Minimal change collapses to the unrelaxed tower
ProvedTranscendenceTower.MinimalChange.copy_is_minimumformalizing-contextlogicnonmonotonic-logic
For every relaxation set , the copy context belongs to , has empty change set, is the unique relaxed model with empty change set, and is therefore the -minimum of the relaxed models.
This is a negative result about the obvious proposal. Preferring contexts that change as little as possible returns the unrelaxed tower no matter what was relaxed: circumscribing the abnormality recovers precisely the case McCarthy calls pointless, in which each context faithfully copies the one below.
So minimal change is not the mechanism that makes relaxation meaningful. Permitting a context to drop an assumption never forces it to, and the minimal-change reading always declines the invitation.
Preamble
import Definitions.Def_TranscendenceTowerMinimalChange set_option autoImplicit false open TranscendenceTower TranscendenceTower.Relaxation
Formal statement
namespace TranscendenceTower.MinimalChange
theorem copy_is_minimum (v : Atom 0 → Bool) (A : Set (Lang 0)) :
copy v ∈ Mod v A ∧ Delta v (copy v) = ∅
∧ (∀ N ∈ Mod v A, Delta v N = ∅ → N = copy v)
∧ (∀ N ∈ Mod v A, Delta v (copy v) ⊆ Delta v N) := by sorry
end TranscendenceTower.MinimalChange
Source
Mateo Estrada, Harmless and Pointless: Consistency, rigidity, and the indeterminacy of context transfer in McCarthy's formalization of context, 7 September 2026, Section 5, Theorem 10.