CONCEPT
The Myth of the Given
Wilfrid Sellars’s name for the deepest error in empiricist epistemology—the belief that knowledge rests on a stratum of raw, uninterpreted fact delivered to the mind prior to any concept—and the foundational fiction of AI objectivity.
The Myth of the Given is the idea that somewhere beneath our theories and concepts there is a bedrock of pure, uninterpreted experience—bare facts the world simply hands over, self-authenticating and theory-free, waiting to be read off and used to build knowledge.
Wilfrid Sellars named and attacked this idea in his 1956 masterwork “Empiricism and the Philosophy of Mind,” arguing that no such stratum can exist: to register anything as a red patch, a warm sensation, or a round object already requires possessing the concepts of red, warm, and round, which requires knowing how they contrast with other concepts, when they reliably apply, and what would make one wrong. The supposedly bare given is already saturated with conceptual structure; the building goes all the way down. In artificial intelligence this myth reappears as the foundational fiction of
machine learning objectivity: the idea that data is the given—raw material the world supplies uninterpreted, free of human bias—and that