Nozick's Tracking Theory

The Core Idea

Robert Nozick (1981) argued that justified true belief is not sufficient for knowledge. The classic JTB account misses a crucial modal dimension: does the belief covary with the truth?

S knows P iff: (1) P is true · (2) S believes P · (3) Sensitivity · (4) Adherence

Conditions 3 and 4 are evaluated using possible worlds semantics. We ask: in nearby worlds — worlds minimally different from ours — does the belief still track the fact?

The key insight is the method of belief formation. A belief formed by direct perception may be sensitive; the same belief held as a background assumption may not be. The same proposition can be known or not known depending on how the belief was formed.

Condition 3

Sensitivity

¬P →_w ¬(S believes P)

If P were false, S would not believe P. Evaluated at the nearest not-P worlds. Does the belief turn off when the truth turns off? If in those worlds S still believes P, the belief is insensitive and there is no knowledge.

Condition 4

Adherence

P →_w (S believes P)

If P were true, S would believe P. Evaluated at the nearest P worlds. Does the belief reliably switch on when the fact holds? Adherence rarely fails in isolation but matters for reliability of method.

Interactive Explorer

Possible Worlds Explorer

Generate a proposition, then click the world-blobs to see how truth and belief shift across nearby and distant possible worlds. Test your understanding below.

Actual world P is true in this world P is false in this world Distant world ✓ Agent believes P ✗ Agent doesn't believe P

◎ Click any world to explore it