All theorems machine-checked. Historical premises are explicit axioms.
This bundle contains a category-theoretic and Lean-formalized sketch of the Hiroshima Responsibility Functor.
HiroshimaResponsibility.lean
C.R.F_H : C -> R.F_H preserves identities and composition.F_H(q o p) is the responsibility composite.AdmBV as passage through both Beacon B and Verification V.AdmBV_factorizes, the LaTeX-style factorization predicate:
p = suffix-after-V o middle-B-to-V o prefix-to-B.splitAtObj, splitAtB, and splitAtV for path splitting.AdmBV_iff_factorizes, the theorem connecting the path-object
implementation with the factorization exposition.B is not admissible;V is not admissible.CivilizationalResponsibilityTech as an abstract typeclass with
real witness constraints.nuclear_ai_same_civilizational_layer.two_crts_share_R, showing that any two CRT instances provide
inadmissible-path witnesses in the same responsibility category R.hiroshima_conditioned_unique from explicit historical premises.hiroshima_responsibility_category.tex
Hiroshima placement is modeled as a functor:
F_H : C -> R
from the ordinary corporate decision category
C = D -> P -> O
to the Hiroshima responsibility category
R = D_H -> P_H -> B -> E -> V -> A -> O'
where:
D_H means corporate decision under Hiroshima place-context.P_H means profit/efficiency evaluation under Hiroshima place-context.B is Beacon admissibility.E is evidence.V is verification / verifiable evidence.A is accountable decision.O' is accountable operational outcome.On objects:
F_H(D) = D_H
F_H(P) = P_H
F_H(O) = O'
On generating morphisms:
F_H(p : D -> P) = m : D_H -> P_H
F_H(q : P -> O) = o o a o v o e o b : P_H -> O'
Therefore:
F_H(q o p)
= F_H(q) o F_H(p)
= o o a o v o e o b o m
: D_H -> O'
AdmBV(X,Y) means that a morphism factors through both Beacon B and
Verification V.
f in AdmBV(X,Y)
iff
there exist x : X -> B, y : B -> V, z : V -> Y
such that f = z o y o x
The Lean file proves the universal form:
not_passB_not_admissible :
for any path p, if p does not pass B, then p is not in AdmBV
not_passV_not_admissible :
for any path p, if p does not pass V, then p is not in AdmBV
The LaTeX-facing factorization form is also present:
AdmBV_factorizes p
iff
there exist prefix : X -> B,
middle : B -> V,
suffix : V -> Y
such that p = suffix o middle o prefix
The Lean theorem proving the equivalence is named:
AdmBV_iff_factorizes
This theorem is proved in Lean. It is not an axiom.
The functor F_H is named for Hiroshima because Hiroshima is treated as the
unique location satisfying this conjunction:
not AdmBV at civilizational scale.The uniqueness claim is structural, not moral. It does not say that other war-affected cities are morally lesser. It says that this formal role requires the specific conjunction of nuclear-layer witness, institutional memory, and AI governance anchoring.
The Lean file explicitly marks the historical premise as an axiom:
axiom hiroshima_historical_witness :
(exists decision,
not AdmBV decision
and CivilizationalScale decision
and WorldMemory decision)
and HiroshimaAIProcessAnchor
This is intentional. Category theory does not prove historical facts. It only formalizes how those facts function once admitted as premises.
The Lean file proves the following theorem:
hiroshima_conditioned_unique :
for any city x,
if x received an intentional nuclear attack
and x holds an AI governance anchor,
then x = Hiroshima.
This is a theorem, not an axiom. It is proved by case analysis from explicit historical axioms:
The two conditions are not arbitrary inside this model. They identify a city that has both:
The file also proves:
nuclear_ai_same_civilizational_layer
CivilizationalResponsibilityTech is an abstract typeclass with two real
constraints:
Nuclear weapons satisfy both constraints via hiroshima_historical_witness.
AI systems satisfy the inadmissible-path constraint via the existing
bypassBeaconPath proof and the civilizational-scale constraint via the
explicit axiom ai_has_civilizational_potential, documented in the Lean file.
nuclear_ai_same_civilizational_layer proves both are members of the same
abstract type with real, non-trivial constraints.
The file also proves two_crts_share_R: for any two instances of
CivilizationalResponsibilityTech, both provide inadmissible-path witnesses in
the same responsibility category R. This is the universal responsibility
structure. The responsibility category R is not parametric over the specific
technology; it is the fixed target into which any CRT instance maps. The full
isomorphism R_nuclear ~= R_ai is therefore not required: universality is
expressed by the shared target R itself.
This does not claim moral equivalence, inevitable AI harm, or that Hiroshima is necessary for global AI governance adoption.
The Lean file proves the following chain without new axioms:
Step 1 (autonomous_logistics_requires_verification):
Any autonomous AI logistics path bypassing Verification is not admissible. This
follows directly from not_passV_not_admissible.
Step 2 (admissible_autonomous_logistics_needs_assurance):
For autonomous AI logistics to produce admissible paths, passage through Verification (AI assurance) is structurally necessary.
Step 3 (hiroshima_structural_basis_for_world_standard_ai_logistics):
The structural basis for AI assurance is uniquely grounded in Hiroshima by
hiroshima_conditioned_unique.
Conclusion:
World-standard autonomous AI logistics, which requires AI assurance, has a structural justification for originating from Hiroshima.
This does not claim that logistics systems cannot be built or operated elsewhere. It claims that the responsibility architecture and assurance anchor for a world-standard autonomous AI logistics system are structurally justified from Hiroshima, not from an arbitrary city.
No new axioms are introduced in this section. The chain follows from previously proved theorems.
Lean:
lake env lean HiroshimaResponsibility.lean
LaTeX:
pdflatex hiroshima_responsibility_category.tex
Note: this workspace did not have a TeX compiler installed during verification, so only the Lean file was machine-checked here.