The Shadow Price of Distinguishability, Volume 2: What Is Structural, What Is Conditional, and What Cannot Be Tested

A formal framework in which phenomenal state is identified with an endogenous operational code and two conjugate shadow prices of unrealized distinguishability — the cost of not knowing at the moment of action (Πact) and the residual unpredictability of the world (Πpred) — under a hard resource constraint m + k = B. Volume 1 (18 pages, 32 references) establishes the framework (AFF quotient, PIP* postulate), derives closed-form consequences in three toy models, and proposes three falsifiable human protocols with specific numerical inequalities. Published as Zenodo DOI: 10.5281/zenodo.18745079. Volume 2 (23 pages, 35 references) revises and strengthens Volume 1 through closed-form analysis over 1100 parameter configurations and 100 stochastic seeds. Key results: (i) Theorem 1 proves universal monotonicity of nuance degradation under agentive load via Blackwell ordering — tunnel vision is structural, not parameter-dependent; (ii) a generalized dissociation theorem identifies three regimes (competitive, pure, and boundary dissociation) depending on hint informativeness, replacing the narrow R(D) > 1 statement; (iii) sensitivity analysis shows curvature of degradation is parameter-conditional, falsifying the earlier universal concavity claim; (iv) prediction horizon invariance across H = 1..10 confirms the H = 1 choice is non-arbitrary; (v) 100-seed growing-budget simulation with Wilson confidence intervals exhibits a sharp survival phase transition; (vi) MDP-homomorphism zombie divergence theorem with explicit scope caveat; (vii) coherence table across five phenomenological facts and six competing theories; (viii) two new sections on underdetermination of bridging principles (PIP* included) and a viability no-go result.