Knowledge in the Head and in the World
Central claim: precise behavior can emerge from imprecise knowledge because knowledge is distributed — partly in the head, partly in the world. We combine both because neither alone suffices. (Demonstration: people can't draw a penny accurately yet use money fine — Nickerson & Adams.)
Four reasons precise behavior emerges from imprecise knowledge:
- Knowledge is both in the head and in the world.
- Great precision is not required — only enough to distinguish the right choice from the alternatives.
- Natural (physical) constraints exist in the world.
- Cultural constraints and conventions exist in the head.
Two kinds of knowledge
- Knowledge of (declarative) — facts and rules ("stop at red"; "key out of Saab requires reverse"). Easy to write and teach; need not be true; knowing a rule ≠ following it.
- Knowledge how (procedural) — skills (returning a tennis serve). Largely subconscious (behavioral level), hard or impossible to verbalize; learned by practice, not description.
The tradeoff
- Knowledge in the world: easy to use without learning, but only useful when present; can clutter; relies on being perceivable (signifiers, constraints, mappings are knowledge in the world).
- Knowledge in the head: efficient and fast once learned, available anywhere, but requires learning and can be forgotten.
- When precision is unexpectedly required the combined knowledge can fail — e.g., new coins confused with old (US Susan B. Anthony dollar; French 10-franc) because people store only partial descriptions, precise enough only to discriminate the items they're used to.
Design takeaway
Put enough knowledge in the world that non-experts succeed without prior learning, while still allowing experts to operate from the head — a key principle for Designing for Error.
Confidence
1.0 — explicit chapter framework.
Sources: ["DOET ch.3 'Precise Behavior from Imprecise Knowledge' p.74-109"]