, Brooks built small mobile robots that navigated cluttered real environments robustly, achieving what he called insect-level intelligence with no symbolic representation at all. The claim was not merely that this was an efficient engineering trick but that representation-free, embodied interaction is closer to how real intelligence is actually built up, from the bottom.
The paper helped split AI into camps and seeded the modern fields of behavior-based and embodied robotics. Brooks went on to co-found iRobot and Rethink Robotics, carrying the bottom-up, build-it-and-test-it philosophy from the lab into commercial machines.