Context
Exposure measures how many extra levels of finite input information are needed to predict an action at a chosen output resolution.
Definitions
- The quotients satisfy , and equality in every quotient separates points of .
- Exposure means for every .
Hypotheses and scope
- The exact state space is compact Hausdorff.
- Each finite observation space is finite discrete and each quotient map is continuous and onto.
Proof or evidence
These assumptions make the canonical map from X into the inverse limit continuous and bijective; compactness and Hausdorffness then make it a homeomorphism.
Verification notes
The definition and completeness assumptions were reconciled with the M22 specification, proof, and closure audit. On 9 September 2026, a worked binary-digit example was checked in separate same-model passes for exact exposure, continuity, tower dependence, and boundary cases; this is not external expert validation.
Limitations
- Continuity alone need not give one finite exposure bound across all levels.
- The filtration depends on the selected observation tower.
Notes
Exposure is an upper bound and is monotone in e. Least exposure, when it exists, is a derived invariant. The definition records a uniform additive observation budget rather than an arbitrary continuity modulus.