theorem
proved
Turing_incomplete
show as:
view math explainer →
open explainer
Read the cached plain-language explainer.
open lean source
IndisputableMonolith.Complexity.ComputationBridge on GitHub at line 161.
browse module
All declarations in this module, on Recognition.
explainer page
depends on
formal source
158 }
159
160/-- Turing incompleteness: the model ignores recognition cost -/
161theorem Turing_incomplete (TM : TuringModel) :
162 ∃ (problem : Type) (LC : LedgerComputation),
163 -- The ledger model captures costs Turing ignores (existence of a hard measurement instance)
164 (∃ (n : ℕ) (M : Finset (Fin n)) (hM : M.card < n),
165 ¬ (∀ b R, LC.measure (LC.encode (BalancedParityHidden.enc b R).toList) M = b)) ∧
166 -- Turing counts only evolution, not measurement
167 TM.recognition_complexity = fun _ => 0 := by
168 -- Witness: any problem with balanced-parity output
169 let LC : LedgerComputation := {
170 states := Unit -- placeholder
171 evolve := id
172 encode := fun _ => ()
173 measure := fun _ _ => false
174 flux_conserved := fun _ => rfl
175 measurement_bound := by
176 intro n M hM
177 -- Apply the balanced-parity lower bound
178 classical
179 intro h
180 -- Instantiate the universal claim at `b = true` with any mask `R`.
181 -- Our `measure` always returns `false`, so it cannot equal `true`.
182 have h' := h true (fun _ => false)
183 simpa using h'
184 }
185 use Unit, LC
186 -- Provide a concrete hard instance using the bound and trivial size witness.
187 refine ⟨?_, ?_⟩
188 · refine ⟨1, (∅ : Finset (Fin 1)), by decide, ?_⟩
189 -- Instantiate the universal impossibility from the `measurement_bound` field.
190 simpa using (LC.measurement_bound 1 (∅) (by decide))
191 · funext n