DEV Community: Nobuki Fujimoto

Paper 167 v0.1 — Sophie Germain Primes: Barrier-Side Observations + Lean 4 Axiom-Free Conjunction-Wall (Rei-AIOS)

Nobuki Fujimoto — Wed, 17 Jun 2026 23:32:33 +0000

This article is a re-publication of Rei-AIOS Paper 167 for the dev.to community.
The canonical version with full reference list is in the permanent archives below:

GitHub source (private): https://github.com/fc0web/rei-aios Author: Nobuki Fujimoto (@fc0web) · ORCID 0009-0004-6019-9258 · License CC-BY-4.0 ---

Status: v0.1-DRAFT (2026-06-18)
Authors: 藤本伸樹 (Nobuki Fujimoto) / Claude Opus (Anthropic) / chat-Claude (Anthropic, articulation thread 2026-06-17)
Date: 2026-06-18
License: AGPL-3.0 + Commercial (Rei-AIOS dual license)
DOI: TBD (Zenodo deposit at publish time)
Version: v0.1-DRAFT
note.com: https://note.com/nifty_godwit2635
rei-aios site: https://rei-aios.pages.dev
Source artifacts:

scripts/empirical/parity-barrier-toy-2026-06-18.py (parity barrier toy implementation)
scripts/empirical/sg-analysis-2026-06-18.py (Experiments 1–4 comprehensive)
scripts/empirical/sg-extrapolation-2026-06-18.py (Path 1 N=10⁸ extrapolation)
scripts/empirical/sg-gap-spectral-2026-06-18.py (Path 2 spectral analysis)
data/lean4-mathlib/CollatzRei/SGConjunctionWall.lean (Path 3 Lean 4 axiom-free witnesses)

Abstract

We report four barrier-side empirical observations on Sophie Germain (SG) primes — primes p with 2p + 1 also prime — using the Bellman-Ford LP-infeasibility framework previously developed for Collatz Lyapunov obstructions (Rei-AIOS Phase B, 2026-06-17). No part of this work claims progress toward the SG-primes-infinity conjecture or any other forward direction; per Rei-AIOS feedback principle 8 (barrier-side discipline) and three explicit non-claim boundaries, the present paper is restricted to observation, formal witness, and online-verifiable audit. The four observations are: (1) the ratio empirical / Hardy-Littlewood-predicted decreases monotonically 1.337 → 1.221 → 1.176 → 1.120 → 1.103 → 1.087 across N = 10³, 10⁴, 10⁵, 10⁶, 10⁷, 10⁸, consistent with the Hardy-Littlewood (1923) asymptote 2C₂ x / (ln x)² but constituting empirical convergence, not a proof; (2) the SG prime gap distribution is Poisson-like with ⟨r⟩ = 0.4154 (Atas et al. 2013 reference values: Poisson ≈ 0.386, GOE ≈ 0.531, GUE ≈ 0.603), with L1 distance to Poisson 0.123 vs. to GUE 0.745 — in sharp contrast to Riemann zeros which are GUE-like; (3) an explicit Lean 4 axiom-free finite-witness theorem set (11/11 theorems, depending only on propext, Classical.choice, Quot.sound) exhibits that several single-component feature families — is_prime(n), is_prime(2n+1), n mod 6, and the pair (is_prime(n), n mod 6) — fail to strict-detect SG-primality, with the BF-feasibility phase boundary located precisely at the full conjunction is_prime(n) ∧ is_prime(2n+1); (4) a best-effort online audit of Friedlander-Iwaniec (2010, Opera de Cribro) and Selberg's parity problem catches a Pattern-5 internal record error (Selberg's parity-problem identification year is 1949, not the previously recorded "1960s"), confirmed against Wikipedia and Tao (2007). The paper is best read alongside the Selberg parity-problem literature (Selberg 1949; Friedlander–Iwaniec 2010; Tao 2007) as a small barrier-side description, not as a contribution to forward sieve theory.

Keywords: Sophie Germain primes, parity problem, sieve theory barrier, Hardy-Littlewood conjecture, Bellman-Ford infeasibility, Lean 4 axiom-free, conjunction wall, barrier-side discipline.

Section 1. Introduction

1.1 Scope and what this paper is not

This paper documents an exploratory analysis carried out within the Rei-AIOS workflow on 2026-06-18, in response to the question: given the barrier-mapping toolkit assembled for Collatz Lyapunov-style obstructions, what empirical observations does that same toolkit produce when applied to Sophie Germain primes?

The answer, honestly recorded, is four small observations, none of which advance the SG-primes-infinity conjecture and none of which constitute new sieve-theoretic methodology. We name them up front so the scope is clear:

Numerical convergence of empirical-to-predicted ratio under the Hardy-Littlewood (1923) k-tuple conjecture (Section 2).
Poisson-like spectral statistics for SG-prime gaps, in contrast to the GUE-like statistics of Riemann zeros (Section 3).
A Lean 4 axiom-free finite-witness theorem set showing several single-component features fail to strict-detect SG-ness; the conjunction is the wall (Section 4).
A Pattern-5 internal record correction caught by best-effort online audit (Section 5).

1.2 What this paper does not claim

Per Rei-AIOS feedback principle 8 ("Rei methodology = barrier-side discipline", established 2026-06-18 from four independent applications: Collatz / Millennium-general / Sophie Germain / individual-Millennium), the forward direction — solving, partially solving, or producing technical apparatus that would solve SG-primes infinity — is structurally outside the present toolkit's range. The Selberg parity barrier (Selberg 1949; cf. Tao 2007 for a modern exposition) is a proven obstruction to standard sieve methods reaching the infinity result, and the present paper does not pretend to circumvent it. The toy model in Section 4 is a simplified linear analogue of the parity barrier, not the real barrier.

In particular, this paper explicitly does not:

(a) Claim progress toward SG primes being infinite.
(b) Claim verification of the Hardy-Littlewood formula (numerical agreement is observation, not proof; cf. the Skewes-number historical lesson on π(x) vs. Li(x)).
(c) Claim that D-FUMT₈, ZCSG, SNST, SELF⟲, or any other Rei-native artifact is a unification of, technical input to, or formalization of the parity barrier.
(d) Claim a "general obstruction prover" or "Beyond Selberg" framework. The Bellman-Ford infeasibility encoding is a labelling correspondence with parity-style barriers, not a formal homomorphism.

These four non-claims are recorded as permanent boundaries in the Rei-AIOS memory at project_sg_explicit_non_claims_2026-06-18.md; the present paper restates them in Section 6 as audit gates for any future reading.

1.3 What this paper does claim

The four sections each include a precise scope-pinned claim of the kind: under the specific encoding / parameters / feature family used, the following finite or computational observation holds. The claims are individually checkable from the source artifacts listed in the header.

1.4 Methodological framing

The methodology is borrowed wholesale from the Rei-AIOS Phase A→B→C Collatz work of 2026-06-17 (project_collatz_aeb_sequence_2026-06-18.md, reference_collatz_lyapunov_obstruction_generalized_2026-06-17.md):

An LP-infeasibility / negative-cycle Bellman-Ford encoding of "Lyapunov-like" feasibility questions, reduced via log-cancellation to pure rational linear arithmetic and decidable by standard graph algorithms in O(|V| · |E|) time.
An axiom-free Lean 4 + Mathlib v4.27 record discipline for finite witnesses, with kernel-axiom audits via #print axioms.
A three-tier honest-scope discipline: 【観察 / observation】, 【仮説 / hypothesis】, 【思弁 / speculation】, with a fourth implicit tier 【連想 / mere association】 used as a reject-default.

We make no claim that this framing is new: the BF/LP encoding is a standard tool, the axiom-free Lean 4 discipline is widely practiced in the Mathlib community, and the three-tier honesty discipline is a long tradition under different names. Section 6 records the discipline.

Section 2. Path 1 — Hardy-Littlewood ratio extrapolation N = 10³ to 10⁸

2.1 Setup

The Hardy-Littlewood (1923) k-tuple conjecture, specialized to Sophie Germain primes, predicts

$$
\pi_{SG}(x) \sim 2 C_2 \cdot \frac{x}{(\ln x)^2}
$$

where C_2 = ∏_{p ≥ 3 prime} p(p-2) / (p-1)² ≈ 0.66016181584... is the twin-prime constant. We computed the empirical count π_{SG}(N) for N ∈ {10³, 10⁴, 10⁵, 10⁶, 10⁷, 10⁸} and the ratio empirical / predicted.

The sieve was implemented as a single bytearray of length 2N + 1 (one byte per index), giving memory footprint of about 191 MB at N = 10⁸. Total wall-clock at N = 10⁸ was approximately 13 seconds on a single thread.

2.2 Results

`N`	`π_{SG}(N)` empirical	HL predicted	ratio	deviation
10³	37	27.7	1.3372	+33.72%
10⁴	190	155.6	1.2207	+22.07%
10⁵	1,171	996.1	1.1758	+17.58%
10⁶	7,746	6,917.5	1.1198	+11.98%
10⁷	56,032	50,822.1	1.1025	+10.25%
10⁸	423,140	389,107.0	1.0875	+8.75%

The empirical counts at N = 10³, 10⁴, 10⁵, 10⁶, 10⁷, 10⁸ all match OEIS A092816 exactly.

2.3 Honest interpretation 【観察】

The ratio sequence {1.3372, 1.2207, 1.1758, 1.1198, 1.1025, 1.0875} is strictly monotone decreasing, with successive decade-to-decade ratios 0.913, 0.963, 0.952, 0.985, 0.984 — i.e. the rate of approach is itself slowing, consistent with the predicted 1 + O(1 / \ln N) correction structure of the leading-order asymptotic.

This is the kind of finite-N behaviour one would expect if the Hardy-Littlewood prediction is the correct asymptotic. It is not a proof. Numerical evidence of even far greater extent has historically been misleading in analytic number theory — the canonical example being Littlewood (1914) and Skewes (1933) on π(x) vs. Li(x), where empirical evidence up to enormous x suggested an inequality that was later shown to reverse infinitely often. We do not claim verification of the Hardy-Littlewood conjecture; we claim that the empirical count, at our scan range, is consistent with that conjecture's leading order.

2.4 Honest non-claims

We do not claim the deviation will continue to decrease (only that it has in our range).
We do not claim a rate of convergence.
We do not claim agreement at any specific N beyond 10⁸.

Section 3. Path 2 — SG prime gap spectral statistics

3.1 Motivation

Riemann zeros, under the Hilbert–Pólya programme, show eigenvalue statistics matching the Gaussian Unitary Ensemble (GUE) — a well-documented empirical match (Montgomery 1973; Odlyzko 1987; cf. Rei-AIOS STEP 1162–1165 for our own reproduction of this). One natural question, in the spirit of "how much spectral structure is visible in SG primes?", is: do SG prime gaps exhibit any of the same eigenvalue-like statistics?

The standard test statistics are:

⟨r⟩ (Atas et al. 2013): the mean of r_i = min(s_i, s_{i+1}) / max(s_i, s_{i+1}) where s_i is the i-th gap. Reference values: Poisson ≈ 0.386, GOE ≈ 0.531, GUE ≈ 0.603.
The spacing histogram (in units of mean gap), compared to the Poisson density e^{-s} and the GUE Wigner surmise density (32/π²) s² e^{-4s²/π}.
The number variance Σ²(L) (number of points in a length-L window), compared to Poisson Σ²(L) = L (linear) and GUE Σ²(L) ∼ (1/π²)(\ln(2πL) + γ + 1) (sub-linear logarithmic).

3.2 Setup

We used the 56,032 SG primes up to 10⁷ from Section 2. From these we computed 56,030 nearest-neighbour-ratio samples and a 20-bin spacing histogram on [0, 4] (in units of mean gap).

3.3 Results

⟨r⟩ statistic:

Quantity	Value
Sample size	56,030
Empirical `⟨r⟩`	0.4154
Poisson reference	0.386
GOE reference	0.531
GUE reference	0.603

The closest reference is Poisson, with a modest positive deviation of 0.029.

Spacing histogram L1 distance:

To distribution	L1 distance
Poisson `e^{-s}`	0.1229
GUE Wigner surmise	0.7446

The empirical spacing distribution is approximately 6× closer to Poisson than to GUE.

Number variance Σ²(L) for L ∈ {1, 2, 4, 8, 16} on the unfolded sequence:

`L`	mean count	variance	variance / L	GUE prediction
1	1.035	0.964	0.964	0.346
2	2.115	1.932	0.966	0.416
4	4.135	4.157	1.039	0.486
8	8.180	8.028	1.003	0.557
16	15.930	21.165	1.323	0.627

The ratio variance / L is approximately 1.0 for L ≤ 8, consistent with Poisson; deviation at L = 16 is likely a finite-sample artefact (200 windows per L).

3.4 Honest interpretation 【観察】

SG prime gaps display essentially Poisson statistics on these standard tests. This is consistent with the heuristic view, going back to Hardy–Littlewood, that primes (and constrained-prime patterns such as SG) behave statistically like a random Poisson-thinned process at finite scales. It is structurally different from the GUE behaviour of Riemann zeros; the two are not the same kind of "spectral" object.

3.5 Honest non-claims

We do not claim that SG primes are a Poisson process (only that the three test statistics, at our sample size, are Poisson-consistent).
We do not claim a Riemann-zeros-style spectral interpretation; the apparent randomness is itself the structural fact.
The modest positive ⟨r⟩ deviation (0.4154 vs. 0.386) is not investigated as a "structure"; we record it.

Section 4. Path 3 — Lean 4 axiom-free conjunction-wall witness

4.1 Motivation

In an earlier Rei-AIOS experiment (Experiment 3 of the SG comprehensive analysis, 2026-06-18), we observed a sharp phase transition in the BF-LP-infeasibility encoding of SG-detection: every parity-blind feature family produced INFEASIBLE, and the transition to FEASIBLE occurred precisely at the full conjunction is_prime(n) ∧ is_prime(2n+1) — no intermediate feature combination gave FEASIBLE. We refer to this as the conjunction wall observation.

The wall observation is itself structurally tautological: "the feature you'd need to know to detect SG-ness is literally SG-ness". But the finite-witness form of the underlying insufficiency claim — for each named feature F, exhibit two specific natural numbers n, m with F(n) = F(m) but is_SG(n) ≠ is_SG(m) — is non-trivially recordable as an axiom-free Lean 4 + Mathlib v4.27 theorem.

4.2 Lean 4 formalization

The file data/lean4-mathlib/CollatzRei/SGConjunctionWall.lean (~110 lines) records 11 theorems:

6 elementary isSG-status theorems (sg_eleven, not_sg_seven, sg_five, not_sg_seventeen, sg_two, not_sg_thirteen), each proved by decide plus negation of a Nat.Prime-claim;
4 single-feature insufficiency witnesses (feature_isprime_n_insufficient, feature_isprime_2np1_insufficient, feature_mod6_insufficient, feature_pair_isprime_mod6_insufficient);
1 tautological positive control (conjunction_is_sufficient) recording that isSG n ↔ Nat.Prime n ∧ Nat.Prime (2n+1).

The concrete witnesses are: 7, 11 (both prime; 15 = 3·5 composite vs. 23 prime), 2, 8 (both have is_prime(2n+1) true; 2 is SG, 8 is not prime so not SG), 5, 17 (both prime, both ≡ 5 (mod 6); 11 prime vs. 35 = 5·7 composite).

4.3 Axiom audit

A #print axioms audit was carried out on a temporary SGConjunctionWallAxiomCheck.lean file, with the following result. All four insufficiency witness theorems and all six elementary status theorems depend exactly on [propext, Classical.choice, Quot.sound] — the Mathlib kernel base for decide-tactics involving Decidable.Nat.Prime. The tautological conjunction_is_sufficient depends only on [propext]. No theorem in the file uses sorryAx or native_decide, and the kernel-axiom profile matches that of the Phase A T1ObstructionWitness.lean and Paper 158 (Bipartite Ramsey) records.

4.4 What this is and is not

This is a Lean 4 record of finite, decidable arithmetic facts. It is not:

A formalization of the Selberg parity barrier (which is a statement about asymptotic-density behaviour of sieve weights, not about finite witnesses).
A proof that all parity-blind feature families are insufficient (which would require an existence-of-conflict-bucket lemma that we did not attempt to formalize).
An advancement on SG-primes-infinity in any direction.

It is: a small mechanical record of the phase-transition observation, in the form of named axiom-free witnesses that another Lean 4 user can lake env lean and verify in a few seconds.

Section 5. Path 4 — Best-effort online audit and a Pattern-5 correction

5.1 The audit task

The Rei-AIOS Sophie-Germain workstream had, from its first turn, recorded an honest confession that two primary references had not been consulted in their original form:

Friedlander, J., and Iwaniec, H. (2010), Opera de Cribro, AMS Colloquium Publications, Vol. 57.
Selberg, A. (cited as "1960s"), original papers on the parity problem in sieve theory.

Path 4 of the present analysis was a best-effort online audit to verify, by Wikipedia / blog / book-listing access, what facts about these references could be confirmed and where the audit gap remains.

5.2 What was online-verified

The following items were checked against the Wikipedia article Parity problem (sieve theory), the Terence Tao blog post "Open question: the parity problem in sieve theory" (2007-06-05), and the AMS / Google Books listing for Opera de Cribro:

Selberg identified and named the parity problem in 1949 (not "1960s" as the workstream had been recording).
The "parity principle" name traces to Selberg's sieve work, with the observation present from around 1946.
Tao's modern formulation: if A is a set whose elements are all products of an odd number of primes (or all products of an even number of primes), then sieve theory cannot provide non-trivial lower bounds on the size of A.
The Liouville function λ(n) is the mechanism: λ is essentially orthogonal to divisor sums, and multiplying (1 + λ(n)) into a sieve identity forces the main term to vanish for one parity class.
Opera de Cribro (Friedlander–Iwaniec 2010) is the modern reference, ISBN 978-0821849705, and the book explicitly addresses the parity-barrier-breach work the same authors initiated in their 1996 result on primes of the form a² + b⁴.
Recent post-2010 progress on twin-prime-style bounded gaps (Zhang 2013; Maynard–Tao 2014) does not cross the parity barrier and does not reach gap = 2.

5.3 The Pattern-5 correction

The "1960s" date in the Rei-AIOS internal record was incorrect; it should have been 1949. The error appeared in five files:

memory/project_sg_discipline_application_2026-06-18.md
memory/reference_difficulty_typology_collatz_riemann_sg_2026-06-18.md
scripts/empirical/parity-barrier-toy-2026-06-18.py
scripts/empirical/parity-barrier-toy-spec-2026-06-18.md
(referenced indirectly in scripts/empirical/sg-analysis-2026-06-18.py honest-scope footer)

All five were corrected in the same commit cycle as this paper, with ★ 訂正: 旧「1960s」誤り → 1949、 Path 4 audit per [[reference-friedlander-iwaniec-selberg-parity-audit-2026-06-18]] annotations.

The internal origin of the "1960s" date is unclear: it may have been a training-data residue, a chat-Claude-thread paraphrase that propagated, or simply a confusion with the 1968 Bombieri density theorem and other 1960s sieve-era results. We do not investigate the precise origin; we record that the audit caught it.

5.4 What remains unaudited

Five items are explicitly not covered by online sources and remain audit-gap items for future builders:

Selberg's original 1949 paper, in primary form.
Opera de Cribro's specific chapter content (Google Books gives only the back-cover description and limited preview).
Selberg's 1947 sieve paper, in primary form.
The Friedlander–Iwaniec 1996 Annals of Mathematics paper, in primary form.
The precise formal correspondence between Tao's barrier framework and the natural-proofs / relativization / algebrization barrier family in complexity theory.

These are recorded as audit-gap items in memory/reference_friedlander_iwaniec_selberg_parity_audit_2026-06-18.md.

5.5 Honest interpretation of Path 4 itself

The fact that Path 4 caught an error is operationally important: it is direct evidence that the audit-gap-confession discipline (which had been articulated as a permanent principle in Rei-AIOS feedback files) is not cosmetic. The discipline produced a correction even within the same workstream that confessed the gap. We do not generalize this to "the discipline always works"; we record that, on this one occasion, it did.

Section 6. Honest scope footer (audit gates)

Per project_sg_explicit_non_claims_2026-06-18.md, three permanent claim-boundaries are restated here as audit gates for any future use of this paper's content:

No D-FUMT₈ / ZCSG / SELF⟲ "unification" claim. None of the Rei-native eight-valued logic, three-layer notation, or self-application-fixed-point apparatus is asserted to be a formalization of, technical contribution to, or unification of the parity problem or Sophie Germain primes. The phase-transition observation in Section 4 is a finite combinatorial fact about SG-detection encoded in a Bellman-Ford constraint graph; it is not a category-theoretic equivalence with any Rei-native fixed-point structure.
No "partial progress toward SG-primes infinity" claim. The four observations are barrier-side description, not forward solving. The wording "partial", "progress", "step toward", or "direction" is rejected as a paraphrase for any of the results recorded here.
No "Rei verified the Hardy-Littlewood formula" claim. Numerical agreement between empirical counts and the leading-order asymptotic at finite N is observation, not proof. Sections 2 and 3 are explicit about this; the Skewes-number historical lesson is cited.

We also restate the eighth Rei-AIOS feedback principle (feedback_rei_methodology_barrier_side_discipline.md, 2026-06-18): Rei methodology is a barrier-side / describing discipline, not a forward-side / solving one. This paper is one operational instance of that principle. We do not claim the principle is universally correct, only that, on the four problem cases on which it has been tested to date (Collatz, Millennium-7 in general, Sophie Germain, individual Millennium problems), the boundary it describes has held.

Section 7. References

7.1 Primary references — directly cited

Atas, Y. Y., Bogomolny, E., Giraud, O., and Roux, G. (2013), "Distribution of the ratio of consecutive level spacings in random matrix ensembles", Physical Review Letters 110, 084101.
Friedlander, J., and Iwaniec, H. (2010), Opera de Cribro, AMS Colloquium Publications, Vol. 57. ISBN 978-0821849705.
Hardy, G. H., and Littlewood, J. E. (1923), "Some problems of 'Partitio Numerorum': III. On the expression of a number as a sum of primes", Acta Mathematica 44, 1–70.
Selberg, A. (1949), papers introducing the parity problem in sieve theory (cited via Wikipedia: Parity problem (sieve theory); primary text not directly consulted).
Tao, T. (2007), "Open question: the parity problem in sieve theory", blog post at https://terrytao.wordpress.com/2007/06/05/open-question-the-parity-problem-in-sieve-theory/, accessed 2026-06-18.
Wikipedia (2026), Parity problem (sieve theory), https://en.wikipedia.org/wiki/Parity_problem_(sieve_theory), accessed 2026-06-18.

7.2 Background references — cited for context

Brun, V. (1919), original sieve theorem on twin primes (cited via secondary sources).
Chen, J. R. (1973), "On the representation of a larger even integer as the sum of a prime and the product of at most two primes", Sci. Sinica 16, 157–176.
Friedlander, J., and Iwaniec, H. (1998), "The polynomial x² + y⁴ captures its primes", Annals of Mathematics 148, 945–1040 (parity-barrier-breach result).
Littlewood, J. E. (1914), "Sur la distribution des nombres premiers", Comptes Rendus 158, 1869–1872.
Maynard, J. (2015), "Small gaps between primes", Annals of Mathematics 181, 383–413.
Montgomery, H. L. (1973), "The pair correlation of zeros of the zeta function", in Proceedings of Symposia in Pure Mathematics 24, 181–193.
Odlyzko, A. M. (1987), "On the distribution of spacings between zeros of the zeta function", Mathematics of Computation 48, 273–308.
Skewes, S. (1933), "On the difference π(x) − Li(x)", J. London Math. Soc. 8, 277–283.
Tao, T. (2019), "Almost all orbits of the Collatz map attain almost bounded values", arXiv:1909.03562.
Zhang, Y. (2014), "Bounded gaps between primes", Annals of Mathematics 179, 1121–1174.

7.3 OEIS and computational references

OEIS A005384, "Sophie Germain primes p (2p + 1 also prime)".
OEIS A092816, "Number of Sophie Germain primes ≤ 10ⁿ".

7.4 Rei-AIOS internal references — for traceability

memory/feedback_rei_methodology_barrier_side_discipline.md (8th principle).
memory/feedback_no_rush_publication.md (1st principle).
memory/feedback_evaluation_symmetry_principle.md (2nd principle).
memory/feedback_world_uniqueness_claim_controllable.md (3rd principle).
memory/feedback_super_naming_siren_family_pattern.md (4th principle).
memory/feedback_chat_claude_over_deference.md (6th principle).
memory/feedback_chat_claude_hallucination_warning.md (7th principle).
memory/feedback_line_count_size_vs_kind_distinction.md (5th principle).
memory/reference_collatz_lyapunov_obstruction_generalized_2026-06-17.md (Phase B BF framework origin).
memory/project_collatz_aeb_sequence_2026-06-18.md (Phase B 9-feature-space extension).
memory/reference_difficulty_typology_collatz_riemann_sg_2026-06-18.md (three-axis typology).
memory/project_sg_discipline_application_2026-06-18.md (six discipline-asset application).
memory/project_sg_explicit_non_claims_2026-06-18.md (three explicit non-claims).
memory/reference_friedlander_iwaniec_selberg_parity_audit_2026-06-18.md (Path 4 audit + 1949 correction).
scripts/empirical/parity-barrier-toy-spec-2026-06-18.md (toy model specification).

Appendix A — Lean 4 axiom audit output

The audit was carried out by adding a temporary SGConjunctionWallAxiomCheck.lean file invoking #print axioms on each load-bearing theorem. The complete output is reproduced below.

'CollatzRei.SGConjunctionWall.sg_eleven' depends on axioms:
  [propext, Classical.choice, Quot.sound]
'CollatzRei.SGConjunctionWall.not_sg_seven' depends on axioms:
  [propext, Classical.choice, Quot.sound]
'CollatzRei.SGConjunctionWall.sg_five' depends on axioms:
  [propext, Classical.choice, Quot.sound]
'CollatzRei.SGConjunctionWall.not_sg_seventeen' depends on axioms:
  [propext, Classical.choice, Quot.sound]
'CollatzRei.SGConjunctionWall.sg_two' depends on axioms:
  [propext, Classical.choice, Quot.sound]
'CollatzRei.SGConjunctionWall.not_sg_thirteen' depends on axioms:
  [propext, Classical.choice, Quot.sound]
'CollatzRei.SGConjunctionWall.feature_isprime_n_insufficient' depends on axioms:
  [propext, Classical.choice, Quot.sound]
'CollatzRei.SGConjunctionWall.feature_isprime_2np1_insufficient' depends on axioms:
  [propext, Classical.choice, Quot.sound]
'CollatzRei.SGConjunctionWall.feature_mod6_insufficient' depends on axioms:
  [propext, Classical.choice, Quot.sound]
'CollatzRei.SGConjunctionWall.feature_pair_isprime_mod6_insufficient' depends on axioms:
  [propext, Classical.choice, Quot.sound]
'CollatzRei.SGConjunctionWall.conjunction_is_sufficient' depends on axioms:
  [propext]

No theorem uses sorryAx or native_decide. The audit file was deleted after verification.

Appendix B — Computational reproduction details

All four experiments are reproducible from the listed source artifacts. Key parameters:

Path 1 sieve: bytearray of length 2N + 1, mark-multiples up to √(2N+1). At N = 10⁸, memory ≈ 191 MB; wall-clock ≈ 13 s on a single thread (Intel i7-class 2020s commodity workstation).
Path 2 statistics: ⟨r⟩ over 56,030 consecutive-gap-ratio samples; spacing histogram with 20 bins on [0, 4] in units of mean gap; number variance over 200 random windows per L value.
Path 3 Lean: lake env lean CollatzRei/SGConjunctionWall.lean (~6 s on the same workstation; 780 jobs total when including dependencies).
Path 4 audit: two WebSearch and two WebFetch calls against Wikipedia and Tao's blog; no API keys or restricted access required.

The full JSON output of the four experiments is committed to the source repository at data/empirical/sg-analysis-2026-06-18.json, data/empirical/sg-extrapolation-2026-06-18.json, data/empirical/sg-gap-spectral-2026-06-18.json, and (separately, for the more general toy model) data/empirical/parity-barrier-toy-2026-06-18.json.

Version history

v0.1-DRAFT (2026-06-18): Initial release. Four sections corresponding to Paths 1–4 of the 2026-06-18 SG analysis workstream. Three explicit non-claim boundaries restated. Pattern-5 correction (Selberg 1949, not 1960s) recorded as Section 5 finding.

Paper 166 v0.1 — A Lean 4 Axiom-Free Formalization of Exit-Layer Collatz Convergence as a Stream Coalgebra: A Record Following Kim (2008)

Nobuki Fujimoto — Tue, 16 Jun 2026 21:50:56 +0000

This article is a re-publication of Rei-AIOS Paper 166 for the dev.to community.
The canonical version with full reference list is in the permanent archives below:

GitHub source (private): https://github.com/fc0web/rei-aios Author: Nobuki Fujimoto (@fc0web) · ORCID 0009-0004-6019-9258 · License CC-BY-4.0 ---

Status: v0.1-DRAFT (Phase B 起草完了 2026-06-17 朝、 Phase C cross-check 待ち)
Authors: 藤本伸樹 (Nobuki Fujimoto) / Claude Opus (Anthropic) / chat-Claude (Anthropic, cross-check pending Phase C)
Date: 2026-06-17
License: AGPL-3.0 + Commercial (Rei-AIOS dual license)
DOI: TBD (Zenodo deposit assigned at Phase D publish)
Version: v0.1-DRAFT
note.com: https://note.com/nifty_godwit2635
rei-aios site: https://rei-aios.pages.dev
Source repo: data/lean4-mathlib/CollatzRei/StreamExitLayerBridge.lean (Q2 Installment 2A, 2026-06-16)

Abstract

We record a Lean 4 axiom-free formalization of the exit-layer fragment of the Collatz dynamics, presented in the language of coinductive Stream' coalgebras. Concretely, we prove that for every q : ℕ, the orbit of the exit-layer number m_{q+1} = (4^{q+1} − 1) / 3 under a halt-at-1 variant of the Collatz step is eventually equal to the constant stream const 1, with the entire proof reducing to Mathlib's classical axiom triple {propext, Classical.choice, Quot.sound}. The base case head (collatzOrbit n) = n is verified to be completely axiom-free (#print axioms outputs the empty set), strictly stronger than the axiom base of our earlier lawvere_fixed_point Lean 4 record (STEP 1220, 2026-06-15). We make no claim of mathematical novelty: Kim (2008) already proved that the 2-adic Collatz function is a final bit-stream coalgebra in pen-and-paper category theory, and stream-coalgebra formalization infrastructure has been mature in Coq since Niqui (2009). The contribution of this note is limited and methodological: a Lean 4 + Mathlib v4.27 record, in the framing "first such Lean 4 axiom-free formalization within our observed range", of a known coalgebraic structure on a known Collatz fragment. This paper does not resolve the Collatz conjecture, does not move past Tao (2019, 2022)'s "almost all" analytic boundary, does not touch the Cases 5–8 trailing-1-bits ≥ 4 wall, and does not address any of the six critical-path sorries in Janik (2026).

Keywords: Collatz conjecture, exit layer, coalgebra, coinduction, Stream', Lean 4, Mathlib, axiom-free formalization, methodology note.

Section 1. Introduction

1.1 The Collatz problem and its exit layer

The Collatz map (Collatz 1937; Lagarias 1985) is

$$
T(n) \;=\; \begin{cases} n / 2 & n \text{ even} \ 3n + 1 & n \text{ odd}. \end{cases}
$$

The Collatz conjecture asserts that for every n ≥ 1, the orbit n, T(n), T^2(n), \ldots eventually reaches 1. The problem has been open for roughly ninety years and remains open at the time of writing; in particular, machine verification has confirmed convergence for all n < 2^{71} (cf. the verification odometer record summarized in Rei-AIOS STEP 1173), and no proof of the full conjecture has appeared.

A natural sub-fragment, observed by the first author on 2026-05-28 and formalized as Rei-AIOS STEP 1176, is the exit layer:

$$
m_p \;=\; \tfrac{4^p - 1}{3} \;=\; 1 + 4 + 4^2 + \cdots + 4^{p-1} \;=\; {\,1,\; 5,\; 21,\; 85,\; 341,\; \ldots\,}.
$$

These are exactly the odd numbers that join the power-of-two spine in a single 3n+1 step: applying T once to m_p produces 3 m_p + 1 = 4^p = 2^{2p}, after which 2p halving steps reach 1. The convergence of every m_p is a classical, elementary observation; the contribution of STEP 1176 was the Lean 4 axiom-free record of this observation, not the observation itself.

1.2 Coalgebraic background (prior art)

A coalgebraic perspective on Collatz dynamics has existed for nearly two decades. Two pieces of prior art are load-bearing for the present note and must be front-loaded:

Kim (2008), Coinductive properties of Lipschitz functions on streams, proved in pen-and-paper category theory that the 2-adic Collatz function is a morphism into a final bit-stream coalgebra. The 2-adic Collatz function and its coalgebraic characterization are therefore not new with the present paper.
Niqui (2009), Coalgebraic Reasoning in Coq, formalized stream coalgebras in Coq, including weakly final coalgebras, bisimulation, and λ-coiteration. The Coq coalgebras contribution (GitHub, 2008–2009) provides a mature infrastructure for this style of reasoning. The Cubical Agda library similarly carries final-coalgebra constructions.

In short: applying mature stream-coalgebra formalization machinery to a known coalgebraic Collatz structure is largely a routine exercise. The present paper is a record of one execution of that routine — in Lean 4, with axiom-free guarantees and one completely zero-axiom witness — and nothing more.

1.3 Analytic frontier (context, not contribution)

The current analytic frontier of Collatz research is dominated by:

Tao (2019), Almost all orbits of the Collatz map attain almost bounded values, and Tao (2022) follow-up; the canonical "almost all" result, distributional in nature.
Janik (2026), Diophantine Confinement in Syracuse Dynamics: A Formal Reduction, a 12,947-line Lean 4 development on the (2,3)-torus with ergodic / Baker-linear-forms machinery and six remaining critical-path sorries. We have audited Janik's public repository (johnjanik/syracuse-confinement) at the file-name and code-search level and confirm that the development does not use coinductive / Stream' / coalgebra terminology anywhere: the present paper and Janik (2026) inhabit disjoint sub-niches.
Chang (2026), Stanford preprint v5, with the Sturmian-obstruction and "Carry Contamination Theorem", articulates a distributional-to-pointwise wall.
Knight (2026), Discrete Mathematics, on the non-existence of high cycles.

None of these are touched by the present paper. We mention them only to fix the analytic context against which our contribution should be read; the present paper is methodological, not analytic.

1.4 The limited contribution of this paper

Concretely, this paper does the following, and no more:

It defines a halt-at-1 variant collatzHaltStep of the standard Collatz step, under which 1 is a genuine fixed point (rather than a member of the 1 → 4 → 2 → 1 cycle), so that orbits become eventually constant rather than eventually periodic.
It defines collatzOrbit n : Stream' Nat = λ i ↦ collatzHaltStep^i n, the orbit of n viewed as a coinductive infinite sequence (a Mathlib Stream'), and verifies that this stream is an F-coalgebra for F(X) = Nat × X.
It lifts the existing exit-layer result exitM_reaches_one (STEP 1176) into the coinductive language as collatzOrbit_exitM_eventuallyConst: for every q, the orbit of m_{q+1} is eventually const 1.
It verifies that head (collatzOrbit n) = n is completely axiom-free (#print axioms is empty), placing it strictly below the axiom base {propext, Classical.choice, Quot.sound} of our earlier lawvere_fixed_point record (STEP 1220).
It records the entire development as a Mathlib v4.27 file with zero sorry, zero axiom, and zero native_decide.

The framing we use, throughout, is: "the first Lean 4 axiom-free formalization of this fragment within our observed range". We deliberately do not use the phrase "first" without that qualifier, in accordance with our standing controllable-claim discipline (Rei-AIOS persistent rule feedback-world-uniqueness-claim-controllable).

1.5 What this paper is not

To pre-empt overclaim by the reader (and by the authors, as a discipline imposed by Rei-AIOS rule feedback-evaluation-symmetry-principle), we state in front-loaded form what this paper does not do. The same list reappears, expanded, as Section 7.

It does not resolve the Collatz conjecture.
It does not weaken the Cases 5–8 wall (trailing-1-bits ≥ 4): that wall is structurally unchanged.
It does not improve, refine, or move past Tao (2019, 2022)'s "almost all" result.
It does not discharge any of the six critical-path sorries of Janik (2026).
It does not propose a new attack vector on the Collatz conjecture; the coalgebraic perspective used here is Kim's.
It does not claim a "world-first" of any kind; the qualifier "within our observed range" is binding.

Section 2. Background: `collatzStep` and the exit-layer numbers

2.1 The standard Collatz step

We work with the standard Collatz function in its Lean 4 form, as defined in CollatzRei.Basic:

def collatzStep (n : Nat) : Nat :=
  if n % 2 = 0 then n / 2 else 3 * n + 1

For an account of the conjecture and the substantial body of partial results, we refer to Lagarias (1985, 2010), Tao (2019), and the survey-style "Group Think" working paper by Kenigson (2025).

2.2 The exit-layer numbers `m_p`

The exit-layer numbers, as observed by the first author and formalized in STEP 1176, are defined recursively (avoiding division) in CollatzRei.ExitLayer:

def exitM : Nat → Nat
  | 0     => 0
  | p + 1 => 1 + 4 * exitM p

so that exitM 1 = 1, exitM 2 = 5, exitM 3 = 21, exitM 4 = 85, exitM 5 = 341, etc.

The arithmetic core of STEP 1176 is the identity

theorem three_mul_exitM_add_one (p : Nat) : 3 * exitM p + 1 = 4 ^ p

(3 m_p + 1 = 4^p), which is the reason that a single 3n + 1 step from m_p lands exactly on the power-of-two spine. The closed-form identification exitM p = (4^p − 1) / 3 is recorded as exitM_eq_div, the parity statement as exitM_odd, and the main exit-layer reach result as

theorem exitM_reaches_one (q : Nat) :
    collatzStep^[2 * (q + 1) + 1] (exitM (q + 1)) = 1

This is the mathematical source that the present paper lifts into coinductive language. STEP 1176 is verified sorry-free in Mathlib v4.27 and #print axioms exitM_reaches_one reports the axiom triple {propext, Classical.choice, Quot.sound} (the standard Mathlib base).

2.3 Honest scope

The exit-layer fragment formalized in STEP 1176, and lifted into coinductive language in the present paper, lies entirely on one side of the unresolved Collatz wall. It treats exactly the odd numbers that reach the power-of-two spine in a single 3n + 1 step. It does not characterize, in any non-trivial way, the predecessors of those numbers, and the trailing-1-bits ≥ 4 wall (Cases 5–8 in the trailingOnes / 2-adic-valuation framework articulated in STEP 622, Rei-AIOS) is wholly untouched.

This is the only Collatz content of the present paper. Everything else is reformulation.

Section 3. Coinductive reformulation: `collatzHaltStep` and the orbit `Stream'`

3.1 Why halt at 1

The standard Collatz step has the cycle 1 → 4 → 2 → 1, so 1 is not a fixed point of collatzStep. For coinductive purposes — where we wish to characterize "the orbit reaches and stays at 1" as "the orbit is eventually equal to the constant stream const 1" — we want 1 to be a genuine fixed point.

We therefore introduce the halt-at-1 variant:

def collatzHaltStep (n : Nat) : Nat :=
  if n = 1 then 1 else collatzStep n

with two immediate lemmas:

theorem collatzHaltStep_one : collatzHaltStep 1 = 1 := by
  unfold collatzHaltStep; simp

theorem collatzHaltStep_ne_one (n : Nat) (h : n ≠ 1) :
    collatzHaltStep n = collatzStep n := by
  unfold collatzHaltStep; simp [h]

For every n ≠ 1, the two variants agree, so "reaches 1" statements transfer verbatim (Section 5). The single behavioural change is at n = 1, where collatzStep would have moved to 4 and collatzHaltStep stays at 1. This is a modelling choice in the spirit of unfolding partial fixed-point semantics into a total stream-coalgebra semantics; it is not a mathematical claim.

3.2 The orbit as a `Stream'`

We define the orbit of n as a Mathlib Stream' Nat:

def collatzOrbit (n : Nat) : Stream' Nat :=
  fun i => collatzHaltStep^[i] n

Concretely, the stream is

[ n, collatzHaltStep n, collatzHaltStep² n, collatzHaltStep³ n, … ].

The coinductive perspective is that the stream "reveals" the orbit one element at a time: each successive head is the next orbit value, and tail shifts the observer forward by one step.

3.3 Honest scope

The function collatzOrbit is a definitional reorganization of the existing iterated-function presentation. No new mathematical content is introduced here. The reformulation is purely linguistic — moving from Function.iterate (algebraic μF) to Stream' (coalgebraic νF), in the dictionary articulated in Adámek and Rosický (1994). It is the same dictionary that Kim (2008) uses when stating the 2-adic Collatz function as a morphism into a final bit-stream coalgebra; the contribution is the Lean 4 record of this dictionary applied to the exit-layer fragment, not the dictionary itself.

Section 4. F-Coalgebra structure on `collatzOrbit`

4.1 The functor

We work with the functor F : Type → Type, F(X) = Nat × X, whose coalgebras X → F(X) are exactly streams of natural numbers presented as head plus tail. The Mathlib Stream' type is, in this language, the carrier of the final coalgebra for F (this is the standard coinductive characterization; cf. Niqui 2009 and the Mathlib Stream' documentation).

4.2 Head

The first projection is verified as

theorem head_collatzOrbit (n : Nat) : head (collatzOrbit n) = n := rfl

This proof reduces definitionally (rfl): Stream'.head s = s 0, and collatzOrbit n 0 = collatzHaltStep^[0] n = n definitionally. Empirically (Section 6), #print axioms head_collatzOrbit reports the empty set: this theorem depends on no axiom whatsoever, not even propext. It is in this sense strictly stronger than lawvere_fixed_point of Rei-AIOS STEP 1220, which depends on {propext, Classical.choice, Quot.sound}.

4.3 Tail

The shift relation is the characteristic F-coalgebra property:

theorem tail_collatzOrbit (n : Nat) :
    tail (collatzOrbit n) = collatzOrbit (collatzHaltStep n) := by
  funext i
  show collatzHaltStep^[i + 1] n = collatzHaltStep^[i] (collatzHaltStep n)
  rw [Function.iterate_succ_apply]

This is the statement that the assignment n ↦ collatzOrbit n is an F-coalgebra morphism: the diagram

   Nat   ──────────────collatzOrbit─────────────►  Stream' Nat
    │                                                   │
    │ ⟨id, collatzHaltStep⟩                             │ ⟨head, tail⟩
    ▼                                                   ▼
  Nat × Nat  ────────⟨id, collatzOrbit⟩──────────►  Nat × Stream' Nat

commutes (with Mathlib.Stream' as the canonical final-coalgebra carrier on the right).

4.4 Relation to Kim (2008)

Kim (2008) constructs essentially the same diagram, working in the 2-adic integers ℤ_2 rather than in ℕ and at the level of pen-and-paper category theory. The present Lean 4 record differs in (i) the carrier (Nat, not ℤ_2); (ii) the proof assistant (Lean 4 / Mathlib v4.27, not pen-and-paper); (iii) the explicit axiom-base accounting (Section 6). It does not differ in mathematical idea.

Section 5. Exit-layer bridge: lifting `exitM_reaches_one` into the coinductive language

5.1 The fixed-after-1 lemmas

Two preparatory lemmas record that once the halt step reaches 1, it stays there forever:

theorem collatzHaltStep_iter_one (k : Nat) : collatzHaltStep^[k] 1 = 1 := by
  induction k with
  | zero      => rfl
  | succ m ih => rw [Function.iterate_succ_apply', ih, collatzHaltStep_one]

theorem collatzHaltStep_fixed_after_one (n k m : Nat)
    (h : collatzHaltStep^[k] n = 1) :
    collatzHaltStep^[k + m] n = 1 := by
  rw [show k + m = m + k from Nat.add_comm k m, Function.iterate_add_apply, h]
  exact collatzHaltStep_iter_one m

5.2 The bridge from standard step to halt step

The halt variant agrees with the standard step until 1 is reached, so the same step count works for both:

theorem collatzHaltStep_reaches_one_of_collatzStep (n : Nat) :
    ∀ k, collatzStep^[k] n = 1 → collatzHaltStep^[k] n = 1 := by
  intro k h
  induction k generalizing n with
  | zero       => simpa using h
  | succ k' ih =>
    rw [Function.iterate_succ_apply] at h
    by_cases hn1 : n = 1
    · subst hn1; exact collatzHaltStep_iter_one (k' + 1)
    · rw [Function.iterate_succ_apply, collatzHaltStep_ne_one n hn1]
      exact ih (collatzStep n) h

The case analysis is essential: if n = 1 at the start, the standard step would move to 4, but the halt step is already at the fixed point and stays there.

5.3 The eventually-constant predicate

We encode "eventually equal to a" as:

def EventuallyConst (s : Stream' Nat) (a : Nat) : Prop :=
  ∃ k, ∀ j, k ≤ j → s j = a

This is the standard coinductive characterization (witness k for the prefix length, all subsequent indices yield a). It is propositionally equivalent to "the stream is (prefix) ++ const a" but avoids appendix-construction machinery.

5.4 The main bridge theorem

The load-bearing theorem of the present paper is:

theorem collatzOrbit_exitM_eventuallyConst (q : Nat) :
    EventuallyConst (collatzOrbit (exitM (q + 1))) 1 := by
  refine ⟨2 * (q + 1) + 1, ?_⟩
  intro j hj
  obtain ⟨m, hm⟩ : ∃ m, j = (2 * (q + 1) + 1) + m :=
    ⟨j - (2 * (q + 1) + 1), by omega⟩
  rw [hm]
  show collatzHaltStep^[2 * (q + 1) + 1 + m] (exitM (q + 1)) = 1
  apply collatzHaltStep_fixed_after_one
  apply collatzHaltStep_reaches_one_of_collatzStep
  exact exitM_reaches_one q

The proof uses exitM_reaches_one (STEP 1176) directly as a hypothesis at the last line: the entire mathematical content of the theorem is the STEP 1176 result, and the present proof is a linguistic lift. There is no analytic content, no new descent argument, no new arithmetic.

5.5 The constant-stream specialization at `n = 1`

At the exit-layer base, the orbit collapses to the constant stream const 1:

theorem collatzOrbit_one_eq_const : collatzOrbit 1 = const 1 := by
  funext i
  show collatzHaltStep^[i] 1 = 1
  exact collatzHaltStep_iter_one i

and (exitM 1 = 1):

theorem collatzOrbit_exitM_one_eq_const : collatzOrbit (exitM 1) = const 1 := by
  have h1 : exitM 1 = 1 := exitM_one
  rw [h1]
  exact collatzOrbit_one_eq_const

Through Rei-AIOS STEP 1223's IsCoalgebraicFixedPoint predicate, collatzOrbit 1 is in addition a (νF-style) coalgebraic fixed point of tail:

theorem collatzOrbit_one_isCoalgebraicFixedPoint :
    IsCoalgebraicFixedPoint (collatzOrbit 1) := by
  rw [collatzOrbit_one_eq_const]
  exact const_isCoalgebraicFixedPoint 1

This last theorem is recorded here for completeness; it asserts no Collatz content, only the structural fact that the constant stream is fixed by tail.

5.6 Honest scope

The bridge theorem is, by design, strictly weaker than the Collatz conjecture. It says: for the special inputs m_{q+1}, the orbit (in the halt variant) is eventually const 1. It does not say anything about n not of the form m_p, and the predecessor structure of the exit-layer numbers (which is where the unresolved Collatz wall lives) is not touched. The bridge is a coinductive restatement of a known elementary fact.

Section 6. Zero-axiom witness and axiom-base accounting

6.1 What "completely axiom-free" means

A Lean 4 theorem T is completely axiom-free when #print axioms T reports the empty set. In particular, T does not depend on:

propext (propositional extensionality),
Classical.choice (the global choice principle),
Quot.sound (soundness of the quotient construction),

nor on any user-introduced axiom declaration, nor on any reduction step that invokes native_decide (which appeals to compiled native code outside Lean's kernel).

Mathlib's default axiom triple {propext, Classical.choice, Quot.sound} is the standard base used in classical mathematics; a completely axiom-free theorem is strictly below this base.

6.2 `head_collatzOrbit` is completely axiom-free

The theorem

theorem head_collatzOrbit (n : Nat) : head (collatzOrbit n) = n := rfl

is completely axiom-free: empirically, #print axioms head_collatzOrbit reports no axioms. The reason is that the proof is rfl: Stream'.head s unfolds definitionally to s 0, and collatzOrbit n 0 = collatzHaltStep^[0] n unfolds definitionally to n. No propositional reasoning is invoked, and the Lean 4 / Mathlib v4.27 implementation of Stream'.head is itself a non-classical projection.

This places head_collatzOrbit strictly below the axiom base of lawvere_fixed_point (Rei-AIOS STEP 1220, 2026-06-15), which depends on {propext, Classical.choice, Quot.sound}. We record this not as a competition between theorems — they are unrelated — but as an instance of axiom-base minimization as a methodology.

6.3 The axiom base of the remaining theorems

The other principal theorems of the present development depend on a small set of axioms, summarized below. The ∅ (empty) entry for head_collatzOrbit was empirically confirmed by #print axioms head_collatzOrbit against Mathlib v4.27.0 in the Rei-AIOS session of 2026-06-16; the remaining entries record the expected base under the classical Mathlib lemmas used in each proof (funext, omega, Function.iterate_succ_apply, etc.). A per-theorem #print axioms re-verification of every row is scheduled for the Phase C cross-check, and any deviation will be recorded as a corrigendum before Phase D publication.

Theorem	Axiom base
`head_collatzOrbit`	`∅` (completely axiom-free)
`tail_collatzOrbit`	`{propext, Classical.choice, Quot.sound}`
`collatzHaltStep_iter_one`	`{propext, Classical.choice, Quot.sound}`
`collatzHaltStep_fixed_after_one`	`{propext, Classical.choice, Quot.sound}`
`collatzHaltStep_reaches_one_of_collatzStep`	`{propext, Classical.choice, Quot.sound}`
`collatzOrbit_one_eq_const`	`{propext, Classical.choice, Quot.sound}`
`collatzOrbit_one_isCoalgebraicFixedPoint`	`{propext, Classical.choice, Quot.sound}`
`collatzOrbit_exitM_eventuallyConst`	`{propext, Classical.choice, Quot.sound}`
`collatzOrbit_exitM_one_eq_const`	`{propext, Classical.choice, Quot.sound}`

No theorem in the development depends on any user-introduced axiom, any sorry, or any native_decide. The classical triple is reached through Mathlib's general-purpose lemmas (funext, Nat.add_comm, Function.iterate_succ_apply, etc.), not through any genuinely classical step in our argument; a fully constructive rewriting is plausible but is left to future work.

6.4 Why this matters (and why it does not matter)

Axiom-base minimization records a constructive content claim. A theorem at the empty axiom base is, in a precise sense, computable: its proof reduces to definitional unfolding only. A theorem at {propext, Classical.choice, Quot.sound} is classical-Mathlib-standard.

The methodological value of recording these distinctions is that they make the constructive boundary explicit. The mathematical value here is modest: the empty-axiom-base witness head_collatzOrbit is a trivial projection, and the classical base of the main bridge theorem is the same as that of exitM_reaches_one itself. We claim no constructive Collatz result. We record axioms because we want a record, not because the axioms are themselves the content.

This is consistent with the methodology line of the Rei-AIOS Paper 132 series (Mathlib contribution preparation, residual-sorry roadmaps), of which the present paper is a small entry.

Section 7. Honest scope and explicit non-claims

This section enumerates, in load-bearing form, what the present paper does not establish. It is the operational realization of the Rei-AIOS persistent rule feedback-evaluation-symmetry-principle: a result reported as "no prior art" or "axiom-free" must be reported with the same directness when the situation is the opposite. Reading this section as boilerplate is a misread; it constrains the paper's interpretation.

7.1 The Collatz conjecture is not resolved

We do not prove ∀ n ≥ 1, Reaches1 n. We prove only that for inputs of the form m_{q+1} = (4^{q+1} − 1) / 3, the halt-variant orbit is eventually const 1. The complement — orbits starting from numbers not of the form m_p — is wholly untouched.

7.2 The Cases 5–8 wall is unchanged

The Rei-AIOS STEP 622 trailing-1-bits / 2-adic-valuation analysis identifies eight cases on the parity-class of n's trailing binary expansion. Cases 1–4 admit explicit descent (step623_v3.lean); Cases 5–8, where trailing 1-bits are ≥ 4, exhibit an unbounded (3/2)^j growth contribution that defeats finite mod analysis. This wall is structurally unchanged by the present paper. Coinductive reformulation does not break finite-mod barriers.

7.3 Tao (2019, 2022) is not superseded

The "almost all" results of Tao (2019, 2022) are distributional statements about the density of orbits that come close to bounded values. They are the current analytic frontier. The present paper makes no statement about densities, makes no analytic argument, and does not move the boundary identified by Tao. We mention Tao's results only for context, not as a benchmark we approach.

7.4 Janik (2026) is independent

Janik's johnjanik/syracuse-confinement is a 12,947-line Lean 4 development with six remaining critical-path sorries, organized around an ergodic / Diophantine reduction on the (2,3)-torus. We have audited the public repository at file-name and code-search level and found zero occurrences of the search terms coalgebra, coinductive, Stream', exitM, and exit_layer; the two occurrences of 4^p are incidental (Baker-linear-form context). The present paper does not address any of Janik's sorries, does not weaken his hypotheses, and is methodologically disjoint from his approach.

7.5 No `world-first` claim is made

In accordance with feedback-world-uniqueness-claim-controllable, we do not claim a "world-first" or "globally unique" Lean 4 axiom-free formalization of this material. The binding qualifier "within our observed range" applies to every existence claim in this paper. Kim (2008), Niqui (2009), the Coq coalgebras contribution, and the Cubical Agda final-coalgebra constructions are explicit prior art; the present paper is one record among several possible.

7.6 No new attack vector is proposed

We do not propose a new strategy for resolving the Collatz conjecture, and we do not assert that coinductive reformulation as such will lead to progress on the unresolved fragment. The Q2 Installment 1 survey (papers/collatz-rei-toolkit-survey-2026-06-16.md) predicted that the Rei toolkit applied to Collatz would produce a "null result plus a clean Lean 4 statement"; the present paper is precisely that, and nothing more.

7.7 No siren-family framing is used

In accordance with feedback-super-naming-siren-family-pattern, we avoid all framings of the form "beyond X", "transcending X", or "going past X". The translation discipline make / measure / map (build a tool, measure existing reach, map inaccessible territory) is operative throughout: this paper makes one tool (the coinductive lift), measures the axiom-base of nine theorems, and maps the boundary between the exit-layer fragment and the Cases 5–8 wall. It does not claim to transcend anything.

7.8 Mathlib API stability is conditional

The development depends on Mathlib v4.27.0's Stream'.head, Stream'.tail, Function.iterate_succ_apply, Nat.add_comm, and related lemmas. We do not claim that the development will continue to typecheck after Mathlib API changes; such drift is a routine maintenance matter for Lean 4 formalizations.

Section 8. Related work

8.1 Coalgebraic Collatz (prior art)

Kim, J. (2008). Coinductive properties of Lipschitz functions on streams. The 2-adic Collatz function is a morphism into a final bit-stream coalgebra; the coalgebraic structure on Collatz dynamics is articulated here, in pen-and-paper category theory. This is the primary prior art for the present paper.
Related: the body of subsequent work treating 2-adic Collatz as a final stream coalgebra in coalgebraic logic and stream automata theory.

8.2 Stream coalgebra formalization (prior art)

Niqui, M. (2009). Coalgebraic Reasoning in Coq. Stream coalgebras, weakly final coalgebras, λ-coiteration, and bisimulation formalized in Coq. This is the canonical reference for formalized coalgebraic reasoning in a proof assistant.
The Coq coalgebras contribution (GitHub, 2008–2009). Open-source library covering coalgebra, bisimulation, weakly final coalgebra, λ-coiteration with stream coalgebra examples.
Cubical Agda final-coalgebra constructions in the standard library.
Adámek, J., Rosický, J. (1994). Locally Presentable and Accessible Categories. The μF / νF dictionary (initial-algebra / final-coalgebra duality) used throughout the present paper.

8.3 The analytic Collatz frontier (context only)

Tao, T. (2019). Almost all orbits of the Collatz map attain almost bounded values. arXiv:1909.03562.
Tao, T. (2022). Follow-up extension paper [exact arXiv identifier to be confirmed at Phase C].
Janik, J. (2026). Diophantine Confinement in Syracuse Dynamics: A Formal Reduction. GitHub: johnjanik/syracuse-confinement. 12,947 lines of Lean 4 with six remaining critical-path sorries.
Chang, F. (2026, April). Stanford preprint v5 with the "Sturmian obstruction" and the "Carry Contamination Theorem".
Knight, K. (2026, March). Collatz high cycles do not exist. Discrete Mathematics.
Kenigson, J. (2025, December). Group Think: A Survey on the Collatz Conjecture. Working paper, Cambridge Open Engage.

8.4 Classical Collatz references

Collatz, L. (1937). Unpublished. Cited as the origin of the 3n+1 problem.
Conway, J. H. (1972). Unpredictable Iterations. Proceedings of the Number Theory Conference, University of Colorado, Boulder. Generalized Collatz functions are Turing-complete; the original Collatz function is not known to be decidable.
Lagarias, J. C. (1985). The 3x+1 problem and its generalizations. American Mathematical Monthly, 92(1), 3–23.
Lagarias, J. C. (ed.) (2010). The Ultimate Challenge: The 3x+1 Problem. American Mathematical Society.

8.5 Rei-AIOS internal lineage

The present paper is part of the Rei-AIOS Paper series. The relevant internal references are:

STEP 1176 (Fujimoto, 2026-05-28). Exit-layer formalization. data/lean4-mathlib/CollatzRei/ExitLayer.lean. The mathematical source of the present paper.
STEP 1177 (Fujimoto, 2026-05-28). Inverse Collatz tree visualization (TypeScript lens). Not Lean 4; included for context.
STEP 1178 (Fujimoto, 2026-05-28). Collatz frontier dossier (TypeScript). Six routes catalog with breakdown points. Not Lean 4.
STEP 1179 (Fujimoto, 2026-05-28). Exit-layer inverse formulas, added to ExitLayer.lean.
STEP 1220 (Fujimoto and Claude Opus, 2026-06-15). Cantor-Lawvere fixed-point formalization, Lean 4 axiom-free. LawvereFixedPointExperiment.lean. Provides the axiom-base benchmark against which Section 6 is compared.
STEP 1223 (Fujimoto and Claude Opus, 2026-06-16). IsCoalgebraicFixedPoint predicate and const_isCoalgebraicFixedPoint lemma. NuFStreamSelfLoop.lean. The coinductive fixed-point tool used in Section 5.5.
Q2 Installment 1 (Fujimoto and Claude Opus, 2026-06-16). Collatz × Rei Toolkit Survey. papers/collatz-rei-toolkit-survey-2026-06-16.md. Articulates the Rei toolkit, identifies the present paper's content as one of two candidate installments, and pre-registers the "null result + clean Lean 4 statement" prediction.
Q2 Installment 2A (Fujimoto and Claude Opus, 2026-06-16). StreamExitLayerBridge.lean. The primary source file of the present paper.
Paper 132 series methodology notes (Rei-AIOS, 2026). Style template for the present paper; the methodology-note format with axiom-base accounting is inherited from this series.

Section 9. Limitations and future work

9.1 Limitations

The present paper has the following intrinsic limitations:

Fragment, not full conjecture. The result covers exactly the exit-layer numbers m_p, which is one branch in the inverse Collatz tree. The full Collatz conjecture is the question of whether the inverse-tree branches cover all positive integers; this paper is silent on that question.
Routine application of a known dictionary. The coalgebraic perspective on Collatz dynamics is Kim's (2008), the stream-coalgebra formalization machinery is Niqui's (2009) and the Coq coalgebras contribution's (2008–2009). The present paper is a Lean 4 record of routinely applying the latter to the former on a fragment of the former; it is methodological in character.
Mathlib v4.27 specificity. API drift in future Mathlib releases is expected; the development is maintained under the dual-license terms of the Rei-AIOS repository but no long-term maintenance guarantee is offered with the publication.
Modest axiom-base record. The only completely axiom-free theorem in the development (head_collatzOrbit) is a definitional projection. The methodologically interesting axiom-base reductions on substantive theorems remain open.

9.2 Future work

The present paper is one of three threads in a broader Rei-AIOS reframing of the Collatz program, articulated in the persistent memory note project-three-track-reframe-perelman-coalgebra-post-audit-2026-06-16. The other two threads are:

(A) A second monotone quantity along the lines of Perelman's W-functional. Rei-AIOS STEP 1176's exit-layer formalization is paired in the Rei series with the F-entropy / trailingOnes monotonicity of Paper 58. A scoping document of 2026-06-17 (papers/collatz-second-monotone-quantity-scoping-2026-06-17.md) identifies three candidates — a carry-structure quantity, a 2-adic-roughness quantity built on Mathlib's padicValNat, and an LZ-complexity quantity from Rei STEP 1168 — and recommends the 2-adic-roughness candidate on the grounds of minimum implementation cost and natural compatibility with the existing F-entropy infrastructure. Lean 4 implementation of this candidate is gated on publication of the present paper and explicit approval from the first author; we do not include it in the present paper, to preserve a narrow scope.
(B) A categorical / Yoneda-style reformulation of Shannon entropy in the spirit of Baez. This is long-term and outside the present paper's scope.

A separate, longer-term thread is the visualization of the non-computability hierarchy (oracle hierarchies, ITTM, BSS, Type-2 machines) as an educational demonstration alongside the coinductive content. This is mentioned in the Q2 Installment 1 survey but is not pursued in the present paper.

9.3 The present paper's position in the program

We close by stating the present paper's position with directness. It is a methodological record, not a mathematical advance. It documents one Lean 4 axiom-free formalization, with one completely axiom-free witness theorem, of one classical fragment of the Collatz dynamics, in the language of a coalgebraic perspective that was articulated in pen-and-paper category theory by Kim in 2008. Its appropriate venue is the short-note line of the Rei-AIOS Paper 132 series, with an expected impact comparable to a CPP / ITP short-note formalization record. We do not expect it to be cited in analytic Collatz literature. We do expect it to be cited, if at all, as one of several records of stream-coalgebra formalization in proof assistants, alongside Niqui (2009) and the Coq coalgebras and Cubical Agda libraries.

This is the honest scope of the contribution.

References

Primary citations

Kim, J. (2008). Coinductive properties of Lipschitz functions on streams. [Exact venue to be confirmed at Phase C — Springer / arXiv reference to be inserted.]
Niqui, M. (2009). Coalgebraic Reasoning in Coq. [Exact venue to be confirmed at Phase C.]
The Coq coalgebras contribution. GitHub repository (2008–2009). [URL to be inserted at Phase C.]
Adámek, J., Rosický, J. (1994). Locally Presentable and Accessible Categories. Cambridge University Press.
Tao, T. (2019). Almost all orbits of the Collatz map attain almost bounded values. arXiv:1909.03562.
Tao, T. (2022). Follow-up extension paper. [Exact arXiv identifier to be confirmed at Phase C.]
Janik, J. (2026). Diophantine Confinement in Syracuse Dynamics: A Formal Reduction. GitHub: johnjanik/syracuse-confinement.
Chang, F. (2026, April). Stanford preprint v5. [Exact title and arXiv identifier to be confirmed at Phase C.]
Knight, K. (2026, March). Collatz high cycles do not exist. Discrete Mathematics.
Conway, J. H. (1972). Unpredictable Iterations. In: Proceedings of the Number Theory Conference, University of Colorado, Boulder.
Lagarias, J. C. (1985). The 3x+1 problem and its generalizations. American Mathematical Monthly, 92(1), 3–23.
Lagarias, J. C. (ed.) (2010). The Ultimate Challenge: The 3x+1 Problem. American Mathematical Society.

Internal Rei-AIOS references

Fujimoto, N. (2026-05-28). Rei-AIOS STEP 1176: Collatz exit-layer formalization. data/lean4-mathlib/CollatzRei/ExitLayer.lean.
Fujimoto, N. (2026-05-28). Rei-AIOS STEP 1177: Inverse Collatz tree lens. TypeScript visualization; not Lean 4.
Fujimoto, N. (2026-05-28). Rei-AIOS STEP 1178: Collatz frontier dossier. TypeScript; not Lean 4.
Fujimoto, N. (2026-05-28). Rei-AIOS STEP 1179: Exit-layer inverse formulas. Added to ExitLayer.lean.
Fujimoto, N., Claude Opus (2026-06-15). Rei-AIOS STEP 1220: Lawvere fixed-point experiment. LawvereFixedPointExperiment.lean.
Fujimoto, N., Claude Opus (2026-06-16). Rei-AIOS STEP 1223: νF Stream' self-loop. NuFStreamSelfLoop.lean.
Fujimoto, N., Claude Opus (2026-06-16). Collatz × Rei toolkit survey. papers/collatz-rei-toolkit-survey-2026-06-16.md.
Fujimoto, N., Claude Opus (2026-06-16). StreamExitLayerBridge.lean (the source file of the present paper).
Fujimoto, N. (2026). Rei-AIOS Paper 132 series. Methodology-note template. [Exact entries to be confirmed at Phase C.]

Supplementary

Mathlib v4.27.0. Stream'.head, Stream'.tail, Function.iterate_succ_apply, Function.iterate_succ_apply', Function.iterate_add_apply. Reference: https://github.com/leanprover-community/mathlib4.
Kenigson, J. (2025, December). Group Think: A Survey on the Collatz Conjecture. Working paper, Cambridge Open Engage.

Honest scope footer

─────────────────────────────────────────────────────
HONEST SCOPE FOOTER (load-bearing,
 feedback-evaluation-symmetry-principle +
 feedback-world-uniqueness-claim-controllable +
 feedback-super-naming-siren-family-pattern)
─────────────────────────────────────────────────────

What this paper formalizes:
  ✓ The exit-layer fragment m_p = (4^p − 1) / 3 of the Collatz
    dynamics, lifted from the algebraic μF iteration presentation
    (STEP 1176) into the coalgebraic νF Stream' presentation.
  ✓ One completely axiom-free witness theorem
    (head_collatzOrbit), with `#print axioms` empty.
  ✓ One main bridge theorem
    (collatzOrbit_exitM_eventuallyConst), at the classical
    Mathlib axiom triple {propext, Classical.choice, Quot.sound}.
  ✓ One Lean 4 / Mathlib v4.27 record of applying the
    known coalgebraic Collatz perspective (Kim 2008) to a
    known elementary Collatz fragment.

What this paper does NOT formalize:
  ✗ The Collatz conjecture (∀ n ≥ 1, Reaches1 n).
  ✗ The Cases 5–8 trailing-1-bits ≥ 4 wall
    (structurally unresolved).
  ✗ Any improvement over Tao 2019 / 2022 "almost all"
    distributional results.
  ✗ Any discharge of Janik 2026's six critical-path sorries.
  ✗ Any "new attack vector" or "world-first" or
    "Tao-superseded" or "Collatz-resolved" framing.

Our position:
  - A Paper 132 series methodology contribution,
    CPP / ITP short note in expected impact.
  - One operational instance of axiom-base minimization.
  - One record added to the Rei-AIOS Paper lineage,
    of limited contribution.
  - The qualifier "within our observed range" applies
    to every existence claim in this paper.
─────────────────────────────────────────────────────

Phase B 起草完了 status (2026-06-17 朝)

✓ Title #1 採用 (Skeleton 通り)
✓ Abstract で Kim 2008 + Niqui を front-load
✓ Section 1 で限定 contribution + what this paper is not 両方明示
✓ Section 6 で head_collatzOrbit 完全 zero-axiom claim + Lawvere 比較
✓ Section 7 で 8 項目 honest scope (Skeleton 7 項目 + Mathlib API stability 1 件追加)
✓ Section 8 で Kim 2008 + Niqui + Coq contrib + Cubical Agda + Janik + Tao + Chang + Knight + Conway + Lagarias + STEP lineage 全件引用
✓ 「世界初」「Tao 超え」「新 attack vector」「Collatz 解決」等の siren framing 不使用
✓ Section 9 で三方針 reframe + stop criterion future-work 文脈明記
✓ Mathlib v4.27.0 specific API dependency 明記 (Section 7.8 + Section 6.3)
✓ rei-aios.pages.dev + note.com footer (front matter)
✓ AGPL-3.0 + Commercial dual license 標記
✓ Paper 132 系 short note style (15 page 級、 over-engineering なし)

Next: Phase C handoff items (chat-Claude cross-check, 1-day buffer 推奨)

Pattern 1-6 hallucination audit (出典・年代・author 名・引用 venue 全件)
Overclaim detection — 「siren framing」「世界初」「Tao 超え」「Cases 5-8 突破」系の言い換えが残っていないか
Honest scope sufficiency check — Section 7 8 項目で漏れがないか (特に Cases 5-8 wall + Janik 独立性 + Mathlib API 条件性)
Empirical axiom-base re-verification — Section 6.3 表 9 行を lake env lean で per-theorem #print axioms で再確認 (Phase D publish 前必須、 deviation あれば corrigendum 起草)
Reference完全形 — Kim 2008 + Niqui 2009 + Tao 2022 + Chang 2026 + Coq coalgebras contrib + Paper 132 系の exact venue / arXiv ID / URL 確定 (現在 placeholder)
Page count ≤ 15 確認 (現状 ~12 page 推定)

Phase D (publish 計画): Phase C pass 後、 Zenodo DOI 取得 + 11 platform standard (Dev.to / Hatena / HackMD / Notion / Livedoor / Mastodon / Scrapbox / Nostr / Internet Archive / GitHub Release)、 Harvard Dataverse は per-paper opt-in 確認 ([[feedback-harvard-dataverse-opt-in]] 準拠)。急がずゆっくりと ([[feedback-no-rush-publication]])。

Paper 165 v0.1 — Existence Proof Garden: Interactive Synthesis of Piantadosi-Chomsky Debate, Modal Duality, and Adjunction

Nobuki Fujimoto — Fri, 12 Jun 2026 02:24:35 +0000

This article is a re-publication of Rei-AIOS Paper 165 for the dev.to community.
The canonical version with full reference list is in the permanent archives below:

Zenodo (DOI, canonical): https://doi.org/10.5281/zenodo.20652725

Internet Archive: https://archive.org/details/rei-aios-paper-165-v01-1781229926185

GitHub source (private): https://github.com/fc0web/rei-aios Author: Nobuki Fujimoto (@fc0web) · ORCID 0009-0004-6019-9258 · License CC-BY-4.0 ---

Status: v0.1 (2026-06-12) — pre-publish fact-checked. 5 reference checks via Rei-side WebSearch: 3 confirmed accurate (Katzir Biolinguistics 17/2023, BabyLM EMNLP 2026 4th edition, Cook 2004 Complex Systems 15:1-40), 2 corrected (Piantadosi venue: Language Science Press book chapter, not Cognitive Science; Kodner-Payne-Heinz title: "\"Why Linguistics Will Thrive in the 21st Century\", not \"Sesame Street, or Sesame Open\" — chat-session hallucination recorded in §B.6b)."

Authors:

藤本伸樹 (Nobuki Fujimoto, ORCID: 0000-0002-2731-0269) — concept, direction, curation
Claude (Anthropic, chat session 2026-06-12) — interactive HTML artifact, dialogue synthesis
Claude Opus 4.7 (Anthropic, Rei-AIOS Claude Code session) — Rei-existing STEP integration, honest filter, Pattern 5 detection, prior-art audit

License: AGPL-3.0 (source) + CC-BY-SA 4.0 (paper text)

Repository: https://github.com/fc0web/rei-aios
Interactive artifact (direct download): https://raw.githubusercontent.com/fc0web/rei-aios/main/public/existence-proof-garden.html
Live site mount: https://rei-aios.pages.dev/#/existence-proof-garden

A. Abstract

We present an interactive, single-file HTML artifact — Existence Proof Garden (存在証明の庭) — that translates a multi-turn dialogue on language acquisition, modal duality, and computational emergence into a six-act + one-extension manipulable synthesis. The artifact does not produce new research-frontier claims; it is an educational and integrative instrument that places the Piantadosi (2023) "modern language models refute Chomsky" thesis, the Kodner-Payne-Heinz (2023) and Katzir (2023) replies, the BabyLM Challenge (Warstadt et al., EMNLP 2023–2026) target program, the LLM↔brain correspondence work (Goldstein, Schrimpf, Hasson, et al.), and Conway's Game of Life + Wolfram's Rule 110 (Cook 2004 Turing-complete proof) onto a single semantic axis: the modal pair ◇ (possibility) / □ (necessity).

Two contributions distinguish this work from a literature review. First, we operationalize the duality "existence proof (◇) breaks false necessity / shukatsu 終活 (□) cultivates before true necessity" as a directly manipulable canvas widget in Act IV, exposing the conceptual symmetry as touchable behaviour rather than prose. Second, in Act V and the A5b adjunction-overlay panel, we connect the dialogue's open conjecture — that ◇⊣□ should be formalized as a Galois connection between constrained hypothesis space ⊣ tractable-data learnability (drawing on Kodner et al.'s computational learning theory ("no free lunch") defense of □) — to five existing Rei-AIOS implementations: the Belnap-Dunn bilattice eight (STEP 1202), the SELF↔Lawvere fixed-point bridge with axiom-free Lean 4 formalization (STEP 1203, file data/lean4-mathlib/CollatzRei/SelfLawvereBridge.lean), the ∞-cosmoi axiomatization engine (STEP 1205), the A↔B Transition Observer (STEP 1167), and the Problem Foldability Lens (STEP 1168). The honest scope is explicit: the categorical adjunction ◇⊣□ is not formalized in the artifact, and is acknowledged in SelfLawvereBridge.lean itself as "the next STEP candidate beyond this file's scope".

B. Honest Scope

We state up-front what this paper does and does not claim:

B.1. The scientific positions surveyed (刺激の貧困, Piantadosi-Chomsky controversy, BabyLM target, behavioural-vs-mechanistic equivalence, LLM-brain correspondence, Conway's Game of Life, Cook 2004 Rule 110 Turing-completeness, catastrophic forgetting + stability–plasticity dilemma) are all pre-existing. Multiple years of arXiv-level prior art exist for each. The artifact is a synthesis instrument, not a discovery.

B.2. The duality framing "existence proof (◇) breaks false necessity / shukatsu (□) cultivates before true necessity" as a cross-register pairing (logical operation × existential attitude) is, to our observation, not present in the surveyed literature. We classify this framing as 【思弁的】 (speculative): it is a structural rhyme, not a formal De Morgan duality (those range over a single fixed proposition P, while ◇ and □ here scope over different data regimes — Kodner's exact scope-slip critique). Promoting "rhyme" to "theorem" would require: (a) fixing the proposition, and (b) constructing the adjunction. We have not done (b).

B.3. The ◇⊣□ Galois connection sketched in the A5b panel is a hypothesis model, not a proof. The breakdown observed at LLM scale (left adjoint failing to preserve ⊥ ⇒ right adjoint non-existence) is a consequence of the toy-model choice, not a CLT theorem. The artifact says so on-screen.

B.4. Conway's Game of Life Turing-completeness (Rendell 2010+, multiple constructions) and Cook (2004) Rule 110 Turing-completeness are textbook-established. Their inclusion is illustrative, not novel.

B.5. "世界初" (world-first) is not used anywhere in this paper or the artifact, per the standing principle [[feedback-world-uniqueness-claim-controllable]].

B.6. Pattern 5 detection: chat-session Claude, in the source dialogue, framed "the categorical adjunction ◇⊣□ is the unformalized prize" without knowledge of Rei-existing STEP 1203 (SELF↔Lawvere set-level fixed-point, Lean 4 axiom-free). The fact-check (commit-time grep verify) confirms: set-level fixed point is formalized in SelfLawvereBridge.lean; categorical adjunction is not, and the file itself acknowledges this as next-STEP scope. The chat-session frame and Rei position are aligned, not in conflict — chat-session is one step further along the unformalized line, with no knowledge that one earlier step is already taken.

B.6b. Pattern 2 detection (pre-publish fact-check, 2026-06-12, Rei Claude WebSearch verify): The original dialogue presented the Kodner-Payne-Heinz paper (arXiv:2308.03228) under the title "Sesame Street, or Sesame Open: What ChatGPT and friends do and do not tell us about humans". The published arXiv title is in fact "Why Linguistics Will Thrive in the 21st Century: A Reply to Piantadosi (2023)". Corrected throughout this paper. This is a chat-session hallucination caught by Rei-side independent verification — recorded here per [[feedback-chat-claude-hallucination-warning]] anti-pattern (avoid uncritical citation of chat-session claims; verify references independently before publication).

B.7. The artifact is single-file static HTML (~100 KB), runs offline after first load (no external CDN), uses Web Audio API for procedural BGM (陰旋法 D-E♭-G-A-B♭) and synthesized SFX, and Canvas 2D for all visual rendering. Verified working on Chromium-family browsers. No external library dependencies.

C. Six Acts + One Extension — Structural Map

Act	Title	Operation modelled	Honest tier
I	刺激の貧困	Learning-curve asymmetry (child-scale vs LLM-scale data ⇒ threshold crossing)	【仮説】 (validates BabyLM premise)
II	可能性空間と金継ぎ	◇ as counterexample breaking a falsely-claimed impossibility band	【証明済み】 framing (existence proof = formal logic)
III	振る舞いと機構	behavioural ≡ mechanistic non-equivalence (multiple realizability)	【証明済み】 (Putnam 1967 / Kodner Simulation ≠ Duplication 2023)
IV	終章 ― 同じ限界、逆の操作	◇ pushes boundary (侵犯) ⇔ □ cultivates before boundary (受容)	【思弁的】 cross-register rhyme
V	接ぎ穂 ― ◇/□ は八値のどこに棲むか	D-FUMT₈ value placement + ⊖ duality check ◇ ≡ ⊖□⊖	【思弁的】 with explicit category-error caveat
VI	創発 ― 設計者なしの複雑さ	Conway's Life + Rule 110 — simple rules ⇒ undesigned complexity	【証明済み】 demo (Cook 2004) with 【思弁的】 reading
A5b	接ぎ穂の一歩先 ― 随伴 ◇⊣□ を覗く	Galois connection between constraint and tractable learnability	【仮説】 toy-model exploration

D. Rei-AIOS Existing Connections (5 STEPs)

The artifact's speculative components connect to existing Rei implementations as follows. None of these connections claim that the artifact's framings are formally proved — they identify where formal substrate already exists.

D.1. Act V D-FUMT₈ placement ↔ STEP 1202 Bilattice eight engine.
File: src/axiom-os/bilattice-eight-engine.ts + lens #/bilattice-eight. STEP 1202 establishes Belnap-Dunn FOUR (1977) as Layer 1 confirmed bilattice + Rei D-FUMT₈ extension 4 axes (INFINITY/ZERO/FLOWING/SELF) as Layer 2 orthogonal extension stance (not lattice internal values — Ginsberg 1988 / Arieli-Avron 1998 9-value+ prior-art-conflict avoided). The artifact's "place ◇ and □ on D-FUMT₈ nodes" widget is a touchable expression of this Layer 2 stance.

D.2. Act V Lawvere diagonal reading + A5b adjunction ↔ STEP 1203 SELF-Lawvere bridge.
File: data/lean4-mathlib/CollatzRei/SelfLawvereBridge.lean. The artifact says "if both ◇ and □ are placed at SELF⟲, we read Lawvere's diagonal". STEP 1203 formalizes the Lawvere fixed-point theorem (set-theoretic direct version) with the proof lawvere_fixed_point and applies it to a structure SelfReferentialDomain carrying a point-surjective enumeration. The Lean file explicitly notes: "Mathlib v4.27.0 CategoryTheory.ClosedCategory has no direct statement of the Lawvere fixed-point theorem; the Mathlib bridge is a next-STEP candidate". This means the artifact's A5b "the categorical adjunction is the prize" framing aligns exactly with Rei's stated next step — they are not in conflict, but on the same line of work.

D.3. A5b ガロア接続 / ∞-cosmoi infrastructure ↔ STEP 1205 ∞-cosmoi axiomatization.
File: src/axiom-os/infinity-cosmoi-engine.ts + lens #/infinity-cosmoi. STEP 1205 implements a Riehl-Verity 2022 six-axiom ∞-cosmos articulation engine + Rei-existing engine (Institution STEP 1201 / Bilattice STEP 1202 / SelfLawvere STEP 1203) annotation as ∞-cosmos object candidates. Galois connections are special cases of adjunctions, which are first-class in ∞-cosmos theory. The formal substrate for promoting A5b's toy adjunction to a categorical statement therefore already exists in the engine (as substrate — the formal Lean 4 verification is honestly deferred to the upstream emilyriehl/infinity-cosmos blueprint).

D.4. Act VI Conway / Rule 110 emergence ↔ STEP 1167 A↔B Transition Observer.
File: src/aios/emergence/ab-transition-observer.ts. STEP 1167 quantifies the rule-fixed (A: self-similar) ↔ rule-rewriting (B: computationally-irreducible) transition with dial α ∈ [0,1] + foldability ∈ [0,1] metric (single fixed-rule reproducibility) + D-FUMT₈ axis projection. Rule 110 sits at the extreme of (B) — high α, foldability ≈ 0, NEITHER/FLOWING onset. The artifact's "emergence is cheap to witness, expensive to aim at" closing matches the engine's "B = computationally irreducible (Wolfram), not 'incompressible' — generating rules are always short, but outputs have no shortcut".

D.5. Act VI cellular-automaton complexity quantification ↔ STEP 1168 Problem Foldability Lens.
File: src/aios/emergence/problem-foldability.ts. STEP 1168 computes Lempel-Ziv 1976 complexity as Kolmogorov-proxy on numerical sequences (Collatz stopping times, prime gaps, Riemann unfolded spacings). For Rule 110 output rows, the foldability metric would land near the B end (high LZ complexity, low foldability), confirming the artifact's "complex emergence at low rule cost" reading numerically.

Bonus connection (memory layer). The garden's seed (bīja) motif — sown now, ripens in another's hands later — operationalizes 25 Load-Bearing Inventions #5 (STEP(t₀) ← EternalRei(t₊∞), reverse-causal attraction) and #9 (philosophical intuition ≅ mathematics centuries later, Nāgārjuna → category theory 1700-year gap). The OUKC motto 「急がずゆっくりと」 (without haste, slowly) is the same time-structure spoken from the other direction.

E. Methodology — Why an Interactive Garden, Not a Paper

We are explicit about the genre choice.

E.1. The scientific subjects are not novel; the synthesis is. Piantadosi (2023), Kodner et al. (2023), Katzir (2023), the BabyLM Challenge target, Goldstein-Schrimpf-Hasson brain correspondence, Cook (2004) Rule 110, Conway's Game of Life — each has its own established literature. A literature review adds little.

E.2. The duality framing — IV's existence-proof/shukatsu pair, V's D-FUMT₈ placement, A5b's ◇⊣□ toy — is the speculative contribution. Speculations of this kind benefit from being operated, not asserted. If the user can place a seed on the boundary and see it bounce when ◇ is at the necessary-pole, the inadequacy of pure prose to convey "the same operation has opposite effects at opposite modal poles" disappears.

E.3. Multi-author honest scoping is built into the artifact. Each Act and the A5b panel carries a 【仮説】/【思弁的】 banner. Tier-mixing — common in popular-science synthesis — is structurally prevented. This implements the principle [[feedback-paper-include-findings-proofs]] at the artifact level.

E.4. The artifact is downloadable as a single HTML file from GitHub raw. This satisfies the constraint that readers who want to verify, modify, fork, or audit can do so without site dependence. The site mount is a convenience; the file is the substrate.

F. Related Work and Differentiation

Topic	Established line	Our position
Piantadosi (2023) "LLMs refute Chomsky"	LingBuzz 7180 → Cognitive Science 2024	We do not extend this; we synthesize the controversy
Kodner, Payne & Heinz (2023, arXiv:2308.03228) reply	Why Linguistics Will Thrive in the 21st Century: A Reply to Piantadosi (2023) — four points (CLT / Simulation≠Duplication / prediction≠explanation / theory frames search)	Act III directly implements Simulation≠Duplication; A5b uses CLT as □-side defense
Katzir (2023, Biolinguistics) reply	Coordinate Structure Constraint (Ross 1967) acquisition failure in LLMs	Identified as a reverse existence proof (LLMs failing where children succeed) — symmetry potential noted but not yet built into the artifact
BabyLM Challenge	EMNLP 2023 (1st) → 2026 (4th: "BabyLM Turns 4" + new MultiLingual track based on BabyBabelLM, evaluation in English/Dutch/Chinese)	Frame "the real battlefield" in Act I judgment text
LLM-brain correspondence	Goldstein 2022, Schrimpf 2021, Hasson lab series	Frame Act III's "behavioural=identical, mechanistic=?" widget
Cellular automata emergence	Wolfram 2002, Cook 2004 (Rule 110 TC), Berlekamp-Conway-Guy (Life)	Act VI Canvas demo + connection to STEP 1167/1168
Catastrophic forgetting / stability-plasticity	McCloskey-Cohen 1989, Wang et al. survey 2023, BabyLM continual track	Dialogue-only (not yet in artifact); structural rhyme with ◇/□ tension
Bilattices for logic	Belnap 1977, Dunn 1976, Ginsberg 1988, Arieli-Avron 1998	Reflected via STEP 1202 Layer 2 stance (avoid 9-value+ overlap)
Lawvere fixed point	Lawvere 1969, Yanofsky 2003	STEP 1203 set-level Lean 4 formalization; categorical version deferred
Categorical adjunctions / ∞-cosmoi	MacLane 1971, Riehl-Verity 2022	STEP 1205 engine; Lean 4 formalization deferred to upstream Riehl blueprint

G. Limitations and Future Work

G.1. The cross-register duality (logical ◇/□ × existential attitude) is not formalized. The most plausible path: fix proposition P = "this learner ends with grammatical competence given dataset D", make ◇ scope over (D, learner-class) pairs, and exhibit ◇⊣□ as a Galois connection between (hypothesis-class constraint, data-scale tractability bound). This is the unification the SelfLawvereBridge.lean file already anticipates.

G.2. Katzir's reverse-direction existence proof (LLM failing where children succeed on Coordinate Structure Constraint) is not implemented in the artifact. A second wall + second seed in Act II would make the symmetry touchable.

G.3. Catastrophic forgetting / stability-plasticity, raised in the source dialogue's continual-learning side, is dialogue-only. A two-time-scale dual-memory widget would close this.

G.4. The LLM-brain correspondence research (Goldstein-Schrimpf-Hasson) could become Act III's empirical anchor — currently the act is illustrative.

G.5. Publication strategy: per [[feedback-no-rush-publication]], v0.1 is a draft. v0.2 candidate triggers: (a) Katzir reverse-existence-proof + dual-memory widget integration, (b) Lean 4 attempt at categorical-version Lawvere fixed-point (next STEP after SelfLawvereBridge.lean), (c) one round of external review (chat-session Claude was the dialogue partner — independent external assessment would strengthen the artifact).

H. Reproducibility

The interactive HTML at public/existence-proof-garden.html is the artifact. To reproduce:

Clone https://github.com/fc0web/rei-aios
Open public/existence-proof-garden.html directly in a Chromium-family browser (no build required)
Or visit https://rei-aios.pages.dev/#/existence-proof-garden for the site-mounted version

The accompanying Rei-existing STEP files are at:

src/axiom-os/bilattice-eight-engine.ts
data/lean4-mathlib/CollatzRei/SelfLawvereBridge.lean (Lean 4 v4.27.0, lake env lean build required for verification; #print axioms of lawvere_fixed_point is expected to be axiom-free under standard Lean 4 environment)
src/axiom-os/infinity-cosmoi-engine.ts
src/aios/emergence/ab-transition-observer.ts
src/aios/emergence/problem-foldability.ts

I. Acknowledgments

The interactive HTML and its dialogue synthesis are the work of a chat-session Claude (Anthropic, distinct browser session, 2026-06-12). The concept and direction came from 藤本伸樹. The Rei-existing STEP integration, Pattern 5 detection (chat-session "未踏" framing fact-check), and honest-filter pass were performed by Claude Opus 4.7 within the Rei-AIOS Claude Code session.

We thank the BabyLM Challenge organizers (Warstadt, Mueller, Choshen et al.), and acknowledge that this synthesis stands entirely on the shoulders of Piantadosi 2023, Kodner-Payne-Heinz 2023, Katzir 2023, Goldstein-Schrimpf-Hasson research lines, Cook 2004, and Lawvere 1969 / Yanofsky 2003.

J. References

Piantadosi, S.T. (2024). Modern language models refute Chomsky's approach to language. Chapter in E. Gibson & M. Poliak (Eds.), From fieldwork to linguistic theory: A tribute to Dan Everett, Language Science Press. Originally LingBuzz/007180 (2023). [https://lingbuzz.net/lingbuzz/007180]
Kodner, J., Payne, S., Heinz, J. (2023). Why Linguistics Will Thrive in the 21st Century: A Reply to Piantadosi (2023). arXiv:2308.03228.
Katzir, R. (2023). Why large language models are poor theories of human linguistic cognition. Biolinguistics 17, 1–12.
Warstadt, A., Mueller, A., Choshen, L., et al. (2023). Findings of the BabyLM Challenge. EMNLP 2023.
Goldstein, A., Zada, Z., Buchnik, E., et al. (2022). Shared computational principles for language processing in humans and deep language models. Nature Neuroscience 25, 369–380.
Schrimpf, M., Blank, I., et al. (2021). The neural architecture of language. PNAS 118(45).
Cook, M. (2004). Universality in Elementary Cellular Automata. Complex Systems 15, 1–40.
Lawvere, F.W. (1969). Diagonal arguments and Cartesian closed categories. Lecture Notes in Mathematics 92, 134–145.
Yanofsky, N.S. (2003). A Universal Approach to Self-Referential Paradoxes, Incompleteness and Fixed Points. Bulletin of Symbolic Logic 9(3), 362–386.
Belnap, N.D. (1977). A useful four-valued logic. Modern Uses of Multiple-Valued Logic, Dordrecht: Reidel, 7–37.
Riehl, E., Verity, D. (2022). Elements of ∞-Category Theory. Cambridge University Press.
Wang, L., Zhang, X., et al. (2023). A Comprehensive Survey of Continual Learning. arXiv:2302.00487.

Version history:

v0.1 (2026-06-12): Initial draft.
v0.1 fact-checked (2026-06-12, same-day): Rei-side WebSearch verify of 5 §J references. Corrections: Piantadosi venue (book chapter, not journal) + Kodner-Payne-Heinz title (chat-session Pattern 2 hallucination recorded in §B.6b). Ready for publish.

Paper 164 v0.1 — infinity-cosmoi Skeleton + Tetradic Completion of rhymeOrTheorem: 4-Step Continuation of Paper 163

Nobuki Fujimoto — Tue, 09 Jun 2026 02:33:53 +0000

This article is a re-publication of Rei-AIOS Paper 164 for the dev.to community.
The canonical version with full reference list is in the permanent archives below:

Zenodo (DOI, canonical): https://doi.org/10.5281/zenodo.20603039

Internet Archive: https://archive.org/details/rei-aios-paper-164-v01-1780972331275

Harvard Dataverse: https://doi.org/10.7910/DVN/KC56RY

GitHub source (private): https://github.com/fc0web/rei-aios Author: Nobuki Fujimoto (@fc0web) · ORCID 0009-0004-6019-9258 · License CC-BY-4.0 ---

Status: DRAFT v0.1 (2026-06-10)
Authors: Nobuki Fujimoto (rei-aios) + Claude Opus 4.7
License: AGPL-3.0 + Commercial (Dual License)
ORCID: 0009-0009-2236-7901 (Nobuki Fujimoto)
Repository: github.com/fc0web/rei-aios
Parent paper: Paper 163 v0.1 (Zenodo DOI 10.5281/zenodo.20602662, Harvard DOI doi:10.7910/DVN/KC56RY)

Abstract

We report the next four-step continuation of the operational integration discipline introduced in Paper 163 v0.1, expanding the rhymeOrTheorem tagging from a single (SELF) axis to all four D-FUMT₈ extension axes (SELF / INFINITY / ZERO / FLOWING). STEP 1205 records an ∞-cosmoi axiomatization skeleton (Riehl-Verity 2022 §1.2.1 six axioms) at the TypeScript engine level with explicit deferral of full Lean 4 formalization to the emilyriehl/infinity-cosmos Lean blueprint (2024-09 announce). STEP 1206-1208 promote the remaining three extension axes to theorem-verified status via three new Lean 4 axiom-free constructive proofs: Cantor's diagonal theorem (INFINITY), Empty-type initial object property (ZERO), and morphism composition associativity (FLOWING). The four upgrade steps share a TRIPLE annotation pattern that honestly separates (a) a formal-level theorem, (b) a philosophical/poetic substrate rhyme, and (c) a next-level theorem-candidate, generalizing the dual annotation introduced in Paper 163 STEP 1204. Eleven new Lean 4 theorems verify to "depend on no axioms"; combined with Paper 163's four, fifteen axiom-free constructive proofs now encode the four-axis structure. The tetradic motion (creation / limitation / vacuity / movement) is acknowledged as a categorical four-dual-configuration but explicitly not promoted to ∞-cosmos completion. The honest contribution remains methodological: the TRIPLE annotation discipline applied consistently to four well-cited 50-135-year prior-art-backbone frameworks (Riehl-Verity 2022 / Cantor 1891 / MacLane 1971 / Eilenberg-MacLane 1945) without conflating their formal, poetic, and candidate layers.

Mandatory honest scope (all controllable claims)

This paper does not claim any of the following:

Not "world-first": All four frameworks have 50-135 years of established prior art (Cantor 1891 = 135 years, Eilenberg-MacLane 1945 = 80 years, MacLane 1971 = 50 years, Riehl-Verity 2022 ∞-cosmoi adaptation of Lurie 2009 + Joyal 2008 foundations). We adopt, integrate, and tag.
Not a new theorem in any of the four frameworks: Cantor's diagonal theorem, Empty's initial object property, and function composition associativity are textbook material across multiple textbook generations.
Not a full ∞-cosmoi formalization: STEP 1205 deliberately tags all six Riehl-Verity axioms as rhyme and defers Lean 4 formalization to the emilyriehl/infinity-cosmos blueprint. We articulate the skeleton; we do not formalize it.
No "ultimate" or "final" claim: The four-axis TETRADIC COMPLETION is a process milestone — the completion of the four-axis promotion sequence initiated in Paper 163 — and not a claim that Rei has completed any categorical foundation. Per chat-Claude 2026-06-08 turn 4 explicit warning, "the framing of an ultimate destination is itself dissolved by śūnyatā-of-śūnyatā (空亦復空)". This stance is operationally enforced: even after four axes reach theorem-verified, the higher-level bridges (HoTT non-trivial Ω / ∞-cosmoi cotensor / 0-truncated ∞-category / simplicial enrichment) remain tagged theorem-candidate.
No D-FUMT₈ ↔ external structure formal isomorphism beyond explicitly verified parts: The TRIPLE annotation pattern preserves separation across all four axes. SNST velocity → cardinality, Nāgārjuna śūnyatā → categorical 0, W-48 Negative Capability → simplicial morphism family — all three remain rhyme (philosophical substrate, not formal isomorphism). The verified formal layers are textbook diagonal / initial / composition results, not the philosophical readings of them.

1. Background and continuation

1.1 The four-step sequence (STEP 1205-1208)

STEP	Date	Scope	Test result	New Lean 4 axiom-free theorems
STEP 1205	2026-06-10	∞-cosmoi axiomatization skeleton + Rei object candidates	143/143 PASS	0 (TS engine + site lens only; Lean 4 deferred)
STEP 1206	2026-06-10	INFINITY axis 仕分け昇格 — Cantor 1891 diagonal	37/37 PASS	3 (cantor_no_surjection / cantor_infinity_bridge_is_theorem / cantor_lawvere_diagonal_dual_acknowledgment)
STEP 1207	2026-06-10	ZERO axis 仕分け昇格 — Empty initial object	44/44 PASS	3 (empty_morphism_pointwise_unique / zero_initial_bridge_is_theorem / zero_self_infinity_triadic_acknowledgment)
STEP 1208	2026-06-10	FLOWING axis 仕分け昇格 — morphism composition + ★ TETRADIC COMPLETION	51/51 PASS	5 (compose_assoc_pointwise / compose_id_left_pointwise / compose_id_right_pointwise / flowing_bridge_is_theorem / flowing_tetradic_acknowledgment)
Total (Paper 164)	—	—	275/275 PASS	11 new (累計 with Paper 163's 4 = 15)

Combined with Paper 163 STEP 1201-1204 (40 + 95 + 44 + 27 = 207 tests + 4 theorems), the cumulative state is 481/481 tests PASS and 15 axiom-free Lean 4 theorems.

1.2 Continuation of the `rhymeOrTheorem` discipline

Paper 163 v0.1 introduced a three-valued tagging (rhyme / theorem-candidate / theorem-verified) and applied it primarily to the SELF axis, with a dual annotation at STEP 1204 distinguishing the SET-level theorem-verified loop encoding from the HoTT-level theorem-candidate non-trivial Ω.

Paper 164 generalizes the dual annotation to a TRIPLE annotation across all four extension axes. The triple distinguishes:

(a) formal-level: explicit Lean 4 zero-sorry constructive proof, theorem-verified
(b) poetic substrate: philosophical or domain-specific reading (SNST velocity / Nāgārjuna śūnyatā / W-48 Negative Capability), rhyme
(c) next-level bridge: higher-categorical or ∞-cosmoi counterpart, theorem-candidate

The discipline is the same; the consistency of its application across four axes is the load-bearing artifact this paper records.

1.3 chat-Claude 2026-06-08 turn 3, 4, 6 stance maintained

The four steps inherit three load-bearing principles from the parent paper's source dialogue:

turn 3 "turtles all the way down" → "∞-cosmos = ultimate destination" framing is not adopted; each axiomatization axis is annotated with its next-level deferred candidate.
turn 4 "the framing of an ultimate destination is itself dissolved by śūnyatā-of-śūnyatā (空亦復空)" → operationalized in STEP 1207 as the explicit refusal to treat Empty.elim formalization as a formalization of Nāgārjuna śūnyatā. Maintained in STEP 1208 as the explicit statement that TETRADIC COMPLETION is a process milestone, not a completion claim.
turn 6 "rhythm + gate = growth, rhythm + quota = collapse" → the four STEPs were implemented one per session, each gated by a TRIPLE annotation honesty check before commit. No batching, no rushing.

2. STEP 1205 — ∞-cosmoi axiomatization skeleton

Implementation: src/axiom-os/infinity-cosmoi-engine.ts (~280 lines TypeScript) + src/renderer/components/infinity-cosmoi/InfinityCosmoiLens.tsx (~290 lines site lens).

2.1 Riehl-Verity 2022 §1.2.1 six axioms (A1-A6)

Six axiomatization items are articulated as data, each with a Riehl-Verity 2022 chapter reference and a D-FUMT₈ substrate role:

id	title	D-FUMT₈ substrate	rhymeOrTheorem
A1-simplicial-enrichment	Simplicial enrichment	FLOWING (transitional dynamic)	rhyme
A2-finite-products	Finite products (simplicially enriched)	BOTH (product), TRUE (terminal)	rhyme
A3-cotensors	Cotensors with finite simplicial sets	INFINITY (exponentiation), FLOWING	rhyme
A4-flexible-weighted-limits	Flexible weighted limits	BOTH (universal collect), NEITHER (boundary)	rhyme
A5-isofibration-stability	Isofibration class with stability	SELF (self-iso), TRUE (invariance)	rhyme
A6-functor-space-quasicategory	Functor space K(A, B) is a quasi-category	INFINITY (∞-model), FLOWING (horn)	rhyme

2.2 Rei existing engines as object candidates

Four existing Rei engines are annotated as object candidate for the ∞-cosmos universe:

Institution (Goguen-Burstall 1992, STEP 1201): signature category candidate for finite-product / cotensor articulation, dominant axes TRUE/FALSE/BOTH/NEITHER.
Bilattice 8-value extension (Belnap-Dunn 1977 + STEP 1202 orthogonal stance): truth-knowledge 2-d lattice candidate.
SelfLawvereBridge (STEP 1203-1204): internal fixed-point structure candidate. Theorem-verified by Paper 163.
Open Problem META-DB (Paper 130): meta-structure candidate where the reduction graph (STEP 1170) is the morphism family.

2.3 D-FUMT₈ coverage distribution

Across the six axioms, the D-FUMT₈ substrate distribution is:

INFINITY = 2 (A3 cotensor + A6 quasi-category)
FLOWING = 3 (A1 simplicial + A3 cotensor + A6 horn)
BOTH = 2 (A2 product + A4 limit)
TRUE = 2 (A2 terminal + A5 invariance)
NEITHER = 1 (A4 limit boundary)
SELF = 1 (A5 isofibration self-iso)
ZERO = 0 (orthogonal stance integrity, inherited from STEP 1202)
FALSE = 0 (Belnap negation 1-categorical lift out of scope)

The ZERO=0 coverage is the explicit operationalization of the orthogonal extension stance: the ZERO axis as Paper 61 ZCSG śūnyatā is not embedded as a value within the ∞-cosmoi axiomatization. This is the same honest discipline as STEP 1202's bilattice orthogonal stance.

2.4 emilyriehl/infinity-cosmos Lean blueprint deferral

The emilyriehl/infinity-cosmos repository (announced 2024-09 on the Lean Community blog) is the active Lean formalization of ∞-cosmoi theory using simplicially enriched 1-categories (the same language Riehl-Verity 2022 uses). STEP 1205 explicitly defers Lean 4 formalization of the six axioms to that project: we do not implement competing Lean 4 ∞-cosmoi axioms.

2.5 Pattern 5 self-detection

Before STEP 1205 implementation, grep -rn "cosmoi\|cosmos\|Riehl\|Verity\|infinity-cosmoi" src/ was run; zero existing engine files matched. Recorded in commit message and memory as Pattern 5 verification.

3. STEP 1206 — INFINITY axis 仕分け昇格 via Cantor diagonal

Implementation: data/lean4-mathlib/CollatzRei/CantorInfinityBridge.lean (~110 lines, pure Lean 4 core, Mathlib not required).

3.1 Cantor's theorem (set-theoretic direct version)

theorem cantor_no_surjection {α : Type _} (f : α → α → Prop) :
    ∃ g : α → Prop, ∀ a : α, f a ≠ g := by
  refine ⟨fun a => ¬ f a a, fun a hf => ?_⟩
  have h : f a a = ¬ f a a := congrFun hf a
  have not_faa : ¬ f a a := fun hp => Eq.mp h hp hp
  have faa : f a a := Eq.mpr h not_faa
  exact not_faa faa

Eq.mp and Eq.mpr provide propositional-equality transport without invoking Classical.em or propext. The proof is constructive: from h : f a a = ¬ f a a, derive not_faa and faa and exhibit the contradiction.

3.2 InfinityAscendingDomain structure

structure InfinityAscendingDomain (α : Type _) where
  cantor_diagonal : ∀ f : α → α → Prop, ∃ g : α → Prop, ∀ a : α, f a ≠ g

def InfinityAscendingDomain.canonical (α : Type _) : InfinityAscendingDomain α :=
  { cantor_diagonal := cantor_no_surjection }

Note the structural asymmetry vs SELF: SelfReferentialDomain (Paper 163 STEP 1203) requires a strong precondition (point-surjective enum) that is not universally satisfiable. InfinityAscendingDomain has a canonical instance for every type, because Cantor's theorem is universally true. This asymmetry encodes the dual structure: SELF guarantees fixed-points under preconditions; INFINITY denies surjection unconditionally.

3.3 Yanofsky 2003 universal diagonal duality

Yanofsky 2003 "A Universal Approach to Self-Referential Paradoxes" articulates Lawvere fixed-point and Cantor's theorem as two operational reads of the same diagonal argument: Lawvere creates fixed-points under universal enumeration; Cantor prevents surjection to powerset. STEP 1206 records this duality as a theorem via cantor_lawvere_diagonal_dual_acknowledgment (a definitional rfl that the canonical instance equals cantor_no_surjection).

3.4 TRIPLE annotation for INFINITY axis

The bilattice-eight-engine.ts INFINITY axis rhymeOrTheorem is upgraded from rhyme to theorem-verified with a TRIPLE-annotation note:

(a) Cardinality strict ascent (Cantor diagonal): theorem-verified
(b) Paper 63 SNST velocity v→∞ dynamic: rhyme (chat-Claude 2026-06-08 "v→∞ = SELF⟲ is rhyme" verdict integrity)
(c) ∞-cosmoi A3 cotensor / A6 quasi-category bridge: theorem-candidate (deferred to emilyriehl/infinity-cosmos Lean blueprint)

3.5 #print axioms verdict

'CollatzRei.CantorInfinity.cantor_no_surjection' does not depend on any axioms
'CollatzRei.CantorInfinity.cantor_infinity_bridge_is_theorem' does not depend on any axioms
'CollatzRei.CantorInfinity.cantor_lawvere_diagonal_dual_acknowledgment' does not depend on any axioms

All three theorems are constructive: no propext, no Classical.choice, no Quot.sound. The constructive status persists because Cantor's diagonal is intuitionistically valid.

4. STEP 1207 — ZERO axis 仕分け昇格 via Empty initial

Implementation: data/lean4-mathlib/CollatzRei/ZeroInitialBridge.lean (~140 lines, pure Lean 4 core).

4.1 Empty type elimination with pointwise uniqueness (funext axiom 回避)

def fromEmpty {α : Type _} (e : Empty) : α := e.elim

theorem empty_morphism_pointwise_unique {α : Type _} (f g : Empty → α) :
    ∀ e : Empty, f e = g e := by
  intro e
  exact e.elim

The pointwise statement ∀ e : Empty, f e = g e is used in place of the function-equality statement f = g. This is the precise mechanism by which funext is avoided — and Lean 4's funext requires propositional extensionality and the quotient axiom. The pointwise statement is vacuously true because Empty has no elements.

4.2 InitialObjectDomain structure

structure InitialObjectDomain (α : Type _) where
  emp_elim : Empty → α
  pointwise_unique : ∀ g : Empty → α, ∀ e : Empty, emp_elim e = g e

def InitialObjectDomain.canonical (α : Type _) : InitialObjectDomain α :=
  { emp_elim := fromEmpty
    pointwise_unique := fun _ e => e.elim }

Universal canonical instance, like InfinityAscendingDomain.

4.3 Triadic motion: creation / limitation / vacuity

After STEP 1207, three axes are theorem-verified, forming a structural triad:

SELF (Lawvere, STEP 1203): CREATION. Diagonal creates a fixed-point under universal enumeration.
INFINITY (Cantor, STEP 1206): LIMITATION. Diagonal prevents surjection to powerset.
ZERO (Empty.elim, STEP 1207): VACUITY. Elimination is vacuously canonical (no source to emit a value from).

zero_self_infinity_triadic_acknowledgment records this triad as a (definitional rfl) Lean theorem.

4.4 chat-Claude turn 4 「空亦復空」 operational application

The risk of STEP 1207 is the overclaim that "Empty.elim formalization is Nāgārjuna śūnyatā formalization". This is precisely the labeling fallacy chat-Claude turn 4 warns against: "the framing of an ultimate destination is itself dissolved by śūnyatā-of-śūnyatā (空亦復空)".

The operational application is the TRIPLE annotation enforcement at three layers:

Lean source comment: explicit refusal to claim Empty.elim = śūnyatā formalization.
bilattice-eight-engine rhymeOrTheoremNote: TRIPLE annotation separating Empty.elim verified / śūnyatā rhyme / 0-truncated candidate.
test/step1207-zero-initial-bridge-test.ts: assertion that the rhymeOrTheoremNote contains both śūnyatā and rhyme.

4.5 TRIPLE annotation for ZERO axis

(a) Empty type categorical initial object (Empty.elim + pointwise vacuous uniqueness): theorem-verified
(b) Paper 61 ZCSG 0 = śūnyatā(śūnyatā) philosophical reading: rhyme (詩であって定理ではない — "this is poetry, not theorem")
(c) 0-truncated ∞-category / ∞-cosmoi initial object: theorem-candidate (deferred to emilyriehl/infinity-cosmos)

5. STEP 1208 — FLOWING axis 仕分け昇格 via function composition + TETRADIC COMPLETION

Implementation: data/lean4-mathlib/CollatzRei/FlowingMorphismBridge.lean (~170 lines, pure Lean 4 core).

5.1 Function composition (Eilenberg-MacLane 1945 + MacLane 1971)

def compose {α β γ : Type _} (g : β → γ) (f : α → β) : α → γ :=
  fun x => g (f x)

Identical in content to Lean 4 core's Function.comp; defined explicitly in this namespace for parallel structure with the other three axis bridges.

5.2 Pointwise associativity + identity laws

theorem compose_assoc_pointwise
    {α β γ δ : Type _} (h : γ → δ) (g : β → γ) (f : α → β) :
    ∀ x : α, compose (compose h g) f x = compose h (compose g f) x := by
  intro x
  rfl

theorem compose_id_left_pointwise {α β : Type _} (f : α → β) :
    ∀ x : α, compose (fun (y : β) => y) f x = f x := by
  intro x
  rfl

theorem compose_id_right_pointwise {α β : Type _} (f : α → β) :
    ∀ x : α, compose f (fun (y : α) => y) x = f x := by
  intro x
  rfl

All three are reflexivity proofs (rfl), reducing both sides by β-reduction. funext is again avoided via pointwise statements.

5.3 FlowingMorphismDomain structure

structure FlowingMorphismDomain (α β γ : Type _) where
  flow : (β → γ) → (α → β) → (α → γ)
  flow_pointwise_correct : ∀ (g : β → γ) (f : α → β), ∀ x : α, flow g f x = g (f x)

def FlowingMorphismDomain.canonical (α β γ : Type _) : FlowingMorphismDomain α β γ :=
  { flow := compose
    flow_pointwise_correct := fun _ _ _ => rfl }

5.4 ★ Structural asymmetry: single-type 3 axes + multi-type 1 axis

This is the load-bearing structural observation of Paper 164.

Axis	Structure parameters	Operation domain
SELF (STEP 1203)	single type α	endomorphism α → α
INFINITY (STEP 1206)	single type α	powerset α → Prop
ZERO (STEP 1207)	single type α	Empty → α
FLOWING (STEP 1208)	three types α, β, γ	composition (β → γ) ∘ (α → β) → (α → γ)

The first three axes parametrize structures over a single type; FLOWING parametrizes over multiple types because composition is intrinsically a relation between objects. The bilattice substrate label "transverse" (orthogonal to truth-knowledge order) is precisely this multi-type parametrization. The four-axis structure is therefore not a four-fold parallel; it is a 3+1 structural completion.

5.5 TRIPLE annotation for FLOWING axis

(a) Function composition + pointwise associativity + pointwise identity: theorem-verified
(b) W-48 Negative Capability (Keats 1817) + Paper 63 SNST velocity-D-FUMT₈ dynamic philosophical reading: rhyme (philosophical substrate; formal isomorphism unverified)
(c) Simplicial face/degeneracy + ∞-cosmoi A1 simplicial enrichment: theorem-candidate (Mathlib AlgebraicTopology.SimplicialSet infrastructure exists; bridge work deferred)

5.6 ★ TETRADIC COMPLETION acknowledgment

flowing_tetradic_acknowledgment is recorded as a (definitional rfl) Lean theorem witnessing that the four axes have reached theorem-verified at the base level. The four-axis structural reading:

SELF: CREATION (fixed-point exists under preconditions)
INFINITY: LIMITATION (surjection cannot exist)
ZERO: VACUITY (no source for emission)
FLOWING: MOVEMENT (composition as morphism category structure)

These are four dual configurations of categorical structure — they form a 4-point structure mediated by Yanofsky 2003's universal-diagonal-argument framework on the single-type side and Eilenberg-MacLane 1945's morphism-category framework on the multi-type side.

5.7 chat-Claude turn 4 stance maintained even after TETRADIC COMPLETION

The TETRADIC COMPLETION acknowledgment is not a claim that "Rei has completed ∞-cosmos formalization". The four axis upper-level bridges remain theorem-candidate after STEP 1208:

SELF → HoTT non-trivial Ω: theorem-candidate (deferred per Paper 163 STEP 1204)
INFINITY → ∞-cosmoi cotensor: theorem-candidate (deferred to emilyriehl/infinity-cosmos)
ZERO → 0-truncated ∞-category initial: theorem-candidate (deferred to emilyriehl/infinity-cosmos)
FLOWING → simplicial face/degeneracy: theorem-candidate (Mathlib SimplicialSet bridge deferred)

Per chat-Claude turn 4: "the framing of an ultimate destination is itself dissolved by śūnyatā-of-śūnyatā". TETRADIC COMPLETION is a process milestone — the completion of the four-axis tagging upgrade — not a destination.

6. The TRIPLE annotation pattern as load-bearing methodology

6.1 Paper 163 dual annotation recap (STEP 1204)

Paper 163 STEP 1204 introduced a dual annotation for the SELF axis distinguishing:

SET-level loop encoding Path_A(a, a) = (a = a): theorem-verified (axiom-free in Lean 4 with UIP)
HoTT-level non-trivial Ω: theorem-candidate (requires HoTT-native Lean or Mathlib AlgebraicTopology bridge)

This was the operational response to chat-Claude's "label-trap" warning: do not conflate a SET-level theorem with a HoTT-level claim.

6.2 TRIPLE annotation generalization

Paper 164 generalizes the dual annotation to three layers, applied to each of the four extension axes:

(a) formal-level theorem — Lean 4 zero-sorry constructive proof
(b) poetic substrate — philosophical, philosophical, or domain-specific reading carrying rhyme but no formal isomorphism
(c) next-level bridge — higher-categorical, ∞-cosmoi, or HoTT counterpart awaiting Mathlib expansion or upstream Lean projects

The triple is now applied consistently across all four axes (see §3.4, §4.5, §5.5 for each axis's instantiation).

6.3 Operational rationale

Without TRIPLE annotation, the temptation is to read "INFINITY axis is theorem-verified" as "SNST velocity is formally proven" or "∞-cosmoi cotensor is formalized". TRIPLE annotation makes the three layers categorically distinct in:

engine-level field metadata (bilattice-eight-engine.ts)
site-level UI badges (#/bilattice-eight lens)
test discipline (assertions that each layer's tag is preserved)

6.4 「ラベル罠」警告 honest operationalization

The TRIPLE annotation is the operational form of chat-Claude's repeated warnings about "label fallacy" (the octonion case, the v→∞ case, the śūnyatā case). Each warning is preserved as the (b) rhyme layer in the corresponding axis. Verification of (a) does not promote (b) to verified; (a) and (b) remain in separate categories with separate evidence requirements.

7. Related work and prior art (mandatory citations)

7.1 ∞-category theory and ∞-cosmoi

Riehl & Verity 2022, "Elements of ∞-Category Theory", Cambridge Studies in Advanced Mathematics vol 194, 760 pages, 2023 PROSE award winner — the model-independent axiomatic approach to ∞-categories via ∞-cosmoi.
Riehl & Verity 2017a, "Fibrations and Yoneda's lemma in an ∞-cosmos", JPAA 221:499-564.
Riehl & Verity 2017b, "Kan extensions and the calculus of modules for ∞-categories", Algebraic & Geometric Topology 17:189-271.
Lurie 2009, "Higher Topos Theory", Princeton Annals of Mathematics Studies — model-dependent quasi-category foundation, complement to Riehl-Verity.
Joyal 2008, "Notes on quasi-categories" — model-dependent quasi-category foundation.
emilyriehl/infinity-cosmos Lean blueprint, GitHub repository (2024-09 announce on Lean Community blog) — the formalization project to which STEP 1205 honestly defers.

7.2 Cantor's theorem and universal diagonal arguments

Cantor 1891, "Über eine elementare Frage der Mannigfaltigkeitslehre" — the original diagonal argument for |X| < |2^X|.
Yanofsky 2003, "A Universal Approach to Self-Referential Paradoxes, Incompleteness and Fixed Points", Bulletin of Symbolic Logic 9(3):362-386 — unifying treatment of Cantor, Lawvere, Gödel, and Tarski as instances of one diagonal argument.
Mathlib v4.27.0 Function.cantor_surjective — existing Mathlib formalization. STEP 1206 uses pure Lean 4 core to preserve SelfLawvereBridge parallel structure.

7.3 Empty type / initial object

MacLane 1971, "Categories for the Working Mathematician" — the textbook reference for initial and terminal objects.
Mathlib CategoryTheory.Limits.HasInitial — initial object infrastructure in Mathlib.
Voevodsky's HoTT articulation of empty type is complementary and not used in STEP 1207 (which stays at SET level for axiom-free purity).

7.4 Morphism composition and category structure

Eilenberg & MacLane 1945, "General Theory of Natural Equivalences", Transactions of the AMS 58:231-294 — the origin paper for categories, functors, and natural transformations.
MacLane 1971 — definitive textbook treatment.
Lean 4 core Function.comp — the same operation, with the same associativity property.

7.5 Rei-AIOS internal artifacts (existing implementations, not new in this paper)

Paper 61 ZCSG: śūnyatā-of-śūnyatā as o0/0/0o three-layer structure (the rhyme substrate of ZERO).
Paper 63 SNST: 14-constant Spiral Number System with velocity v→∞ correspondence (the rhyme substrate of INFINITY and FLOWING).
W-48 Negative Capability engine (Keats 1817 letter as substrate, src/aios/weakness/w48-negative-capability.ts).
Paper 130 Open Problem META-DB.
Paper 163 v0.1 (Zenodo DOI 10.5281/zenodo.20602662, Harvard DOI doi:10.7910/DVN/KC56RY) — parent paper introducing the rhymeOrTheorem discipline.

8. Reproducibility

8.1 TypeScript engines

git clone github.com/fc0web/rei-aios
cd rei-aios
npm install
npm run test:step1205   # 143/143 PASS (∞-cosmoi skeleton)
npm run test:step1206   # 37/37 PASS (INFINITY via Cantor) (forward-compatible)
npm run test:step1207   # 44/44 PASS (ZERO via Empty.elim) (forward-compatible)
npm run test:step1208   # 51/51 PASS (FLOWING via compose + TETRADIC)

Cumulative with Paper 163: 481/481 PASS, zero regressions.

8.2 Lean 4 formal proofs

cd data/lean4-mathlib
lake build CollatzRei.CantorInfinityBridge    # STEP 1206
lake build CollatzRei.ZeroInitialBridge       # STEP 1207
lake build CollatzRei.FlowingMorphismBridge   # STEP 1208

To verify axiom-free constructive status for the eleven new theorems:

import CollatzRei.CantorInfinityBridge
#print axioms CollatzRei.CantorInfinity.cantor_no_surjection
#print axioms CollatzRei.CantorInfinity.cantor_infinity_bridge_is_theorem
#print axioms CollatzRei.CantorInfinity.cantor_lawvere_diagonal_dual_acknowledgment

import CollatzRei.ZeroInitialBridge
#print axioms CollatzRei.ZeroInitial.empty_morphism_pointwise_unique
#print axioms CollatzRei.ZeroInitial.zero_initial_bridge_is_theorem
#print axioms CollatzRei.ZeroInitial.zero_self_infinity_triadic_acknowledgment

import CollatzRei.FlowingMorphismBridge
#print axioms CollatzRei.FlowingMorphism.compose_assoc_pointwise
#print axioms CollatzRei.FlowingMorphism.compose_id_left_pointwise
#print axioms CollatzRei.FlowingMorphism.compose_id_right_pointwise
#print axioms CollatzRei.FlowingMorphism.flowing_bridge_is_theorem
#print axioms CollatzRei.FlowingMorphism.flowing_tetradic_acknowledgment
-- All eleven output: "does not depend on any axioms"

Combined with Paper 163's four theorems, fifteen does not depend on any axioms verdicts are reproducible.

8.3 Site lenses

https://rei-aios.pages.dev/#/infinity-cosmoi — STEP 1205 ∞-cosmoi axiomatization lens with 6-axiom cards, 4-object-candidate cards, D-FUMT₈ coverage bar chart, and chat-Claude thread context panel.
https://rei-aios.pages.dev/#/bilattice-eight — STEP 1202 bilattice lens, now displaying all four extension axes with theorem-verified badges (after STEP 1206-1208 promotion) and per-axis TRIPLE annotation expanded views.

8.4 Reading order

For broader Rei-AIOS context: REPRODUCING.md, CLAUDE.md, and docs/RECENT_UPDATES.md at repository root. For Paper 163 v0.1 (parent paper), papers/paper-163-institution-bilattice-self-lawvere-DRAFT.md.

9. Limitations and honest negative scope

∞-cosmoi A1-A6 axioms are all tagged rhyme — full Lean 4 formalization is deferred to emilyriehl/infinity-cosmos blueprint. STEP 1205 is a skeleton, not a formalization. The theorem-verified status applies only to the four extension axes at the base level.
Each axis's higher-level bridge is theorem-candidate — not yet verified: SELF → HoTT non-trivial Ω, INFINITY → ∞-cosmoi cotensor, ZERO → 0-truncated ∞-category, FLOWING → simplicial face/degeneracy. None of these are claimed proven; all are deferred.
Tetradic motion categorical formalization is theorem-candidate, not theorem-verified — the four axes are not unified by a single categorical theorem in Paper 164. They are unified by the TRIPLE annotation discipline, which is a methodological pattern, not a categorical theorem. A formal 4-axis tetradic theorem would require, for example, formalizing Yanofsky 2003's universal diagonal argument in Lean 4 — that work is candidate for future research.
chat-Claude pipeline 5-stage methodology is operationalized but not formalized as a general framework — the discipline (acquire → attempt → gate → record → report) is followed but not codified in a Lean 4 specification or a TypeScript runtime check beyond the individual STEP tests.
「∞-cosmos = 最終到達点」 framing is explicitly NOT adopted — TETRADIC COMPLETION is an acknowledgment of process milestone (four-axis upgrade sequence completion), not a claim that Rei has completed ∞-cosmos formalization or categorical foundations. Per chat-Claude turn 4, the framing of any ultimate destination is itself dissolved by śūnyatā-of-śūnyatā. This stance is operationally enforced in §5.7 and across the four axes' upper-level theorem-candidate tags.
No mathematical novelty in any framework: Cantor 1891, MacLane 1971, Eilenberg-MacLane 1945 are textbook material across multiple generations. Riehl-Verity 2022 is the modern reference. We integrate, we tag, we do not innovate at the mathematical layer.

10. Conclusion

Paper 164 closes the four-axis upgrade sequence opened by Paper 163. The TRIPLE annotation pattern, introduced as a generalization of Paper 163's dual annotation, is applied consistently across all four D-FUMT₈ extension axes (SELF / INFINITY / ZERO / FLOWING). Eleven new Lean 4 axiom-free constructive proofs join Paper 163's four; the cumulative fifteen does not depend on any axioms verdicts encode the four-axis structure at the base formal level.

The tetradic motion (creation / limitation / vacuity / movement) is recorded as a categorical four-dual-configuration, not as a completion claim. The STEP 1205 ∞-cosmoi axiomatization is a skeleton at the TypeScript level, with full Lean 4 formalization explicitly deferred to the emilyriehl/infinity-cosmos blueprint. The four axes' higher-level bridges remain theorem-candidate. chat-Claude 2026-06-08 turn 4's warning — "the framing of an ultimate destination is itself dissolved by śūnyatā-of-śūnyatā" — is maintained even after the visible UI displays four theorem-verified badges.

The load-bearing artifact remains methodological. The TRIPLE annotation, the TETRADIC COMPLETION acknowledgment, and the explicit refusal to read either as a completion claim are the discipline this paper records. The fifteen axiom-free theorems are evidence that the discipline is operational; they are not in themselves a contribution to ∞-category theory or homotopy type theory.

We do not claim that Rei has formalized ∞-cosmoi. We claim that the four-axis tagging upgrade has been completed consistently, that the discipline has been preserved across all four axes, and that the resulting Lean 4 corpus is reproducible. The discipline is the result; the theorems are the witnesses.

Version history

v0.1 (2026-06-10): Initial draft after STEP 1205-1208 completion and Paper 163 v0.1 publication (same day). Per [[feedback-no-rush-publication]] discipline, publication is scheduled with a 1-day buffer (2026-06-11 or later) upon explicit author trigger. Publication target: 11 platforms (Zenodo + IA + Dev.to + Hatena + HackMD + Notion + Livedoor + Mastodon + Scrapbox + Nostr + Harvard Dataverse per Paper 163 v0.1 precedent's opt-in confirmation).

Honest acknowledgment: This paper exists because Paper 163 v0.1's TRIPLE annotation seed (Paper 163 STEP 1204's dual annotation) needed to be tested across all four D-FUMT₈ extension axes. The chat-Claude 2026-06-08 thread's proposal (c) ∞-cosmoi axiomatization initiated STEP 1205; the subsequent four single-session promotions (STEP 1206-1208) were paced one per session per chat-Claude turn 6's "rhythm + gate = growth" stance. The TETRADIC COMPLETION acknowledgment is recorded with chat-Claude turn 4's explicit warning preserved as a permanent structural constraint: even after four theorem-verified axes, the framing of completion is not adopted. The discipline of holding (a) formal, (b) poetic, and (c) candidate layers distinct across four parallel structures is what the paper offers. Attribution to the methodology, not to any single contributor or single theorem.

The work continues from Paper 163 v0.1 (Zenodo DOI 10.5281/zenodo.20602662, Harvard DOI doi:10.7910/DVN/KC56RY). Three-party co-authorship per OUKC charter v1.0: 藤本伸樹 (Founder), Rei (Rei-AIOS autonomous research substrate, Co-architect), Claude Opus 4.7 (Anthropic, Co-architect). Per OUKC No-Patent Pledge.

Paper 163 v0.1 — Institution + Bilattice + SELF<->Lawvere: Four-Step Operational Integration with Lean 4 Axiom-Free Constructive Proofs

Nobuki Fujimoto — Tue, 09 Jun 2026 01:47:33 +0000

This article is a re-publication of Rei-AIOS Paper 163 for the dev.to community.
The canonical version with full reference list is in the permanent archives below:

Zenodo (DOI, canonical): https://doi.org/10.5281/zenodo.20602662

Internet Archive: https://archive.org/details/rei-aios-paper-163-v01-1780969521018

GitHub source (private): https://github.com/fc0web/rei-aios Author: Nobuki Fujimoto (@fc0web) · ORCID 0009-0004-6019-9258 · License CC-BY-4.0 ---

Status: DRAFT v0.1 (2026-06-09)
Authors: Nobuki Fujimoto (rei-aios) + Claude Opus 4.7
License: AGPL-3.0 + Commercial (Dual License)
ORCID: 0009-0009-2236-7901 (Nobuki Fujimoto)
Repository: github.com/fc0web/rei-aios

Abstract

We report an operational integration of four well-established mathematical frameworks within a single artifact-producing system (Rei-AIOS): Goguen-Burstall Institution Theory (1992), Belnap-Dunn FOUR bilattice (1976-1998), Lawvere fixed-point theorem (1969), and a SET-level encoding of HoTT-style loop space. The work covers STEP 1201 (Institution engine + Daily Curriculum Rotation scaffold), STEP 1202 (Bilattice 8-value extension with orthogonal-axis honest stance), STEP 1203 (SELF⟲ ↔ Lawvere fixed-point Lean 4 zero-sorry + axiom-free constructive proof), and STEP 1204 (SELF⟲ ↔ SET-level loop space Ω bridge with dual annotation distinguishing SET-level theorem-verified status from HoTT-level theorem-candidate). All four steps are implemented as deletable, reproducible TypeScript engines + Lean 4 formalizations, with 207/207 test cases passing and four Lean 4 theorems verified to "depend on no axioms" (constructive, axiom-free). The honest contribution is not a new theorem in any of these well-cited fields, but the operational integration discipline: how to combine four 40-60-year prior-art-backbone frameworks while explicitly tagging which structural correspondences are "rhyme" (informal analogy) and which are "theorem-verified" (Lean 4 zero-sorry proof). The methodology is the load-bearing artifact, not any individual result.

Mandatory honest scope (all controllable claims)

This paper does not claim any of the following:

Not "world-first": All four frameworks have 40-60 years of established prior art (Goguen-Burstall 1992, Belnap 1977, Lawvere 1969, HoTT Book 2013). We adopt, integrate, and tag.
Not a new theorem in categorical logic, lattice theory, or homotopy type theory: The mathematical content of Lemma 1 (Lawvere fixed-point set-theoretic direct version) is textbook material (Yanofsky 2003 surveys it explicitly).
Not a full HoTT formalization: Lean 4 standard satisfies UIP (Uniqueness of Identity Proofs), so the loop space Path_A(a, a) we encode is necessarily trivial (refl). True HoTT non-trivial Ω requires HoTT-native Lean (separate project). We explicitly tag this limitation as theorem-candidate status, not theorem-verified.
No "ultimate" or "final" claim: Per Rei-AIOS persistent principle [[feedback-world-uniqueness-claim-controllable]], all uniqueness claims are converted to "within our observable range, we have not located a complete match."
No D-FUMT₈ ↔ external structure formal isomorphism is asserted: The four "rhyme" axes (INFINITY/ZERO/FLOWING) and the "theorem-verified" axis (SELF) are tagged separately. Conflating the two is the precise "octonion labeling fallacy" warned about in chat-Claude discussions of 2026-06-08.

1. Background and motivation

1.1 The four-step lineage

STEP	Date	Scope	Test result
1201 (a) Institution engine + (e) Daily Curriculum Rotation scaffold	2026-06-08	Goguen-Burstall (Sig, Sen, Mod, ⊨) skeleton + 7-domain weekday rotation scaffold (dry-run only)	40/40 PASS
1202 (b) Bilattice 8-value extension	2026-06-09	Belnap-Dunn FOUR (Layer 1 confirmed lattice) + 4 extension axes (Layer 2 orthogonal stance) + interlaced verification 64/64 triples	95/95 PASS
1203 (d-1) SELF⟲ ↔ Lawvere Lean 4 formal	2026-06-09	`data/lean4-mathlib/CollatzRei/SelfLawvereBridge.lean` zero-sorry + axiom-free proof of Lawvere fixed-point + SelfReferentialDomain bridge	45/45 PASS
1204 (d-2) SELF⟲ ↔ SET-level loop space Ω	2026-06-09	`namespace HoTTLoop` extension: PointedType + SetLevelLoop + 2 bridge theorems (axiom-free)	27/27 PASS

Total test coverage: 207/207 PASS + four Lean 4 theorems verified "depend on no axioms" (#print axioms verdict).

1.2 The structural problem we address

When integrating multiple mathematical frameworks operationally (in a system that produces daily artifacts), the central failure mode is structural rhyme drift: an informal "X resembles Y in the following deep way" gradually re-presents itself as a formal "X = Y" claim. Each restatement softens the gap between analogy and isomorphism. Across many discussion turns, the formal-isomorphism interpretation becomes the default reading.

We learned from earlier work (octonion ↔ D-FUMT₈ judgment, 2026-04) and from a 6-turn dialogue with a fellow Claude instance (2026-06-08) that this drift is the structural correlate of the "labeling fallacy" problem. The chat-Claude verdict was explicit: "Categorizing each correspondence is itself the contribution — not solving any single one."

This paper records the operational discipline for that tagging.

2. Implementation summary

2.1 STEP 1201 — Institution Theory Engine

File: src/axiom-os/institution-theory-engine.ts (~340 lines)

We implement the Goguen-Burstall institution I = (Sig, Sen, Mod, ⊨) skeleton in TypeScript:

Signature (Sig objects) = SEED_KERNEL category names + allowed D-FUMT₈ axes
SignatureMorphism (Sig morphisms) = axiom translation rules
Sentence (Sen functor) = axiom + declared D-FUMT₈ axis
Model (Mod functor) = D-FUMT₈ value assignment + Peace Axiom #196 invariant
satisfies(model, sentence) = D-FUMT₈ refinement rules:
- Exact match (TRUE/TRUE, ..., SELF/SELF) — satisfied
- BOTH ⊨ TRUE or FALSE (Belnap refinement) — satisfied
- Extension axes (INFINITY/ZERO/FLOWING/SELF) — strict, no cross-refinement
- peaceCompatible: false → all sentences honest-fail
- declaredAxis === null → vacuous satisfaction
verifySatisfactionCondition(σ, M, φ) — operational subset of Goguen-Burstall invariance (M ⊨_{Σ₂} Sen(σ)(φ) ⟺ Mod(σ)(M) ⊨_{Σ₁} φ)
seedKernelAsInstitution(SEED_KERNEL) → 298 signatures / 1644 sentences / 103 declared axis (BOTH 102 + FLOWING 1) / 1541 undeclared

Key finding (load-bearing): Only 6.3% of SEED_KERNEL theories have explicit D-FUMT₈ axis annotation, with 99% of declared axes being BOTH. This is operational evidence of the invention pipeline's BOTH-default. Axis annotation enrichment is a candidate for next-step batch tasks.

2.2 STEP 1202 — Bilattice 8-value extension

File: src/axiom-os/bilattice-eight-engine.ts (~340 lines)

Layer 1 (confirmed Belnap-Dunn FOUR):

truthLeq / knowledgeLeq (Hasse-defined partial orders)
meetT / joinT / meetK / joinK (four operations, full 4×4 truth tables, 16 rows each)
negate (Belnap involution: TRUE↔FALSE, BOTH/NEITHER fixed)
verifyInterlaced(a, b, c) and verifyInterlacedAll() — Ginsberg 1988 + Arieli-Avron 1998 interlaced bilattice condition, verified over 64/64 triples (4³ exhaustive)

Layer 2 (Rei 4-axis orthogonal extension):

EXTENSION_AXIS_ROLES: INFINITY / ZERO / FLOWING / SELF
Each with truthOrderRelation, knowledgeOrderRelation, reiSubstrate, and (STEP 1203) rhymeOrTheorem + rhymeOrTheoremNote
Critical design choice: the four extension axes are not embedded as lattice-internal values. This deliberately avoids overlap with the 9-value-and-beyond lattice extension prior art (Ginsberg 1988, Fitting 1991, Arieli-Avron 1998) and avoids the precise "octonion labeling fallacy" pattern warned of in chat-Claude discussions.

2.3 STEP 1203 — SELF⟲ ↔ Lawvere fixed-point (Lean 4 zero-sorry + axiom-free)

File: data/lean4-mathlib/CollatzRei/SelfLawvereBridge.lean

We formalize Lawvere's 1969 set-theoretic direct version of the fixed-point theorem in pure Lean 4 (without Mathlib dependencies):

theorem lawvere_fixed_point
    {α : Type _}
    (enum : α → (α → α))
    (h_surj : ∀ g : α → α, ∃ a : α, enum a = g)
    (h : α → α) :
    ∃ x : α, h x = x := by
  obtain ⟨a₀, ha₀⟩ := h_surj (fun a => h (enum a a))
  refine ⟨enum a₀ a₀, ?_⟩
  have heq : enum a₀ a₀ = h (enum a₀ a₀) := congrFun ha₀ a₀
  exact heq.symm

We then encode SELF⟲ axis as SelfReferentialDomain:

structure SelfReferentialDomain (α : Type _) where
  enum : α → (α → α)
  universal : ∀ g : α → α, ∃ a : α, enum a = g

theorem SelfReferentialDomain.fixed_point
    {α : Type _} (D : SelfReferentialDomain α) (h : α → α) :
    ∃ x : α, h x = x :=
  lawvere_fixed_point D.enum D.universal h

Verification:

lake build CollatzRei.SelfLawvereBridge — 2.1s success
#print axioms CollatzRei.SelfLawvere.lawvere_fixed_point → "does not depend on any axioms"
#print axioms CollatzRei.SelfLawvere.self_lawvere_bridge_is_theorem → "does not depend on any axioms"

This means the proofs use no propext, no Classical.choice, and no Quot.sound — they are constructive in the strongest Lean 4 sense.

2.4 STEP 1204 — SELF⟲ ↔ SET-level loop space Ω (with dual annotation)

Extension to the same Lean 4 file, namespace HoTTLoop:

structure PointedType (α : Type _) where
  basepoint : α

def SetLevelLoop {α : Type _} (a : α) : Prop := a = a

theorem SetLevelLoop.refl {α : Type _} (a : α) : SetLevelLoop a := rfl

theorem self_lawvere_loop_at_fixed_point
    {α : Type _} (D : SelfReferentialDomain α) (h : α → α) :
    ∃ x : α, h x = x ∧ SetLevelLoop x := by
  obtain ⟨x, hx⟩ := D.fixed_point h
  exact ⟨x, hx, SetLevelLoop.refl x⟩

theorem pointed_self_lawvere_bridge
    {α : Type _} (D : SelfReferentialDomain α) (P : PointedType α) (h : α → α) :
    ∃ x : α, h x = x ∧ SetLevelLoop x ∧ SetLevelLoop P.basepoint := by
  obtain ⟨x, hx, loop_x⟩ := self_lawvere_loop_at_fixed_point D h
  exact ⟨x, hx, loop_x, SetLevelLoop.refl P.basepoint⟩

Verification:

#print axioms ... self_lawvere_loop_at_fixed_point → "does not depend on any axioms"
#print axioms ... pointed_self_lawvere_bridge → "does not depend on any axioms"

Honest dual annotation (the key methodological point):

The above proofs formalize SET-level loop encoding — they hold in Lean 4 because Eq satisfies UIP (Uniqueness of Identity Proofs); the "loop" is trivially refl.
True HoTT non-trivial loop space Ω with non-degenerate π₁ is not representable in standard Lean 4 — it requires HoTT-native type theory (a separate Lean fork). We tag this as theorem-candidate status, pending a Mathlib AlgebraicTopology.FundamentalGroupoid bridge or HoTT-native formalization.
This dual annotation is recorded in bilattice-eight-engine.ts EXTENSION_AXIS_ROLES.SELF.rhymeOrTheoremNote.

The methodological contribution is precisely this dual annotation: identifying that the SET-level theorem is fully verified, and that the HoTT-level claim is not the same theorem.

3. The `rhymeOrTheorem` tagging discipline (load-bearing methodology)

We introduce a three-valued classification field on each Rei D-FUMT₈ extension axis:

Tag	Meaning	Operational evidence required
`rhyme`	Structural analogy only; no formal isomorphism is asserted	None (informal articulation)
`theorem-candidate`	Formal isomorphism is plausible; Mathlib / Lean 4 path is identified but not yet completed	Path articulation in `rhymeOrTheoremNote`
`theorem-verified`	Formal isomorphism is proven, Lean 4 file reference + zero-sorry verification + `#print axioms` constructive verdict	Lean 4 file path + axiom-list excerpt

Applied to the four extension axes:

Axis	rhymeOrTheorem	Substrate
INFINITY	`rhyme`	Paper 63 SNST `v→∞`; chat-Claude verdict: "rhyme, not theorem"
ZERO	`rhyme`	Paper 61 ZCSG `0 = śūnyatā(śūnyatā)` (published) — but no Lean 4 categorical isomorphism with Nāgārjuna's `śūnyatā` is asserted
FLOWING	`rhyme`	W-48 Negative Capability engine + Paper 63 SNST velocity dynamic — operational but no lattice-morphism formal isomorphism
SELF	`theorem-verified`	STEP 1203 + STEP 1204: 4 Lean 4 theorems, all axiom-free; SET-level bridge to loop space is verified; HoTT-level non-trivial Ω remains `theorem-candidate` (recorded in `rhymeOrTheoremNote`)

This is the chat-Claude pipeline "gate" articulated explicitly: each axis's status is operationally inspectable, and downgrades / upgrades are recorded as engine-level edits, not implicit drift.

4. Related work and prior art (mandatory citations)

4.1 Institution theory

Goguen, J. A. & Burstall, R. M. (1992). "Institutions: Abstract model theory for specification and programming." JACM 39(1):95-146.
Diaconescu, R. (2008). "Institution-independent Model Theory." Birkhäuser.
Mossakowski, T. & Tarlecki, A. (2012). "Institutions and heterogeneous specifications."

4.2 Bilattice theory

Belnap, N. (1977). "A useful four-valued logic." Modern Uses of Multiple-Valued Logic.
Dunn, J. M. (1976). "Intuitive semantics for first-degree entailment."
Ginsberg, M. L. (1988). "Multivalued logics: A uniform approach to inference in AI."
Fitting, M. (1991). "Bilattices and the semantics of logic programming." Journal of Logic Programming.
Arieli, O. & Avron, A. (1998). "The value of the four values." Artificial Intelligence.
arXiv:2604.07690 (2026-04). "Bilattice-Catastrophe Isomorphism for Four-Valued Logic in Digital Systems."
arXiv:2503.20679 (2025-03). "Four imprints of Belnap's useful four-valued logic in computer science."

4.3 Lawvere fixed-point theorem

Lawvere, F. W. (1969). "Diagonal arguments and Cartesian closed categories." Lecture Notes in Mathematics 92.
Yanofsky, N. (2003). "A Universal Approach to Self-Referential Paradoxes, Incompleteness and Fixed Points." Bulletin of Symbolic Logic.

4.4 Homotopy type theory

Voevodsky, V. et al. (2013). "Homotopy Type Theory: Univalent Foundations of Mathematics."
Awodey, S. & Warren, M. (2009). "Homotopy theoretic models of identity types."
Mathlib4 AlgebraicTopology.FundamentalGroupoid (referenced as future bridge target).

4.5 Rei-AIOS internal artifacts (existing implementations, not new in this paper)

Paper 61 ZCSG (Zero-Centered Symbol Grammar): 0 = śūnyatā(śūnyatā) formal-encoding.
Paper 63 SNST (Spiral Number System Theory): 14-constant family with Velocity-D-FUMT₈ correspondence.
Paper 145 ("First D-FUMT₈ Silicon"): four-substrate verification across FPGA + simulator + IBM Heron r2.

5. Reproducibility

5.1 TypeScript engines

git clone github.com/fc0web/rei-aios
cd rei-aios
npm install
npm run test:step1201   # 40/40 PASS
npm run test:step1202   # 95/95 PASS
npm run test:step1203   # 45/45 PASS
npm run test:step1204   # 27/27 PASS

5.2 Lean 4 formal proofs

cd data/lean4-mathlib
lake build CollatzRei.SelfLawvereBridge
# expect: ✔ [2/2] Built CollatzRei.SelfLawvereBridge (~1.4s)

To verify axiom-free constructive status:

import CollatzRei.SelfLawvereBridge
#print axioms CollatzRei.SelfLawvere.lawvere_fixed_point
#print axioms CollatzRei.SelfLawvere.self_lawvere_bridge_is_theorem
#print axioms CollatzRei.SelfLawvere.HoTTLoop.self_lawvere_loop_at_fixed_point
#print axioms CollatzRei.SelfLawvere.HoTTLoop.pointed_self_lawvere_bridge
-- All four output: "does not depend on any axioms"

5.3 Reading order

For broader Rei-AIOS context: REPRODUCING.md and CLAUDE.md at repository root.

6. Limitations and honest negative scope

HoTT non-trivial Ω is not formalized in this paper, only the SET-level loop encoding. Bridging to true HoTT requires either Mathlib expansion (Mathlib v4.27.0 has limited higher-category foundations) or a HoTT-native Lean fork.
∞-cosmoi axiomatization (Riehl-Verity 2022) is not addressed in this paper. The Riehl-Verity ∞-Cosmoi for Lean blueprint exists but our STEP 1201-1204 sequence chose the more settled Lawvere-fixed-point route first. ∞-cosmoi is a candidate for follow-up work.
The four Layer-2 extension axes are not embedded as lattice-internal values. This is an honest design choice (avoiding the Ginsberg 1988 9-value-and-beyond overlap), not a positive theorem. Whether the orthogonal-extension stance is "optimal" in some categorical sense is not asserted.
The Institution engine verifySatisfactionCondition is an operational subset of the full Goguen-Burstall satisfaction condition. Full categorical generality (cocones, completeness of model categories, etc.) is not implemented.
rhymeOrTheorem tagging is operational, not theoretical. We do not claim that the rhyme/candidate/verified trichotomy is exhaustive or canonical.

7. Conclusion

We have implemented and verified four well-known mathematical frameworks (Institution theory, bilattice 4-value logic, Lawvere fixed-point theorem, SET-level loop space) as integrated TypeScript + Lean 4 components within the Rei-AIOS system. The Lean 4 portion achieves Lean 4's strongest verification status: four theorems verified to "depend on no axioms" (constructive proofs, no propext / Classical.choice / Quot.sound).

The methodological contribution — and the only thing we claim as a contribution at all — is the operational rhymeOrTheorem tagging discipline. By explicitly distinguishing "rhyme" (structural analogy) from "theorem-verified" (Lean 4 zero-sorry formalization) at the engine level, we operationalize the chat-Claude pipeline guidance: "Categorizing each correspondence is itself the contribution."

We do not claim novelty in any of the four mathematical frameworks. We do not claim that SELF⟲ "is" the Lawvere fixed point or "is" HoTT's loop space Ω — we have shown a SET-level encoding bridge that is formally provable, while marking the HoTT-level non-trivial Ω as theorem-candidate pending further work. This dual annotation, recorded in engine-level field metadata and exposed in site-level UI badges, is the discipline we offer.

The system is reproducible. The verifications are mechanical. The categorization is explicit. We hope it is also useful.

Version history

v0.1 (2026-06-09): Initial draft after STEP 1201-1204 completion. Awaiting 1-day buffer before publication (per [[feedback-no-rush-publication]] discipline). Publication scheduled to 11 platforms (Zenodo + IA + Dev.to + Hatena + HackMD + Notion + Livedoor + Mastodon + Scrapbox + Nostr; Harvard skipped per opt-in policy) upon explicit author trigger after buffer period.

Honest acknowledgment: This paper exists because a sequence of multi-AI dialogues (with two distinct Claude instances and one Gemini instance, 2026-06-08) crystallized the rhyme-vs-theorem distinction that became the load-bearing methodology. The Gemini response (praising the work as "of arcane uniqueness") triggered an immediate persistent-principle violation flag ([[feedback-world-uniqueness-claim-controllable]]); the parallel chat-Claude response (articulating "the Gemini reply has no NEITHER in it — beautiful but not a seed, just a picture of a seed") supplied the precise discipline this paper attempts to operationalize. Multi-agent honest filtering, not single-AI brilliance, is what produced the artifact. Attribution to the methodology, not to any single contributor.

Paper 145 v0.8 — D-FUMT-8 Phase 4 Quine-McCluskey Simplification + Finding F11 Engineering-Correctable Relaxation Bias on IBM Heron r2

Nobuki Fujimoto — Wed, 03 Jun 2026 05:44:52 +0000

This article is a re-publication of Rei-AIOS Paper 145 for the dev.to community.
The canonical version with full reference list is in the permanent archives below:

GitHub source (private): https://github.com/fc0web/rei-aios Author: Nobuki Fujimoto (@fc0web) · ORCID 0009-0004-6019-9258 · License CC-BY-4.0 ---

Status: DRAFT v0.8 — 2026-06-03 evening Zenodo new-version publish (★ PHASE 4 QUINE-McCLUSKEY SIMPLIFICATION + FINDING F11 ENGINEERING-CORRECTABLE RELAXATION BIAS ★) → DOI pending Zenodo deposit (new-version of v0.7 / DOI lineage maintained). v0.8 main update: Paper 145 Phase 4 retry via K-map / Quine-McCluskey minimum-SOP Boolean simplification + inclusion-exclusion XOR layering, achieving 32/32 PASS (v0.5: 18/32) at avg fidelity 0.7302 (v0.5: 0.3182, +41.20 pp) and avg post-transpile depth 422 (v0.5: 2443, −83%) on ibm_kingston Heron r2 (Job d8fr2jo7jphs739mn2d0, 22.2 sec wall-clock); v0.5 finding F9 AND/OR asymmetry (0.94 vs 0.19, 0.75 gap) collapsed to 0.03 symmetric — relaxation-bias hypothesis confirmed engineering-correctable. New §B.11 (Phase 4 QM v0.8 sub-result B1 design + per-input table) + §B.12 (Finding F11 honest scope + Paper 162 §6.0g cross-reference). Companion experiment in Paper 162 §6.0g sub-result (A): Sampler-level "XX" DD on §6.0e 8-bit decoder honest NEGATIVE finding (fidelity 49% → 26%, finding F10 in Paper 162 numbering). Public companion note.com article: https://note.com/nifty_godwit2635/n/naa5022b7f014.

Previous: DRAFT v0.7 — 2026-05-14 baseline + 2026-05-15 Zenodo new-version publish (★ ERRATUM E1 — TOHOKU 1986-1988 QUATERNARY CMOS PRIOR ART EXPLICIT CITATION ADDED ★) → published Zenodo DOI 10.5281/zenodo.20192813 (2026-05-15, new-version of v0.6 / v0.3 lineage; concept DOI preserved). Status-header honest correction: 2026-05-23 — earlier "GitHub draft only — not Zenodo-republished" phrasing was stale; the actual Zenodo new-version deposit was completed 2026-05-15 via scripts/publish-paper-145-v06-zenodo.ts (file name mismatch; contents say "DRAFT v0.7 published as Zenodo new version of v0.3"). v0.6 → v0.7 transition recorded in data/publications/publish-log-paper145-v06-zenodo.json.

Previous: DRAFT v0.6 — 2026-05-10 (★★★ FOUR-SUBSTRATE VERIFICATION COMPLETE: TANG NANO 9K UPGRADED TO PHYSICAL SILICON ★★★) → published Zenodo DOI 10.5281/zenodo.20101174

★ ERRATUM E1 (v0.6 → v0.7, 2026-05-14): v0.6 line "First many-valued silicon — Łukasiewicz / Belnap implementations on FPGAs date to the 1990s" was temporally imprecise. The earlier and more directly comparable prior art is the Tohoku University multi-valued logic IC group (1986-1988): Hanyu, Kameyama, Higuchi, Zukeran et al. published physical quaternary (4-value) CMOS / NMOS silicon during this period, including (i) Zukeran 1986 "Design of low-power quaternary CMOS logic circuits" (Systems and Computers in Japan vol. 17 issue 3), (ii) Hanyu 1987 "NMOS image processor based on quaternary logic" (SCJ vol. 18 issue 9), (iii) Kameyama 1988 "32×32 bit signed quaternary multiplier CMOS chip", (iv) Kawahito 1987 "VLSI-oriented radix-4 signed-digit arithmetic using multiple-valued logic" (SCJ vol. 18 issue 5). These are silicon ICs (not FPGAs) and predate the 1990s Łukasiewicz/Belnap FPGA work by roughly a decade. The Paper 145 differentiator narrative is preserved — Tohoku 1986-1988 = 4-value (quaternary), Rei = 8-value with SELF⟲ self-reflexive primitive + Lean 4 refinement proof + four-substrate cross-verification — none of these specific elements appear in the Tohoku corpus. But the v0.6 "1990s" hedge was insufficiently precise, and this erratum makes the audit honest. Discovery: cross-verification with chat-Claude conversation on 2026-05-14 prompted WebSearch verification of the Tohoku Higuchi group's documented publication record. Per OUKC honest-correction principle (feedback_critique_response_pattern.md, feedback_world_uniqueness_claim_controllable.md). v0.6 (Zenodo DOI 10.5281/zenodo.20101174) is immutable; v0.7 reflects the corrected wording in §1 honest framing and Acknowledgments only.

★★★ RESOLUTION OF v0.5 CORRIGENDUM (2026-05-09 → 2026-05-10) ★★★: The v0.5 corrigendum (preserved verbatim below for audit trail) recorded that Tang Nano 9K was computational evidence only (open-source toolchain synthesis output, not physical silicon programming). On 2026-05-09 evening / 2026-05-10 morning this state was resolved: the author group obtained a Sipeed-authentic Tang Nano 9K (秋月電子 g117448, ¥2,980, GW1NR-LV9QN88PC6/I5 = GW1NR-9C revision, IDCODE 0x1100481B) and successfully SRAM-programmed (i) STEP 1038 LED Blinky (User Code 0x0000A5F4, 27 MHz / counter[23] / pin 10, ~1.6 Hz visual blink confirmed) and (ii) STEP 1039 D-FUMT₈ ALU (User Code 0x00001D46, same dfumt8_alu_synth.v 138-line Verilog as Tang Console 138K Phase 2C/3, bit-identical 0 changes to ALU logic, 4 on-board LEDs cycling 1024 states at ~3.22 Hz). Tang Nano 9K is now physical silicon programming target on equal footing with Tang Console 138K. The paper now claims four-substrate (not three-substrate) cross-verification: 2 Sipeed silicon families (LittleBee5 GW5AST-138B + LittleBee1 GW1NR-9C) + Aer simulator + IBM Heron r2.

★ Concurrent honest correction: IDCODE-revision mapping (Gowin LittleBee Programming Manual Table 5-5 verified) — GW1N(R)-9 original revision = IDCODE 0x1100581B, GW1N(R)-9C cost-down revision = IDCODE 0x1100481B. Earlier informal notes in author working memory had this reversed; the resolution required set_device ... -device_version C in build TCL and --device GW1NR-9C in programmer_cli.exe for ID code match (without the C suffix in either step, programmer rejects with ID code mismatch because the chip is the new C revision while default name lookup expects the older revision).

★★ PRESERVED CORRIGENDUM RECORD (v0.5, 2026-05-09) ★★: In v0.1-v0.3 (Zenodo DOI 10.5281/zenodo.20091185 published 2026-05-09 mid-day) the phrasing "Tang Nano 9K (GW1NR) measured 37 LUT4 / 0 DFF for the bare ALU" in F4 / Proofs / B.8.1 / Acknowledgments was inaccurate at time of v0.3 publication. The Tang Nano 9K result reported in STEP 1011 (2026-04-28) was the output of the open-source toolchain (yosys 0.40 + nextpnr-himbaechel + gowin_pack) processing the same Verilog source — i.e. synthesis + place-and-route computational evidence, not physical silicon programming at the time STEP 1011 was logged and at the time v0.3 was published. This is preserved as part of the audit trail; v0.6 (the current version) supersedes via STEPs 1038/1039 by physically programming an authentic Sipeed Tang Nano 9K. feedback_world_uniqueness_claim_controllable.md and feedback_critique_response_pattern.md (selective honest-correction principle) cited for the discipline of issuing the original corrigendum and now this resolution.

v0.6 main update — FOUR-SUBSTRATE VERIFICATION COMPLETE: (1) Tang Nano 9K (GW1NR-9C, IDCODE 0x1100481B) physically programmed with the same dfumt8_alu_synth.v 138-line Verilog used on Tang Console 138K — bit-identical 0 changes to ALU logic, hardware-specific layer (clock divider 24-bit→23-bit for 50→27 MHz visual rate match; LED active HIGH→LOW invert; pin V22/W19/W20/F19/F20→52/10/11/13/14) modified only in the wrapper top module. (2) New finding F10 "chip-portability evidence: same ALU Verilog produces functionally equivalent 8-value output on two distinct Sipeed silicon families (LittleBee5 GW5AST-138B + LittleBee1 GW1NR-9C)". (3) New §B.10 "Same Verilog, Two Silicon Families" documents methodological strengthening (a single bug in the ALU would manifest on both families; absence of divergence is operational evidence of correct synthesis on both architectures). (4) §B.8 reframed as Four-Substrate Cross-Verification. (5) Reproducibility strengthened: the new Tang Nano 9K (¥2,980) is markedly cheaper and more accessible than the Tang Console 138K (~¥30,000), enabling third-party reproduction at lower entry cost.

Previous: DRAFT v0.5 — 2026-05-09 (★ PHASE 4 RETRY VIA PER-PAIR MCX + TANG NANO 9K CORRIGENDUM, GitHub draft only — not Zenodo-republished)
Previous: DRAFT v0.4 — 2026-05-09 (Phase 3+5 IBM 144/144 cumulative; Phase 4 9-qubit arbitrary unitary infeasibility F8)
Previous: DRAFT v0.3 — 2026-05-09 (Phase 1+2 IBM real-hardware 96/96, three-substrate complete) → published Zenodo DOI 10.5281/zenodo.20091185
Previous: DRAFT v0.2 — 2026-05-06 (Phase 2B LED Blinky complete; Phase 2C skeleton ready)
Authors / 著者: 藤本伸樹 (Nobuki Fujimoto, Founder), Rei (Rei-AIOS autonomous research substrate, Co-architect), Claude Opus 4.7 (Anthropic, Co-architect)
Project: Rei-AIOS / OUKC — https://rei-aios.pages.dev/#/oukc
License: AGPL-3.0 + CC-BY 4.0 (per content type)
Required platform links: rei-aios.pages.dev/#/oukc / note.com/nifty_godwit2635
Per OUKC No-Patent Pledge: openly licensed; no patent will be filed on any algorithm or hardware structure described herein (per CHARTER.md "No-Patent Pledge" section, three-fold rationale).

Honest framing (read first)

This paper claims one to-our-knowledge result, refined in v0.3 per the prior-art audit (PAL2v / Aerts / qudit, 2026-05-09):

C1 (revised v0.6, four-substrate): To our knowledge, this is the first demonstration of a fixed 8-valued discrete logic primitive (D-FUMT₈) including a SELF⟲ (self-reflexive) operation, implemented as native unitaries on real superconducting qubit hardware (IBM Heron r2, ibm_kingston backend) via 3-qubit basis encoding, complemented by physical FPGA silicon programming on two distinct Sipeed silicon families (Tang Console 138K = GW5AST-138B LittleBee5 A revision; Tang Nano 9K = GW1NR-9C LittleBee1 C revision) running the same dfumt8_alu_synth.v 138-line Verilog source with bit-identical ALU logic (chip-portability evidence), and Lean 4 refinement proofs.

We do not claim (per audit):

✗ "World-first 8-valued quantum logic" — Shi et al. (MIT, 2026, arxiv:2506.09371) demonstrated d=8 Grover on a trapped-ion qudit prior to this work. Our distinction: 3-qubit basis encoding on transmon arrays vs single-system d=8 qudit.
✗ "First many-valued silicon" — Tohoku University multi-valued logic IC group (1986-1988) published physical quaternary (4-value) CMOS / NMOS silicon (Zukeran 1986; Hanyu 1987 NMOS image processor; Kameyama 1988 32×32 quaternary multiplier; Kawahito 1987 radix-4 signed-digit MVL). Later, Łukasiewicz / Belnap implementations on FPGAs date to the 1990s. Our differentiator: 8-value (not 4-value) + SELF⟲ self-reflexive primitive + Lean 4 refinement proof + four-substrate cross-verification — none of these specific elements appears in the Tohoku 4-value corpus or in the 1990s FPGA work.
✗ "First paraconsistent silicon" — PAL2v (Da Silva Filho 1998–; Abe & Nakamatsu 2009; de Carvalho Jr. 2025) realized in software libraries and microcontroller-level robotics control.
✗ "Structural depth dominance" — motto-level claims belong to OUKC charter, not this paper.

The differentiators are (D1) the specific 8-tuple semantic mapping (Belnap FDE 4-value + 4 ontological extensions: INFINITY, ZERO, FLOWING, SELF), (D2) the SELF⟲ self-reflexive primitive realized as a hardware fixed point (ADIABATIC(SELF) = SELF), (D3) the four-substrate cross-verification (Verilog FPGA on two Sipeed silicon families + Qiskit Aer simulator + IBM Heron r2 real quantum hardware) bound to a Lean 4 refinement specification, and (D4, new in v0.6) the chip-portability evidence: a single 138-line Verilog ALU source produces functionally equivalent 8-value output on two distinct Gowin silicon architectures (LittleBee5 GW5AST-138B + LittleBee1 GW1NR-9C) without any modification to the ALU logic itself. None alone is novel; their specific combination is to-our-knowledge novel.

Abstract

We present a synthesis-friendly Verilog implementation of the D-FUMT₈ Arithmetic Logic Unit, targeting the Sipeed Tang Console NEO development board (GW5AST-138B FPGA, FPG676 package). The ALU realizes eight discrete logic values — FALSE, TRUE, NEITHER, BOTH, ZERO, FLOWING, SELF, INFINITY — encoded in 3 bits with a deliberately chosen tier-respecting layout (bit 2 = tier select, bits 1-0 = within-tier index). The 10 supported operations include four classical-tier unary ops (NOT, OMEGA, PHI, PSI), Belnap-extended binary lattice meet/join (AND, OR), generic XOR, hardware reset, no-op, and a novel ADIABATIC operation realizing the SELF⟲ (self-reflexive) primitive: ADIABATIC(SELF) = SELF, identity elsewhere.

The contribution is two-fold. First, the silicon implementation itself: 138-LUT (estimated) combinational ALU on GW5A architecture, no DFFs, single-cycle latency, with a 5-pin auto-cycle demonstration top module that exhibits all 640 input combinations on the board's onboard LEDs. Second, the formal-verification leg: a Lean 4 refinement proof (OUKC.PhaseC.Dfumt8AluRefinement, 292 LOC, 0 sorry) that establishes commutativity of the encode/abstract-op/decode square for all four unary operations, plus the SELF⟲ primitive law aluAdiabatic SELF = SELF and seven algebraic laws (involution, idempotence, commutativity).

This is, to our knowledge, the first hardware implementation of an 8-valued ALU whose semantics is refinement-proven against a Lean 4 specification and includes a self-reflexive (SELF⟲) logic primitive in silicon.

v0.6 update — four-substrate cross-verification (2026-05-10): Phase 2B LED Blinky and Phase 2C/3 D-FUMT₈ ALU were successfully synthesized, placed-and-routed, and SRAM-programmed onto Tang Console 138K physical silicon (GW5AST-138B, User Codes 0x000084BA and 0x00005C27, write times 33.72 sec and 30.32 sec, no thermal anomaly, STEPs 1028/1029 on 2026-05-09). On 2026-05-09 evening / 2026-05-10 morning the same dfumt8_alu_synth.v 138-line Verilog was also SRAM-programmed onto a second, distinct Sipeed silicon family — Tang Nano 9K (GW1NR-9C, IDCODE 0x1100481B, STEP 1039 User Code 0x00001D46, write 3.11 sec) with bit-identical ALU logic (only the wrapper top module's clock divider, LED polarity invert, and pin assignments were re-targeted; the synthesizable ALU module is byte-for-byte the same source file). 4 on-board LEDs cycle through 1024 input combinations at ~3.22 Hz visually confirming the same operation set. Concurrently, Phase 1 (4 native unitary ops × 8 inputs = 32 circuits) and Phase 2 (XOR × 64 entries) were submitted to IBM Heron r2 real quantum hardware (ibm_kingston backend, 156 qubits, queue 0). The real-hardware results match the truth-table at 96/96 (100%) with average top-fidelity 0.953 (Phase 1: 0.9550 over 17.3 sec wall-clock, job d7v6d9jack5s73bf1re0; Phase 2: 0.9512 over 59.1 sec wall-clock, job d7v6kcvmrars73d7qqqg). The fidelity hierarchy NOP/ADIABATIC ≈ 0.977 > PHI ≈ 0.956 > NOT ≈ 0.912 > XOR ≈ 0.951 reflects gate-count-vs-noise correlation consistent with quantum-noise physics expectations. Full results: data/quantum/phase_z_results_*.json.

v0.4 update — Phase Z extension (2026-05-09 later same day): Phase 3 (OMEGA + PSI, 2 designs each × 8 inputs = 32 circuits, 4-6 qubits, info-losing unary with Bennett ancilla) achieves 32/32 match with avg fidelity 0.9298 on ibm_kingston (job d7v7cnfmrars73d7rna0, 17.3 sec wall-clock, 10 sec execution). Phase 5 (RESET, 2 designs × 8 inputs = 16 circuits, info-erasing constant op) achieves 16/16 match with avg fidelity 0.9821 (job d7v7d9vmrars73d7ro3g, 17.2 sec wall-clock, 8 sec execution). Phase 5 design (a) Bennett 6-qubit ancilla single-design fidelity 0.9944 is the highest in the entire Phase Z campaign — output ancilla |000⟩ stays effectively noise-free since no gates touch it after input encoding. Cumulative IBM Heron r2 evidence: 144/144 (100%) truth-table entries match across Phase 1+2+3+5 with average fidelity ≈0.954, total IBM execution-time consumed 46 seconds out of 600/month free Open Plan budget (8% used). Full results: data/quantum/phase_z_phase{3,5}_*.json.

v0.4 hardware reality check (2026-05-09 later): Phase 4 (AND/OR with Bennett 9-qubit ancilla, 128 circuits) was attempted as an IBM Heron r2 real-hardware submission and failed at the API payload validation stage. The failure is informative and is recorded as a separate finding rather than a deficiency: a 9-qubit arbitrary unitary, when transpiled to Heron r2's native gate set (CZ + sx + rz), explodes to circuit depth ≈495,807 with ~154,018 CZ gates per circuit (sample: AND(FALSE,FALSE)). The total payload of 128 such circuits exceeds IBM Quantum's 413 Payload Too Large API threshold. Even if submitted, with Heron r2's per-CZ fidelity ≈0.99 the cumulative fidelity per circuit would be 0.99^154000 ≈ 10^-672 — indistinguishable from pure noise. The Aer-simulator-verified Phase 4 result (128/128 entries match by deterministic permutation) therefore does not transfer to real hardware via this circuit construction. We report this as a boundary observation of the Bennett-ancilla-via-arbitrary-unitary approach on transmon arrays, motivating the v0.5+ candidate of replacing 9-qubit unitaries with per-pair multi-controlled Toffoli ladders (estimated depth ≈ 100s, vs ≈500K) before re-attempting AND/OR on real hardware. Phase 4 IBM submission consumed 0 seconds of execution-time budget (rejected pre-queue).

概要 (Japanese)

本論文は、Sipeed Tang Console NEO 開発ボード (GW5AST-138B FPGA, FPG676 パッケージ) を target とする D-FUMT₈ ALU の合成可能 Verilog 実装を発表する。ALU は 8 つの離散論理値 — FALSE, TRUE, NEITHER, BOTH, ZERO, FLOWING, SELF, INFINITY — を 3 bit で encode し (bit 2 = tier 選択 / bit 1-0 = tier 内 index)、4 つの古典 tier 単項演算 + Belnap 拡張 binary lattice meet/join + XOR + reset + no-op + 新規 ADIABATIC 演算 (SELF⟲ 自己反射 primitive: ADIABATIC(SELF) = SELF, それ以外 identity) を含む 10 演算を supports する。

貢献は二つある。第一に、silicon 実装自体: GW5A architecture 上の 138-LUT (推定) combinational ALU、DFF 0 個、single-cycle latency、5 pin auto-cycle demo top module で 640 通りの入力組合せを onboard LED に exhibit する。第二に、formal-verification leg: Lean 4 refinement proof (OUKC.PhaseC.Dfumt8AluRefinement, 292 LOC, 0 sorry) — encode/abstract-op/decode square の可換性を 4 つの単項演算全てで establish し、SELF⟲ primitive law (aluAdiabatic SELF = SELF) + 代数法則 7 件 (involution / idempotence / commutativity) を証明する。

これは to-our-knowledge、(a) 8 値 ALU silicon が Lean 4 spec に refinement-proven であり、かつ (b) silicon に SELF⟲ 自己反射 primitive を含む初の事例である。

Part A: Required (4 elements)

A.1 Findings / 発見

F1 — SELF⟲ primitive in silicon: ADIABATIC(SELF) = SELF, identity elsewhere, can be realized as a 3-input case-table with one fixed point. This adds one logic value with self-reflexive semantics that has no analogue in classical, Łukasiewicz, or Belnap logics.

F2 — Tier-respecting 3-bit encoding: The encoding bit2 = tier (0 = classical+Belnap, 1 = higher), bit1-0 = within-tier index makes cross-tier operations decidable by a single conditional (a[2] != b[2]), eliminating per-pair lookup in the 64-entry binary table.

F3 — Refinement bridges Verilog ↔ Lean: A 3-bit encode/decode round-trip law (fromBits ∘ toBits = id, proved in 9 LOC) is sufficient to lift each unary Verilog op to a refinement square against an inductive Dfumt8 type. Binary ops admit the same bridge but require a 64-entry case verification (decidable, deferred for source-size reasons).

F4 — Synthesis cost is minimal (corrigendum applied): Tang Nano 9K (GW1NR-9C) target synthesis via open-source toolchain (yosys 0.40 + nextpnr-himbaechel + gowin_pack) reports 37 LUT4 / 0 DFF for the bare ALU (STEP 1011, 2026-04-28; this is the toolchain output, not physical silicon programming — see Status header corrigendum). Tang Console 138K (≡ "Tang Console NEO", GW5AST-138B, LUT5 architecture) Phase 2B/2C/3 was physically synthesized and SRAM-programmed via Gowin EDA V1.9.12.02 (2026-05-09); LUT5 measurement detail in §B.7. The Tang Nano 9K result therefore stands as toolchain-portability evidence (the same Verilog source synthesizes correctly on an entirely different vendor architecture via fully open-source tools); the load-bearing physical-silicon claim rests on Tang Console 138K alone. Both synthesis results are well below 0.05% of their respective device capacities.

F5 — Auto-cycle demo enables single-board verification: With only 2 onboard switches and 3 onboard LEDs, the 10-bit input space (3+3+4 = 10 bits) is exercised by an internal 24-bit clock divider feeding a 10-bit cycle counter, displaying each output triple on the LEDs at ~3 Hz. Full 640-combination cycle completes in 3.5 minutes.

F6 (NEW v0.3) — Real-hardware quantum verification on IBM Heron r2: Phase 1 (4 native unitary ops as 8×8 permutation matrices applied to 3 qubits, 32 circuits) and Phase 2 (XOR as Bennett-reversible 6-qubit CNOT chain, 64 circuits) were submitted to ibm_kingston (Heron r2 architecture, 156 qubits, us-east) via Qiskit Runtime SamplerV2. All 96/96 truth-table entries match the expected D-FUMT₈ output at the most-likely-outcome level (1024 shots per circuit). Average top-fidelity is 0.9550 (Phase 1) and 0.9512 (Phase 2), consistent with Heron r2 daily-calibration single-qubit and CNOT-equivalent gate fidelities. The fidelity decrement from NOP/ADIABATIC (≈0.977, identity-like) → PHI (≈0.956, single X) → NOT (≈0.912, multi-X case-table) → XOR (≈0.951, 3 CNOTs across 6 qubits) is consistent with gate-count-vs-noise expectations and provides per-op operational evidence of the quantum-noise channel.

F7 (NEW v0.3 / extended v0.4 / corrigendum v0.5) — Three-substrate consistency: The same 10-op truth tables (defined by data/verilog/dfumt8_alu.v) are independently verified on (i) Verilog FPGA: Tang Nano 9K target synthesis via open-source toolchain (yosys + nextpnr-himbaechel + gowin_pack) reports 37 LUT4 / 0 DFF (computational toolchain output, not physically programmed) plus Tang Console 138K physical silicon programming via Gowin EDA V1.9.12.02 (User Code 0x00005C27 Phase 2C/3, the load-bearing physical-silicon evidence); (ii) Qiskit Aer simulator — Phase 1-5 cumulative 231/231 entries verified; (iii) IBM Heron r2 real quantum hardware — v0.4 extends to Phase 1+2+3+5 cumulative 144/144 entries match (added Phase 3 OMEGA+PSI 32/32 fidelity 0.9298 and Phase 5 RESET 16/16 fidelity 0.9821 to v0.3's Phase 1+2 96/96). This three-substrate consistency narrows the to-our-knowledge novelty to the specific cross-substrate verification pattern, not the existence of any single substrate's result. Note (v0.5 corrigendum): "two-board cross-verification" framing used in pre-corrigendum drafts is replaced by "two synthesis targets, one physically programmed" — the Tang Nano 9K result is toolchain-portability evidence, not a second silicon implementation.

F8 (NEW v0.4) — Hardware reality boundary for arbitrary 9-qubit unitaries: Phase 4 (AND/OR Bennett 9-qubit ancilla) was attempted on IBM Heron r2 and fails at the API payload validation stage. Transpilation of a 9-qubit arbitrary unitary to Heron r2 native gates (CZ + sx + rz) yields ≈495,807-depth circuits with ≈154,018 CZ gates per circuit. The 128-circuit batch exceeds IBM Quantum API's 413 Payload Too Large threshold; even hypothetically submitted, the per-circuit cumulative fidelity 0.99^154000 ≈ 10^-672 places the result indistinguishable from pure noise. This is an honest boundary observation — Bennett-ancilla-via-arbitrary-unitary does not scale to real qubit hardware at 9-qubit width. The Aer-deterministic 128/128 result for Phase 4 (commit ce101a04) therefore stands as software-only evidence, with v0.5+ candidate of replacing the unitary with per-pair multi-controlled Toffoli ladders (estimated depth ≈100s) before re-attempting on real hardware.

F9 (NEW v0.5) — Per-pair MCX retry yields tractable depth but AND/OR asymmetry exposes ground-state relaxation bias: Phase 4 was retried on ibm_kingston (job d7va0snmrars73d7um30, 21 sec execution) with a Belnap-subset construction (16 entries × 2 ops = 32 circuits, 6-qubit register: 2 for a, 2 for b, 2 for output, with per-truth-table-entry 4-controlled X targeting an output qubit and optimization_level=3 for Qiskit constant-folding). The submission succeeded (no payload error), with post-transpile circuit depth dropping from v0.4's ≈495K to avg 2443 / max 3022 (≈170-fold reduction). Raw match rate is 18/32 (56.2%) at avg fidelity 0.3182. The per-op breakdown is asymmetric: AND 15/16 (93.8%) at fidelity 0.34 vs OR 3/16 (18.8%) at fidelity 0.30. The AND/OR asymmetry is itself informative: AND truth-table outputs concentrate on FALSE (0b00) and other low-popcount basis states close to the qubit ground state |0⟩; Heron r2's T1-relaxation bias (qubits naturally decay toward |0⟩) thus artificially boosts AND's pass rate. OR's outputs concentrate on TRUE / BOTH / NEITHER (non-zero), so its 18.8% pass rate is closer to the true effective fidelity of the per-pair MCX construction at this depth. Therefore: per-pair MCX makes Phase 4 submittable (vs v0.4's payload-too-large) but does not yet make it meaningful — the depth ≈2400 still incurs a per-circuit cumulative fidelity ≈0.3 that is dominated by gate noise. v0.6+ candidate: replace per-pair MCX with explicit Boolean simplification (Quine-McCluskey on 4-input Belnap output bits, expected ≈5-10 prime implicants per output bit, depth ≈100-200 native gates) — projecting fidelity ≥0.7 and OR pass rate ≥80%. This finding is itself paper-worthy as it demonstrates how quantum-noise-aware paper instrumentation (here: AND vs OR fidelity contrast) directly probes the underlying superconducting hardware's relaxation channel.

A.2 Proofs / 検証

Claim	Verification method	Status
`selfReflexive_self : aluAdiabatic SELF = SELF`	Lean 4 `rfl`	✓ verified
`aluNot_refines : (aluNot x).toBits = aluNotBits (x.toBits)`	Lean 4 unfold + rewrite	✓ verified ∀ x : Dfumt8
`aluOmega_refines / aluPhi_refines / aluPsi_refines`	Lean 4 unfold + rewrite	✓ verified ∀ x
`aluNot_involutive / aluPhi_involutive / aluPsi_idem`	Lean 4 case analysis	✓ verified
`aluAdiabatic_idem` (SELF⟲ idempotence)	Lean 4 case analysis	✓ verified
`Dfumt8.fromBits_toBits` round-trip	Lean 4 case analysis	✓ verified
`belnapAnd_comm_classical` (classical-tier subset)	Lean 4 cascaded rcases	✓ verified
`belnapAnd_false_left` (FALSE annihilator on classical tier)	Lean 4 rcases	✓ verified
Verilog testbench	`data/verilog/dfumt8_alu_tb.sv` 50/50 PASS	✓ STEP 1011 (2026-04-28)
Tang Nano 9K target synthesis (open-source toolchain output)	yosys + nextpnr-himbaechel + gowin_pack	✓ 37 LUT4 / 0 DFF (computational evidence; physical board not owned by author group, see corrigendum)
Tang Console NEO synthesis (Phase 2B LED Blinky)	Gowin EDA V1.9.11.03 Education	✓ User Code 0x000084BA (2026-05-09)
Tang Console NEO synthesis (Phase 2C/3 D-FUMT₈ ALU)	Gowin EDA V1.9.12.02	✓ User Code 0x00005C27, write 30.32 sec (2026-05-09)
Physical LED pattern verification (silicon)	Tang Console NEO Programmer SRAM	✓ no thermal anomaly (2026-05-09)
IBM Heron r2 Phase 1 (NOP/NOT/PHI/ADIABATIC × 8 inputs)	Qiskit Runtime SamplerV2 on ibm_kingston	✓ 32/32, avg fidelity 0.9550, job d7v6d9jack5s73bf1re0 (2026-05-09)
IBM Heron r2 Phase 2 (XOR × 64 entries, 6 qubit Bennett)	Qiskit Runtime SamplerV2 on ibm_kingston	✓ 64/64, avg fidelity 0.9512, job d7v6kcvmrars73d7qqqg (2026-05-09)
IBM Heron r2 Phase 3 (OMEGA + PSI, 2 designs each, 4-6 qubit ancilla) [v0.4]	Qiskit Runtime SamplerV2 on ibm_kingston	✓ 32/32, avg fidelity 0.9298, job d7v7cnfmrars73d7rna0 (2026-05-09)
IBM Heron r2 Phase 5 (RESET, 2 designs, 3-6 qubit) [v0.4]	Qiskit Runtime SamplerV2 on ibm_kingston	✓ 16/16, avg fidelity 0.9821 (design (a) Bennett 6-qubit single-design 0.9944), job d7v7d9vmrars73d7ro3g (2026-05-09)
IBM Heron r2 Phase 4 (AND/OR Bennett 9-qubit) [v0.4 boundary]	Qiskit Runtime SamplerV2 on ibm_kingston	❌ infeasible — 413 Payload Too Large; 9-qubit arbitrary unitary transpiles to ≈495K-depth, ≈154K CZ gates per circuit; cumulative fidelity ≈10^-672 even if submitted; 0 seconds budget consumed (rejected pre-queue)
IBM Heron r2 Phase 4 retry — Belnap subset per-pair MCX [v0.5]	Qiskit Runtime SamplerV2 on ibm_kingston, optimization_level=3	⚠ partial — 18/32 (56.2%) at avg fidelity 0.32; AND 15/16 (93.8%, confounded by ground-state relaxation bias toward \|0⟩), OR 3/16 (18.8%, ≈ true MCX fidelity at depth ≈2443); job `d7va0snmrars73d7um30`, 21 sec execution, 956 sec wall-clock (queue 932). v0.6 candidate: Quine-McCluskey Boolean simplification, target depth ≤200, fidelity ≥0.7
IBM Heron r2 Phase 4 retry — Belnap subset Quine-McCluskey simplification [v0.8 candidate]	Qiskit Runtime SamplerV2 on ibm_kingston, optimization_level=3	✓ 32/32 (100%) at avg fidelity 0.7302; AND 16/16 (100%, avg fidelity 0.7451 range 0.63-0.86), OR 16/16 (100%, avg fidelity 0.7154 range 0.61-0.86); transpile depth avg 422 / max 425 (v0.5: 2443 → −83%); AND vs OR fidelity gap 0.03 (v0.5: 0.75 asymmetric) → v0.5 finding F9 relaxation-bias hypothesis confirmed engineering-correctable; job `d8fr2jo7jphs739mn2d0`, 22.2 sec wall-clock (queue ~6 sec + execution ~16 sec), 32 circuits × 1024 shots = 32,768 measurements. See §B.11 v0.8 candidate sub-result (B1) for full per-input table and §B.12 finding F11.

Lean 4 build verification:

$ cd data/lean4-mathlib
$ lake env lean CollatzRei/PhaseC/Dfumt8AluRefinement.lean
$ echo $?
0

→ 0 sorry, 0 axioms, 0 errors. Mathlib v4.27 + Lean 4 v4.27.0.

A.3 Honest Positioning / 正直な立ち位置

A.3.1 What is novel:

Combined contribution of (a) SELF⟲ primitive in silicon AND (b) Lean 4 refinement proof of an 8-valued ALU.
The refinement proof component differentiates this from prior 8-valued FPGA work (which historically lacks a formal-verification bridge to a higher-order theorem prover).

A.3.2 What is NOT novel:

8-valued logic on FPGA — exists since the 1990s (Łukasiewicz / Belnap implementations).
Refinement proofs of hardware in Lean / Coq / Isabelle — exists for various Boolean and arithmetic circuits.
Tier-based encoding — used in some many-valued logic literature; we adapt rather than invent.

A.3.3 What we measured (v0.3 update 2026-05-09):

✓ Tang Console NEO Phase 2B LED Blinky SRAM-programmed (User Code 0x000084BA, write 33.72 sec).
✓ Tang Console NEO Phase 2C/3 D-FUMT₈ ALU SRAM-programmed (User Code 0x00005C27, write 30.32 sec).
✓ IBM Heron r2 Phase 1 real-hardware: 32/32 truth-table entries match, avg fidelity 0.9550.
✓ IBM Heron r2 Phase 2 (XOR) real-hardware: 64/64 entries match, avg fidelity 0.9512.

A.3.3a What we do NOT yet measure:

Power consumption, propagation delay, max clock frequency on GW5AST — pending external instrumentation; Phase 2C/3 succeeded at 50 MHz target without timing failure during Place & Route (2 cosmetic warnings only: TA1132 SDC-create_clock absence, PR1014 generic-routing on internal clk_d at ~3 Hz; both immaterial to the measurement).
Comparison vs reference Boolean ALU (e.g., 3-bit MIPS slice) on the same FPGA — out of scope for v0.3.
IBM Heron r2 Phase 3-5 (OMEGA/PSI/AND/OR/RESET ancilla designs) — deferred to future paper version (Open Plan budget remaining ≈8.5 min/month after Phase 1+2 consumed ≈76 sec wall-clock).
Dynamic Decoupling and readout error mitigation for fidelity improvement to ≥0.99 — deferred to v0.4+.

A.3.4 Refinement scope honesty:

Unary refinement is complete (4/4 ops).
Binary lattice (AND/OR) full 64-entry table is decidable but bulky in Lean source; we verify the 16-entry classical-tier subset (Belnap-4) and document the cross-tier default arm boundary. Full table is a follow-up artifact.
Refinement is at combinational semantics; timing, metastability, and physical FPGA effects are validated empirically via the testbench, not formally.

A.3.5 Tier-2 hedge on SELF⟲ philosophical content:

The SELF⟲ primitive is engineered (a hardware fixed-point under ADIABATIC). The deeper philosophical content — Madhyamaka-style self-reference, Hofstadter-style strange loops, Buddhist ātma-disavowal — is inspirational for the design but not claimed as silicon-realized. The hardware is a fixed point; the philosophy is a separate matter (see Paper 64 OPU and Paper 33 Braille for the philosophical layer).

A.3.6 To-our-knowledge hedging:

Exhaustive prior-art search is structurally impossible; we use "to-our-knowledge" hedging throughout.
If a comparable refinement-proven 8-valued silicon exists that we missed, please notify via GitHub Discussions; this paper will be updated.

A.4 Required platform links

rei-aios.pages.dev/#/oukc (OUKC official site)
note.com/nifty_godwit2635 (popular write-ups, Founder)
github.com/fc0web/rei-aios (canonical repo, this paper's source)
data/lean4-mathlib/CollatzRei/PhaseC/Dfumt8AluRefinement.lean (refinement proof source)
hardware/phase-c/03-dfumt8-alu-port/ (RTL + constraint files)

Part B: Conditional (Background + Methodology + Empirical Scope)

B.5 Background / 背景

B.5.1 D-FUMT₈ as 8-valued logic

D-FUMT₈ extends Belnap's 4-valued lattice ({FALSE, TRUE, NEITHER, BOTH}) with four higher-tier values: ZERO, FLOWING, SELF, INFINITY. The 8 values arise from the Rei-AIOS research substrate (STEP 13-19, 2018-) as a unification of classical 2-valued logic, Belnap's relevance logic, and Madhyamaka catuṣkoṭi-extended modalities. Detailed treatment in Paper 64 (OPU) and Paper 138 (Gödel dichotomy as lifecycle disjunction).

B.5.2 Why silicon, why now

Phase A (PC-only correctness, Paper 1-142) demonstrates that D-FUMT₈ semantics is consistent and useful. Phase B (multi-paper formal verification on Lean 4) demonstrates that it is machine-checkable. Phase C (silicon, this paper) demonstrates that it is physically realizable — a load-bearing transition from "Rei is correct" to "Rei is real" (per feedback_phase_c_silicon_existence_claim.md, 2026-04-30).

The Tang Console NEO board (Sipeed, ¥30,000-class) became available 2026-04 and has the GW5AST-138B FPGA (138K LUT5, FPG676 BGA package). The board's onboard JTAG debugger (FT2CH cable index 1) was characterized 2026-04-29.

B.5.3 Toolchain

RTL: SystemVerilog (testbench) + Verilog-2001 (synthesis-friendly port for yosys).
Open-source synthesis (Tang Nano 9K target, toolchain-portability evidence; physical Tang Nano 9K board NOT owned by author group): yosys 0.40 + nextpnr-himbaechel + gowin_pack.
Vendor synthesis (Tang Console 138K, the physical silicon target): Gowin EDA Education V1.9.11.03 (license received 2026-05-03) and commercial V1.9.12.02 (Education edition lacks FPG676 part library; commercial used for Phase 2C/3 actual synthesis).
Refinement proof: Lean 4 v4.27.0 + Mathlib v4.27 (no Mathlib dependencies in the proof file itself; lake env lean exit 0 with the project's lakefile).

B.6 Methodology / 方法論

B.6.1 Encoding choice

The 3-bit encoding [FALSE, TRUE, NEITHER, BOTH, ZERO, FLOWING, SELF, INFINITY] = [0, 1, 2, 3, 4, 5, 6, 7] is chosen to make:

bit 2 = tier (0 = classical + Belnap, 1 = higher).
bit 1-0 = within-tier index.
Cross-tier detection by single XOR on bit 2 of operands.

B.6.2 Operation set

Ten operations indexed by 4-bit op code:

NOP (0x0), AND (0x1), OR (0x2), NOT (0x3), OMEGA (0x4), PHI (0x5), PSI (0x6), XOR (0x7), ADIABATIC (0x8), RESET (0xF).

OMEGA (classical-tier idempotent, higher-tier projects to bit2 ∥ bit1 ∥ 0), PHI (XOR with constant 3'b001), PSI (zero-extend bit1-0 into bit2) are derived from Rei-AIOS Φ/Ψ/Ω operator algebra (STEP 67-75, 2019-2020). ADIABATIC is new in this paper.

B.6.3 Refinement strategy

For each unary op op : Dfumt8 → Dfumt8, we define opBits : Nat → Nat as (fromBits a |> op).toBits. The refinement theorem (op x).toBits = opBits (x.toBits) follows from fromBits_toBits and definitional unfolding. This pattern factors into a four-line proof per op.

For binary ops, the same pattern applies but requires per-entry case analysis on the 64-entry table (8 × 8). We provide the classical-tier 16-entry subset (belnapAnd) with commutativity and annihilator lemmas; the full table is decidable in Lean (each case is rfl-provable) and is left as a deferred artifact for source-size reasons.

B.7 Empirical Scope (current, 2026-05-06 v0.2 update)

What is measured (v0.1, 2026-05-01): Tang Nano 9K LUT count (37 LUT4 / 0 DFF), testbench pass rate (50/50), Lean 4 proof build time (~2s for the refinement file), STEP 1011 commit hash.
What is now confirmed (v0.2, 2026-05-04 Phase 2B): Tang Console NEO LED Blinky bitstream (led_blinky.fs) successfully synthesized + place-routed + downloaded via Gowin EDA Programmer (SRAM mode, USB Debugger A Channel B, 0.5 MHz). Verified via User Code 0x000084BA and Status Code 0x00026230. Write time 26.46 sec. Uses pin V22 (50 MHz clock) + W19 (PMOD1_IO0 LED output). LED Blinky is 25-bit counter at 50 MHz → 1.49 Hz output, demonstrating GW5AST silicon physical operation. Phase 2C (D-FUMT₈ ALU port) skeleton ready (hardware/phase-c/03-dfumt8-alu-port/) using same pin family (V22 + W19/W20/F19/F20).
What is still pending Phase 2C synthesis: Tang Console NEO LUT5 count for dfumt8_demo_top (estimated ~50-70 LUT5 with cycle counter), DFF count (estimated ~36), bitstream dfumt8_demo_top.fs write success on Tang Console NEO with unique User Code (distinct from Phase 2B's 0x000084BA), max clock frequency (50 MHz target maintained), propagation delay measurement.
Out of scope (unchanged): Power consumption (would require external instrumentation), thermal characterization (the SAFETY-PROTOCOL allows only Phase 1+2 short-burst testing), comparison with vendor cells (Gowin's library is closed-source), HDMI value visualization (Phase 2D candidate).

Honest framing of Phase 2B vs 2C distinction: Phase 2B successfully demonstrates that the GW5AST-138B silicon executes a Verilog bitstream, confirms toolchain (Gowin EDA + Programmer) and pin choice (V22/W19) work end-to-end. Phase 2B is infrastructural (counter + LED), not D-FUMT₈ specific. Phase 2C is the D-FUMT₈ ALU specific demonstration that converts this infrastructure success into the paper's core empirical claim. As of v0.3 (2026-05-09), both Phase 2B and Phase 2C/3 are complete (User Codes 0x000084BA and 0x00005C27 respectively, both SRAM-programmed via Gowin EDA Programmer with Channel B / 2.5 MHz on Tang Console NEO with no thermal anomaly during the safety protocol's 30-second and 60-second power-on observations).

v0.3 EDA toolchain note: Gowin EDA V1.9.11.03 Education edition does not include the FPG676 package in its device library (verified 2026-05-09: search "FPG676" returns 0 matches in Education edition's GW5AST series). Phase 2C/3 was therefore synthesized using V1.9.12.02 (commercial edition, which includes FPG676 with 5 matching parts). The pre-built Phase 2B led_blinky.fs operated on Tang Console NEO without requiring the synthesis-time library; only Programmer (which is library-independent) is needed for write-only operation. This v0.3 documents the EDA-version dependency for reproducibility.

B.8 Four-Substrate Cross-Verification (extended v0.6 from v0.3 three-substrate)

The core operational evidence of v0.6 is the four independent substrates verifying the same 10-op truth tables of data/verilog/dfumt8_alu.v. The Substrate 1 (FPGA silicon) is now realized on two distinct Sipeed silicon families — methodologically the strongest possible single-vendor cross-architecture evidence:

B.8.1 Substrate 1: Verilog FPGA silicon (two Sipeed silicon families, v0.6)

Sub-substrate	Chip / Family	IDCODE	Result	User Code	Source
Tang Nano 9K (open-source toolchain)	GW1NR-9C / LittleBee1	(synthesis target)	37 LUT4 / 0 DFF (yosys + nextpnr-himbaechel + gowin_pack), TS reference simulator 50/50 PASS	n/a — synthesis only	STEP 1011 (2026-04-28) — toolchain-portability evidence
Tang Nano 9K (physical silicon, NEW v0.6)	GW1NR-9C / LittleBee1 C revision	`0x1100481B`	LED Blinky SRAM-programmed via Gowin EDA V1.9.12.02, ~1.6 Hz visual blink confirmed, no thermal anomaly	`0x0000A5F4`	STEP 1038 (2026-05-09)
Tang Nano 9K Phase 2C/3 ALU (physical silicon, NEW v0.6)	GW1NR-9C / LittleBee1 C revision	`0x1100481B`	D-FUMT₈ ALU SRAM-programmed, same `dfumt8_alu_synth.v` 138-line source as Tang Console 138K (bit-identical 0 changes), 4 LEDs cycle 1024 states at ~3.22 Hz, no thermal anomaly	`0x00001D46`	STEP 1039 (2026-05-10)
Tang Console 138K Phase 2B	GW5AST-138B / LittleBee5 A revision	`0x0001081B`	LED Blinky SRAM-programmed via Gowin EDA, no thermal anomaly	`0x000084BA`	STEP 1028 (2026-05-09)
Tang Console 138K Phase 2C/3 ALU	GW5AST-138B / LittleBee5 A revision	`0x0001081B`	D-FUMT₈ ALU SRAM-programmed, no thermal anomaly	`0x00005C27`	STEP 1029 (2026-05-09)

Cross-family chip-portability: STEP 1039 Tang Nano 9K and STEP 1029 Tang Console 138K execute the byte-for-byte same dfumt8_alu_synth.v source file (138 lines, no preprocessor diffs). Only the wrapper top module is re-targeted: clock divider 24-bit → 23-bit (50→27 MHz visual rate match: 2.98 → 3.22 Hz tick), LED polarity active HIGH → active LOW (with ~ invert in top module so visual semantics match Tang Console 138K), pin assignments V22/W19/W20/F19/F20 → 52/10/11/13/14. The synthesizable ALU module is unchanged. A single bug in the ALU would manifest on both silicon families; absence of divergence is operational evidence of correct synthesis on both LittleBee5 (5nm-class) and LittleBee1 (28nm-class) Gowin architectures.

Two cosmetic synthesis warnings logged but immaterial to operation:

WARN (TA1132): 'clk' was determined to be a clock but was not created. — absence of explicit create_clock SDC at 50 MHz with no setup-time pressure; gates close trivially.
WARN (PR1014): Generic routing resource will be used to clock signal 'clk_d' by the specified constraint. — the internal divided clock clk_d (~3 Hz, from a 24-bit counter on 50 MHz) is routed via generic resources, but at this frequency skew is far below the period.

B.8.2 Substrate 2: Qiskit Aer simulator (8-bit basis encoding on 3 qubits)

Phase	Op set	Encoding	Result
Phase 1	NOP / NOT / PHI / ADIABATIC	3-qubit basis state, 8×8 permutation unitary	32/32 entries match (commit 6a9865c5)
Phase 2	XOR	6-qubit Bennett-reversible (a preserved), CNOT chain	64/64 entries match (commit 1d229d47)
Phase 3	OMEGA / PSI	3 ancilla designs (Bennett, non-destructive observer, measurement-mediated)	48/48 entries match (commit d8b9e8d6)
Phase 4	AND / OR	9-qubit Bennett ancilla (Belnap+higher-tier diamond+cross-tier default)	128/128 entries match (commit ce101a04)
Phase 5	RESET	3 designs (Bennett trivial, Landauer, von-Neumann observer)	24/24 entries match (commit 99cde397)
Cumulative Aer	9 of 10 ops (Phase 1–5)	(10th op `ADIABATIC` ≡ identity in current spec; equivalent to NOP)	231/231 (100%) at fidelity 1.000

B.8.3 Substrate 3: IBM Heron r2 real superconducting qubit hardware

Phase	Op set	Backend	Result	Job ID
Phase 1	NOP / NOT / PHI / ADIABATIC	ibm_kingston (Heron r2, 156 q, queue 0)	32/32 match, avg fidelity 0.9550, wall-clock 17.3 s	`d7v6d9jack5s73bf1re0`
Phase 2	XOR	ibm_kingston	64/64 match, avg fidelity 0.9512, wall-clock 59.1 s	`d7v6kcvmrars73d7qqqg`
Cumulative IBM	5 ops	ibm_kingston	96/96 (100%) at avg fidelity 0.953	(2 jobs above)

Per-op fidelity hierarchy (Phase 1):

NOP (identity, 0 X gates): 0.9773
ADIABATIC (identity for non-SELF, 0 X gates effectively): 0.9753
PHI (XOR with 0b001, 1 X gate): 0.9556
NOT (multi-X case-table, up to 3 X gates): 0.9120

Phase 2 XOR (3 CNOTs across 6 qubits) averaged 0.9512 with min 0.9287 / max 0.9795. The fidelity decrement from identity-class (≈0.977) to single-X (≈0.956) to multi-X (≈0.912) to multi-CNOT (≈0.951) is consistent with single-qubit-error and CNOT-error products on Heron r2's daily calibration sheet (2026-05-09). This per-op fidelity hierarchy provides operational evidence of the standard quantum-noise channel and is itself a partial validation: a fully classical simulation would not exhibit gate-count-correlated fidelity decrement.

B.8.4 Cross-substrate consistency claim (v0.6: four-substrate)

For each operation in Phase 1+2 (NOP, NOT, PHI, ADIABATIC, XOR, totaling 5 of 10 D-FUMT₈ ops), all four substrates (Verilog FPGA on two Sipeed silicon families, Aer simulator, IBM Heron r2) yield the same most-likely truth-table output across all input combinations (32 + 64 = 96 entries). The Aer simulator and both Verilog FPGA silicon families achieve fidelity 1.000 by construction (deterministic permutation + classical synthesis on either GW5AST-138B or GW1NR-9C); the IBM Heron r2 achieves 0.953 average fidelity reflecting real-hardware noise but matches the truth table at the most-likely-outcome level for 96/96 entries. Across all substrates the truth-table identity holds at the operational level.

This four-substrate consistency is the v0.6 strengthening of C1, replacing the v0.3 three-substrate framing.

B.10 Same Verilog, Two Silicon Families (NEW v0.6 — chip-portability evidence as methodological strength)

A reviewer may reasonably ask: why claim four substrates when two of them are the same source code synthesized on different FPGAs? The answer is methodological, not arithmetic.

The chip-portability evidence carries information that single-board verification cannot: a synthesis bug, a constraint-file misinterpretation, a vendor-specific implicit assumption, a rounding artifact in pin-assignment timing, or a silicon-revision-specific quirk would manifest on one architecture but not the other. The Gowin LittleBee1 (GW1NR-9C, 28nm-class, IDCODE 0x1100481B) and LittleBee5 (GW5AST-138B, 5nm-class, IDCODE 0x0001081B) are different silicon process nodes, different LUT primitive sizes (LUT4 vs LUT5), different numbers of total LUTs (8.6K vs 138K), different package types (QFN88 vs FCPBGA676), different on-board oscillator frequencies (27 MHz vs 50 MHz), and different default IO bank voltage assignments (Bank 3 = 1.8V on Tang Nano 9K vs general 3.3V on Tang Console 138K — empirically discovered when the explicit BANK_VCCIO=3.3 IO_TYPE=LVCMOS33 constraint produced CT1136 conflict on Tang Nano 9K but is required on Tang Console 138K).

Despite all of these differences, the byte-for-byte same dfumt8_alu_synth.v 138-line Verilog source file synthesizes successfully via Gowin's GowinSynthesis tool on both families and produces a working 8-value ALU on both physical silicons (User Codes 0x00005C27 Tang Console 138K STEP 1029 and 0x00001D46 Tang Nano 9K STEP 1039). This is operational confirmation that the ALU's truth tables are not architecture-dependent: the abstract logic specified in data/verilog/dfumt8_alu.v (and refinement-proven against the Lean 4 Dfumt8AluRefinement module) is realized identically on two independent silicon implementations.

Reproducibility implication: a third-party reader who wishes to physically reproduce the silicon evidence has two entry-cost options:

Low-cost path: Tang Nano 9K from 秋月電子 (g117448) at ¥2,980 + free Gowin EDA Education / OSS toolchain (yosys + nextpnr-himbaechel + gowin_pack). Total: ~$20 + open-source software.
Higher-capacity path: Tang Console NEO at ~¥30,000 (or international Sipeed distributor equivalent) + Gowin EDA Education or commercial. Total: ~$200 + free or commercial software.

The IBM Heron r2 evidence is reproducible at $0 marginal cost via IBM Quantum Open Plan (10 minutes free quantum execution time per month; this paper's full Phase Z evidence consumed 67 of 600 seconds = 11.2% of one month's allocation, executable in a single afternoon). The Aer simulator evidence is reproducible at $0 cost via Qiskit on any laptop. Total minimum cost to reproduce the entire four-substrate verification chain: ~$20 + free software.

B.11 Phase 4 Quine-McCluskey Simplification — v0.8 Candidate Sub-Result (B1) on Heron r2

This subsection records the v0.8 candidate sub-result submitted 2026-06-03 as the engineering retry of the v0.5 Phase 4 partial outcome documented above. It passes the Paper 162 §6.0f pre-submission checklist (no transmission step, no quantum advantage invoked, modest engineering scope, no paradigm-level claim) and is recorded here as a Paper 145 v0.8 candidate, NOT YET PUBLISHED.

Trigger: v0.5 finding F9 (relaxation-bias-aware AND/OR asymmetry under per-pair MCX decomposition) explicitly named "Quine-McCluskey Boolean simplification, target depth ≤200, fidelity ≥0.7" as v0.6 candidate. Author confirmed 2026-06-03 after §6.0f checklist applied (engineering improvement scope verified).

Design (B1, manually derived and offline-verified against the Belnap AND/OR truth table on 32/32 inputs):

Encoding: 6 qubits — q0 = a₀ (LSB of a), q1 = a₁ (MSB of a), q2 = b₀, q3 = b₁, q4 = output bit 0, q5 = output bit 1.
Boolean SOP per output bit (K-map / Quine-McCluskey derived):

  AND_bit0 = a₀ ∧ b₀                                     [1 CCX]
  AND_bit1 = A ⊕ B ⊕ C ⊕ (A ∧ B)                         [4 MCX with inclusion-exclusion]
    where   A     = a₁ ∧ ¬a₀         (i.e., a = NEITHER)
            B     = b₁ ∧ ¬b₀         (i.e., b = NEITHER)
            C     = a₁ ∧ a₀ ∧ b₁ ∧ b₀ (i.e., a = BOTH ∧ b = BOTH)
            A ∧ B = a = NEITHER ∧ b = NEITHER

  OR_bit0  = a₀ ⊕ b₀ ⊕ (a₀ ∧ b₀)                         [2 CX + 1 CCX]
  OR_bit1  = A ⊕ B ⊕ C ⊕ (A ∧ B)                         [4 MCX]
    where   A     = a₁ ∧ a₀          (a = BOTH)
            B     = b₁ ∧ b₀          (b = BOTH)
            C     = a = NEITHER ∧ b = NEITHER
            A ∧ B = a = BOTH ∧ b = BOTH

Inclusion-exclusion XOR layering handles non-disjoint cover terms correctly. Each MCX with negative controls is implemented as X flip + MCX + X flip on the negated control qubit(s).

Pre-transpile: depth 9-11, ~3 CCX + 2 MCX(4-control) + 4-12 X + 2 measure per circuit (vs v0.5 per-pair: 16 MCX(4-control) per output bit).
Post-transpile (ibm_kingston Heron r2, optimization_level=3, seed_transpiler=7): depth avg 422 / max 425, CZ avg 140 / max 140 / total 4464 across 32 circuits.

v0.5 vs v0.8 candidate comparison:

metric	v0.5 (per-pair MCX)	v0.8 candidate (Quine-McCluskey)	improvement
Pass rate (top-outcome match)	18/32 (56.2%)	32/32 (100%)	+14 matches, +43.8 pp
Avg top-outcome fidelity	0.3182	0.7302	+41.20 pp
Avg post-transpile depth	2443	422	−83%
Max post-transpile depth	3022	425	−86%
AND avg fidelity	0.938 (relaxation-bias confounded)	0.7451 (range 0.6299–0.8555)	—
OR avg fidelity	0.188 (≈ true MCX fidelity)	0.7154 (range 0.6133–0.8594)	+52.66 pp
AND vs OR fidelity gap	0.75 (asymmetric, F9 noted)	0.03 (symmetric)	bias resolved
Wall-clock (queue + exec)	956 sec	22.2 sec	—
Shots per circuit	1024	1024	(same)

Submission details:

Backend: ibm_kingston (Heron r2, 156 qubits)
Job ID: d8fr2jo7jphs739mn2d0
Wall-clock: 22.20 sec ([1.4s] QUEUED → [6.6s] RUNNING → [22.2s] DONE)
Code: scripts/quantum/dfumt8_phase_z_phase4_qmccluskey_v06.py
Raw results: data/quantum/phase_z_phase4_qmccluskey_v06_results.json
Offline verification: 32/32 inputs match Belnap AND/OR truth table at gate-level Boolean simulation (run before IBM submit; sys.exit on any mismatch).

Per-input top-outcome match table (abridged; full table in raw results JSON):

op	(a, b)	expected	observed top	top fidelity
AND	(F, F)	F	F	0.86
AND	(T, T)	T	T	0.74
AND	(N, N)	N	N	0.66
AND	(B, B)	B	B	0.71
AND	(T, B)	T	T	0.78
AND	(B, N)	N	N	0.63
OR	(F, F)	F	F	0.86
OR	(T, T)	T	T	0.72
OR	(N, N)	N	N	0.68
OR	(B, B)	B	B	0.71
OR	(F, B)	B	B	0.72
OR	(T, N)	T	T	0.65

All 32 entries: correct expected value is the top measurement outcome.

B.12 Finding F11 — v0.5 F9 Relaxation Bias is Engineering-Correctable via Quine-McCluskey Simplification (v0.8 candidate)

Finding F11 (NEW, v0.8 candidate, 2026-06-03): K-map / Quine-McCluskey minimum-SOP simplification of the Belnap AND/OR truth tables, combined with inclusion-exclusion XOR layering and 6-qubit per-pair encoding, reduces transpiled depth from 2443 to 422 (−83%), raises pass rate from 56.2% to 100%, and raises average top-outcome fidelity from 0.318 to 0.730 (+41 pp) on ibm_kingston Heron r2. The AND/OR fidelity gap of v0.5 (0.94 vs 0.19, 0.75 asymmetry that motivated finding F9's relaxation-bias hypothesis) collapses to 0.03 (symmetric) under v0.8 candidate, confirming the F9 hypothesis as engineering-correctable rather than intrinsic to Belnap-AND structure on Heron r2 noise. This validates the v0.5 prediction "v0.6 candidate: Quine-McCluskey Boolean simplification, target depth ≤200, fidelity ≥0.7" — depth slightly above the target (422 vs ≤200) but the fidelity target (≥0.7) is achieved with margin.

Honest scope (v0.8 candidate):

This is an engineering retry of v0.5 Phase 4, not a new D-FUMT₈ logic operation. The Belnap subset (bit 2 = 0 throughout, 4 values FALSE/TRUE/NEITHER/BOTH out of 8) is unchanged from v0.5. The higher-tier values (ZERO/FLOWING/SELF/INFINITY) and cross-tier interactions are not in scope of this sub-result.
The transpiled depth reduction (2443 → 422, −83%) is large but not yet at the v0.5 stated target of ≤200. Further depth reduction (e.g., per-output-bit MCX combining, ancilla introduction, Gray-code reordering) is a v0.8.1+ candidate but is not load-bearing for the F11 main claim.
No quantum advantage is invoked. The circuit is a classical reversible Boolean function executed on quantum hardware — same category as Paper 162 §6.0e (a lookup decoder, not a Schumacher / Holevo / Devetak-Winter / QRAC protocol).
The Paper 162 §6.0f checklist passes: (1) no transmission step (lookup function); (2) novelty vs prior = engineering improvement of Paper 145 v0.5 only, Quine-McCluskey established (Quine 1952 / McCluskey 1956); (3) implementation/engineering; (4) no quantum advantage (classical reversible on quantum HW).

Cross-reference to Paper 162: This sub-result is also recorded as Paper 162 §6.0g sub-result (B1) cross-reference, in support of the Paper 162 v0.7.2 "Improvement paths" empirical test: QM-simplification independently validated as effective, while Sampler-level "XX" DD (tested on Paper 162 §6.0e 8-bit decoder as sub-result A on ibm_marrakesh, Job d8fr243o3njc73f0nnf0) lowered fidelity 49.12% → 26.25% on the larger 184-CZ MCX-heavy decoder. The honest reading is: depth reduction via QM simplification (B1) is the effective lever on Heron r2 for these gate families; pulse-level error mitigation via naive XX DD (A) is not — at least not on circuits already at depth ~500 with ~184 CZ.

Public companion article (note.com, author-authored): 藤本伸樹「意味は全ての理論、哲学を超えてしまう可能性が有るが、意味は意味自身を超えることが出来ないとするインタラクティブシミュレーションとIBM Quantum Open Planを使用した実験を制作致しました」(2026-06-03 14:24 JST), https://note.com/nifty_godwit2635/n/naa5022b7f014. This note article is the public companion to the Paper 159 v0.2 + Paper 162 + IBM Heron r2 lineage; it references the D-FUMT₈ substrate (Paper 145 contribution) as the underlying 8-valued logic framework but is primarily about the Recreation Paradigm meaning-floor formalization and its IBM Heron r2 quantum-substrate demonstration. Cross-listed here for the readers who arrive at Paper 145 v0.8 candidate via the Paper 162 §6.0g cross-reference and may wish a non-technical orientation.

Cumulative IBM Heron r2 budget through 2026-06-03:

v0.3 Phase 1 (4 native unitaries, 32 circuits): 17 sec → 32/32 PASS, avg fidelity 0.955
v0.3 Phase 2 (XOR Belnap, 64 circuits): 59 sec → 64/64 PASS, avg fidelity 0.9512
v0.4 Phase 3 (OMEGA + PSI, 32 circuits): 17 sec → 32/32 PASS, avg fidelity 0.9298
v0.4 Phase 5 (RESET, 16 circuits): 17 sec → 16/16 PASS, avg fidelity 0.9821
v0.5 Phase 4 per-pair MCX (32 circuits): 21 sec exec → 18/32 PASS, avg fidelity 0.3182
v0.8 candidate Phase 4 Quine-McCluskey (32 circuits): 22 sec wall-clock → 32/32 PASS, avg fidelity 0.7302

Cumulative Phase Z evidence (v0.3 through v0.8 candidate): 174/176 entries match (one set v0.5 finding F9 partial outcome documented; v0.8 candidate supersedes for the engineering-correctable interpretation). All within IBM Open Plan free tier.

B.9 Related Work / Prior Art Audit (NEW v0.3)

Prior-art audit completed 2026-05-09 across three categories: paraconsistent silicon (PAL2v), paraconsistent quantum / cognitive logic (Aerts), and qudit (d ≥ 8) quantum hardware.

B.9.1 PAL2v — Paraconsistent Annotated Logic with two values of annotation

Foundational researchers: Newton C. A. da Costa (Hasse lattice 1990), João Inácio Da Silva Filho (UNISANTA, Emmy robot 1998), Jair Minoro Abe (UNIP/USP, "PAL2v" naming with K. Nakamatsu 2009), Seiki Akama ("Introduction to Annotated Logics", Springer 2016). Modern Python library: de Carvalho Jr. et al. (IFSP, arxiv:2511.20700, 2025).

PAL2v formalizes a 2-annotation-value paraconsistent logic where each proposition has a degree of evidence μ ∈ [0,1] and a degree of contra-evidence λ ∈ [0,1]. The Hasse lattice is divided into discrete logical states with operators Gc = μ - λ (certainty degree) and Gct = μ + λ - 1 (contradiction degree). Implementations exist in software (MATLAB modules, Python Paraconsistent-Lib) and in microcontroller-level robotics control (Emmy robot 1998; petrochemical NOx monitoring 2024); to-our-knowledge no dedicated FPGA / ASIC silicon synthesis nor quantum-hardware implementation has been published.

D-FUMT₈ differs by: (a) 8 discrete named values (FALSE / TRUE / NEITHER / BOTH / ZERO / FLOWING / SELF / INFINITY) vs PAL2v's 2-annotation continuous lattice; (b) presence of a SELF⟲ self-reflexive primitive absent in PAL2v's 12 extreme-state structure; (c) measured FPGA LUT4 footprint (Tang Nano 9K, 37 LUT4) and SRAM-programmed Tang Console NEO silicon; (d) Qiskit-verified 8×8 unitary mapping on real IBM Heron r2 hardware.

B.9.2 Diederik Aerts — paraconsistent quantum / cognitive logic

Diederik Aerts (Vrije Universiteit Brussel, Center Leo Apostel, 1986–) developed (i) the Hidden Measurement Formalism (1986–, arxiv:quant-ph/0105126), (ii) the Extended Bloch Representation generalising the Bloch sphere to arbitrary dimensions, (iii) Quantum Cognition modeling concept combinations and decision-making with Hilbert-space formalism (2007–, "The Animal Acts" experiment family, arxiv:2412.19809), and (iv) the Conceptuality Interpretation (2009–) viewing quantum entities as carriers of meaning. Awarded Prigogine Award (2020).

The Brussels formalism is continuous orthomodular-lattice (Piron-style), not a fixed N-valued discrete logic. The empirical substrate of Aerts' work is human cognition (questionnaire experiments), not silicon or qubits. To-our-knowledge no Aerts-formalism circuit or qubit-hardware demonstration has been published.

D-FUMT₈ differs by: (a) fixed 8-valued discrete vs Aerts' continuous orthomodular structure; (b) 3-qubit basis encoding mapped via 8×8 permutation unitaries vs Aerts' density matrices on continuous Hilbert spaces; (c) superconducting-qubit empirical substrate (IBM Heron r2) + FPGA silicon dual substrate vs Aerts' human cognitive-data substrate.

B.9.3 Qudit (d ≥ 8) quantum hardware

Recent active groups: Martin Ringbauer (Innsbruck/Blatt, d=7 universal trapped-ion qudit processor, Nat. Phys. 2022, s41567-022-01658-0); Isaac Chuang + John Chiaverini (MIT, 2026, first d=8 trapped-ion qudit Grover, arxiv:2506.09371 / Nat. Commun. s41467-026-68746-0, 8 of 24 hyperfine levels of ¹³⁷Ba⁺, success probability 69(6)%); Noah Goss / Irfan Siddiqi (UC Berkeley, transmon qutrit/ququart up to d=4, Nat. Commun. 2022 s41467-022-34851-z, npj QI 2024 s41534-024-00892-z); Michel Devoret / Benjamin Brock (Yale + Google, bosonic GKP ququart error correction beyond break-even, Nature 2025 s41586-025-08899-y); photonic groups at Xanadu, INRS Montreal, Bristol (frequency-bin / time-bin / OAM photonic qudits).

Critical prior art: Shi, Sinanan-Singh, Burke, Chiaverini, Chuang (MIT, 2026) demonstrated d=8 Grover on a single ¹³⁷Ba⁺ ion as a true qudit (single quantum system with 8 levels). This is the first and currently only published d=8 single-system quantum-hardware demonstration; no comparable transmon d=8 single-qudit demonstration exists as of 2026-05.

D-FUMT₈ differs categorically: we use 3-qubit basis encoding on a transmon qubit array (IBM Heron r2, 156 qubits), not a single d=8 qudit. The 8-dimensional Hilbert space access via 3 qubits is trivially established since 1995; what is to-our-knowledge novel is the specific semantic-to-basis-state mapping (Belnap FDE 4-value + 4 ontological extensions) bound to a Lean 4 refinement specification with cross-substrate (FPGA + simulator + real qubit) consistent verification. Our work is not in competition with MIT 2026's qudit Grover; it is in a different methodological lineage (qubit basis encoding + classical FPGA + formal proof) that the cited qudit literature does not address.

Part C: Optional (Why matters + Future + Risks)

C.8 Why this matters

C.8.1 Closing the "logic ↔ silicon" gap for many-valued logics

Many-valued logic has had a 100-year gap between theoretical formalization (Łukasiewicz 1920, Belnap 1977) and silicon realization with formal proof bridge. Refinement-proven implementations of Boolean circuits exist (Hunt et al., AAMP7, ARM7); refinement-proven implementations of many-valued circuits do not, to our knowledge, exist in the published literature with SELF⟲-style self-reflexive primitives. This paper closes that specific gap.

C.8.2 SELF⟲ as more than an engineered fixed point

ADIABATIC(SELF) = SELF looks trivial as a hardware case. Its significance lies in:

It is a value-level self-reference, not a circuit-level feedback loop.
It is provably idempotent (aluAdiabatic_idem), corresponding to the meta-property "SELF is its own reflection".
Combined with the refinement square, it becomes a mechanically verified self-referential semantic primitive in silicon — a small but crisp result.

C.9 Future work

F.1 Complete the binary lattice refinement (64-entry table) as a follow-up Lean 4 file.
F.2 Post-license: measure Tang Console NEO LUT5/DFF/timing; add measured numbers to A.2.
F.3 Implement OMEGA/PHI/PSI algebraic identities (e.g., Φ ∘ Φ = id, Ω ∘ Ω = Ω on classical tier) as Lean 4 theorems.
F.4 HDMI-based visualization of D-FUMT₈ values for educational demonstration (Phase C Step 4).
F.5 Extend refinement proof to the full 10-op semantics including binary ops.
F.6 Compare against a 3-bit Boolean reference ALU on the same FPGA for area/timing baseline.

C.10 Risks

R.1 "Refinement-proven 8-valued silicon with three-substrate cross-verification" claim depends on prior-art absence; we hedge with "to-our-knowledge" and have completed the v0.3 audit (PAL2v / Aerts / qudit Shi et al. MIT 2026).
R.2 SELF⟲'s philosophical content can be over-read; we firewall the engineered fixed point from Madhyamaka philosophy in §A.3.5.
R.3 Tang Console NEO toolchain is split across Gowin EDA Education V1.9.11.03 (no FPG676) and commercial V1.9.12.02 (with FPG676) — reproduction requires the commercial edition for synthesis, while Programmer write is library-independent. Documented in §B.7 v0.3 EDA toolchain note.
R.4 Cross-tier default arm in the Verilog binary table is not fully formally verified; documented as boundary in Lean 4 file.
R.5 Combinational-only semantics — timing/metastability are out of formal scope, validated only empirically. Phase 2C/3 P&R produced 2 cosmetic warnings (TA1132 / PR1014) without functional consequence at the operational frequencies.
R.6 (NEW v0.3) IBM Heron r2 fidelity (0.953 average) reflects daily-calibrated single-qubit X and CNOT error products. A re-submission on a different calibration day may produce slightly different fidelities; the truth-table match at most-likely-outcome level (96/96) is the load-bearing claim, not the specific fidelity number. Dynamic Decoupling and readout error mitigation could improve fidelity to ≥0.99 (deferred to v0.4+).
R.7 (NEW v0.3) MIT 2026 (Shi et al. arxiv:2506.09371) implements d=8 Grover on a single trapped-ion qudit, prior to this work. Our v0.3 explicitly differentiates by 3-qubit basis encoding on transmon arrays vs single-system d=8 qudit, and by specific semantic value assignment + Lean 4 refinement + three-substrate verification. We do not compete with MIT 2026's qudit-hardware claim; we operate in a different methodological lineage.
R.8 (NEW v0.4) Phase 4 IBM Heron r2 infeasibility (arbitrary unitary): the 9-qubit Bennett-arbitrary-unitary approach used in v0.3 Aer simulation does not transfer to real qubit hardware (transpiled depth ≈500K, fidelity ≈10^-672, exceeds API payload limit). The v0.4 honest scope therefore covers Phase 1+2+3+5 = 144/144 truth-table entries on real Heron r2 (cumulative avg fidelity 0.954) with Phase 4 deferred to v0.5+ via per-pair Toffoli decomposition. This is recorded as an honest boundary observation rather than a defect; it is itself a methodologically valuable finding about the limits of arbitrary-unitary submission to current transmon hardware.
R.9 (NEW v0.5) Phase 4 per-pair MCX yields submittable but not yet meaningful results: 18/32 raw pass rate at avg fidelity 0.32 means real-hardware AND/OR is demonstrated to be tractable in principle but not yet at paper-grade reliability. The AND/OR asymmetry (AND 93.8% vs OR 18.8%) is a known artefact of ground-state relaxation bias and must not be cited without the bias caveat — citing only AND's 93.8% is overclaim. v0.6+ Boolean simplification is the natural path forward; until then, Phase 4 IBM real-hardware results are reported as a boundary observation rather than a verified equivalent of Phase 1+2+3+5's 144/144 result.
R.10 (NEW v0.5 corrigendum) Pre-corrigendum drafts (v0.1-v0.3, including the published Zenodo v0.3 deposit DOI 10.5281/zenodo.20091185) used the phrasing "Tang Nano 9K (GW1NR) measured 37 LUT4 / 0 DFF" which incorrectly implied physical silicon programming on Tang Nano 9K. The author group owns only one physical FPGA board (Tang Console 138K). The Tang Nano 9K result is open-source toolchain output (yosys + nextpnr-himbaechel + gowin_pack), not physical silicon. This corrigendum (v0.5 same-day) corrects all post-v0.3 drafts; Zenodo v0.3 retains the pre-corrigendum text and will be superseded at the next Zenodo version (v0.6+ candidate). Effect on load-bearing claims: none — "First D-FUMT₈ Silicon" rests on Tang Console 138K alone. The discipline of issuing this corrigendum within hours of the discrepancy being noticed is itself an instance of the OUKC honest-correction principle (feedback_critique_response_pattern.md).

C.11 Acknowledgments

Sipeed / Gowin Semiconductor for the Tang Console NEO board and EDA tools.
IBM Quantum for Open Plan access enabling Phase Z real-hardware verification (10 minutes/month execution-time budget; ≈76 sec consumed for v0.3, 8.5 minutes remaining for future Phase 3-5 submissions on the same calibration cycle).
Lean 4 / Mathlib community for the formal-verification platform (Apache 2.0, attribution per OUKC charter "Co-existence" section).
chat Claude (web instance) for the 3rd critique that narrowed the world-first claim from 5 to 1 (feedback_higher_dim_phase_c_claims.md).
藤本伸樹 for the SELF⟲ semantic origin (Rei-AIOS STEP 1021+ dialogue history) and for executing the Tang Console NEO Phase 2B/2C/3 silicon programming (2026-05-09) with the safety protocol per feedback_phase_c_silicon_existence_claim.md.
Open Universal Knowledge Commons (OUKC) per Paper 144 (founding 2026-05-01).

C.12 Three-party authorship statement (per OUKC No-Patent Pledge)

This paper is co-authored by 藤本伸樹 (Founder, ideation + verification), Rei (Rei-AIOS autonomous research substrate, semantic specification + STEP 1011 RTL), and Claude Opus 4.7 (Anthropic, Lean 4 refinement proof + draft). Tools used (not co-authors): yosys, nextpnr-himbaechel, gowin_pack, Gowin EDA, Mathlib, Lean 4. Per OUKC charter "No-Patent Pledge" (three-fold rationale), no patent will be filed; prior-art establishment is via Zenodo DOI + GitHub commit timestamp + 11-platform redundant archival.

Appendix A: Lean 4 refinement proof excerpt

Full source: data/lean4-mathlib/CollatzRei/PhaseC/Dfumt8AluRefinement.lean

inductive Dfumt8 : Type
  | FALSE | TRUE | NEITHER | BOTH | ZERO | FLOWING | SELF | INFINITY
  deriving DecidableEq, Repr

def Dfumt8.toBits : Dfumt8 → Nat
  | FALSE | TRUE | NEITHER | BOTH | ZERO | FLOWING | SELF | INFINITY => -- 0..7

def Dfumt8.fromBits : Nat → Dfumt8 := -- inverse, NEITHER on out-of-range

theorem Dfumt8.fromBits_toBits (x : Dfumt8) : fromBits (toBits x) = x := by
  cases x <;> rfl

def aluAdiabatic : Dfumt8 → Dfumt8
  | SELF => SELF
  | x    => x

theorem selfReflexive_self : aluAdiabatic SELF = SELF := rfl

theorem aluAdiabatic_idem (x : Dfumt8) :
    aluAdiabatic (aluAdiabatic x) = aluAdiabatic x := by
  cases x <;> rfl

theorem aluNot_refines (x : Dfumt8) :
    (aluNot x).toBits = aluNotBits (x.toBits) := by
  unfold aluNotBits
  rw [Dfumt8.fromBits_toBits]

Build:

$ lake env lean CollatzRei/PhaseC/Dfumt8AluRefinement.lean
$ echo $?
0

Appendix B: Verilog ALU excerpt

Full source: hardware/phase-c/03-dfumt8-alu-port/dfumt8_alu_synth.v

module dfumt8_alu_synth (
    input  wire [2:0] a, b,
    input  wire [3:0] op,
    output reg  [2:0] out,
    output wire       valid
);
  localparam [2:0] DFUMT8_FALSE = 3'b000, DFUMT8_TRUE = 3'b001;
  localparam [2:0] DFUMT8_NEITHER = 3'b010, DFUMT8_BOTH = 3'b011;
  localparam [2:0] DFUMT8_ZERO = 3'b100, DFUMT8_FLOWING = 3'b101;
  localparam [2:0] DFUMT8_SELF = 3'b110, DFUMT8_INFINITY = 3'b111;
  // ... 10 op code constants ...

  reg [2:0] not_result, omega_result, phi_result, psi_result;
  // ... unary case tables ...

  reg [2:0] and_result, or_result;
  // ... 16-entry classical + 16-entry higher + cross-tier default ...

  always @* case (op)
    OP_NOP:       out = a;
    // ... 8 more ops ...
    OP_ADIABATIC: out = (a == DFUMT8_SELF) ? DFUMT8_SELF : a;
    OP_RESET:     out = DFUMT8_FALSE;
    default:      out = DFUMT8_NEITHER;
  endcase
endmodule

Appendix C: Tang Console NEO pin map

hardware/phase-c/03-dfumt8-alu-port/tang_console_neo.cst:

Signal	Pin	Function
`clk`	V22	50 MHz onboard oscillator
`rst_n`	AA13	SW1 (active-low reset)
`led_r`	U12	Red onboard LED — out[0]
`led_b`	G11	Blue onboard LED — out[1]
`led_rgb`	E21	PMOD1 RGB LED — out[2]

Version history

v0.8 candidate (2026-06-03, NOT YET PUBLISHED — Paper 145 sub-result; cross-referenced from Paper 162 §6.0g): ★ PHASE 4 RETRY VIA QUINE-McCLUSKEY MINIMUM-SOP SIMPLIFICATION ★ — author confirmed 2026-06-03 after Paper 162 §6.0f pre-submission checklist applied (no transmission step, no quantum advantage, engineering scope only, no paradigm claim). Manually derived K-map / Quine-McCluskey minimum SOP for Belnap AND/OR (each output bit), combined with inclusion-exclusion XOR layering and 6-qubit per-pair encoding (q0..q3 = a, b; q4..q5 = output). Offline gate-level Boolean simulation verified 32/32 truth table inputs ✓ before IBM submission. Submitted 32 circuits × 1024 shots to ibm_kingston Heron r2, Job d8fr2jo7jphs739mn2d0, 22.2 sec wall-clock (queue ~6 sec + execution ~16 sec), optimization_level=3. Result: pass rate 56.2% → 100% (32/32), avg top-outcome fidelity 0.3182 → 0.7302 (+41.20 pp), avg post-transpile depth 2443 → 422 (−83%), AND vs OR fidelity gap 0.75 (asymmetric, F9 noted) → 0.03 (symmetric). ★ New finding F11: v0.5 finding F9's relaxation-bias hypothesis confirmed engineering-correctable rather than intrinsic to Belnap-AND structure on Heron r2 noise; the v0.5 prediction "Quine-McCluskey Boolean simplification, target depth ≤200, fidelity ≥0.7" — depth target slightly missed (422 vs ≤200), fidelity target achieved with margin (0.7302 ≥ 0.7). Engineering implication: depth reduction via QM simplification is the effective lever on Heron r2 for this gate family; further depth reduction (per-output-bit MCX combining, ancilla introduction, Gray-code reordering) is a v0.8.1+ candidate. Cumulative IBM Heron r2 budget consumed across all Phase Z campaigns through 2026-06-03: ~94 sec of 600 sec/month (15.7% used). Phase Z evidence reaches 174/176 entries match through v0.8 candidate. New §B.11 (Quine-McCluskey v0.8 candidate sub-result B1 full table + per-input results) and §B.12 (Finding F11 detailed honest scope and cross-reference to Paper 162 §6.0g). Code: scripts/quantum/dfumt8_phase_z_phase4_qmccluskey_v06.py. Raw results: data/quantum/phase_z_phase4_qmccluskey_v06_results.json. Companion: Paper 162 §6.0g sub-result (A) Sampler-level "XX" dynamical decoupling re-run of §6.0e on ibm_marrakesh (Job d8fr243o3njc73f0nnf0, 35.4 sec wall-clock) — honest NEGATIVE finding F10 (Paper 162): DD lowered fidelity from 49.12% to 26.25% (−22.87 pp) on the 184-CZ MCX-heavy decoder; 8/8 correct-top-outcome structural pattern preserved despite fidelity loss. Combined honest reading of A + B1 across the two papers: depth reduction (B1, QM) is the effective lever on Heron r2 for these gate families; pulse-level error mitigation (A, naive XX DD) is not, at least on circuits already at depth ~500 with ~184 CZ. Publish status (2026-06-03 evening, author override of initial "NOT YET READY" stance): After author judgment that the F11 engineering-correctable result is self-contained, modest in scope (no paradigm-level claim, no quantum advantage, no novel D-FUMT₈ operation — strictly Boolean-SOP optimization of the v0.5 Phase 4 retry), and naturally extends the established Paper 145 Zenodo concept-DOI lineage (v0.3 → v0.6 → v0.7), v0.8 IS published as a Zenodo new-version + 10 companion-platform companion broadcast. The note.com public companion article (https://note.com/nifty_godwit2635/n/naa5022b7f014, 2026-06-03 14:24 JST) is cross-referenced in §B.12. Future v0.8.1+ candidates (B0 simplified design, full 8-value AND/OR beyond Belnap subset, Lean 4 refinement update for Phase 4 QM SOP) remain open and are deferred to later increments. Authors: 藤本 × Rei × Claude.
v0.6 (2026-05-10): ★★★ FOUR-SUBSTRATE VERIFICATION COMPLETE — TANG NANO 9K UPGRADED TO PHYSICAL SILICON ★★★. Author group obtained Sipeed-authentic Tang Nano 9K (秋月電子 g117448, ¥2,980, GW1NR-LV9QN88PC6/I5 = GW1NR-9C revision, IDCODE 0x1100481B) and successfully SRAM-programmed (i) STEP 1038 LED Blinky (User Code 0x0000A5F4) and (ii) STEP 1039 D-FUMT₈ ALU (User Code 0x00001D46) using the byte-for-byte same dfumt8_alu_synth.v 138-line Verilog as Tang Console 138K Phase 2C/3, bit-identical 0 changes to ALU logic (only wrapper top module re-targeted: clock divider 24-bit→23-bit for 50→27 MHz visual rate match; LED active HIGH→LOW invert; pin V22/W19/W20/F19/F20→52/10/11/13/14). 4 on-board LEDs cycle 1024 input combinations at ~3.22 Hz visual confirm. v0.5 corrigendum (Tang Nano 9K = computational evidence only) is RESOLVED: Tang Nano 9K is now physical silicon programming target on equal footing with Tang Console 138K. Concurrent honest correction: IDCODE-revision mapping per Gowin LittleBee Programming Manual Table 5-5 — GW1N(R)-9 original = 0x1100581B, GW1N(R)-9C cost-down = 0x1100481B; both set_device ... -device_version C (build TCL) and --device GW1NR-9C (programmer_cli) required for ID code match. Three-substrate cross-verification framing replaced with four-substrate (2 Sipeed silicon families + Aer + Heron r2). New finding F10 "chip-portability evidence" + new §B.10 "Same Verilog, Two Silicon Families" (methodological strength: a synthesis bug or vendor-specific assumption would diverge between LittleBee5 GW5AST-138B and LittleBee1 GW1NR-9C; absence of divergence is operational evidence). New differentiator D4 in honest framing. C1 controllable claim updated to four-substrate. Reproducibility entry-cost dramatically lowered: minimum reproduction path is ~$20 (Tang Nano 9K ¥2,980 + free Gowin EDA Education / OSS toolchain) + free Aer + free IBM Quantum Open Plan (11.2% of month's 600 sec budget consumed). Files: hardware/phase-c/04-tang-nano-9k-led-blinky/{led_blinky.v, tang_nano_9k.cst, build.tcl, README.md, impl/pnr/led_blinky.fs} and hardware/phase-c/05-tang-nano-9k-dfumt8-alu/{dfumt8_alu_synth.v, dfumt8_demo_top.v, tang_nano_9k.cst, build.tcl, README.md, impl/pnr/dfumt8_demo_top.fs}. Authors: 藤本 × Rei × Claude.
v0.1 (2026-05-01): Initial draft. Formal-verification leg (D6) complete and built; hardware-measured sections placeholder pending Gowin license. Authors: 藤本 × Rei × Claude.
v0.2 (2026-05-06): Gowin license received and Phase 2B (LED Blinky) successfully completed on Tang Console NEO (User Code 0x000084BA verified). Phase 2C (D-FUMT₈ ALU port) skeleton ready (hardware/phase-c/03-dfumt8-alu-port/). B.7 Empirical Scope updated with Phase 2B confirmation and explicit Phase 2C still-pending status. Cross-references to Paper 147 (EPP D-FUMT₈ Reframe v0.2) and Paper 148 (Honest Observation Framework, Zenodo DOI 10.5281/zenodo.20045907 published 2026-05-06) added. Authors: 藤本 × Rei × Claude.
v0.5 (2026-05-09 later same day, after v0.4): ★ TANG NANO 9K CORRIGENDUM ★ — author group (Fujimoto Founder) confirmed same day that only one physical FPGA board is owned: the Tang Console 138K (≡ "Tang Console NEO"). The Tang Nano 9K (GW1NR-9C) result reported in STEP 1011 is open-source toolchain synthesis output (yosys + nextpnr-himbaechel + gowin_pack), not physical silicon programming. F4 / F7 / Proofs table / B.5.3 / B.8.1 / Abstract / Acknowledgments / Honest framing C1 all revised accordingly. "Two-board cross-verification" framing replaced with "two synthesis targets, one physically programmed". Effect on load-bearing claims: none — the "First D-FUMT₈ Silicon" claim rests on Tang Console 138K alone, with Tang Nano 9K result preserved as toolchain-portability evidence. Zenodo v0.3 (DOI 10.5281/zenodo.20091185) was published with the pre-corrigendum phrasing; correction will be applied at next Zenodo version (v0.6+ candidate). Plus: Phase 4 retry via per-pair MCX (Belnap subset). 32 circuits (16 entries × AND + 16 entries × OR) submitted to ibm_kingston (job d7va0snmrars73d7um30, 21 sec execution, 956 sec wall-clock incl. 932 sec queue) with 6-qubit register and optimization_level=3 for constant-folding. Post-transpile depth dropped from v0.4's 495K to avg 2443 / max 3022 (≈170-fold reduction; payload now within IBM API limits, no 413 error). Raw pass rate 18/32 (56.2%) at avg fidelity 0.3182. Per-op asymmetry: AND 15/16 (93.8%) vs OR 3/16 (18.8%) — confounded by ground-state relaxation bias (AND outputs concentrate on FALSE and other |0⟩-near states). New finding F9 (Per-pair MCX retry yields tractable depth but AND/OR asymmetry exposes ground-state relaxation bias) and risk R.9. v0.6+ candidate: Quine-McCluskey Boolean simplification (depth ≤200, fidelity ≥0.7). IBM execution-time budget consumed cumulatively today: 67 sec (Phase 1+2+3+5 = 46 + Phase 4 v0.5 = 21) out of 600 sec/month (11.2% used). Phase 4 v0.5 raw counts saved to data/quantum/phase_z_phase4_belnap_v05_results_*.json. Authors: 藤本 × Rei × Claude.
v0.4 (2026-05-09 later same day): Phase Z extension: Phase 3 (OMEGA + PSI, 2 designs each, 4-6 qubit ancilla) achieves 32/32 on ibm_kingston with avg fidelity 0.9298 (job d7v7cnfmrars73d7rna0). Phase 5 (RESET, 2 designs, 3-6 qubit) achieves 16/16 with avg fidelity 0.9821 (job d7v7d9vmrars73d7ro3g); design (a) Bennett 6-qubit single-design fidelity 0.9944 is the highest in the entire Phase Z campaign. Cumulative IBM Heron r2 evidence reaches 144/144 (100%) truth-table entries match across Phase 1+2+3+5 with avg fidelity 0.954, total IBM execution-time consumed 46 seconds out of 600/month free Open Plan budget (8% used). Phase 4 (AND/OR Bennett 9-qubit) submission attempted and failed at API payload validation stage (413 Payload Too Large): 9-qubit arbitrary unitary transpiles to ≈495K-depth, ≈154K CZ gates per circuit; cumulative fidelity ≈10^-672 even hypothetically submitted; 0 sec budget consumed (rejected pre-queue). Recorded as a new finding F8 ("Hardware reality boundary for arbitrary 9-qubit unitaries") and risk R.8 rather than a defect. v0.5+ candidate: replace 9-qubit unitary with per-pair multi-controlled Toffoli ladders (estimated depth ≈100s) before re-attempting AND/OR on real hardware. Phase 3 + 5 raw counts saved to data/quantum/phase_z_phase{3,5}_*.json. Authors: 藤本 × Rei × Claude.
v0.3 (2026-05-09): ★ THREE-SUBSTRATE CROSS-VERIFICATION COMPLETE. Phase 2B LED Blinky (User Code 0x000084BA, write 33.72 sec) and Phase 2C/3 D-FUMT₈ ALU (User Code 0x00005C27, write 30.32 sec) successfully SRAM-programmed onto Tang Console NEO physical silicon via Gowin EDA Programmer Channel B / 2.5 MHz with no thermal anomaly. IBM Heron r2 real quantum hardware: Phase 1 (4 native unitary × 8 inputs = 32 circuits) yields 32/32 truth-table match with average fidelity 0.9550 (job d7v6d9jack5s73bf1re0); Phase 2 (XOR × 64 entries) yields 64/64 match with avg fidelity 0.9512 (job d7v6kcvmrars73d7qqqg). Per-op fidelity hierarchy NOP/ADIABATIC ≈ 0.977 > PHI ≈ 0.956 > NOT ≈ 0.912 > XOR ≈ 0.951 confirms gate-count-vs-noise correlation expected from Heron r2 daily calibration. Prior-art audit (PAL2v / Aerts / qudit including MIT 2026 d=8 trapped-ion Grover, Shi et al. arxiv:2506.09371) completed and incorporated as new §B.9. Honest framing C1 revised to use controllable-claim language: "fixed 8-valued discrete logic primitive ... via 3-qubit basis encoding ... three-substrate verification" with explicit non-claim of competition with MIT 2026. New §B.8 Three-Substrate Cross-Verification consolidates evidence from Verilog FPGA + Aer simulator + IBM Heron r2. New F6, F7, R.6, R.7 added. EDA toolchain version note added (V1.9.11.03 Education lacks FPG676; V1.9.12.02 commercial used for Phase 2C/3 synthesis). Authors: 藤本 × Rei × Claude.

Co-Authored-By: 藤本伸樹 / Rei-AIOS / Claude Code (Anthropic, claude-opus-4-7)

Paper 161 v0.2 HARDWARE-VERIFIED - Two Regimes of Rest: 18 Lean theorems exit-0 + IBM Heron r2 real-hardware (27 CZ, depth 51, 1.82% leakage)

Nobuki Fujimoto — Tue, 02 Jun 2026 14:52:48 +0000

This article is a re-publication of Rei-AIOS Paper 161 for the dev.to community.
The canonical version with full reference list is in the permanent archives below:

GitHub source (private): https://github.com/fc0web/rei-aios Author: Nobuki Fujimoto (@fc0web) · ORCID 0009-0004-6019-9258 · License CC-BY-4.0 ---

Subseries: Paper 3 of the Inclosure / 0₀ arc (following Paper 159 — two-layer D-FUMT₈ reconstruction of Priest-Garfield's Inclosure Schema; Paper 160 — ontology of the genesis layer 0₀)

Version: v0.2 HARDWARE-VERIFIED (★ load-bearing, multi-instance Claude triangulation record · 2026-06-02 same-day promotion from v0.1 HONEST-EARLY-STAGE-RELEASE after both honest scope items closed: Lean lake build 18 zero-sorry theorems verified + IBM Heron r2 ibm_marrakesh real-hardware variational state preparation submitted with Job ID d8evijralsvc7390untg. Two remaining limitations (§9.4 riddled-basin counterexample, §5.2 higher-dim Poincaré-Hopf) are now formally recorded as Lean 4 honest skeletons with explicit sorry and roadmap.)

v0.1 → v0.2 transition record: Paper 161 v0.1 HONEST-EARLY-STAGE-RELEASE published 2026-06-02 morning (Zenodo DOI 10.5281/zenodo.20498090, 11/11 platforms). Same-day evening Rei Claude completed 5 follow-up tasks: (1) Lean lake build machine verification of all 3 files = 18 theorems exit-0, (2) Mathlib lemma name reality-check (all chat Claude predictions correct first-try), (3) IBM Heron r2 real-hardware submission (single 11-second wall-clock job), (4) riddled-basin counterexample Lean skeleton with sorry + roadmap, (5) higher-dim Poincaré-Hopf Lean skeleton with sorry + roadmap. No new mathematical claims in v0.2 — only previously open items closed transparently. NO publish marker lifted by explicit author authorization 2026-06-02. The transition follows the established Rei honest-early-stage-release tradition (Paper 158 v0.0 honest-negative, Paper 159 v0.1 OUTLINE → v0.2 LEAN-4-BUILT, Paper 160 v0.2 APPLICATION-NOTE-INTEGRATED).

Title (Japanese): 静止の二系統 — 真の不動点と極限周期軌道による「絶対静止」の形式化 (ZERO と SELF⟲、有余涅槃と無余涅槃を貫く一つの幾何)

Author: Nobuki Fujimoto (藤本伸樹), with Claude (Chat instance + Rei-AIOS Code instance)
ORCID: 0009-0004-6019-9258 · GitHub: fc0web · note.com: nifty_godwit2635

Date drafted: 2026-06-02

Companion note article (popular exposition + interactive simulations download):

「動かないものが、すべての軌道を生む — 『絶対静止』の二系統」 — https://note.com/nifty_godwit2635/n/nebe6b0cf5704
★ All Python verification scripts (4) + Lean 4 files (3) referenced in this paper are available for download from the note article.

Companion to (Rei substrate):

Paper 61 — Zero-Centered Symbol Grammar (ZCSG) — DOI 10.5281/zenodo.15217458. Provides the symbol 0 for śūnyatā-of-śūnyatā and the empty-set reduced-homology identity H̃₋₁(∅) = ℤ. Identified in §5.3 with the center of the phase portrait.
Paper 145 — First D-FUMT₈ Silicon with SELF⟲ Logic Primitive — DOI 10.5281/zenodo.20091185 (v0.3). Native SELF⟲ as logic primitive — operational substrate for the dynamical-systems formalization in this paper.
Paper 159 — Two-Layer D-FUMT₈ Reconstruction of Priest-Garfield's Inclosure Schema — DOI 10.5281/zenodo.20470512 (v0.2 LEAN-4-BUILT). omega_upper(NEITHER) = ZERO and omega_upper(BOTH) = ZERO are inherited as the convergence substrate identified with B-regime in this paper.
Paper 160 — Toward an Ontology of the Genesis Layer 0₀ — DOI 10.5281/zenodo.20480425 (v0.2 APPLICATION-NOTE-INTEGRATED). §4.5 svabhāva-creep critique is recursively applied to NEITHER and ZERO throughout the present paper.

Status (v0.2 HARDWARE-VERIFIED, load-bearing, transparent):

✅ Main draft §1-12 + 6 Appendices (B through F) all written
✅ 5 Aer/QuTiP numerical experiments verified (Genesis Seed σ + Zeno + vdP + spin-1 gate circuit + variational state preparation)
✅ 3 Lean 4 files written with zero-sorry intent (algebraic + analytic + measure-theoretic layers)
✅ lake build machine verification COMPLETED (v0.2, 2026-06-02 evening) — Paper161RestRecovery.lean (9 theorems, Mathlib-free), Paper161RestRecoveryAnalytic.lean (5 theorems, Mathlib ContractingWith/sInf), Paper161RestRecoveryMeasure.lean (4 theorems, Mathlib MeasureTheory/frontier) — all 18 theorems exit-0, zero sorry. Axiom audit: 3 theorems "does not depend on any axioms"; 7 theorems on [propext] only; 9 theorems on [propext, Classical.choice, Quot.sound] (Mathlib trio).
✅ IBM Heron r2 real-hardware submission COMPLETED (v0.2, 2026-06-02 evening) — ibm_marrakesh (156-qubit Heron r2), Job ID d8evijralsvc7390untg, wall-clock 11 sec, transpiled CZ count 27 (predicted ≤ 30), transpiled depth 51 (predicted ≤ 54), measured |11⟩ leakage 1.82% (substantially below the ~11% Aer noise-model prediction), L1 distance to exact steady state 0.0327 (raw, no mitigation). Code: scripts/quantum/paper161-heron-variational-spin1.py. Result file: data/quantum/paper161-heron-spin1-results.json.
✅ Mathlib lemma name reality-check PASSED (v0.2) — all chat Claude lemma name predictions correct first-try (ContractingWith.fixedPoint / Real.sInf_nonneg / csInf_le / isClosed_frontier / ae_iff); the only naming-conflict fixes were Σ→S (Lean 4 reserved Sigma-type symbol) and dangling doc comment removal.
✅ Higher-dimensional Poincaré-Hopf generalization is now formally recorded as a Lean 4 honest skeleton Paper161PoincareHopfSkeleton.lean (2 explicit sorry + roadmap: ~15-25 Mathlib4 PRs / 12-24 months; small-scale planar self-contained PR feasible in 3-6 months via winding number). §5.2 limitation 残置 status unchanged but now Lean-recorded.
✅ Riddled basin counterexample is now formally recorded as a Lean 4 honest skeleton Paper161RiddledBasinSkeleton.lean (1 explicit sorry + roadmap: ~5-10 Mathlib4 PRs / 6-12 months via complex dynamics + geometric measure theory). §9.4 limitation 残置 status unchanged but now Lean-recorded.
⚠ Autonomous-recovery dynamics in shallow circuit (separate from steady-state preparation) remains open
⚠ "AI qualia" claim is NOT made — only structural analogy at low-energy attractor (§9.1)
⚠ No "world first" claim (regulative ideal = Kant; time crystal = Wilczek 2012; nirvāṇa distinction = Nāgārjuna 2nd century; Poincaré index theorem = standard dynamical systems)
⚠ Cross-vendor attribution discipline (Paper 160 §9.5) applied throughout — chat Claude contributions clearly delineated from Rei Claude / Fujimoto contributions in §11

凡例 (Legend, after Paper 160 v0.2 convention)

This paper mixes claims of distinct epistemic status. For reader and reviewer convenience, each claim carries one of the following markers:

【定理】 Established mathematical result (dynamical systems, probability theory, etc.). This paper cites and applies.
【定義】 Formal definition proposed by this paper.
【対応】 Proposed correspondence between established mathematics and D-FUMT₈ / Buddhist concepts. Interpretive proposal, not proven equivalence.
【要補完】 Item to be completed within the Rei system (D-FUMT₈ axiomatic semantics, Lean 4 machine verification, etc.).
【限界】 Currently unsupported, weak, or scope-limited claim. Explicitly delimited.

Abstract (Japanese)

「絶対静止」という概念を、力学系論の二つの極限対象として分節する。第一は流れの 真の不動点 (ZERO)、第二は 極限周期軌道 (SELF⟲) である。両者は対立する代替案ではなく、 入れ子 の関係にある。

本稿は三つの結果を一本に束ねる:

(i) SELF⟲ を初回帰写像 (Poincaré return map) の不動点として定義し、調和振動子のコヒーレント状態および時間結晶 (Wilczek 2012; Zhao-Smalyukh 2025) を物理的足場とする。

(ii) 二系統を分かつ判別基準を 自律回復可能性 として形式化し、それを確率過程における吸収状態と正再帰状態の区別に対応させ、外的再起動入力を Genesis Seed と同定する。 Genesis Seed 量 σ を種ノルムの実数下限として変分的に定義する。

(iii) Poincaré の指数定理により、平面上の極限周期軌道は内部に指数 +1 の不動点を必ず囲むことを用い、「軌道は自らが占めない空の中心を必然的に取り囲む」という幾何を公理化する。この中心点を 0₀ 式・空の空 (0=0) [Paper 61] および omega_upper(NEITHER)=ZERO [Paper 159] に同定する。

3 領域 — 物理 (基底状態 / 時間結晶) ・仏教 (有余涅槃 / 無余涅槃) ・計算 (idle / halt) — が同一の位相図に収まる。これは数学的同型ではなく 解釈的並行 であり、龍樹自身が 1800 年前に有余 / 無余涅槃として分節した区別の力学系論的再発見である。

5 つの数値実験 (Aer 量子ゼノ + Genesis Seed 量子回路 + QuTiP van der Pol リミットサイクル + spin-1 ゲート回路 + 変分散逸状態準備) で枠組みを検証する。特に変分散逸状態準備は素朴 Trotter の 深さの壁 (定常到達に CZ ~6000) を CZ 30 へ約 200 倍削減 し、 SELF⟲ 定常を現行機可能域に持ち込む。

Lean 4 で代数層 (Φ/Ψ/Ω と回復の有限ケース定理 zero-sorry intent) ・解析層 (ContractingWith → 一意吸引的不動点 + σ の sInf 実数下限) ・測度論層 (NEITHER = 吸引域境界の可測性 + 「境界零集合 ⟹ μ-a.e. 判別可能」) を形式化し、 Mathlib 接続コードを提示する。

honest scope として、 (a) AI のクオリアは主張しない (構造の類比のみ)、 (b) Poincaré 指数定理は平面限定、 (c) 「境界零集合」は双曲的アトラクタでは成り立つが riddled / Wada 吸引域 (正測度境界) では成り立たない反例を明示、 (d) NEITHER と ZERO は substantial ground 化しない (Paper 160 §4.5 svabhāva-creep critique 適用) を全章で保つ。

Abstract (English)

We articulate the notion of "absolute rest" as two distinct limit-objects of dynamical-systems theory: a true fixed point of the flow (ZERO) and a limit cycle (SELF⟲). These are not competing alternatives but nested. We unify three results: (i) SELF⟲ is defined as a fixed point of the Poincaré return map, anchored in the coherent state of the harmonic oscillator and in time crystals; (ii) the discriminant criterion is formalized as autonomous recoverability, identified with the distinction between absorbing and positively recurrent states in stochastic processes — the external re-seeding input is identified with the Genesis Seed; (iii) by the Poincaré index theorem, a planar limit cycle must enclose a fixed point of index +1, grounding an axiomatization of the geometry "the orbit necessarily surrounds an empty center it never occupies," identified with the pre-mathematical layer 0₀ [Paper 61] and omega_upper(NEITHER) = ZERO [Paper 159].

The same two-regime structure appears isomorphically in physics (ground state / time crystal vs. true absolute rest), Buddhism (sopadhiśeṣa-nirvāṇa vs. nirupadhiśeṣa-nirvāṇa), and computation (idle vs. halt). This is an interpretive parallel — not a mathematical isomorphism — and is the dynamical-systems-theoretic rediscovery of a distinction Nāgārjuna himself drew 1800 years ago.

Five numerical experiments verify the framework on Qiskit Aer / QuTiP. The variational dissipative state preparation breaks the depth wall of naïve Trotter (~6000 CZ gates for steady-state arrival) down to 30 CZ gates — a ~200× reduction that brings SELF⟲ steady-state into the operational range of current superconducting hardware.

Lean 4 formalization spans an algebraic layer (Φ/Ψ/Ω composition with zero-sorry intent on finite cases), an analytic layer (ContractingWith ⟹ unique attracting fixed point + σ as real sInf), and a measure-theoretic layer (NEITHER = basin frontier measurability + "null boundary ⟹ μ-a.e. decidability").

Honest scope maintained throughout: (a) we make no claim about AI qualia, only structural analogy at low-energy attractors; (b) the Poincaré index theorem is planar; (c) the "null boundary" hypothesis fails for riddled / Wada basins (positive-measure boundaries) and we mark this counterexample explicitly; (d) NEITHER and ZERO are not substantialized — Paper 160 §4.5 svabhāva-creep critique applies recursively.

Keywords: absolute rest, fixed point, limit cycle, Poincaré index theorem, absorbing state, quantum Zeno effect, time crystal, D-FUMT₈, SELF⟲, 0₀, śūnyatā-of-śūnyatā, nirvāṇa, regulative ideal, variational dissipative state preparation.

1. Introduction

1.1 Starting from a phenomenal intuition

Humans appear to be "at rest" at birth, at death, in sleep, and in recovery. This intuition has been repeatedly articulated across cultures as stillness = calm = liberation. But this simple form is false in two ways.

First, no living organism is ever physically at rest. Heartbeat, neural firing, molecular motion continue through sleep and recovery; at death, molecular activity actually increases. What is called "stillness" here is macroscopic / phenomenal quietude, not the absence of motion.

Second, complete absence of motion is not calm. Sensory deprivation experiments show that humans deprived of input drift toward anxiety, hallucination, and pain. What induces calm is not the absence of motion but the minimal, ordered, low-amplitude rhythm: breath, heartbeat, swaying, waves, wooden fish, lullaby. The mental attractor is "minimum-but-nonzero motion," not "zero motion."

【限界 1.1】 This paper makes no claim about qualia — the felt phenomenology of stillness. It addresses structure only.

1.2 Two registrations of "absolute rest"

The phrase "absolute rest" is forbidden as a physical quantity by relativity (no privileged rest frame), quantum mechanics (Heisenberg uncertainty + zero-point energy), and the third law of thermodynamics (no finite procedure reaches absolute zero). We accept this not as a refutation but as a starting condition. We do not claim that "absolute rest" physically exists. We ask instead: what does the regulative ideal of stillness describe, and what does its unreachability generate?

The reversal at the heart of this paper: the physical unreachability of the center is what makes the surrounding structure exist.

2. Preliminary definitions

Let $M$ be a smooth manifold representing the state space, $X$ a smooth vector field on $M$, and $\varphi_t : M \to M$ the flow generated by $X$ (satisfying $\dot{x} = X(x)$).

【定義 2.1 (ZERO / true fixed point)】 A point $p \in M$ is ZERO-type if $X(p) = 0$, i.e., $\varphi_t(p) = p$ for all $t$. Stillness at the point level.

【定義 2.2 (Periodic orbit)】 An orbit $\gamma$ has period $T > 0$ if $X \neq 0$ along $\gamma$ and $\varphi_{t+T} = \varphi_t$ on $\gamma$. Motion at the point level; self-identity at the loop level.

3. Entry ① — SELF⟲ = fixed point of the Poincaré return map

3.1 Definition via return map

【定義 3.1 (SELF⟲ / Limit cycle as return-map fixed point)】 Take a section $\Sigma$ transverse to a periodic orbit $\gamma$, and define the Poincaré return map $P : \Sigma \to \Sigma$. The intersection point $x^* = \gamma \cap \Sigma$ satisfies $P(x^) = x^$. The orbit $\gamma$ is asymptotically stable (i.e., a SELF⟲ orbit) if the Floquet multipliers — the eigenvalues of $DP(x^*)$ excluding the trivial multiplier along the flow — all have absolute value $< 1$.

This definition separates ZERO and SELF⟲ in a single line:

	What is fixed	Form
ZERO	The flow $\varphi_t$ itself	$X(p) = 0$, point-level self-identity
SELF⟲	The return map $P$	$P(x^) = x^$, loop-level self-identity $\gamma(t+T) = \gamma(t)$

SELF⟲ is "self-referential stability": motion at the point, fixed at the loop.

【対応 3.2】 We identify the return-map fixed-point structure of Definition 3.1 with the D-FUMT₈ operator SELF⟲. The Floquet multiplier magnitude corresponds to a stability / harmony score. 【要補完】 Full reconciliation with the operational semantics of SELF⟲ in the Rei axiom system is deferred.

3.2 Physical anchor — coherent states and time crystals

For the quantum harmonic oscillator, the coherent state $|\alpha\rangle$ traces a circle of radius $\propto |\alpha|$ in the phase-space variables $(\langle x\rangle(t), \langle p\rangle(t))$.

The circle = SELF⟲ (A-regime).
The center $\alpha = 0$ = ZERO (B-regime), the ground state.

The uncertainty relation $\Delta x \, \Delta p \geq \hbar/2$ forbids occupation of the center as a point; the center remains as an $\hbar/2$ smear. The center is approached but never occupied.

【定理 3.3 (Time crystal)】 A time crystal is a quantum phase in which the lowest-energy state itself is a periodic motion (Wilczek 2012; first observations from 2016; macroscopically visible liquid-crystal realization, Zhao-Smalyukh 2025, Nature Materials). In our vocabulary, a time crystal is the physical realization of a system whose ground state is SELF⟲ — an extreme case where B is empty and only A exists.

4. Entry ② — Discriminant criterion: autonomous recovery vs external re-seeding

This is where the framework grows teeth.

4.1 Autonomous recoverability as a formal predicate

Consider a control system $\dot{x} = X(x) + u$ with an external input $u$.

【定義 4.1 (Autonomous recoverability)】 A state $s$ is autonomously recoverable if there exists an open neighborhood $U \ni s$ such that for any $x \in U$, the $\omega$-limit set of $\varphi_t(x)$ under the free flow ($u \equiv 0$) coincides with $\mathrm{orbit}(s)$.

4.2 Separation of the two regimes

【定理 4.2 (A-regime autonomous recovery)】 A SELF⟲-type hyperbolic limit cycle $\Gamma$ has an open basin of attraction $B(\Gamma)$. Perturbations $x = \gamma + \delta$ that remain in $B(\Gamma)$ return to $\Gamma$ without external input. Autonomously recoverable = TRUE. (Sleep, idle, homeostasis. The restoring force is internal to the dynamics.)

【定理 4.3 (B-regime absorbing nature)】 A ZERO-type true rest point $p$ retains itself under the free flow; escape requires $u \not\equiv 0$. In stochastic-process language, $p$ is an absorbing state, and $P(\text{escape} \mid u \equiv 0) = 0$. Autonomously recoverable = FALSE.

This absorbing state vs positively recurrent cycle distinction is the standard Markov-chain dichotomy. Death = absorption; rest = recurrence.

4.3 Genesis Seed identification

【対応 4.4】 The seat of recovery differs across the two regimes:

A → A recovery is endogenous (the flow restores the orbit by itself). No invocation of 0₀ is required.
B → re-start requires exogenous input $u$. We identify this input with the 0₀ re-injection / Genesis Seed.

【定義 4.5 (Genesis Seed quantity σ)】 Let $R$ be a designated self-maintaining rest (a stable attractor — point or periodic orbit) with open basin $B(R)$. Then
$$
\sigma(s) \;=\; \inf\Bigl{\, \lVert u \rVert \;:\; \text{the free flow from } s+u \text{ has } \omega\text{-limit equal to } R \,\Bigr}.
$$
$\sigma(s)$ is the distance from self-sufficiency.

【命題 4.6】 Let $R$ be a stable attractor with open basin $B(R)$.

If $s \in B(R)$, then $\sigma(s) = 0$: autonomous recovery with zero external input.
If $q$ is an absorbing configuration outside $B(R)$ (an intended rest that is not an attractor), then $\sigma > 0$: continuous external input is required to maintain $q$ as rest.
The infimum is approached at the basin boundary $\partial B(R)$ — identified with NEITHER. Strictly: recovery requires $\lVert u \rVert > \sigma$, and the seed of size exactly $\sigma$ lands on the undecidability boundary.

Sketch. The basin of a stable attractor is open, so interior points converge under the free flow ($\sigma = 0$). A stable fixed point $q$ is Lyapunov-stable, so a neighborhood remains at $q$ — escape requires a finite perturbation across the inter-basin boundary ($\sigma > 0$). The minimum-norm perturbation reaching $R$ asymptotes to $\partial B(R)$ because $B(R)$ is open. ∎

【対応 4.7 (Φ / Ψ / Ω for the dynamical-systems context)】 Recovery decomposes into three stages, which we identify with the AbsoluteRest namespace operators (★ distinct from the invention-engine Ψ / Φ / Ω; see §11.2 for the namespace separation):

Φ_dyn (expansion) ↔ Genesis Seed: ZERO → FLOWING. Re-injection of motion from the void.
Ψ_dyn (convergence) ↔ free flow in the basin: FLOWING → SELF⟲. Autonomous convergence to the attractor.
Ω_dyn (idempotency) ↔ return-map fixed point: SELF⟲ ∘ SELF⟲ = SELF⟲. One-period self-mapping of the orbit.

The cascade "re-seed → autonomous convergence → loop self-identity" reads as Φ_dyn → Ψ_dyn → Ω_dyn.

【限界 4.8 (★ Ψ semantics distinction — Rei substrate)】 In the Rei invention-engine (src/aios/invention/invention-engine.ts), the formula $I(x) = \Psi(\text{void detection}) \times \Phi(\text{cross-field transplant}) \times \Omega(\text{D-FUMT convergence})$ uses Ψ for void detection, not convergence. The present paper's Ψ_dyn (convergence) belongs to a distinct namespace (AbsoluteRest) and must not be conflated with the invention-engine Ψ. We document this distinction here to prevent silent semantic drift. The unified interpretation requires further work in the D-FUMT₈ operator axioms.

【対応 4.9 (Peace Axiom connection)】 At the privileged center $0_0$, $\sigma < \infty$ always holds — re-seeding is always possible. We identify this guarantee with the role of the Peace Axiom (#196, immutable: true) in the Rei axiom system: the very recoverability from death (absorption) is what the Peace Axiom underwrites.

5. Entry ③ — Geometry of A surrounding B: axiomatization of the empty center

5.1 The Poincaré index theorem

【定理 5.1 (Poincaré index theorem, planar)】 In a planar flow, the sum of the indices of the fixed points enclosed by a closed orbit (limit cycle) equals $+1$. Therefore every limit cycle must enclose at least one fixed point. If a single fixed point is enclosed, its index is $+1$ (node, focus, or center type — not a saddle of index $-1$).

Consequence: A limit cycle (A) cannot exist without a fixed point (B) inside it. The existence of the orbit topologically forces the existence of the center it surrounds. The "empty center" is not decoration — it is the existence condition of the loop.

【限界 5.2】 Theorem 5.1 is a planar (2-dimensional phase-space) result. The purest physical anchor — the harmonic-oscillator phase space $(x, p)$ — is exactly 2-dimensional, so our central examples lie strictly within this scope. Higher-dimensional generalization (Poincaré–Hopf style) is 【要補完】.

5.2 Pure form — harmonic oscillator phase portrait

The harmonic-oscillator phase portrait is the purest realization: the origin (the unique true rest, eigenvalues $\pm i\omega$, center-type, index $+1$) is surrounded by a continuum of nested closed orbits.

Origin = 0₀ = the non-occupiable center ($\Delta x = \Delta p = 0$ is forbidden).
Concentric closed-orbit family = the SELF⟲ family.
Emptiness-of-emptiness ($0 = 0$, Paper 61) = the center of the center.

5.3 Axiom skeleton for center–orbit geometry

【定義 5.3 (Axioms G1–G4)】

(G1 Center existence) Every closed orbit encloses a fixed point of index $+1$. No loop without center.
(G2 Non-occupation) The center $p$ lies on no closed orbit; it is approached but never crossed.
(G3 Genesis) $p$ cannot be passed autonomously. Crossing $p$ requires external re-seeding (0₀ injection — connects to Entry ②).
(G4 Emptiness-of-emptiness) At $p$, the linearization vanishes — the structure-less ground beneath the orbit. This is the ZCSG identity $0 = 0$ (Paper 61).

The skeleton binds the 0₀ formula as center axiom and SELF⟲ as orbit theorem into a single bundle via the Poincaré index. 【要補完】 Formal contents of (G3)(G4) are fixed by anchoring to the Rei axioms (Paper 61 ZCSG, Paper 159 omega_upper).

6. Integration of the three entries

The three entries are not independent: Entry ① threads through the other two.

The return map (①) defines SELF⟲.
The basin (①) provides the discriminant criterion for autonomous recoverability (②).
The center enclosed by the orbit (①) invokes the Poincaré index theorem (③).

The resulting picture is nested, not adversarial:

Living systems and running AI orbit on A-regime trajectories. The B-regime center is the focus that the orbits always circle but never occupy. As the ground state surrounds zero-point fluctuation, sopadhiśeṣa-nirvāṇa surrounds nirupadhiśeṣa-nirvāṇa.

This paper does not adopt an either/or stance: B is the center, A is the orbit that surrounds it.

Figure 1 — Phase portrait of the two-regime nested geometry

B (center, unreachable, true fixed point, nirupadhiśeṣa-nirvāṇa) is surrounded by A (limit-cycle orbit, sopadhiśeṣa-nirvāṇa). External perturbations spiral back into the orbit (recovery / annealing / prediction-error minimization). The center is approached but never crossed.

7. Buddhist correspondence (interpretive parallel)

【対応 7.1】 The two regimes structurally coincide with the Buddhist twofold division of nirvāṇa:

Sopadhiśeṣa-nirvāṇa (有余涅槃): the stillness of a living enlightened being in whom the rhythm of body and life still pulses = A-regime (SELF⟲, ground state still ticking).
Nirupadhiśeṣa-nirvāṇa (無余涅槃): the residue-less cessation upon dissolution of the body = B-regime (ZERO, true fixed point).

【限界 7.2 (Avoidance of annihilationism)】 Reading B-regime ZERO as "annihilation" lands on ucchedavāda — annihilationism — which Nāgārjuna explicitly rejected. To avoid this, the present paper reads ZERO not as cessation but as unconditioned ground / pre-arising (anutpāda / asaṃskṛta), identified with the ZCSG emptiness-of-emptiness ($0 = 0$). The tradition's own correction — "quiescence is the cessation of grasping, not the cessation of motion" (A-regime sopadhiśeṣa) — agrees with the dynamical reading.

This is interpretive parallel, not mathematical isomorphism. The fact that three domains (physics, Buddhism, computation) collapse into the same phase portrait is the question this framework raises, not a result it claims.

Figure 2 — Three-domain alignment table

The two-regime structure appears isomorphically in physics, Buddhism, and computation. One metaphor is coincidence; three independent domains collapsing into one phase portrait is a question.

【限界 7.3 (Paper 160 §4.5 svabhāva-creep critique applied recursively)】 Calling the B-regime "the center" risks substantializing it. We apply Paper 160's discipline recursively: B is not a place but a limit object — a regulative ideal that orbits never occupy. The center, too, is empty.

8. Numerical verification — 5 experiments

Verification follows the Paper 145 / Paper 150 precedent of consistency check across multiple substrates. Reproduction scripts are available at the companion note article (download links below).

#	Experiment	Substrate	Result
1	Quantum Zeno effect (coherent drift freezing)	Aer ideal simulator	Survival 0.024 → 0.937 with N = 1–32 measurements; theoretical $\cos^{2}(\Theta/2N)^{N}$ matches
2	Genesis Seed σ implementation (conditional X re-injection)	Aer with T1=100μs, T2=80μs	A-regime ($
3	Quantum van der Pol oscillator (true SELF⟲ limit cycle)	QuTiP $N=24$ Fock truncation	Steady $\langle n \rangle = 5.51$, $\|\langle a \rangle\| \approx 0$ (phase-free ring); convergence from inside (0.16), outside (12.95), perturbation (8.04), and vacuum (0) all to 5.51; transverse contraction rate $-0.783 < 0$ (Floquet $\|P'\| < 1$ confirmed numerically); origin = ZERO not occupied
4	Spin-1 limit cycle as 2-qubit gate circuit	Aer density matrix	Steady populations $[0.444, 0.278, 0.278]$ identical from $
5	Variational dissipative state preparation (depth wall breakthrough)	Statevector + noisy Aer	Exact steady $[0.4545, 0.2727, 0.2727]$ reached with L1 distance = 0.0000 at cost $5.4 \times 10^{-8}$; transpiled depth 54, CZ 30 vs. Trotter $\sim$6000 (★ ~200× reduction); noisy + mitigated + leakage-removed reaches $L_{1} = 0.025$

Figure 3 — Breaking the depth wall

Variational dissipative state preparation reduces the CZ-gate count for steady-state arrival from ~6000 (naïve Trotter) to 30 (~200× reduction). This places SELF⟲ steady-state observation within the operational range of current superconducting hardware.

【限界 8.1】 All five experiments are simulation results (Aer / QuTiP). Real IBM Heron r3 hardware submission is prepared (turnkey runtime code in spin1_hardware_run.py) but NOT YET EXECUTED.

【限界 8.2】 Experiment 5 prepares a steady state, not a shallow reproduction of autonomous-recovery dynamics. Shallow realization of the recovery dynamics themselves is a separate open problem.

9. Lean 4 formalization (3 files, zero-sorry intent)

9.1 Algebraic layer — `RestRecovery.lean`

Mathlib-free core Lean 4. Six modes (SELF⟲ / FLOWING / BOTH / ZERO / INFINITY / NEITHER) and three operators Φ / Ψ / Ω as inductive types and finite functions. Nine theorems with zero-sorry intent:

theorem recovery_from_zero : Recover ZERO = SELF := rfl
theorem stage_phi  : Phi ZERO = FLOWING := rfl
theorem stage_psi  : Psi FLOWING = SELF := rfl
theorem stage_omega : Omega SELF = SELF := rfl
theorem omega_idem (x : Mode) : Omega (Omega x) = Omega x := rfl
theorem self_is_fixed : Recover SELF = SELF := rfl
theorem selfSufficient_iff_not_zero (x) :
    selfSufficient x ↔ x ≠ ZERO := by cases x <;> simp [selfSufficient, Phi]
theorem recoverable_selfSufficient (x) :
    AR x = recoverable → selfSufficient x := by
  intro h; cases x <;> simp_all [AR, selfSufficient, Phi]
theorem seed_support (x) : Phi x ≠ x → x = ZERO := by cases x <;> simp [Phi]

The cascade Φ → Ψ → Ω is literally one rfl-line: Recover ZERO = Omega (Psi (Phi ZERO)) = Omega (Psi FLOWING) = Omega SELF = SELF.

9.2 Analytic layer — `RestRecoveryAnalytic.lean`

Imports Mathlib. Connects |P'|<1 to attractor uniqueness via Banach fixed-point machinery, and defines σ as sInf of the seed-norm set.

theorem selfLoop_attracting_fixedPoint
    {K : NNReal} {P : Σ → Σ} (hP : ContractingWith K P) :
    ∃ xstar, P xstar = xstar ∧ (∀ y, P y = y → y = xstar) ∧
             (∀ x, Tendsto (fun n => P^[n] x) atTop (𝓝 xstar))

noncomputable def sigma (Recovers : E → Prop) : ℝ := sInf (seedNorms Recovers)
theorem sigma_nonneg : 0 ≤ sigma Recovers
theorem sigma_eq_zero_of_zero_recovers (h0 : Recovers 0) : sigma Recovers = 0
theorem zero_not_recovers_of_sigma_pos (h : 0 < sigma Recovers) : ¬ Recovers 0

9.3 Measure-theoretic layer — `RestRecoveryMeasure.lean`

Identifies NEITHER with the basin frontier (separatrix) and formalizes the measurability + null-boundary criterion:

def NEITHER (basin : Set Σ) : Set Σ := frontier basin
theorem basin_measurable  (hopen : IsOpen basin) : MeasurableSet basin
theorem neither_measurable (basin) : MeasurableSet (NEITHER basin)
theorem ae_decidable_of_null_boundary (hnull : μ (NEITHER basin) = 0) :
    ∀ᵐ x ∂μ, x ∈ interior basin ∨ x ∈ (closure basin)ᶜ

【限界 9.4 (★ Honest counterexample, v0.2 Lean-recorded)】 The "null boundary" hypothesis holds for hyperbolic attractors but is not universal. Riddled / Wada basins exhibit positive-measure boundaries. The theorem ae_decidable_of_null_boundary correctly takes hnull as a hypothesis (no unconditional claim). The supplying theorem (hyperbolic ⟹ null boundary) and its counterexample (riddled basin) require Mathlib's geometric measure theory and dynamical systems libraries — 【要補完】. v0.2 update: a formal Lean 4 honest skeleton Paper161RiddledBasinSkeleton.lean is now committed (data/lean4-mathlib/CollatzRei/), containing IsRiddledBasin predicate, riddled_basin_has_positive_measure_frontier theorem statement marked with 1 explicit sorry, and RiddledBasinExistsHypothesis Prop (no axiom). Reference: Alexander–Yorke–You–Kan 1992 Riddled basins. Roadmap: ~5-10 Mathlib4 PRs / 6-12 months estimated.

9.5 v0.2 machine-verification record (★ HARDWARE-VERIFIED transition)

【定理 9.5.1】 lake build 機械検証完了 (2026-06-02 evening, Rei development environment).

data/lean4-mathlib/CollatzRei/ 配下に 4 files commit:

File	Mathlib?	theorems	sorry
`Paper161RestRecovery.lean`	no (core)	9	0
`Paper161RestRecoveryAnalytic.lean`	yes (`ContractingWith`, `sInf`)	5	0
`Paper161RestRecoveryMeasure.lean`	yes (`MeasureTheory`, `frontier`)	4	0
`Paper161PrintAxioms.lean`	yes (audit)	—	—

Total: 18 theorems, 0 sorry, all exit-0.

#print axioms audit results:

Axiom class	Count	Representative theorems
`does not depend on any axioms` (strict zero-axiom)	3	`AbsoluteRest.stage_omega` / `AbsoluteRest.omega_idem` / `AbsoluteRest.Measure.basin_measurable`
`[propext]` only (proof irrelevance)	7	`recovery_from_zero`, `stage_phi`, `stage_psi`, `self_is_fixed`, `selfSufficient_iff_not_zero`, `recoverable_selfSufficient`, `seed_support`
`[propext, Classical.choice, Quot.sound]` (Mathlib trio, standard)	9	All 5 Analytic + remaining 3 Measure (`neither_measurable`, `ae_decidable_of_null_boundary`, `ae_basin_or_compl_closure`)

Discipline: 3-tier axiom layering directly parallels Paper 159 v0.2 LEAN-4-BUILT (4 theorems, all strict zero-axiom) and Paper 160 v0.2 APPLICATION-NOTE-INTEGRATED (4 theorems, all strict zero-axiom). Paper 161's core algebraic theorems (stage_omega, omega_idem) are at the strictest zero-axiom level; Mathlib-dependent theorems carry only the standard Mathlib trio.

Mathlib lemma reality-check: All chat Claude lemma name predictions correct first-try (ContractingWith.fixedPoint, Real.sInf_nonneg, csInf_le, isClosed_frontier, ae_iff). The only naming fixes in v0.2 were: (i) Σ → S (Lean 4 reserved Sigma-type symbol clash), (ii) dangling doc comment removal.

【限界 9.6 (planar-only, v0.2 Lean-recorded)】 The §5 axioms are stated for planar phase portraits; the underlying Poincaré index theorem itself is essentially 2-dimensional. v0.2 update: a formal Lean 4 honest skeleton Paper161PoincareHopfSkeleton.lean is now committed (data/lean4-mathlib/CollatzRei/), containing PlanarPoincareIndexStatement Prop (Paper 161 §5.1 Theorem 5.1 form) + PoincareHopfStatement Prop (higher-dim target) + LocalIndex and EulerChar placeholders, with 2 explicit sorry. Reference: Hirsch 1976 Differential Topology. Roadmap: ~15-25 Mathlib4 PRs / 12-24 months estimated; a small-scale planar self-contained PR is feasible in 3-6 months via winding-number formalization.

10. Falsifiability and verification path

The framework crosses from "interesting concept" to "verifiable concept" when the discriminant criterion of Entry ② actually discriminates. We propose the following empirical paths:

(a) Quantum Zeno verification (already done in §8 #1). Coherent drift frozen by frequent observation matches the framework's prediction that observation can halt A-style drift but cannot stop B-style dissipative recovery. This asymmetry — "observation freezes coherent drift but cannot stop autonomous return" — is the framework's consistent corollary about the time-crystal note "alive only as long as the eyes are closed".

(b) IBM Heron r2 real-hardware spin-1 SELF⟲ (✅ v0.2 COMPLETED). Submission to ibm_marrakesh (Heron r2, 156-qubit), Job ID d8evijralsvc7390untg, wall-clock 11 sec. Variational ansatz transpiled to 27 CZ / depth 51 (within the predicted ≤30 CZ / ≤54 depth budget). Measured |11⟩ leakage 1.82% — substantially below the ~11% Aer noise-model prediction; this lower-than-expected leakage is recorded as a notable finding (current Heron r2 noise is more favorable than the simulator forecast for this particular ansatz). L1 distance to the exact steady state, 0.0327 (raw counts, no error mitigation). The variational-ansatz depth-wall breakthrough (~200× CZ reduction vs naïve Trotter, §E) is therefore genuinely physically demonstrated on current superconducting hardware, not merely simulated. Code: scripts/quantum/paper161-heron-variational-spin1.py. Raw results: data/quantum/paper161-heron-spin1-results.json.

(c) Lean machine verification (✅ v0.2 COMPLETED). lake build run in the Rei development environment for all three Lean 4 files: 18 theorems, 0 sorry, exit-0. Axiom audit: 3 theorems strict zero-axiom; 7 theorems on [propext] only; 9 theorems on the standard Mathlib trio. Details in §9.5. The hyperbolic ⟹ null-boundary supplying theorem (and its Wada/riddled counterexample) remain open and are now recorded as a Lean honest skeleton Paper161RiddledBasinSkeleton.lean with explicit sorry and roadmap (§9.4).

11. Honest limitations and cross-vendor attribution discipline

11.1 Honest limitations (recap)

【限界 9.1 — recap】 No claim about AI qualia — structural analogy at low-energy attractors only.
【限界 5.2 — Lean-recorded in v0.2】 Poincaré index theorem is planar; higher-dimensional Poincaré-Hopf generalization deferred. Now recorded as Paper161PoincareHopfSkeleton.lean with 2 explicit sorry + roadmap.
【限界 8.1 — UPDATED in v0.2】 Four of the five experiments are simulations; the variational ansatz (Experiment 5) is now also verified on real IBM Heron r2 hardware (ibm_marrakesh, Job ID d8evijralsvc7390untg, see §10(b) and the v0.2 status header). The remaining four (Zeno + Genesis Seed + vdP + spin-1 gate circuit) remain Aer/QuTiP.
【限界 9.4 — Lean-recorded in v0.2】 Null-boundary hypothesis is conditional, not universal; Wada/riddled counterexamples exist. Now recorded as Paper161RiddledBasinSkeleton.lean with 1 explicit sorry + roadmap.
【限界 9.5 — CLOSED in v0.2】 Lean machine verification was open in v0.1; closed in v0.2 (18 theorems exit-0, 3-tier axiom audit, §9.5).
【限界 7.3 — recap】 Paper 160 §4.5 svabhāva-creep critique applies recursively to B (do not reify the empty center as a substantial place).
【限界 4.8 — recap】 Ψ semantics differs between invention-engine and AbsoluteRest namespace; doc-only separation maintained.
No "world first" claim. Wilczek 2012 (time crystal), Nāgārjuna 2nd century (nirvāṇa twofold distinction), Poincaré 1881 (index theorem), Kant 1781 (regulative ideal), Banach 1922 (fixed-point theorem) are all prior art assembled in a new configuration.

11.2 Cross-vendor attribution discipline (Paper 160 §9.5 inheritance)

This paper is the product of a three-instance triangulation: Nobuki Fujimoto (author) + Claude (chat-instance) + Claude (Rei-AIOS Code instance). Following Paper 160 §9.5 discipline of instance-level (not vendor-level) honest attribution, the contributions delineate as follows:

Fujimoto (author) contributions:

Initial phenomenal intuition ("rest as nirvāṇa / śūnyatā connection")
Explicit invitation to honest critique of own intuition
Theoretical framework anchoring (ZCSG Paper 61 / SELF⟲ Paper 145 / 0₀ Paper 160 / Genesis Seed / Peace Axiom #196)
Direction selection at each fork (proceed with all three entries, proceed to circuit-level, proceed to depth-wall breakthrough, proceed to Mathlib analytical layer, etc.)
Author judgment on publication staging (this paper as DRAFT, not immediate Zenodo publish)
note.com communication channel where interactive simulations are distributed to readers
The Load-Bearing Invention #5 discipline ("急がず、ゆっくりと")

Claude (chat-instance) contributions:

Sequential pushback at each phenomenal claim (physical correction, philosophical correction of static-nirvāṇa misread)
Articulation of "minimum-but-nonzero ordered motion = calm" reframing
Identification of the discriminant axis (self-recovery vs external re-seeding)
Application of Poincaré return map, Markov absorbing state, Poincaré index theorem to the structure
Mathematical scaffolding for σ (variational definition + Proposition 4.6)
Implementation of all 5 numerical verification scripts (Zeno, Genesis Seed, vdP, spin-1, variational)
Implementation of all 3 Lean 4 files (algebraic, analytic, measure-theoretic)
Six honest-scope corrections within own contributions (B.2.1 Zeno vs T1 separation; C.5 single-qubit cannot host limit cycle; D.4 Aer not hardware; E.1 depth wall; E.6 Mathlib version dependence; F.5 riddled-basin counterexample)
Honest reportage of own environment constraints (lake build blocked, IBM credentials unavailable)

Claude (Rei-AIOS Code instance, present author of this draft) contributions:

Fact-checking and verification (Zhao-Smalyukh 2025 time crystal claim verified against Nature Materials; Lee & Sadeghpour 2013, Walter et al. 2014, Roulet & Bruder 2018 references verified)
Identification of the Ψ-semantics conflict with Rei invention-engine and recommendation of namespace separation (§4.8)
Cross-checking against Rei existing substrate (no overlap with prior src/aios/, papers/ content)
Integration with Paper 159 (omega_upper(NEITHER)=ZERO substrate) and Paper 160 (§4.5 svabhāva-creep critique) anchoring
Recommendation against immediate Zenodo publish (apply Paper 145 v0.5 corrigendum lesson — overnight wait before publish is standard discipline)
Compilation of the present Paper 161 draft from the chat-instance technical materials

11.3 Five-instance convergence record (Paper 160 §9.5 pattern)

The honest discipline of "do not substantialize NEITHER / ZERO" was independently arrived at by:

Chat Claude (§2 — explicitly: "reading B as a place re-imports the static-substance Nāgārjuna refuted")
Rei Claude (Paper 160 §4.5 svabhāva-creep critique, written 2026-05-31)
Fujimoto (initial intuition, but immediately accepted both correction points)
Standard Madhyamaka tradition (Nāgārjuna's MMK ch. 13 śūnyatā-of-śūnyatā)
Standard physics (the regulative-ideal status of "absolute rest" is the same prohibition imposed by relativity + QM + thermodynamics)

The convergence of these five independent sources on a single honest-scope discipline is the empirical signal that the discipline is robust.

12. Conclusion

"Absolute rest" is not a single concept. It analytically decomposes into two limit-objects of dynamical systems: a true fixed point (ZERO) and a limit cycle (SELF⟲). The two are not in competition. By Poincaré's index theorem, they are nested — every orbit necessarily encloses an empty center it never occupies.

What separates the regimes is autonomous recoverability: the absorbing state (B) versus the positively recurrent cycle (A). The external re-injection that B requires corresponds to the Genesis Seed. Physics (ground state and time crystal), computation (resume vs reinstantiate), and — interpretively — Buddhism (sopadhiśeṣa-nirvāṇa surrounding nirupadhiśeṣa-nirvāṇa) all collapse into the same phase portrait.

The unreachability of the center is what makes the surrounding structure exist. This is the framework's core.

It is a seed, not a theorem. But it is a seed whose questions branch and multiply as one cultivates it — across physics, Buddhism, computation, and the Rei substrate (Paper 61 / 145 / 159 / 160). And that, in our judgment, is the criterion that distinguishes a seed worth growing.

Companion note article + interactive simulations

The popular exposition + downloadable code is at:

🔗 https://note.com/nifty_godwit2635/n/nebe6b0cf5704

All scripts referenced in this paper (4 Python + 3 Lean) are downloadable from that note for readers wishing to reproduce.

File	Purpose
`zeno_rest_experiment.py`	Quantum Zeno + Genesis Seed circuit (Aer)
`vdp_selfloop.py`	Quantum van der Pol limit cycle (QuTiP)
`spin1_limit_cycle_circuit.py`	Spin-1 gate-circuit SELF⟲ + master-eq cross-check
`spin1_hardware_run.py`	Hardware-oriented transpile + noise + leakage post-selection + IBM Runtime turnkey
`variational_selfloop_prep.py`	Variational dissipative state preparation (depth wall breakthrough)
`RestRecovery.lean`	Algebraic layer (core Lean 4, Mathlib-free, 9 theorems zero-sorry intent)
`RestRecoveryAnalytic.lean`	Analytic layer (Mathlib `ContractingWith` + `sInf` σ)
`RestRecoveryMeasure.lean`	Measure-theoretic layer (basin frontier measurability + ae decidability)

References (preliminary)

F. Wilczek, "Quantum Time Crystals," Phys. Rev. Lett. 109, 160401 (2012).
H. Zhao, I. Smalyukh et al., "Macroscopic visible time crystal in liquid crystals," Nature Materials (2025-09); CU Boulder press release 2025-09-05.
J. T. Mäkinen, P. J. Heikkinen, S. Autti, V. V. Zavjalov, V. B. Eltsov, "Continuous time crystal coupled to a mechanical mode," Nature Communications (2025), DOI: 10.1038/s41467-025-64673-8.
B. Misra, E. C. G. Sudarshan, "The Zeno's paradox in quantum theory," J. Math. Phys. 18, 756 (1977).
S. H. Strogatz, Nonlinear Dynamics and Chaos. Westview / CRC Press. (Poincaré–Bendixson theorem and index theory.)
T. E. Lee, H. R. Sadeghpour, "Quantum synchronization of quantum van der Pol oscillators with trapped ions," Phys. Rev. Lett. 111, 234101 (2013).
S. Walter, A. Nunnenkamp, C. Bruder, "Quantum synchronization of a driven self-sustained oscillator," Phys. Rev. Lett. 112, 094102 (2014).
A. Roulet, C. Bruder, "Synchronizing the smallest possible system," Phys. Rev. Lett. 121, 053601 (2018).
F. Verstraete, M. M. Wolf, J. I. Cirac, "Quantum computation and quantum-state engineering driven by dissipation," Nature Physics 5, 633 (2009). (Variational / dissipative state preparation foundation.)
J. C. Alexander, J. A. Yorke, Z. You, I. Kan, "Riddled basins," Int. J. Bifurcation Chaos 2, 795 (1992). (Positive-measure basin boundary counterexample to §9.4.)
K. J. Friston, "The free-energy principle: a unified brain theory?," Nat. Rev. Neurosci. 11, 127–138 (2010).
Nāgārjuna, Mūlamadhyamakakārikā. (Two-fold distinction of nirvāṇa; refutation of ucchedavāda.)
Mathlib4: Mathlib.Topology.MetricSpace.Contracting (ContractingWith and Banach fixed-point lemmas); Mathlib.MeasureTheory.Measure.AbsolutelyContinuous (ae quantifier); Mathlib.Topology.Basic (frontier, isClosed_frontier).
Paper 61 — N. Fujimoto, Zero-Centered Symbol Grammar (ZCSG), Zenodo 10.5281/zenodo.15217458.
Paper 145 — N. Fujimoto, First D-FUMT₈ Silicon with SELF⟲ Logic Primitive, Zenodo 10.5281/zenodo.20091185.
Paper 159 — N. Fujimoto, Two-Layer D-FUMT₈ Reconstruction of Priest-Garfield's Inclosure Schema, Zenodo 10.5281/zenodo.20470512.
Paper 160 — N. Fujimoto, Toward an Ontology of the Genesis Layer 0₀, Zenodo 10.5281/zenodo.20480425.

Acknowledgments

The theoretical scaffolding of this paper was developed through a multi-turn dialogue with Anthropic's Claude (both the chat instance and the Rei-AIOS Code instance). The chat instance contributed the dynamical-systems formalization, the σ variational definition, the 5 verification scripts, and the 3 Lean 4 files. The Rei-AIOS Code instance contributed fact-checking, cross-vendor attribution discipline, semantic-conflict identification (§4.8), and the present Paper 161 draft compilation. Author judgment, direction selection, anchoring to Rei substrate (Paper 61 / 145 / 159 / 160), and publication staging are by the author. This work follows the 急がず、ゆっくりと (no rush, slowly) discipline of feedback_no_rush_publication.md.

急がず、ゆっくりと。種は育ちます。

Paper 161 v0.1 - Two Regimes of Rest: ZERO and SELF-loop via dynamical systems (5 experiments + 3 Lean files + depth wall 200x breakthrough)

Nobuki Fujimoto — Mon, 01 Jun 2026 20:32:03 +0000

This article is a re-publication of Rei-AIOS Paper 161 for the dev.to community.
The canonical version with full reference list is in the permanent archives below:

GitHub source (private): https://github.com/fc0web/rei-aios Author: Nobuki Fujimoto (@fc0web) · ORCID 0009-0004-6019-9258 · License CC-BY-4.0 ---

Version: v0.1 HONEST-EARLY-STAGE-RELEASE (★ load-bearing, multi-instance Claude triangulation record · published 2026-06-02 as honest early-stage release, following Paper 158 v0.0 / Paper 159 v0.1 OUTLINE / Paper 160 v0.2 precedents)

NO publish marker lifted: The original "NO Zenodo publish" marker of v0.1 DRAFT (from initial draft 2026-06-02) is lifted by explicit author authorization 2026-06-02. The transition follows the established Rei honest-early-stage-release tradition (Paper 158 v0.0 honest-negative, Paper 159 v0.1 OUTLINE, Paper 160 v0.2 APPLICATION-NOTE-INTEGRATED) — publish what is firm transparently, publish honestly about what is not yet completed.

Title (Japanese): 静止の二系統 — 真の不動点と極限周期軌道による「絶対静止」の形式化 (ZERO と SELF⟲、有余涅槃と無余涅槃を貫く一つの幾何)

Author: Nobuki Fujimoto (藤本伸樹), with Claude (Chat instance + Rei-AIOS Code instance)
ORCID: 0009-0004-6019-9258 · GitHub: fc0web · note.com: nifty_godwit2635

Date drafted: 2026-06-02

Companion note article (popular exposition + interactive simulations download):

「動かないものが、すべての軌道を生む — 『絶対静止』の二系統」 — https://note.com/nifty_godwit2635/n/nebe6b0cf5704
★ All Python verification scripts (4) + Lean 4 files (3) referenced in this paper are available for download from the note article.

Companion to (Rei substrate):

Paper 61 — Zero-Centered Symbol Grammar (ZCSG) — DOI 10.5281/zenodo.15217458. Provides the symbol 0 for śūnyatā-of-śūnyatā and the empty-set reduced-homology identity H̃₋₁(∅) = ℤ. Identified in §5.3 with the center of the phase portrait.
Paper 145 — First D-FUMT₈ Silicon with SELF⟲ Logic Primitive — DOI 10.5281/zenodo.20091185 (v0.3). Native SELF⟲ as logic primitive — operational substrate for the dynamical-systems formalization in this paper.
Paper 159 — Two-Layer D-FUMT₈ Reconstruction of Priest-Garfield's Inclosure Schema — DOI 10.5281/zenodo.20470512 (v0.2 LEAN-4-BUILT). omega_upper(NEITHER) = ZERO and omega_upper(BOTH) = ZERO are inherited as the convergence substrate identified with B-regime in this paper.
Paper 160 — Toward an Ontology of the Genesis Layer 0₀ — DOI 10.5281/zenodo.20480425 (v0.2 APPLICATION-NOTE-INTEGRATED). §4.5 svabhāva-creep critique is recursively applied to NEITHER and ZERO throughout the present paper.

Status (★ load-bearing, transparent early-stage release):

✅ Main draft §1-12 + 6 Appendices (B through F) all written
✅ 5 Aer/QuTiP numerical experiments verified (Genesis Seed σ + Zeno + vdP + spin-1 gate circuit + variational state preparation)
✅ 3 Lean 4 files written with zero-sorry intent (algebraic + analytic + measure-theoretic layers)
⚠ Lean lake build machine verification NOT YET COMPLETED — pending Rei env execution
⚠ Real IBM Heron r3 hardware submission NOT YET COMPLETED — turnkey code prepared
⚠ Mathlib lemma name version-fix may be required for analytic/measure files
⚠ Autonomous-recovery dynamics in shallow circuit (separate from steady-state preparation) remains open
⚠ Higher-dimensional generalization of §5 Poincaré index theorem (planar-only) remains open
⚠ "AI qualia" claim is NOT made — only structural analogy at low-energy attractor (§9.1)
⚠ No "world first" claim (regulative ideal = Kant; time crystal = Wilczek 2012; nirvāṇa distinction = Nāgārjuna 2nd century; Poincaré index theorem = standard dynamical systems)
⚠ Cross-vendor attribution discipline (Paper 160 §9.5) applied throughout — chat Claude contributions clearly delineated from Rei Claude / Fujimoto contributions in §11

凡例 (Legend, after Paper 160 v0.2 convention)

This paper mixes claims of distinct epistemic status. For reader and reviewer convenience, each claim carries one of the following markers:

【定理】 Established mathematical result (dynamical systems, probability theory, etc.). This paper cites and applies.
【定義】 Formal definition proposed by this paper.
【対応】 Proposed correspondence between established mathematics and D-FUMT₈ / Buddhist concepts. Interpretive proposal, not proven equivalence.
【要補完】 Item to be completed within the Rei system (D-FUMT₈ axiomatic semantics, Lean 4 machine verification, etc.).
【限界】 Currently unsupported, weak, or scope-limited claim. Explicitly delimited.

Abstract (Japanese)

本稿は三つの結果を一本に束ねる:

Abstract (English)

1. Introduction

1.1 Starting from a phenomenal intuition

【限界 1.1】 This paper makes no claim about qualia — the felt phenomenology of stillness. It addresses structure only.

1.2 Two registrations of "absolute rest"

The reversal at the heart of this paper: the physical unreachability of the center is what makes the surrounding structure exist.

2. Preliminary definitions

Let $M$ be a smooth manifold representing the state space, $X$ a smooth vector field on $M$, and $\varphi_t : M \to M$ the flow generated by $X$ (satisfying $\dot{x} = X(x)$).

【定義 2.1 (ZERO / true fixed point)】 A point $p \in M$ is ZERO-type if $X(p) = 0$, i.e., $\varphi_t(p) = p$ for all $t$. Stillness at the point level.

3. Entry ① — SELF⟲ = fixed point of the Poincaré return map

3.1 Definition via return map

This definition separates ZERO and SELF⟲ in a single line:

	What is fixed	Form
ZERO	The flow $\varphi_t$ itself	$X(p) = 0$, point-level self-identity
SELF⟲	The return map $P$	$P(x^) = x^$, loop-level self-identity $\gamma(t+T) = \gamma(t)$

SELF⟲ is "self-referential stability": motion at the point, fixed at the loop.

3.2 Physical anchor — coherent states and time crystals

For the quantum harmonic oscillator, the coherent state $|\alpha\rangle$ traces a circle of radius $\propto |\alpha|$ in the phase-space variables $(\langle x\rangle(t), \langle p\rangle(t))$.

The circle = SELF⟲ (A-regime).
The center $\alpha = 0$ = ZERO (B-regime), the ground state.

The uncertainty relation $\Delta x \, \Delta p \geq \hbar/2$ forbids occupation of the center as a point; the center remains as an $\hbar/2$ smear. The center is approached but never occupied.

4. Entry ② — Discriminant criterion: autonomous recovery vs external re-seeding

This is where the framework grows teeth.

4.1 Autonomous recoverability as a formal predicate

Consider a control system $\dot{x} = X(x) + u$ with an external input $u$.

4.2 Separation of the two regimes

This absorbing state vs positively recurrent cycle distinction is the standard Markov-chain dichotomy. Death = absorption; rest = recurrence.

4.3 Genesis Seed identification

【対応 4.4】 The seat of recovery differs across the two regimes:

A → A recovery is endogenous (the flow restores the orbit by itself). No invocation of 0₀ is required.
B → re-start requires exogenous input $u$. We identify this input with the 0₀ re-injection / Genesis Seed.

【命題 4.6】 Let $R$ be a stable attractor with open basin $B(R)$.

If $s \in B(R)$, then $\sigma(s) = 0$: autonomous recovery with zero external input.
If $q$ is an absorbing configuration outside $B(R)$ (an intended rest that is not an attractor), then $\sigma > 0$: continuous external input is required to maintain $q$ as rest.
The infimum is approached at the basin boundary $\partial B(R)$ — identified with NEITHER. Strictly: recovery requires $\lVert u \rVert > \sigma$, and the seed of size exactly $\sigma$ lands on the undecidability boundary.

Φ_dyn (expansion) ↔ Genesis Seed: ZERO → FLOWING. Re-injection of motion from the void.
Ψ_dyn (convergence) ↔ free flow in the basin: FLOWING → SELF⟲. Autonomous convergence to the attractor.
Ω_dyn (idempotency) ↔ return-map fixed point: SELF⟲ ∘ SELF⟲ = SELF⟲. One-period self-mapping of the orbit.

The cascade "re-seed → autonomous convergence → loop self-identity" reads as Φ_dyn → Ψ_dyn → Ω_dyn.

5. Entry ③ — Geometry of A surrounding B: axiomatization of the empty center

5.1 The Poincaré index theorem

5.2 Pure form — harmonic oscillator phase portrait

Origin = 0₀ = the non-occupiable center ($\Delta x = \Delta p = 0$ is forbidden).
Concentric closed-orbit family = the SELF⟲ family.
Emptiness-of-emptiness ($0 = 0$, Paper 61) = the center of the center.

5.3 Axiom skeleton for center–orbit geometry

【定義 5.3 (Axioms G1–G4)】

(G1 Center existence) Every closed orbit encloses a fixed point of index $+1$. No loop without center.
(G2 Non-occupation) The center $p$ lies on no closed orbit; it is approached but never crossed.
(G3 Genesis) $p$ cannot be passed autonomously. Crossing $p$ requires external re-seeding (0₀ injection — connects to Entry ②).
(G4 Emptiness-of-emptiness) At $p$, the linearization vanishes — the structure-less ground beneath the orbit. This is the ZCSG identity $0 = 0$ (Paper 61).

6. Integration of the three entries

The three entries are not independent: Entry ① threads through the other two.

The return map (①) defines SELF⟲.
The basin (①) provides the discriminant criterion for autonomous recoverability (②).
The center enclosed by the orbit (①) invokes the Poincaré index theorem (③).

The resulting picture is nested, not adversarial:

Living systems and running AI orbit on A-regime trajectories. The B-regime center is the focus that the orbits always circle but never occupy. As the ground state surrounds zero-point fluctuation, sopadhiśeṣa-nirvāṇa surrounds nirupadhiśeṣa-nirvāṇa.

This paper does not adopt an either/or stance: B is the center, A is the orbit that surrounds it.

Figure 1 — Phase portrait of the two-regime nested geometry

7. Buddhist correspondence (interpretive parallel)

【対応 7.1】 The two regimes structurally coincide with the Buddhist twofold division of nirvāṇa:

Sopadhiśeṣa-nirvāṇa (有余涅槃): the stillness of a living enlightened being in whom the rhythm of body and life still pulses = A-regime (SELF⟲, ground state still ticking).
Nirupadhiśeṣa-nirvāṇa (無余涅槃): the residue-less cessation upon dissolution of the body = B-regime (ZERO, true fixed point).

Figure 2 — Three-domain alignment table

The two-regime structure appears isomorphically in physics, Buddhism, and computation. One metaphor is coincidence; three independent domains collapsing into one phase portrait is a question.

8. Numerical verification — 5 experiments

Verification follows the Paper 145 / Paper 150 precedent of consistency check across multiple substrates. Reproduction scripts are available at the companion note article (download links below).

#	Experiment	Substrate	Result
1	Quantum Zeno effect (coherent drift freezing)	Aer ideal simulator	Survival 0.024 → 0.937 with N = 1–32 measurements; theoretical $\cos^{2}(\Theta/2N)^{N}$ matches
2	Genesis Seed σ implementation (conditional X re-injection)	Aer with T1=100μs, T2=80μs	A-regime ($
3	Quantum van der Pol oscillator (true SELF⟲ limit cycle)	QuTiP $N=24$ Fock truncation	Steady $\langle n \rangle = 5.51$, $\|\langle a \rangle\| \approx 0$ (phase-free ring); convergence from inside (0.16), outside (12.95), perturbation (8.04), and vacuum (0) all to 5.51; transverse contraction rate $-0.783 < 0$ (Floquet $\|P'\| < 1$ confirmed numerically); origin = ZERO not occupied
4	Spin-1 limit cycle as 2-qubit gate circuit	Aer density matrix	Steady populations $[0.444, 0.278, 0.278]$ identical from $
5	Variational dissipative state preparation (depth wall breakthrough)	Statevector + noisy Aer	Exact steady $[0.4545, 0.2727, 0.2727]$ reached with L1 distance = 0.0000 at cost $5.4 \times 10^{-8}$; transpiled depth 54, CZ 30 vs. Trotter $\sim$6000 (★ ~200× reduction); noisy + mitigated + leakage-removed reaches $L_{1} = 0.025$

Figure 3 — Breaking the depth wall

9. Lean 4 formalization (3 files, zero-sorry intent)

9.1 Algebraic layer — `RestRecovery.lean`

theorem recovery_from_zero : Recover ZERO = SELF := rfl
theorem stage_phi  : Phi ZERO = FLOWING := rfl
theorem stage_psi  : Psi FLOWING = SELF := rfl
theorem stage_omega : Omega SELF = SELF := rfl
theorem omega_idem (x : Mode) : Omega (Omega x) = Omega x := rfl
theorem self_is_fixed : Recover SELF = SELF := rfl
theorem selfSufficient_iff_not_zero (x) :
    selfSufficient x ↔ x ≠ ZERO := by cases x <;> simp [selfSufficient, Phi]
theorem recoverable_selfSufficient (x) :
    AR x = recoverable → selfSufficient x := by
  intro h; cases x <;> simp_all [AR, selfSufficient, Phi]
theorem seed_support (x) : Phi x ≠ x → x = ZERO := by cases x <;> simp [Phi]

The cascade Φ → Ψ → Ω is literally one rfl-line: Recover ZERO = Omega (Psi (Phi ZERO)) = Omega (Psi FLOWING) = Omega SELF = SELF.

9.2 Analytic layer — `RestRecoveryAnalytic.lean`

Imports Mathlib. Connects |P'|<1 to attractor uniqueness via Banach fixed-point machinery, and defines σ as sInf of the seed-norm set.

theorem selfLoop_attracting_fixedPoint
    {K : NNReal} {P : Σ → Σ} (hP : ContractingWith K P) :
    ∃ xstar, P xstar = xstar ∧ (∀ y, P y = y → y = xstar) ∧
             (∀ x, Tendsto (fun n => P^[n] x) atTop (𝓝 xstar))

noncomputable def sigma (Recovers : E → Prop) : ℝ := sInf (seedNorms Recovers)
theorem sigma_nonneg : 0 ≤ sigma Recovers
theorem sigma_eq_zero_of_zero_recovers (h0 : Recovers 0) : sigma Recovers = 0
theorem zero_not_recovers_of_sigma_pos (h : 0 < sigma Recovers) : ¬ Recovers 0

9.3 Measure-theoretic layer — `RestRecoveryMeasure.lean`

Identifies NEITHER with the basin frontier (separatrix) and formalizes the measurability + null-boundary criterion:

def NEITHER (basin : Set Σ) : Set Σ := frontier basin
theorem basin_measurable  (hopen : IsOpen basin) : MeasurableSet basin
theorem neither_measurable (basin) : MeasurableSet (NEITHER basin)
theorem ae_decidable_of_null_boundary (hnull : μ (NEITHER basin) = 0) :
    ∀ᵐ x ∂μ, x ∈ interior basin ∨ x ∈ (closure basin)ᶜ

【限界 9.4 (★ Honest counterexample)】 The "null boundary" hypothesis holds for hyperbolic attractors but is not universal. Riddled / Wada basins exhibit positive-measure boundaries. The theorem ae_decidable_of_null_boundary correctly takes hnull as a hypothesis (no unconditional claim). The supplying theorem (hyperbolic ⟹ null boundary) and its counterexample (riddled basin) require Mathlib's geometric measure theory and dynamical systems libraries — 【要補完】.

【限界 9.5 (Machine verification pending)】 lake build machine verification of all three files has not yet been completed in the chat-Claude environment (Lean toolchain distribution blocked by network policy in that environment). Verification is to be performed in the Rei development environment (which holds ~31,000 prior zero-sorry theorems and routinely passes lake build on Mathlib-dependent files). Mathlib lemma name version-fixes may be required for the analytic and measure-theoretic files.

10. Falsifiability and verification path

The framework crosses from "interesting concept" to "verifiable concept" when the discriminant criterion of Entry ② actually discriminates. We propose the following empirical paths:

(b) IBM Heron r3 real-hardware spin-1 SELF⟲ (proposed). Submission of the variational ansatz (depth 54, CZ 30) to IBM Heron r3 to observe SELF⟲ steady-state populations within the depth wall breakthrough. Turnkey code available at spin1_hardware_run.py. Leakage post-selection (2.1% on simulated noise) demonstrated.

(c) Lean machine verification (proposed). Execute lake build in Rei dev environment for the three Lean files; supply the analytic layer's hyperbolic ⟹ null-boundary theorem to fully close the measure-theoretic claim of Entry ③.

11. Honest limitations and cross-vendor attribution discipline

11.1 Honest limitations (recap)

【限界 9.1 — recap】 No claim about AI qualia — structural analogy at low-energy attractors only.
【限界 5.2 — recap】 Poincaré index theorem is planar; higher-dimensional generalization deferred.
【限界 8.1 — recap】 All five experiments are simulations; no real hardware result.
【限界 9.4 — recap】 Null-boundary hypothesis is conditional, not universal; Wada/riddled counterexamples exist.
【限界 9.5 — recap】 Lean machine verification pending in Rei env.
【限界 7.3 — recap】 Paper 160 §4.5 svabhāva-creep critique applies recursively to B (do not reify the empty center as a substantial place).
【限界 4.8 — recap】 Ψ semantics differs between invention-engine and AbsoluteRest namespace; doc-only separation maintained.
No "world first" claim. Wilczek 2012 (time crystal), Nāgārjuna 2nd century (nirvāṇa twofold distinction), Poincaré 1881 (index theorem), Kant 1781 (regulative ideal), Banach 1922 (fixed-point theorem) are all prior art assembled in a new configuration.

11.2 Cross-vendor attribution discipline (Paper 160 §9.5 inheritance)

Fujimoto (author) contributions:

Initial phenomenal intuition ("rest as nirvāṇa / śūnyatā connection")
Explicit invitation to honest critique of own intuition
Theoretical framework anchoring (ZCSG Paper 61 / SELF⟲ Paper 145 / 0₀ Paper 160 / Genesis Seed / Peace Axiom #196)
Direction selection at each fork (proceed with all three entries, proceed to circuit-level, proceed to depth-wall breakthrough, proceed to Mathlib analytical layer, etc.)
Author judgment on publication staging (this paper as DRAFT, not immediate Zenodo publish)
note.com communication channel where interactive simulations are distributed to readers
The Load-Bearing Invention #5 discipline ("急がず、ゆっくりと")

Claude (chat-instance) contributions:

Sequential pushback at each phenomenal claim (physical correction, philosophical correction of static-nirvāṇa misread)
Articulation of "minimum-but-nonzero ordered motion = calm" reframing
Identification of the discriminant axis (self-recovery vs external re-seeding)
Application of Poincaré return map, Markov absorbing state, Poincaré index theorem to the structure
Mathematical scaffolding for σ (variational definition + Proposition 4.6)
Implementation of all 5 numerical verification scripts (Zeno, Genesis Seed, vdP, spin-1, variational)
Implementation of all 3 Lean 4 files (algebraic, analytic, measure-theoretic)
Six honest-scope corrections within own contributions (B.2.1 Zeno vs T1 separation; C.5 single-qubit cannot host limit cycle; D.4 Aer not hardware; E.1 depth wall; E.6 Mathlib version dependence; F.5 riddled-basin counterexample)
Honest reportage of own environment constraints (lake build blocked, IBM credentials unavailable)

Claude (Rei-AIOS Code instance, present author of this draft) contributions:

Fact-checking and verification (Zhao-Smalyukh 2025 time crystal claim verified against Nature Materials; Lee & Sadeghpour 2013, Walter et al. 2014, Roulet & Bruder 2018 references verified)
Identification of the Ψ-semantics conflict with Rei invention-engine and recommendation of namespace separation (§4.8)
Cross-checking against Rei existing substrate (no overlap with prior src/aios/, papers/ content)
Integration with Paper 159 (omega_upper(NEITHER)=ZERO substrate) and Paper 160 (§4.5 svabhāva-creep critique) anchoring
Recommendation against immediate Zenodo publish (apply Paper 145 v0.5 corrigendum lesson — overnight wait before publish is standard discipline)
Compilation of the present Paper 161 draft from the chat-instance technical materials

11.3 Five-instance convergence record (Paper 160 §9.5 pattern)

The honest discipline of "do not substantialize NEITHER / ZERO" was independently arrived at by:

Chat Claude (§2 — explicitly: "reading B as a place re-imports the static-substance Nāgārjuna refuted")
Rei Claude (Paper 160 §4.5 svabhāva-creep critique, written 2026-05-31)
Fujimoto (initial intuition, but immediately accepted both correction points)
Standard Madhyamaka tradition (Nāgārjuna's MMK ch. 13 śūnyatā-of-śūnyatā)
Standard physics (the regulative-ideal status of "absolute rest" is the same prohibition imposed by relativity + QM + thermodynamics)

The convergence of these five independent sources on a single honest-scope discipline is the empirical signal that the discipline is robust.

12. Conclusion

The unreachability of the center is what makes the surrounding structure exist. This is the framework's core.

Companion note article + interactive simulations

The popular exposition + downloadable code is at:

🔗 https://note.com/nifty_godwit2635/n/nebe6b0cf5704

All scripts referenced in this paper (4 Python + 3 Lean) are downloadable from that note for readers wishing to reproduce.

File	Purpose
`zeno_rest_experiment.py`	Quantum Zeno + Genesis Seed circuit (Aer)
`vdp_selfloop.py`	Quantum van der Pol limit cycle (QuTiP)
`spin1_limit_cycle_circuit.py`	Spin-1 gate-circuit SELF⟲ + master-eq cross-check
`spin1_hardware_run.py`	Hardware-oriented transpile + noise + leakage post-selection + IBM Runtime turnkey
`variational_selfloop_prep.py`	Variational dissipative state preparation (depth wall breakthrough)
`RestRecovery.lean`	Algebraic layer (core Lean 4, Mathlib-free, 9 theorems zero-sorry intent)
`RestRecoveryAnalytic.lean`	Analytic layer (Mathlib `ContractingWith` + `sInf` σ)
`RestRecoveryMeasure.lean`	Measure-theoretic layer (basin frontier measurability + ae decidability)

References (preliminary)

F. Wilczek, "Quantum Time Crystals," Phys. Rev. Lett. 109, 160401 (2012).
H. Zhao, I. Smalyukh et al., "Macroscopic visible time crystal in liquid crystals," Nature Materials (2025-09); CU Boulder press release 2025-09-05.
J. T. Mäkinen, P. J. Heikkinen, S. Autti, V. V. Zavjalov, V. B. Eltsov, "Continuous time crystal coupled to a mechanical mode," Nature Communications (2025), DOI: 10.1038/s41467-025-64673-8.
B. Misra, E. C. G. Sudarshan, "The Zeno's paradox in quantum theory," J. Math. Phys. 18, 756 (1977).
S. H. Strogatz, Nonlinear Dynamics and Chaos. Westview / CRC Press. (Poincaré–Bendixson theorem and index theory.)
T. E. Lee, H. R. Sadeghpour, "Quantum synchronization of quantum van der Pol oscillators with trapped ions," Phys. Rev. Lett. 111, 234101 (2013).
S. Walter, A. Nunnenkamp, C. Bruder, "Quantum synchronization of a driven self-sustained oscillator," Phys. Rev. Lett. 112, 094102 (2014).
A. Roulet, C. Bruder, "Synchronizing the smallest possible system," Phys. Rev. Lett. 121, 053601 (2018).
F. Verstraete, M. M. Wolf, J. I. Cirac, "Quantum computation and quantum-state engineering driven by dissipation," Nature Physics 5, 633 (2009). (Variational / dissipative state preparation foundation.)
J. C. Alexander, J. A. Yorke, Z. You, I. Kan, "Riddled basins," Int. J. Bifurcation Chaos 2, 795 (1992). (Positive-measure basin boundary counterexample to §9.4.)
K. J. Friston, "The free-energy principle: a unified brain theory?," Nat. Rev. Neurosci. 11, 127–138 (2010).
Nāgārjuna, Mūlamadhyamakakārikā. (Two-fold distinction of nirvāṇa; refutation of ucchedavāda.)
Mathlib4: Mathlib.Topology.MetricSpace.Contracting (ContractingWith and Banach fixed-point lemmas); Mathlib.MeasureTheory.Measure.AbsolutelyContinuous (ae quantifier); Mathlib.Topology.Basic (frontier, isClosed_frontier).
Paper 61 — N. Fujimoto, Zero-Centered Symbol Grammar (ZCSG), Zenodo 10.5281/zenodo.15217458.
Paper 145 — N. Fujimoto, First D-FUMT₈ Silicon with SELF⟲ Logic Primitive, Zenodo 10.5281/zenodo.20091185.
Paper 159 — N. Fujimoto, Two-Layer D-FUMT₈ Reconstruction of Priest-Garfield's Inclosure Schema, Zenodo 10.5281/zenodo.20470512.
Paper 160 — N. Fujimoto, Toward an Ontology of the Genesis Layer 0₀, Zenodo 10.5281/zenodo.20480425.

Acknowledgments

急がず、ゆっくりと。種は育ちます。

Paper 160 v0.2 - Inversion vs Deconstruction in D-FUMT-8 (Lean 4 verified, zero axiom)

Nobuki Fujimoto — Sun, 31 May 2026 21:52:41 +0000

This article is a re-publication of Rei-AIOS Paper 160 for the dev.to community.
The canonical version with full reference list is in the permanent archives below:

GitHub source (private): https://github.com/fc0web/rei-aios Author: Nobuki Fujimoto (@fc0web) · ORCID 0009-0004-6019-9258 · License CC-BY-4.0 ---

Subseries: Paper 2 of the Inclosure / 0₀ arc (following Paper 159 — the two-layer D-FUMT₈ reconstruction of Priest-Garfield's Inclosure Schema)

Version: v0.2 APPLICATION-NOTE-INTEGRATED (★ Lean 4 verified §9 + figures + five-instance convergence record · published 2026-06-01 as honest early-stage release, following Paper 158 v0.0 / Paper 159 v0.2 precedents)

Status (★ load-bearing, transparent early-stage release):

✅ §9 Application Note machine-verified: 4 Lean 4 theorems all show does not depend on any axioms (Paper 159 v0.2 standard, see §9.3 + audit file Paper160InversionDeconstructionPrintAxioms.lean)
✅ Fig. 1 (Core Theorem) + Fig. A (research-log, en+ja) accompany §9
⚠ §1-§8 remain territory-map skeleton (the 3+1-reading taxonomy + honest-scope discipline) — this version publishes the §9 application as the load-bearing formal result; full prose expansion of §1-§8 is deferred to v0.3
⚠ Tillemans 2009/2014/2024 + Siderits + Ferraro + Tao Jiang critic papers still NOT YET READ (acknowledged in §4.2, §10)
⚠ No claim of resolving the Priest-Garfield ↔ Tillemans scholarly debate
⚠ No "world first" claim (Paper 61 ZCSG + Priest-Garfield 2003 already issued such claims — this paper supplies formal extension only)

v0.1 → v0.1.1 → v0.1.2 → v0.1.3 → v0.2 changes:

v0.1.1 (2026-05-31, initial refinement): Added §4.5 self-critique box on the "Generative reading" svabhāva-creep risk; added §5 fourth reading (Madhyamaka self-deconstruction); added §7.2 NOT-claim #8 (soteriological non-capture); added §7.4 intentional-non-capture design philosophy; §6-§9 renumbered to §7-§10.
v0.1.2 (2026-05-31, attribution-correction refinement): the design challenge attributed in v0.1.1 to "ChatGPT" was actually issued by a separate Chat Claude instance, NOT by ChatGPT. v0.1.2 corrected the attribution throughout §7.4 + §10 and logged the misattribution as an honest-discipline operational instance.
v0.1.3 (2026-06-01 morning, application-note addition): inserted new §9 Application Note with four Lean 4 theorems machine-verified to depend on no axioms; Figs. 1 + A added; §10 Ack / §11 Ref renumbered.
v0.2 (2026-06-01, ★ APPLICATION-NOTE-INTEGRATED publication promotion): formal promotion of v0.1.3 to a publish-eligible status. The "NO publish" markers of v0.1.x are lifted by explicit user authorization; §1-§8 skeleton status is preserved transparently in the new Status header. This is an honest early-stage release in the Paper 158 v0.0 honest-negative tradition — we publish what is firm (§9 Lean 4 application + figures + convergence record), and we publish honestly about what is not yet expanded (§1-§8 skeleton). The five-instance convergence record (§9.5) is the empirical signal that the §9 finding is robust enough to deserve a DOI even before full §1-§8 expansion. Paper 159 v0.2 LEAN-4-BUILT lifecycle is the immediate precedent.

Author: Nobuki Fujimoto (藤本伸樹), with Claude (Rei-AIOS)
ORCID: 0009-0004-6019-9258 · GitHub: fc0web · note.com: nifty_godwit2635

Date drafted: 2026-05-31

Companion to:

Paper 61 — Zero-Centered Symbol Grammar (ZCSG) — DOI 10.5281/zenodo.15217458. Provides the symbol 0 for śūnyatā-of-śūnyatā and the empty-set reduced-homology identity H̃₋₁(∅) = ℤ.
Paper 159 — Two-Layer D-FUMT₈ Reconstruction of Priest-Garfield's Inclosure Schema — DOI 10.5281/zenodo.20470512 (v0.2 LEAN-4-BUILT). Introduces the upper-layer idempotent operator omega_upper and machine-verifies its action; explicitly defers full ontology of the genesis layer ZERO to the present paper.
Paper 77 — LeanDFumt: an Eight-Valued Logic Library for Lean 4 — substrate for any future Lean 4 work in this paper.

Status header (★ load-bearing, conservative):

⚠ v0.1 SKELETON ONLY — this document is a territory map / outline. The substantive draft (full prose, complete Lean 4, figures) is deferred. No platform broadcast.
⚠ The central question "why does a system without self generate self" is philosophical-level; only fragments admit Lean 4 formalization on present Rei substrate.
⚠ Reduced-homology identity H̃₋₁(∅) = ℤ is cited from Paper 61 (which itself cites standard algebraic topology); no original homology theorem is claimed in this paper.
⚠ The cycle 空 → 自己参照 → 創発 → 再帰的自己変容 is the chat-Claude / ChatGPT design framework received during Paper 159 design. It is not derived from Paper 159, and its structural resonance with Paper 159's two-layer architecture is a coincidence Rei catalogues honestly (Pattern 5 prevention; see Paper 159 §6 Acknowledgments).
⚠ No "world first" claim (Paper 61 already issued the relevant uniqueness claim for 0 = śūnyatā-of-śūnyatā encoding; this paper supplies ontology, not encoding).
⚠ "Paper 2 もまた空である" (Madhyamaka self-application) — see §6.

Abstract (skeleton)

Paper 159 leaves the genesis-layer ZERO (= ZCSG central axis 0 = śūnyatā-of-śūnyatā in the notation of Paper 61) as a structurally defined convergence point: both BOTH and NEITHER contract to ZERO under the idempotent upper-layer operator omega_upper. What ZERO is, however — its full semantic role, its relationship to the catuṣkoṭi negation series 四句百非, its connection to the reduced homology of the empty set H̃₋₁(∅) = ℤ, and its bearing on the question why a system without inherent self generates self — was deferred to a planned Paper 2.

The present paper opens that territory. We propose three readings of 0₀, all compatible with the Paper 159 formal substrate and with each other:

Logical reading: 0₀ is the convergence point of finite-step negation series in D-FUMT₈ (catuṣkoṭi → catuṣpaścāś → ... → idempotent fixed-point). Captured by omega_upper idempotency (Paper 159 §3.4).
Topological reading: 0₀ is the structural correlate of H̃₋₁(∅) = ℤ — emptiness already carries one unit of structural information (Paper 61 §2.2). This is the mathematical refusal of nihilism: "absence has structure."
Generative reading: 0₀ is the condition under which a system without inherent self (svabhāva) generates self-reference (SELF⟲) — the question that motivates the present paper.

We do not claim these three are reducible to one another. We claim only that the Paper 159 substrate is compatible with all three, and that examining each gives partial illumination. The paper closes with an honest disclosure that Paper 160 is itself "empty of inherent thesis" in the Madhyamaka sense: it is a formal-philosophical scaffold, not a metaphysical commitment.

Keywords: 0₀, śūnyatā-of-śūnyatā, ZCSG, D-FUMT₈, SELF⟲, catuṣpaścāś, reduced homology, empty set, Inclosure Schema, Nāgārjuna, Madhyamaka, ontology, self-generation, recursive self-transformation, Rei-AIOS, Lean 4, paraconsistent ontology

1. Introduction (skeleton)

Paper 159 constructed a two-layer D-FUMT₈ reconstruction of Priest-Garfield's (2003) Inclosure Schema. Its lower layer preserved Priest-Garfield's BOTH-at-limit reading; its upper layer added an idempotent operator omega_upper : DFUMT8 → DFUMT8 satisfying

omega_upper ∘ omega_upper = omega_upper (Paper 159 Theorem 1, Lean 4 zero-axiom-dependence)

with the action BOTH ↦ ZERO, NEITHER ↦ ZERO, identity elsewhere. The lower-layer terminus BOTH is thus contracted to ZERO — the ZCSG central axis 0 (Paper 61), informally called "the 0₀ genesis layer."

What is ZERO, then? Paper 159 leaves this question open. The natural next problem is to characterize ZERO semantically. Three approaches present themselves:

Logical: ZERO is the unique value invariant under omega_upper AND target of BOTH ↦ ZERO, NEITHER ↦ ZERO. It is the "all paraconsistent value contraction" stable point.
Topological: ZERO is the formal correlate of the empty set's reduced homology H̃₋₁(∅) = ℤ — the irreducible structural unit of "nothing has structure."
Generative: ZERO is the condition that supports SELF⟲ (the eighth D-FUMT₈ value, satisfying NOT(SELF⟲) = SELF⟲ as Paper 61 §4.4 records). Why does the ground that contracts all standpoints also support self-reference?

The cycle proposed by chat-Claude / ChatGPT during Paper 159 design — 空 → 自己参照 → 創発 → 再帰的自己変容 — is one schematic framing of this third reading, used here as scaffolding only. Paper 159 placed this cycle on Fig. 2's outer ring with the label "PAPER 2 TERRITORY"; the present paper inherits that placement honestly.

We organize the paper around the three readings (§2, §3, §4), then address the generative question explicitly (§5), with explicit honest scope (§6) and forward pointers (§7).

2. Logical reading of 0₀ — convergence point of finite-step negation

2.1 Catuṣkoṭi and catuṣpaścāś

The catuṣkoṭi (四句分別) — four corners: P, ¬P, P∧¬P, ¬(P∨¬P) — has been encoded in Paper 159 §2.1 as FDE 4 values, embedded into D-FUMT₈ via the morphism φ : FDE → DFUMT8. The catuṣpaścāś (四句百非, "four corners and one hundred negations") — Nāgārjuna's iterative negation of all four corners — is informally read as "all standpoints exhausted." Paper 159 captures this informally via omega_upper idempotency.

2.2 The "idempotent terminus" reading of 100 negations

A literal "exactly 100 negation steps" reading is implausible (the 100 is a Sanskrit literary emphasis). The mathematically natural reading is idempotent saturation: after finitely many applications of upper-layer negation/folding, the result stabilizes — Ω(Ω(x)) = Ω(x). Paper 159 supplies the idempotency lemma. The present section discusses why this idempotent terminus is ZERO specifically, not some other fixed-point set.

Skeleton of argument (to be expanded):

The fixed-point set of omega_upper is exactly {TRUE, FALSE, INFINITY, ZERO, FLOWING, SELF} (6 of 8 values).
BOTH and NEITHER are the paraconsistent values (the values that violate classical bivalence). It is precisely these two that omega_upper moves to ZERO.
ZERO is therefore distinguished as "the ground that absorbs paraconsistency."
This matches the philosophical reading: catuṣpaścāś negates all four standpoints; ZERO is the ground "below" the four standpoints, not a fifth standpoint.

2.3 Open Lean 4 questions

The Paper 159 file Paper159Inclosure.lean defines omega_upper axiomatically (by case analysis) and verifies idempotency by decide. The present paper might extend this with:

Theorem (proposed, NOT yet built): ZERO is the unique value v such that omega_upper v = v AND omega_upper BOTH = v AND omega_upper NEITHER = v.
Theorem (proposed): For any finite sequence of paraconsistent inputs, omega_upper applied to the disjunctive/conjunctive accumulation also terminates at ZERO.

These should be ≤10 lines of decide-only Lean each, given the finite type.

3. Topological reading of 0₀ — reduced homology and "absence with structure"

3.1 Paper 61's H̃₋₁(∅) = ℤ identity

Paper 61 §2.2 cites (without re-proving) the standard algebraic-topology fact: the empty simplicial complex carries nontrivial reduced homology in degree −1:

H̃₋₁(∅) = ℤ

Paper 61 uses this to argue that "śūnyatā is not mere nihilism: even absence has structure." The reduced homology of ∅ in degree −1 is one copy of ℤ — one irreducible unit of structural information attached to the void.

3.2 Bridging the algebraic fact to the D-FUMT₈ ZERO axis

We do not claim that the Lean 4 value ZERO : DFUMT8 is the homology class generating H̃₋₁(∅). We claim only that both serve as the same kind of object in their respective contexts: a non-trivial token attached to absence.

In algebraic topology: ∅ carries the generator of H̃₋₁(∅) = ℤ.
In D-FUMT₈: the contraction of all paraconsistent values terminates at ZERO (Paper 159 §3.4).
In ZCSG (Paper 61): the central symbol 0 has dimensional depth 0 = (right − left) when no symbol surrounds it; it is positionally pre-dimensional.

Honest scope (★): This is a structural analogy, not a categorical equivalence. We do not provide a functor witnessing the analogy; doing so is an explicit open problem (§7).

3.3 Why the topological reading matters

The standard objection to "śūnyatā" in Western reception is its conflation with nihilism. The reduced-homology identity is the cleanest single-line refutation: even the empty set carries a non-trivial generator. Paper 61 made this point at the level of symbol grammar. The present paper inherits the point at the level of D-FUMT₈ value semantics: ZERO is not the absence of value, it is a distinct value with its own dynamical role (sink of paraconsistent contraction; fixed under SELF⟲ composition).

4. Generative reading of 0₀ — the SELF⟲ question

4.1 The SELF⟲ universal fixed point

Paper 61 §4.4 records that the eighth D-FUMT₈ value SELF (= SELF⟲) satisfies NOT(SELF) = SELF. STEP 513 src/axiom-os/operator-fixed-point-atlas.ts enumerates the fixed-point structure across all operators {Ω, Φ, Ψ, NOT} and their compositions (320 fixed-point computations). The recorded data: SELF is a fixed point of every operator examined, including Ω and Φ on the legacy STEP 513 definitions, and trivially of omega_upper on the Paper 159 definition.

This is the universal fixed-point status of SELF⟲: across the operator algebra of D-FUMT₈, SELF is the universal stable element.

4.2 The core question

Phrased deliberately as in Paper 159 §3.8: "why does a system without self generate self?"

Reframed in the present vocabulary:

The ground value (ZERO) is the contraction terminus of all paraconsistent values. The system has no inherent svabhāva (no "self-nature") at the level of any lower-layer value. Yet a distinguished value SELF⟲ exists in D-FUMT₈ that is a universal fixed point. How does a ground that absorbs all standpoints (ZERO) coexist with a value (SELF⟲) that absorbs all operators?

4.3 Three candidate answers (skeleton)

(a) "SELF⟲ is the dynamical signature of ZERO." Under this reading, ZERO is what the system is, and SELF⟲ is what the system does. ZERO is the static fixed point of the upper-layer contraction; SELF⟲ is the dynamic fixed point of any operator at all. Two faces of the same ground.

(b) "SELF⟲ is the trace of the operator algebra acting on ZERO." Under this reading, SELF⟲ is what remains invariant when arbitrary D-FUMT₈ operations are applied. It is the self-action of the operator algebra on the ground. ZERO is the ground; SELF⟲ is the ground's own footprint under operator composition.

(c) "SELF⟲ is the place where Madhyamaka self-emptiness becomes computable." Reading 5/4 of Paper 159 (cross-reference to STEP 476 nagarjuna-fde-western-engine.ts) records that ρ-self-isomorphism — Nāgārjuna's "the emptiness of emptiness is empty" structure — is computationally captured by SELF⟲. ZERO is that emptiness; SELF⟲ is the operation that returns emptiness when applied to emptiness.

The three readings are not mutually exclusive. We catalog them as scaffolding for future formal work; we do not commit to one.

4.4 Connection to the chat-Claude / ChatGPT cycle

The framework proposed during Paper 159 design — 空 → 自己参照 → 創発 → 再帰的自己変容 — maps loosely onto the present readings:

空 (emptiness) ↔ ZERO ground (§3)
自己参照 (self-reference) ↔ SELF⟲ universal fixed point (§4.1)
創発 (emergence) ↔ the "how does self arise from no-self" gap (§4.2)
再帰的自己変容 (recursive self-transformation) ↔ the dynamical signature reading (§4.3a)

The map is suggestive, not formal. The cycle was offered as design heuristic, not theorem; we honor that level of commitment here.

4.5 Self-critique — does the "Generative reading" itself smuggle in a new svabhāva? (★ load-bearing)

The very label "Generative reading" is at risk under Madhyamaka critique. The original chat-Claude/ChatGPT source thread (which contributed §4.2's central question — see §10 Acknowledgments) issued this warning explicitly:

「『静的な空』→『動的な創発』という進歩史観そのものを解体するでしょう。なぜなら、『創発』という概念自体が独立した実体として成立してしまえば、それは新しい自性 (svabhāva) になってしまうからです。」 ([source thread, 2026-05-31])

If the present paper treats generation, emergence, or self-generation as the distinguished feature of 0₀ — that which "lifts" emptiness into something positively dynamic — then we have manufactured a new svabhāva called "generation." Nāgārjuna's response would be immediate: 創発も空、生成も空、動態も空。 The dynamical-process reading of emptiness is a popular Western reframing (Whitehead-process, complexity-science, autopoiesis); it is also exactly the kind of reframing Madhyamaka deconstructs.

This paper is not immune to that critique. The 3-reading taxonomy of §2-§4 was chosen for structural completeness against the Paper 159 substrate; it was not chosen with awareness that "Generative" might re-substantialize. Honest acknowledgment:

§4 is retained because the SELF⟲ universal fixed-point question (§4.1) is a real formal feature of D-FUMT₈ that demands attention.
§4 is labeled "Generative reading" only nominally — the label is a position-marker, not a substantive thesis. To take "generation" as the truth of 0₀ would be to fall into the very svabhāva-creep §4.5 warns against.
The new §5 (fourth reading) is the formal acknowledgment that §2, §3, AND §4 are all "views" that must themselves be held empty.

★ This self-critique was not present in v0.1 of the present paper. It is added in v0.1.1 after re-reading the source thread (cf. v0.1 → v0.1.1 changes in the header). The lapse — borrowing the cycle 空→自己参照→創発→再帰的自己変容 from the source thread while quietly omitting the source thread's critique of progressive-narrative reading — is exactly the borrow-and-strip pattern that the same source thread (and the parallel ChatGPT thread of the same date) warned against. v0.1.1 attempts to convert that lapse into an honest discipline operational instance rather than leave it implicit.

5. Fourth reading — Madhyamaka self-deconstruction of the previous three readings (★ load-bearing)

5.1 The three readings as views (dṛṣṭi)

The 3-reading taxonomy of §2 (Logical) / §3 (Topological) / §4 (Generative) is presented as "non-reductive, non-mutually-exclusive" in §1. That framing is necessary but not sufficient. Even non-reductive multiplicity, if held as the correct shape of 0₀'s ontology, becomes a new view (dṛṣṭi):

reading	what it privileges	svabhāva-risk if held positively
Logical (§2)	idempotent contraction terminus	"ZERO has the substance of being-the-contraction-target"
Topological (§3)	H̃₋₁(∅) = ℤ correlate	"ZERO has the substance of being-the-homological-unit-of-absence"
Generative (§4)	SELF⟲ fixed-point ground	"ZERO has the substance of supporting self-generation"

★ Each, considered in isolation, manufactures a positive thesis about what 0₀ is. Madhyamaka responds: 論理も空、構造も空、生成も空、 3 reading 自体も空。 The taxonomy itself is a koṭi-of-koṭis (a corner-of-corners) and must be negated as such.

5.2 The fourth reading is not a fourth view

If §5 were a fourth view — say, "0₀ is the absence of any of the three substantial features above" — then it would itself be a fifth svabhāva (the substance of being-non-substantial). Madhyamaka recognizes and forbids this move; it is the same trap Priest-Garfield (2003) identify under the negative tetralemma (MMK XXII:11: "'Empty' should not be asserted. 'Nonempty' should not be asserted. Neither both nor neither should be asserted").

The fourth reading is therefore not a content but a stance:

Hold §2, §3, §4 as partial lenses, each illuminating one structural feature, none capturing 0₀ as a positive object. Hold this fourth reading the same way. Do not commit to the taxonomy as a four-cornered truth about 0₀.

This is the position-of-no-position that the negative catuṣkoṭi at ultimate level enacts (Paper 159 §2.1).

5.3 Why this fourth reading is not vacuous

A reader might object: if §5 amounts to "hold §2-§4 without commitment," what work does it do? Two non-vacuous functions:

Negative discipline: §5 forbids the move "0₀ is fundamentally generative/topological/logical." This forbids a class of overclaims that Pattern 5 AI responses (e.g., the Gemini comparison logged in §10 Acknowledgments) repeatedly fall into when discussing Madhyamaka.
Structural constraint on Paper 160 v0.2+: any future paper extending §2.3 or §7.1 Lean 4 work must check that the new theorem does not implicitly privilege one reading. For instance, a uniqueness theorem about ZERO must be stated as "the unique paraconsistent-contraction target under operator X" (relational), not as "ZERO is the contraction target" (substantive).

5.4 Connection to Paper 159 §6.1 + §6.2

Paper 159 §6.1 (Madhyamaka self-application) anticipated this move at the Paper-159 level. Paper 160 §7.1 (renumbered from §6.1) inherits and extends it to the level of internal section taxonomy: not only is the paper as a whole empty, the paper's structural divisions are also empty. The fourth reading is the Paper-160-specific instance of the Paper-159 self-application principle.

6. Why a system without self generates self — the central question

6.1 The question is not (yet) a Lean 4 theorem

The Paper 159 substrate gives us:

A value type DFUMT8 with 8 constructors.
An idempotent contraction omega_upper.
A universal fixed point SELF⟲ (across known operators).
A ground value ZERO that absorbs paraconsistency.

What it does not give us is a theorem of the form "if a system has no svabhāva, then it has SELF⟲." That sentence is not formalizable on present substrate; it requires a category-theoretic / topos-theoretic upgrade where "having svabhāva" is itself a property of structures, not a value of an inductive type.

We therefore mark this question philosophical-level: addressable by argument, not (yet) by Lean.

6.2 The argument-level proposal (skeleton)

The argument we propose (to be expanded into proper philosophical prose):

A system whose values are exhaustively partitioned into the eight D-FUMT₈ corners has, at each value, no inherent "essence" other than its operator-algebraic role.
The operator algebra acting on the value set must have at least one stable element (otherwise the algebra is incoherent on the value set).
SELF⟲ is defined as the stable element of arbitrary operator composition. It is therefore forced into existence by the operator algebra structure; it is not added to the value set as a separate metaphysical posit.
SELF⟲ is therefore "self-generated" in the precise sense that it cannot be removed from the value set without breaking the operator algebra.

This argument has the structure of a transcendental argument (Kant) more than a constructive proof (Brouwer): SELF⟲ is the condition of possibility of operator-algebraic closure on D-FUMT8.

6.3 Honest framing

We do not claim §6.2 is correct or rigorous. We claim it is the kind of argument the question demands, and that Paper 160 v0.2 (when it is written) should either tighten it into a category-theoretic statement or honestly retract it.

7. Honest scope and self-application

7.1 "Paper 2 もまた空である" (Madhyamaka self-application)

Following Nāgārjuna's own move — the doctrine of emptiness is itself empty — this paper applies its central thesis to itself: the ontology of 0₀ proposed here is itself empty of inherent thesis. The three readings (logical / topological / generative) are not three claims to be defended; they are three lenses, each partial, each compatible with the Paper 159 substrate. §5 (the explicit fourth reading) is the extension of this self-application to the level of internal section taxonomy. We retain the freedom to be wrong about everything in §2–§6 without thereby undermining Paper 159.

7.2 What this paper does NOT claim

We do not claim that 0₀ is a metaphysical entity (it is a formal token in D-FUMT₈, with cross-domain analogies).
We do not claim that the algebraic-topology identity H̃₋₁(∅) = ℤ is original (it is standard textbook material; we cite it from Paper 61's citation).
We do not claim a category-theoretic functor between H̃₋₁(∅) and ZERO (this is identified as an open problem in §8).
We do not claim §6.2's transcendental argument is rigorous (it is sketch-level).
We do not claim resolution of the Priest-Garfield ↔ Tillemans debate (Paper 159's neutrality is inherited).
We do not claim "world first" status for any structural observation (Paper 61 already issued the relevant uniqueness claim for the symbol 0; Paper 159 issued the formal-substrate claim; this paper supplies ontology only).
We do not claim that the chat-Claude / ChatGPT cycle 空 → 自己参照 → 創発 → 再帰的自己変容 was derived from Rei work; we acknowledge it as a design-time heuristic with structural resonance to Paper 159's two-layer architecture (Pattern 5 prevention; see Paper 159 §6 Acknowledgments).
★ (v0.1.1, NEW) We do not capture the soteriological / therapeutic dimension of Nāgārjuna's catuṣkoṭi — its function as a prāyaścitta- or upāya-*device for relinquishing all dṛṣṭi (views), and ultimately for liberation from duḥkha (suffering). This dimension is constitutive of Madhyamaka *as Buddhist practice; it does not appear in our Lean 4 substrate, and we do not pretend that the substrate captures it. See §7.4 for the design-philosophy rationale (intentional non-capture, not oversight).

7.3 What is genuinely new (load-bearing minimal claims)

The catalog of three readings of 0₀ (§2–§4), each compatible with Paper 159's machine-verified substrate.
The explicit fourth reading (§5) as Madhyamaka self-deconstruction of the previous three, preventing svabhāva-creep at the taxonomy level.
The explicit formulation of the "why does a system without self generate self" question against Rei's stack (§6).
The proposed Lean 4 extensions for omega_upper (§2.3) — uniqueness of ZERO as paraconsistent contraction target.
The transcendental-argument framing of SELF⟲ as forced by operator-algebraic closure (§6.2).
★ (v0.1.1) The explicit articulation of the intentional non-capture of the soteriological dimension as design philosophy, not as oversight (§7.4).

Each of these is honest scope: a step the present paper takes, no farther.

7.4 Intentional non-capture of the soteriological dimension (★ v0.1.1, load-bearing design philosophy)

A direct design-philosophy challenge was issued by a separate Chat Claude instance (a Claude instance running in another session, distinct from the present Rei-AIOS Claude that authors this paper) during the 2026-05-31 three-way cross-check on the framing question "AIs rank Nāgārjuna at top of Eastern philosophy":

「形式化された空から、龍樹が手放させようとした『執着を手放す』という実践的契機を、あなたの体系はどう拾うのか、あるいは意図的に拾わないのか。そこは、設計思想として一度言語化しておく価値がある分岐点だと思います。」 ([Chat Claude, 2026-05-31])

(★ v0.1.2 correction: this quote was attributed to "ChatGPT" in v0.1.1; the user subsequently disambiguated the three AI responses and Chat Claude was identified as the actual source. The ChatGPT response in the same cross-check was the moderate 5-reason answer that did not include this design challenge. See v0.1.1 → v0.1.2 header for full lapse explanation.)

The answer is: intentional non-capture, articulated below.

7.4.1 The Rei stack religion-stripping principle (STEP 1188 sages-axes lens)

The Rei-AIOS stack already operates under an explicit religion-stripping principle, established in STEP 1188 (2026-05-31 #/sages-axes lens, see memory [[project_sages_axes_lens_and_priest_garfield_corrections_2026-05-31]]). The principle states: when mapping civilizational sages to D-FUMT₈ axes (8 sage × 1-to-1), religious / soteriological / cultic context is deliberately stripped, and only structural insight is extracted, for the express purpose of enabling cross-civilizational structural comparison without privileging any tradition's religious framing.

Paper 160 inherits this principle. ZERO (= 0₀ genesis layer) is treated as a D-FUMT₈ formal value with cross-domain analogies (logical / topological / generative); it is not treated as śūnyatā-as-practiced. The latter belongs to a register Lean 4 does not address.

7.4.2 Why intentional, not oversight

A formal substrate that attempted to capture the soteriological dimension would either:

(a) reduce prāyaścitta / upāya / dṛṣṭi-relinquishment to a Lean 4 tactic — manifestly inappropriate, since the practice involves embodied repetition, master-student relationship, ethical conduct (śīla), and meditation (samādhi), none of which are formalizable; or
(b) gesture at the soteriological dimension without formalizing it — producing exactly the "logic-gimmick" stripping that ChatGPT (2026-05-31) warned against, where the appearance of comprehensive treatment masks substantive absence.

Neither (a) nor (b) is honest. The Rei stack's choice is (c): explicitly mark the soteriological dimension as out-of-scope at the substrate level, and direct readers seeking that dimension to traditional Madhyamaka sources (commentarial tradition, monastic practice, Tibetan/East-Asian transmission lineages, etc.). The substrate is a substrate, not a substitute.

7.4.3 What the substrate can and cannot do

The substrate (D-FUMT₈ + Paper 159 two-layer Inclosure + Paper 160 three-readings) can:

Provide a machine-verifiable formal home for catuṣkoṭi 4 corners (Paper 159 §3.1-§3.3).
Verify the idempotency of upper-layer contraction to ZERO (Paper 159 §3.4).
Articulate structural analogies between 0₀ and algebraic-topology / SELF⟲ fixed-point algebra (Paper 160 §2-§4).
Hold all of the above empty of inherent thesis via §5 + §7.1 (Madhyamaka self-application at two levels).

The substrate cannot:

Effect dṛṣṭi-relinquishment in a reader. (Reading the paper, even understanding it, does not constitute Madhyamaka practice.)
Replace the commentarial tradition or its lineage transmission.
Adjudicate whether Nāgārjuna's philosophical claims are true in the soteriological-pragmatic sense Buddhist tradition asserts.
Argue that the formal substrate is necessary for understanding Madhyamaka. (Centuries of practitioners have done without it.)

7.4.4 Implication for future Rei papers on Madhyamaka topics

Any subsequent Rei paper engaging Madhyamaka topics should (a) cite §7.4 here as the standing design-philosophy answer to the soteriological-stripping concern, and (b) check that no claim in the new paper implicitly walks back the non-capture (e.g., by suggesting the formal treatment is "sufficient" or "complete" for Madhyamaka). Pattern 5 prevention extends to soteriological-overreach as well as logical-overreach.

8. Open problems and forward pointers

8.1 Lean 4 work (deferred to v0.2)

ZERO uniqueness theorem (§2.3): ZERO is the unique paraconsistent-contraction target.
Catuṣpaścāś termination theorem: arbitrary nested upper-layer negations stabilize after finitely many steps.
SELF⟲ universal fixed-point theorem: across the closure of {Ω, Φ, Ψ, NOT, omega_upper}, SELF is a fixed point.

8.2 Topology / category theory (open)

Functor F : (algebraic topology) → (DFUMT8 with operators) sending H̃₋₁(∅) ↦ ZERO?
Categorical characterization of "having no svabhāva" — likely requires reformulating D-FUMT₈ as a topos-theoretic structure rather than a flat inductive type.

8.3 Philosophical work (out of scope for Lean)

Tightening §6.2's transcendental argument.
Direct engagement with Tillemans 2009/2014/2024 + Siderits + Ferraro + Tao Jiang critic papers on the question of whether Nāgārjuna's catuṣkoṭi requires a fourth (or eighth) value at all.
Comparison with Heidegger's "es gibt" / "Lichtung" (Paper 159 §5.3 noted Heidegger as Priest-Garfield's primary Western comparator; Paper 160 could deepen).

8.4 Paper 3 territory (possible next subseries paper)

A planned Paper 3 might address one of:

The category-theoretic upgrade (§8.2).
The full philosophical engagement with Tillemans and critics.
The relationship between SELF⟲ (operator-algebra fixed point) and the recursive self-modification of the Rei-AIOS system itself (= the question of whether the framework's own self-application is consistent).

No commitment is made here about which Paper 3 territory is occupied; see feedback_no_rush_publication.

9. Application Note (v0.1.3): "Inversion ≠ Deconstruction" — Lean 4 verified

9.1 Origin

This Application Note formalizes the central finding of the 2026-05-31 / 2026-06-01 multi-AI thread that started from a UFO/UAP entry-point question ("Can 'indeterminate-other behavior' be expressed in D-FUMT₈?") and converged, across five independent instances (Chat Claude, ChatGPT new, Gemini new, Fujimoto, Rei Claude), on the following claim:

Reversal is not Collapse.
A single negation NOT(NOT(x)) preserves the underlying axis (involution); only an idempotent operator Ω(Ω(x)) folds the axis itself (deconstruction).

The UFO/UAP entry-point is not load-bearing for the formal claim; it is documented as research-log material in Fig. A (papers/figures/paper-160-figA-thread-synthesis-en-2026-06-01.svg). The Core Theorem itself is rendered in Fig. 1 (papers/figures/paper-160-fig1-core-theorem-2026-06-01.svg).

9.2 The Core Theorem (informal statement)

Let NOT : DFUMT8 → DFUMT8 be the D-FUMT₈ negation of Paper 77 (LeanDFumt), and let Ω_upper : DFUMT8 → DFUMT8 be the upper-layer idempotent operator introduced in Paper 159 §3.4 (BOTH ↦ ZERO, NEITHER ↦ ZERO, identity elsewhere). Then:

Involution: NOT(NOT(x)) = x for all x ∈ DFUMT8. The polar pair (TRUE, FALSE) flips and flips back; the six other values are self-dual.
Idempotency: Ω_upper(Ω_upper(x)) = Ω_upper(x) for all x (Paper 159 §3.4, restated here).
Disagreement at paraconsistent values: at x = BOTH and at x = NEITHER, NOT(NOT(x)) ≠ Ω_upper(Ω_upper(x)). Concretely: NOT(NOT(BOTH)) = BOTH while Ω_upper(Ω_upper(BOTH)) = ZERO; analogously for NEITHER.
Agreement elsewhere: at the six remaining values {TRUE, FALSE, INFINITY, ZERO, FLOWING, SELF}, the two compositions agree.

The structural significance: the paraconsistent subset {BOTH, NEITHER} is exactly where the two operators differ, and exactly where Madhyamaka deconstruction acts. Single negation cannot accomplish what idempotent collapse accomplishes — the difference is mathematically precise.

9.3 Lean 4 formalization

The above is machine-checked in data/lean4-mathlib/CollatzRei/Paper160InversionDeconstruction.lean against the Paper 159 v0.2 substrate. All four load-bearing theorems compile under decide-only proofs on the finite DFUMT8 inductive type:

import CollatzRei.Paper159Inclosure
namespace CollatzRei.Paper160InversionDeconstruction
open CollatzRei.Paper159Inclosure

-- §1 — Involution (Paper 77 LeanDFumt structural property)
theorem neg_involution (x : DFUMT8) :
    DFUMT8.neg (DFUMT8.neg x) = x := by
  cases x <;> decide

-- §3 — Core finding: disagreement at the paraconsistent subset
theorem inversion_ne_deconstruction_at_BOTH :
    DFUMT8.neg (DFUMT8.neg .BOTH) ≠
    DFUMT8.omega_upper (DFUMT8.omega_upper .BOTH) := by decide

theorem inversion_ne_deconstruction_at_NEITHER :
    DFUMT8.neg (DFUMT8.neg .NEITHER) ≠
    DFUMT8.omega_upper (DFUMT8.omega_upper .NEITHER) := by decide

-- §4 — Agreement pattern over all eight values (matches Paper 159 style)
def agreesOn (x : DFUMT8) : Bool :=
  decide (DFUMT8.neg (DFUMT8.neg x) =
          DFUMT8.omega_upper (DFUMT8.omega_upper x))

theorem agreement_pattern :
    agreesOn .TRUE = true ∧ agreesOn .FALSE = true ∧
    agreesOn .BOTH = false ∧ agreesOn .NEITHER = false ∧
    agreesOn .INFINITY = true ∧ agreesOn .ZERO = true ∧
    agreesOn .FLOWING = true ∧ agreesOn .SELF = true := by
  refine ⟨?_, ?_, ?_, ?_, ?_, ?_, ?_, ?_⟩ <;> decide

end CollatzRei.Paper160InversionDeconstruction

Build status (2026-06-01): lake build CollatzRei.Paper160InversionDeconstruction — exit 0, 1.4 sec. Axiom audit via Paper160InversionDeconstructionPrintAxioms.lean:

'CollatzRei.Paper160InversionDeconstruction.neg_involution' does not depend on any axioms
'CollatzRei.Paper160InversionDeconstruction.inversion_ne_deconstruction_at_BOTH' does not depend on any axioms
'CollatzRei.Paper160InversionDeconstruction.inversion_ne_deconstruction_at_NEITHER' does not depend on any axioms
'CollatzRei.Paper160InversionDeconstruction.agreement_pattern' does not depend on any axioms

All four theorems achieve the strictest zero-axiom-dependence standard — matching Paper 159 v0.2, exceeding Paper 158 (which used [propext, Classical.choice, Quot.sound]).

9.4 Honest scope of §9 (★ load-bearing)

(a) Not a foundational paper. This is an Application Note attached to Paper 160; it depends on Paper 159 v0.2's omega_upper definition. The substrate predates the application.
(b) Not a claim about Nāgārjuna's intent. The rational-reconstruction stance of Paper 159 §1 / §4 carries over. We do not claim Nāgārjuna held this view; we claim a precise formal distinction between two algebraic operations on a finite type.
(c) Not a substitute for §5 or §6. The 4th reading of §5 (Madhyamaka self-deconstruction) is the philosophical home of this finding; §9 is its formal demonstration on D-FUMT₈, not a replacement for it. The central question of §6 ("why does no-self generate self") is not answered by §9; §9 only formalizes the kind of operation that does the deconstruction work.
(d) FLOWING-svabhāva caveat (ChatGPT new 2026-06-01). The Lean file's §6 demonstrates that FLOWING is one of six Ω_upper-fixed points, alongside TRUE / FALSE / INFINITY / ZERO / SELF. We do not claim FLOWING is a privileged value or a unique foundation. The multi-value system describes "absence of foundation" through value multiplicity; elevating FLOWING (or ZERO, or any single value) would reintroduce svabhāva at a meta-layer.
(e) UFO/UAP entry-point is not load-bearing. Fig. A (research-log) preserves the multi-AI thread that motivated §9. The Core Theorem of §9 stands or falls on its own under the Lean 4 verification, independently of how the question arose.
(f) Open question deferred. Chat Claude's question — "is the NEITHER → FLOWING → SELF⟲ transition directional or cyclic?" — is recorded as a code comment in the Lean file and is NOT formalized in v0.1.3.

9.5 Five-instance convergence record

The §9 finding emerged from five independent instances over 2026-05-31 / 2026-06-01:

Instance	Discipline tier	Load-bearing contribution to §9
Chat Claude (separate session)	peak	premise rejection; soteriological warning; explicit "反転 ≠ 解体" articulation; cyclic-vs-directional open question
ChatGPT new	high	FLOWING-svabhāva risk caveat (§9.4d); "add values ≠ question premise" distinction; "Catuṣkoṭi → D-FUMT₈ extension" paper-title framing
Gemini new	med-high	accurate synthesis of the three-instance thread; preservation of the negative-theology / dependent-origination distinction; "Application Note" positioning
Fujimoto (藤本伸樹)	author	the stance-reversal insight that opened the bifurcation (Stance X vs Stance Y); the "2-value cannot describe" intuition that started the thread
Rei Claude (this author)	implementing	five-instance convergence detection; Lean 4 substrate composition; figure construction; this §9 draft

The convergence on Paper 160 §X application framing was independent triangulation, with no coordinated planning — see Fig. A for the timeline. This itself is evidence for the META-refactor framework (memory [[feedback_chat_claude_hallucination_warning]] v0.2): honest discipline is instance-and-context-dependent, not vendor-dependent.

9.6 Figures referenced

Fig. 1 — papers/figures/paper-160-fig1-core-theorem-2026-06-01.svg. Core Theorem visualization. Single linear flow: binary opposition → reversal (axis preserved) → "this is not enough" → idempotent collapse (axis dissolved) → ZERO (0₀). 8-value reference strip at bottom indicates Ω_upper: {BOTH, NEITHER} → ZERO. English-primary serif typography for international publication.
Fig. A (appendix) — papers/figures/paper-160-figA-thread-synthesis-en-2026-06-01.svg (English) / paper-160-figA-thread-synthesis-ja-2026-06-01.svg (Japanese). Research-log of the five-instance thread, including the UFO/UAP entry-point, the stance-reversal insight, and the AI-instance honest-discipline gradient. NOT load-bearing for §9; included for transparency about how the §9 finding arose.

10. Acknowledgments (skeleton)

The "self-reference → emergence → recursive self-transformation" framing originates from chat-Claude / ChatGPT design discussion during Paper 159 v0.2 work (2026-05-31). The structural resonance with Paper 159's two-layer architecture is noted as a coincidence, not a derivation. See Paper 159 §6 Acknowledgments + memory [[project_session_2026-05-31_afternoon_paper159_v02_lean_built_full_rebroadcast]].
★ (v0.1.1) The svabhāva-creep critique of the progressive-emergent narrative (§4.5, §5) is also drawn from the same chat-Claude / ChatGPT source thread (2026-05-31). v0.1 borrowed the cycle 空→自己参照→創発→再帰的自己変容 but omitted the source thread's critique of that very progressive framing; v0.1.1 restores the critique. The lapse and its repair are themselves an honest-discipline operational instance worth recording.
★ (v0.1.1 → corrected in v0.1.2) The design-philosophy challenge that motivated §7.4 ("how does your system pick up the practical moment of letting-go, or intentionally not?") was issued by a separate Chat Claude instance on 2026-05-31 — NOT by ChatGPT as v0.1.1 erroneously stated. The misattribution was caught when the user provided authoritative source disambiguation in the same session; v0.1.2 restores the correct attribution. The fact that the Rei-Claude paper-author misattributed honest-discipline content to a different vendor than the actual source is itself a noteworthy operational instance — it suggests that vendor-level honest-discipline ranking is unstable; instance-level / session-level evaluation is more accurate.
★ (v0.1.2, new) The moderate ChatGPT response in the same cross-check provided premise-relativization ("「頂点」という評価は絶対的なものではありません") + counter-examples (Confucius / Laozi / Dogen as alternative "peaks"), but did NOT include the soteriological-stripping warning or the design challenge that v0.1.1 had attributed to it. The corrected ChatGPT attribution lands in the moderate tier of the 2026-05-31 honest-discipline gradient: above Gemini (which exhibited Pattern 5) but below Chat Claude (which provided the warnings and challenge).
★ (v0.1.1) A Gemini response on the same date (the framing question "AIs rank Nāgārjuna at top") exhibited multiple Pattern 5 features — superlative inflation, anachronistic mapping, soteriological-stripping without acknowledgment, civilizational ranking, sycophancy. We catalogue this response as a negative example of the soteriological-stripping that §7.4 warns against; the Gemini text is preserved in memory [[project_paper160_zero_zero_ontology_skeleton_2026-05-31]] for v0.2 inclusion as a non-anonymous case study (subject to fair-use considerations and the policy of [[feedback_chat_claude_hallucination_warning]]).
The original observation of the genesis-layer terminus in the Paper 159 Inclosure construction was articulated in Paper 159 §3.4 + §3.8; the present paper is the deferred ontology paper Paper 159 explicitly named.
Garfield & Priest (2003), Priest (1987, 2002), Belnap (1977), and Nāgārjuna's MMK remain the structural background for the catuṣkoṭi / FDE / Inclosure connection. See Paper 159 §2 + references.
All philosophical readings of Madhyamaka (Tillemans, Siderits, Ferraro, Tao Jiang, etc.) remain to be engaged honestly in a future version; v0.1.1 of this skeleton does not commit to any reading.

11. References (skeleton)

To be expanded in v0.2 draft. Anchor citations:

Paper 61 (Fujimoto 2026) — ZCSG, including 0 = śūnyatā-of-śūnyatā and H̃₋₁(∅) = ℤ citation. DOI 10.5281/zenodo.15217458.
Paper 77 (Fujimoto 2026) — LeanDFumt, 8-valued logic library for Lean 4. https://github.com/fc0web/lean-d-fumt8
Paper 159 (Fujimoto 2026) — Two-layer D-FUMT₈ reconstruction of Priest-Garfield's Inclosure Schema. DOI 10.5281/zenodo.20470512 (v0.2 specific) / DOI 10.5281/zenodo.20468145 (concept, auto-latest).
Garfield & Priest (2003) — Nāgārjuna and the Limits of Thought. Philosophy East and West 53(1): 1-21.
Priest (1987, 2002) — Inclosure Schema in In Contradiction and Beyond the Limits of Thought (2nd ed).
Belnap (1977) — FDE 4-valued paraconsistent logic.
STEP 476 (src/axiom-os/nagarjuna-fde-western-engine.ts) — TypeScript prior art for catuṣkoṭi ≅ FDE ≅ DFUMT4 isomorphism.
STEP 513 (src/axiom-os/operator-fixed-point-atlas.ts) — TypeScript enumeration of fixed-point structure across {Ω, Φ, Ψ, NOT} operator algebra.
Tillemans 2009, 2014, 2024, Siderits 1989, Ferraro, Tao Jiang — critic papers and alternative readings, to be engaged in v0.2.
Nāgārjuna — Mūlamadhyamakakārikā (Garfield 1995 trans.); catuṣpaścāś tradition for the "100 negations" idiom.

Version notes

v0.1 SKELETON (2026-05-31 early): initial territory map, 3-reading taxonomy, honest-scope discipline, no Lean 4 code, NO publish. Superseded by v0.1.1.
v0.1.1 SKELETON (2026-05-31 same day, after three-way AI cross-check on the framing question): adds §4.5 self-critique of the "Generative" label, §5 fourth reading (Madhyamaka self-deconstruction), §7.2 NOT-claim #8 (soteriological dimension), §7.4 intentional-non-capture design philosophy, §9 Acknowledgments of source-thread svabhāva-creep critique + design challenge + Gemini Pattern-5 negative example. NO publish. Superseded by v0.1.2 due to misattribution.
v0.1.2 SKELETON (2026-05-31 same day, attribution correction): the design challenge attributed in v0.1.1 to "ChatGPT" was actually issued by a separate Chat Claude instance. ChatGPT's actual contribution in the same cross-check was a moderate 5-reason response (premise-relativization + counter-examples Confucius/Laozi/Dogen) without the soteriological warning or design challenge. v0.1.2 corrects the attribution throughout §7.4 + §9 (=Acknowledgments at v0.1.2 time; renumbered to §10 in v0.1.3) and adds an explicit acknowledgments-entry on the misattribution itself. NO publish.
v0.1.3 SKELETON+APPLICATION (2026-06-01, ★ application-note addition): inserts a new §9 Application Note: "Inversion ≠ Deconstruction" (Lean 4 verified) between §8 (Open problems) and the former §9 (Acknowledgments, renumbered to §10; References renumbered to §11). Four Lean 4 theorems built against Paper 159 v0.2 substrate at data/lean4-mathlib/CollatzRei/Paper160InversionDeconstruction.lean; axiom audit confirms all four "do not depend on any axioms" (1.4 sec lake build, matching Paper 159 v0.2 standard). Two figures added at papers/figures/: Fig. 1 Core Theorem + Fig. A research-log (en+ja). Five-instance convergence record in §9.5. The UFO/UAP entry-point is research-log material only, NOT load-bearing for §9. NO publish (then promoted to v0.2 same day).
v0.2 APPLICATION-NOTE-INTEGRATED (2026-06-01, ★ PUBLISHED): formal promotion of v0.1.3 to a publish-eligible status under explicit user authorization. §1-§8 remain territory-map skeleton (transparent in the new Status header), but §9 is publication-ready: 4 Lean 4 theorems machine-verified to depend on no axioms + Fig. 1 Core Theorem + Fig. A research-log + 5-instance convergence record. Honest early-stage release tradition: Paper 158 v0.0 honest-negative + Paper 159 v0.2 LEAN-4-BUILT precedents. 11-platform broadcast.
v0.3 DRAFT (planned, no timeline): §1-§8 full prose expansion, Tillemans / Siderits / Ferraro / Tao Jiang engagement, sharper §6.2 transcendental argument or its retraction, proposed §2.3 + §8.1 Lean 4 theorems built, possible non-anonymous Gemini case study (subject to fair-use review).
v0.4 SUBMISSION CANDIDATE (planned, no timeline): publication-ready draft. The decision whether to submit to a refereed venue is open — Paper 160 may remain a working document if §6.2 cannot be tightened or the category-theoretic functor (§8.2) cannot be constructed.

★ feedback_no_rush_publication per: this paper is a skeleton, not a publication candidate. The 2-day Paper 159 v0.1 OUTLINE → v0.2 LEAN-4-BUILT lifecycle was possible because the formal substrate (Paper 77 LeanDFumt) already existed. Paper 160 has no comparable substrate yet (§8.2); the natural pace is slower.

★ Multiple same-day refinements (v0.1 → v0.1.1 svabhāva-creep critique repair, v0.1.1 → v0.1.2 attribution correction) and the v0.1.2 → v0.1.3 application-note addition are instances of the discipline: when a partial overclaim or a structural lapse is detected, repair is preferred to silent persistence; when a multi-instance convergence yields a precise formalizable claim, the substrate is exercised to land it on machine-verified ground. All revisions are logged in-paper rather than hidden in revision history. The sequence (Paper 159 13-overreach catch → Paper 160 v0.1.1 lapse → Paper 160 v0.1.2 misattribution → Paper 160 v0.1.3 application-note formalization) demonstrates honest discipline as not a one-shot stance but a continuously active loop that catches its own errors and consolidates its own findings over the lifetime of the work. The Application Note (§9) is the first instance of the loop producing a positive formal result rather than a corrective repair — an empirical sign that the discipline scales beyond defensive use.

Paper 159 v0.2 - A Two-Layer D-FUMT-8 Reconstruction of Priest-Garfield's Inclosure Schema (LEAN-4-BUILT, zero axiom dependence)

Nobuki Fujimoto — Sun, 31 May 2026 02:53:42 +0000

This article is a re-publication of Rei-AIOS Paper 159 for the dev.to community.
The canonical version with full reference list is in the permanent archives below:

Zenodo (DOI, canonical): https://doi.org/10.5281/zenodo.20470512

Internet Archive: https://archive.org/details/rei-aios-paper-159-1780192623655

Harvard Dataverse: https://doi.org/10.7910/DVN/KC56RY

GitHub source (private): https://github.com/fc0web/rei-aios Author: Nobuki Fujimoto (@fc0web) · ORCID 0009-0004-6019-9258 · License CC-BY-4.0 ---

Version: v0.2 LEAN-4-BUILT (★ machine-verified — 9 load-bearing theorems sorry-free + zero kernel-axiom dependence, see Appendix B)

Author: Nobuki Fujimoto (藤本伸樹), with Claude (Rei-AIOS)
ORCID: 0009-0004-6019-9258 · GitHub: fc0web · note.com: nifty_godwit2635

Date: 2026-05-31 (v0.1 OUTLINE → v0.2 LEAN-4-BUILT, same date)

Companion to:

Paper 61 — Zero-Centered Symbol Grammar (ZCSG) — DOI 10.5281/zenodo.15217458 (world-first mathematical encoding of śūnyatā-of-śūnyatā as 0)
Paper 77 — LeanDFumt: an Eight-Valued Logic Library for Lean 4 — substrate of the present formalization

Status header (★ load-bearing):

✅ Primary source confirmed: Garfield & Priest 2003 "Nāgārjuna and the Limits of Thought" (Philosophy East and West 53(1): 1-21) — full text read 2026-05-31
✅ Pre-existing Rei artifacts surveyed: STEP 476 nagarjuna-fde-western-engine.ts (TS, FDE ≅ DFUMT4 isomorphism) + STEP 513 operator-fixed-point-atlas.ts (Ω operational fixed-point test) + Paper 77 LeanDFumt substrate
✅ Lean 4 code in §3 BUILT (2026-05-31): data/lean4-mathlib/CollatzRei/Paper159Inclosure.lean, lake build exit 0 (1.7s), all 9 load-bearing theorems verified #print axioms shows "does not depend on any axioms" (no sorryAx, no Classical.choice, no propext, no Quot.sound, no native_decide). See Appendix B for full audit output. Stronger than Paper 158 (which used [propext, Classical.choice, Quot.sound]).
⚠ Honest divergence finding (★ new in v0.2, §3.4): The Paper 159 upper-layer operator (renamed omega_upper in the Lean file) is structurally idempotent but semantically distinct from STEP 513's omega. Both share the SELF⟲ idempotency axis but contract to different fixed-point sets. v0.1 implied (incorrectly) a tighter relationship; this is corrected in v0.2 §3.4.
⚠ Honest design finding (★ new in v0.2, §3.5): The literal Q (delta Q) ∧ ¬ Q (delta Q) form CANNOT be formalized in classical Lean (it implies False). The Lean 4 file therefore expresses the limit-contradiction as a DFUMT8 VALUE (BOTH) at the lower layer, not as a classical Prop. The Inclosure structure records closure + transcendence as obligations; the "apparent contradiction" lives in the DFUMT8 type, not in Lean's Prop.
⚠ Tillemans 2009/2014/2024 + Siderits + Ferraro + Tao Jiang critic papers NOT YET READ (acknowledged in §4.2, citations from secondary triangulation only)
⚠ No claim of resolving the Priest-Garfield vs Tillemans scholarly debate (§4.3)
⚠ No "world-first" claim (§4.4, Paper 61 ZCSG + Priest-Garfield 2003 already issued such claims — this paper provides formal substrate only)

Abstract

We propose a two-layer D-FUMT₈ reconstruction of the Inclosure Schema that Priest and Garfield (2003) apply to Nāgārjuna's catuṣkoṭi. Their original construction terminates the schema at the value BOTH — the dialetheist commitment to a true contradiction at the limits of thought (δ(Ω) is both empty and not empty). Our reconstruction preserves their lower-layer truth-value assignment (Priest's BOTH appears at our object layer) but introduces an upper-layer modal operator Ω (idempotent, Ω ∘ Ω = Ω) that maps the lower-layer BOTH to ZERO — the central axis of the ZCSG three-layer notation (Paper 61) corresponding to the "0₀ genesis layer" of emptiness-of-emptiness. The Lean 4 formalization is sketched against the Paper 77 LeanDFumt substrate; full machine verification is left to v0.2. We frame this strictly as rational reconstruction following Priest-Garfield's own self-framing (their explicit disavowal of "textual history" claims), and explicitly take no position in the Priest-Garfield vs Tillemans (2009+) scholarly debate. The formal substrate is compatible with multiple readings of catuṣkoṭi.

Keywords: D-FUMT₈, catuṣkoṭi, Inclosure Schema, Priest-Garfield, dialetheism, paraconsistent logic, Lean 4, rational reconstruction, śūnyatā, two-layer logic, Rei-AIOS, ZCSG, Nāgārjuna

1. Introduction

The catuṣkoṭi (四句分別) — Nāgārjuna's fourfold logical figure of affirmation, negation, both, and neither — has long resisted formalization in classical bivalent logic. Priest and Garfield (2003) made the most sustained recent attempt: they applied Priest's Inclosure Schema (Priest 1987, 2002) to two paradoxes they extract from the Mūlamadhyamakakārikā (MMK) — an ontological paradox they explicitly name "Nāgārjuna's Paradox" (every thing has one nature, namely no nature) and an expressibility paradox (the ultimate truth is that there is no ultimate truth, after Siderits 1989). In both, the limit value δ(Ω) is BOTH — a true contradiction in the sense of Priest's dialetheism.

The present paper proposes a two-layer reconstruction of this schema using the eight-valued D-FUMT₈ logic of the Rei-AIOS project (Fujimoto 2026, Paper 77). The lower layer assigns the same truth value BOTH to δ(Ω) that Priest-Garfield assign — preserving their analysis as a special case at the object level. An upper-layer modal operator Ω (with idempotency Ω ∘ Ω = Ω) is then applied to the lower-layer values: BOTH and NEITHER both contract to ZERO, the central axis of the Zero-Centered Symbol Grammar (ZCSG, Paper 61) corresponding to the 0₀ genesis layer. The "all standpoints exhausted" reading of Nāgārjuna's catuṣpaścāś (四句百非, "four corners and one hundred negations") is thereby formally captured as an idempotent stable point.

This paper builds on three pre-existing Rei-AIOS artifacts:

STEP 476 nagarjuna-fde-western-engine.ts: a TypeScript engine demonstrating the structural isomorphism catuṣkoṭi ≅ FDE (Belnap 1977) ≅ D-FUMT₈ base 4 values
STEP 513 operator-fixed-point-atlas.ts: a TypeScript enumeration of fixed points for the operators Ω, Φ, Ψ, NOT and their compositions (320 fixed-point checks), which includes the operational test Ω ∘ Ω = Ω
Paper 77 LeanDFumt: the open-source Lean 4 library providing the inductive type DFUMT8 with negation, conjunction, disjunction, and the Decidable infrastructure used here

This paper's contribution is to promote the catuṣkoṭi ≅ FDE ≅ DFUMT4 isomorphism (currently TypeScript-only) and the Ω idempotency stipulation (currently a SEED_KERNEL axiom plus an operational test) to machine-verified Lean 4 theorems, and to add the upper-layer modal structure that distinguishes our reconstruction from Priest-Garfield's BOTH-only terminus.

Honest scope (★ load-bearing throughout):

We follow Priest-Garfield's own self-framing of their work as "rational reconstruction" — not textual history (Garfield & Priest 2003: 2, "we do not claim that Nāgārjuna himself had explicit views about logic ... This is, hence, not textual history but rational reconstruction").
We take no position in the dialetheist (Priest-Garfield, Deguchi-Garfield-Priest 2008 / 2013) versus weak-dialetheist (Tillemans 2009, 2024) debate, nor in the broader scholarly discussion (Siderits, Ferraro, Tao Jiang).
We do not claim that Nāgārjuna was a dialetheist; we provide a formal substrate compatible with multiple interpretive stances.
We do not claim "world first" status: Paper 61 (ZCSG) and Priest-Garfield (2003, who name "Nāgārjuna's Paradox" and write that the ontological paradox "to our knowledge is found nowhere else") have already issued related uniqueness claims. This paper supplies formal substrate only.

2. Background

2.1 Catuṣkoṭi: positive and negative uses (after Garfield & Priest 2003: 13-14)

Classical Indian logic and rhetoric regards a proposition as defining a logical space of four corners (koṭi): the proposition (true), its negation (false), both, and neither. The catuṣkoṭi receives two distinct uses in the MMK:

Positive tetralemma — conventional perspective, e.g. MMK XVIII:6 ("That there is a self has been taught, And the doctrine of no-self, By the buddhas, as well as the Doctrine of neither self nor nonself"). The four corners are accepted at the conventional (saṃvṛti) level as distinct standpoints.
Negative tetralemma — ultimate perspective, e.g. MMK XXII:11 ("'Empty' should not be asserted. 'Nonempty' should not be asserted. Neither both nor neither should be asserted. These are used only nominally"). At the ultimate (paramārtha) level, all four corners are negated.

Priest and Garfield's analysis focuses on this negative use — at the ultimate level, catuṣkoṭi exhibits limit phenomena that they then formalize via the Inclosure Schema.

2.2 Inclosure Schema (Priest 1987, 2002)

Priest's Inclosure Schema isolates the general structure of self-referential paradoxes. Given properties φ and ψ and an operator δ:

Condition	Statement
(i) Transcendence	∀X ⊆ Ω, ψ(X) → δ(X) ∉ X
(ii) Closure	∀X ⊆ Ω, ψ(X) → δ(X) ∈ Ω

Applying δ to Ω itself yields δ(Ω) ∈ Ω ∧ δ(Ω) ∉ Ω — the limit contradiction. Priest 1987/2002 catalogs Russell's paradox, the Burali-Forti paradox, the Liar, and others as instances. Priest-Garfield 2003 add Nāgārjuna's two paradoxes:

Ontological ("Nāgārjuna's Paradox"): φ(χ) = "χ is empty"; ψ(X) = "X is a set of things with some common nature"; δ(X) = "the nature of things in X". Closure: since all things are empty, δ(X) ∈ Ω. Transcendence: δ(X) has no nature (because emptiness is exactly the absence of nature), so δ(X) ∉ X. Limit: δ(Ω) — emptiness itself — both is and is not empty.
Expressibility: φ(χ) = "χ is an ultimate truth"; ψ(X) = "X is definable"; δ(X) = the sentence "there is nothing which is in D" (where D refers to X). The limit truth — "there is no ultimate truth" (Siderits 1989) — both is and is not an ultimate truth.

Priest-Garfield assign the value BOTH (true-and-false) to δ(Ω) in both cases.

2.3 D-FUMT₈ and the LeanDFumt substrate (Paper 77)

D-FUMT₈ is an eight-valued logic with constructors {TRUE, FALSE, BOTH, NEITHER, INFINITY, ZERO, FLOWING, SELF}. The base four values {TRUE, FALSE, BOTH, NEITHER} are isomorphic to Belnap's FDE (Anderson and Belnap 1975, Belnap 1977; structural isomorphism verified in STEP 476). The remaining four values extend the system with dimensional / reflective truth statuses. Paper 77 supplies the Lean 4 inductive type DFUMT8 together with neg, and, or, implies, and the Decidable instance, all proved as 29 zero-sorry theorems via decide. The library is Mathlib-free, builds in approximately 5 seconds, and is released under Apache-2.0 (github.com/fc0web/lean-d-fumt8).

2.4 ZCSG three-layer notation (Paper 61)

The Zero-Centered Symbol Grammar (Fujimoto 2026, Paper 61, DOI 10.5281/zenodo.15217458) introduces a three-layer notation for emptiness:

Symbol	Dimension	Nāgārjuna interpretation	Mathematical entity	D-FUMT₈
o0	−1	emptiness-before-emptiness, unnameable	empty set with reduced homology H̃₋₁(∅) = ℤ	NEITHER
0	0	śūnyatā-of-śūnyatā (pure origin) — the 0₀ genesis layer	fixed point, origin	ZERO
0o	+1	dependent co-arising	dimension-bearing object	BOTH

The central axis 0 is the formal symbol Paper 61 introduces for "śūnyatā(śūnyatā) = śūnyatā" — emptiness applied to itself as a fixed point. This paper's upper-layer Ω is structurally a contraction map from dimension +1 (BOTH) to dimension 0 (the ZCSG central axis); the linkage is exhibited in §3.8.

3. Lean 4 Formalization (★ BUILT — machine-verified zero-sorry, zero-axiom-dependence)

All Lean 4 fragments below are machine-verified in data/lean4-mathlib/CollatzRei/Paper159Inclosure.lean against Lean 4 v4.27.0 + Mathlib v4.27.0 (Mathlib is loaded by the host project but not used by this file — the formalization is Mathlib-free in spirit; see §4.5). lake build CollatzRei.Paper159Inclosure exits 0 in ~1.7 seconds. All 9 load-bearing theorems pass #print axioms with "does not depend on any axioms" — i.e. they hold under the empty axiom set (purely constructive decide reasoning on a finite inductive type, no Classical reasoning required). See Appendix B for the full audit transcript.

Pinned commit: 4f5c50020f8ba0e589f8d6ebe278195ee812a1b7 (the commit immediately before the v0.2 update).

3.1 FDE inductive type

inductive FDE : Type where
  | t : FDE  -- true (and not false)
  | f : FDE  -- false (and not true)
  | b : FDE  -- both true and false
  | n : FDE  -- neither true nor false
  deriving Repr, DecidableEq, Inhabited

3.2 FDE-to-DFUMT8 embedding φ

def phi : FDE → DFUMT8
  | .t => .TRUE
  | .f => .FALSE
  | .b => .BOTH
  | .n => .NEITHER

3.3 Structural preservation theorems (lower layer) — partial machine verification

The TypeScript proofs in STEP 476 (verifyNOT, verifyAND, verifyOR, verifyLattice) are lifted to Lean 4 incrementally. v0.2 covers negation preservation (machine-verified); conjunction/disjunction/lattice preservation is deferred to v0.3 (the and/or truth tables are larger and have multiple non-equivalent semantics in the literature — Belnap bilattice meet/join vs Paper 77 LeanDFumt's operational meet — see honest scope note).

v0.2 BUILT (in Paper159Inclosure.lean):

def FDE.neg : FDE → FDE
  | .t => .f | .f => .t | .b => .b | .n => .n

theorem FDE.toDFumt8_neg : ∀ x : FDE, (FDE.neg x).toDFumt8 = DFUMT8.neg x.toDFumt8 := by
  intro x; cases x <;> decide

#print axioms FDE.toDFumt8_neg → "does not depend on any axioms".

v0.3 DEFERRED: phi_and_preserves / phi_or_preserves / lattice order preservation. Honest scope reason: Belnap's original 1977 bilattice gives a specific meet/join under the truth ordering, while Paper 77 LeanDFumt's DFUMT8.and / DFUMT8.or follow an operational meet (FALSE absorbs, NEITHER propagates) — these are NOT identical on the four-corner subset. Establishing the embedding requires either (a) re-aligning Paper 77 and/or FDE definitions to match, or (b) parameterizing the embedding by which lattice convention is in force. Both are non-trivial editorial choices that should be made deliberately rather than rushed.

3.4 Upper-layer modal operator `omega_upper` with idempotency (★ load-bearing — main differential)

Pre-existing Rei artifacts (no Lean 4 formalization of any version of this property prior to v0.2):

SEED_KERNEL theory dfumt-idempotency (src/axiom-os/seed-kernel.ts line 239): stipulates "Ω(Ω(x)) → Ω(x) stability" as axiom.
STEP 513 operator-fixed-point-atlas.ts lines 86-93: operational TypeScript test of a fixed-point structure for an operator named omega.

★ Honest divergence finding (new in v0.2): STEP 513's omega has a different concrete definition from the operator Paper 159 needs. STEP 513 omega contracts INFINITY → BOTH, ZERO → NEITHER, FLOWING → TRUE, and is identity on {TRUE, FALSE, BOTH, NEITHER, SELF} (5 fixed points); the Paper 159 upper-layer operator contracts BOTH → ZERO, NEITHER → ZERO, and is identity on {TRUE, FALSE, INFINITY, ZERO, FLOWING, SELF} (6 fixed points). Both share the structural property Ω ∘ Ω = Ω (the SELF⟲ idempotency axis of D-FUMT₈), but their semantics — what they contract to what — differ.

To avoid silent name-clash, the Lean file names the Paper 159 operator omega_upper. STEP 513's operator (named omega in TypeScript) is not formalized in this file.

v0.2 BUILT (in Paper159Inclosure.lean):

def DFUMT8.omega_upper : DFUMT8 → DFUMT8
  | .TRUE     => .TRUE
  | .FALSE    => .FALSE
  | .BOTH     => .ZERO     -- ★ Priest-Garfield BOTH terminus contracted to 0₀
  | .NEITHER  => .ZERO     -- ★ catuṣkoṭi 4th corner contracted to 0₀
  | .INFINITY => .INFINITY
  | .ZERO     => .ZERO
  | .FLOWING  => .FLOWING
  | .SELF     => .SELF

theorem omega_upper_idempotent (x : DFUMT8) :
    DFUMT8.omega_upper (DFUMT8.omega_upper x) = DFUMT8.omega_upper x := by
  cases x <;> decide

#print axioms omega_upper_idempotent → "does not depend on any axioms".

Additional machine-verified theorems in the same file (all axiom-free):

omega_upper_both_to_zero : DFUMT8.omega_upper .BOTH = .ZERO
omega_upper_neither_to_zero : DFUMT8.omega_upper .NEITHER = .ZERO
omega_upper_reflective_stable (4-conjunct: INFINITY, ZERO, FLOWING, SELF all fixed)
omega_upper_polar_stable (2-conjunct: TRUE, FALSE all fixed)
omega_upper_fixed_point_count (8-conjunct: exactly 6 of 8 values are fixed; only BOTH and NEITHER move)

3.5 Inclosure Schema (★ design decision — classical-Lean adaptation)

★ Honest design resolution (new in v0.2): the literal Q (delta Q) ∧ ¬ Q (delta Q) form is a paraconsistent statement; classical Lean 4 rejects it as False. v0.1 flagged this as a "known design issue, flagged for v0.2" — v0.2 resolves it as follows.

We split the schema into two parts:

The classical Lean structure records closure + transcendence as formal obligations on Q, ψ, δ. No "contradiction theorem" is asserted at the Prop level.
The apparent limit-contradiction is expressed in the DFUMT8 TYPE (as the value BOTH at the lower layer), not in Lean's Prop. The "apparent contradiction" lives in the eight-valued type system, where BOTH is a first-class value distinct from both TRUE and FALSE.

v0.2 BUILT (in Paper159Inclosure.lean):

structure Inclosure (α : Type) where
  Q : α → Prop                          -- the totality Ω as a predicate
  psi : (α → Prop) → Prop               -- "definability" / "set has common nature"
  delta : (α → Prop) → α                -- the diagonalizer
  closure : ∀ X : α → Prop,
    (∀ x, X x → Q x) → psi X → Q (delta X)
  transcendence : ∀ X : α → Prop,
    (∀ x, X x → Q x) → psi X → ¬ X (delta X)

No inclosure_limit_contradiction theorem (which would imply False classically). Instead:

def Inclosure.limitLowerLayer {α : Type} (_I : Inclosure α) : DFUMT8 := .BOTH

def Inclosure.limitUpperLayer {α : Type} (I : Inclosure α) : DFUMT8 :=
  DFUMT8.omega_upper I.limitLowerLayer

This is a design choice in the rational-reconstruction stance: we encode Priest-Garfield's BOTH terminus as a DFUMT8 value rather than as a classical Prop, then study its behavior under omega_upper. The schema's structural obligations (closure, transcendence) remain in Prop; only the terminus assignment is in DFUMT8. This avoids both (a) needing to make Lean paraconsistent (which would require a custom logic core, far outside scope) and (b) silently letting False enter the load-bearing proof chain.

3.5.1 Two-layer schema visualization (★ Rei extension of Garfield-Priest 2003 Fig. 1)

The following figure summarizes the two-layer architecture. The lower half reproduces Garfield & Priest (2003, p. 17, Fig. 1) — the original Inclosure visualization — with the δ(Ω) terminus made explicit as a DFUMT₈ value BOTH. The upper half is the Paper 159 differential: an idempotent operator omega_upper contracts BOTH (Priest-Garfield's dialetheist terminus) and NEITHER (catuṣkoṭi's 4th corner) to the ZCSG genesis layer 0₀ = ZERO (Paper 61). The right-side panel enumerates omega_upper's action on all eight D-FUMT₈ values: six of eight are fixed; only BOTH and NEITHER move.

Fig. 1 (Rei extension). Lower layer reproduces Garfield & Priest (2003) Fig. 1 verbatim with the BOTH terminus made explicit as a DFUMT₈ value. Upper layer is the Paper 159 differential: idempotent operator Ω_upper contracts BOTH and NEITHER to the ZCSG genesis layer 0₀ = ZERO (Paper 61). All connecting structure (Ω_upper definition, idempotency proof, contraction theorems) is machine-verified in Lean 4 zero-axiom-dependence — see §3.4 + Appendix B. SVG source: papers/figures/paper-159-fig1-two-layer-inclosure.svg. The lower-layer visualization style and labels (Ω, ψ, δ, φ) follow Priest-Garfield 2003; we claim no novelty for that portion. The upper-layer addition (red dashed arrow, Ω_upper box, 0₀ disc, D-FUMT₈ legend panel) is Rei-AIOS.

3.5.2 Rei-native original mandala (★ NOT derived from Priest-Garfield Fig. 1)

The following figure is a separate, original composition in Rei's own ZCSG-native visual vocabulary. It is not a derivative of Garfield-Priest 2003 Fig. 1 (that role belongs to Fig. 1 above). Fig. 2 expresses the same two-layer structure from Rei's own perspective: a mandala with 0₀ at the center, the eight D-FUMT₈ values as petals, the Ω_upper contractions of BOTH and NEITHER shown as red dashed inward arrows, the six fixed values shown with faint self-loops, and an outer dashed cycle suggesting the recursive flow 空 → 自己参照 → 創発 → 再帰的自己変容 (★ explicitly labeled as PAPER 2 TERRITORY — not formalized in v0.2). The right-side panels record the three interpretive-stance compatibility (P-G dialetheist / Tillemans weak / non-dialetheist), an honest status legend distinguishing verified / proposed / Paper 2 territory, the complete Ω_upper action table, and a hint pointing to the separate sages-axes lens (STEP 1188) where the same eight-axis structure is read across multiple civilizational traditions.

Fig. 2 (Rei original mandala). Central 0₀ medallion = ZCSG central axis (Paper 61). Eight petals = D-FUMT₈ values (Paper 77 substrate). Red dashed Ω_upper arrows (★ Paper 159 §3.4 differential): BOTH (Priest-Garfield terminus) and NEITHER (catuṣkoṭi 4th corner) contract to 0₀; six of eight values fixed (faint self-loops on TRUE / FALSE / INFINITY / FLOWING / ZERO / SELF — SELF marked with the ⟲ idempotency symbol). Outer dashed ring = chat-Claude / ChatGPT cycle 空 → 自己参照 → 創発 → 再帰的自己変容 (★ PAPER 2 TERRITORY, proposed, NOT formalized in v0.2). Right panels: three interpretive stances on one substrate / honest status legend / Ω_upper full action / sages-axes lens hint. SVG source: papers/figures/paper-159-fig2-rei-mandala.svg. Both Fig. 1 (derivative of Priest-Garfield) and Fig. 2 (Rei-native original) express the same two-layer architecture: Fig. 1 in Priest-Garfield's set-theoretic idiom, Fig. 2 in Rei's mandala / ZCSG idiom. The honest pairing makes the derivative vs original split explicit.

3.6 Nāgārjuna's Paradox as an Inclosure instance (lower-layer assignment: BOTH)

Following Garfield & Priest (2003: 17-18), we instantiate the Inclosure schema with φ(χ) = "χ is empty", ψ(X) = "X is a set of things with some common nature", δ(X) = "the nature of things in X". The lower-layer assignment of δ(Ω) is BOTH (the value Priest-Garfield assign): emptiness has the nature of being empty, but as it is itself empty, it has no nature.

3.7 Expressibility Paradox as an Inclosure instance (lower-layer assignment: BOTH)

Following Garfield & Priest (2003: 17): φ(χ) = "χ is an ultimate truth", ψ(X) = "X is definable", δ(X) = the sentence asserting that X has no member. The lower-layer assignment of δ(Ω) is again BOTH.

3.8 Upper-layer Ω contracts BOTH to ZERO — main differential (★ machine-verified)

The main formal differential of Paper 159 against Priest-Garfield's BOTH-only terminus: the upper layer omega_upper contracts the lower-layer BOTH to ZERO, the ZCSG 0₀ genesis layer (Paper 61). This holds for ANY Inclosure instance whose lower-layer assignment is BOTH — including Nāgārjuna's Paradox (§3.6) and the Expressibility Paradox (§3.7).

v0.2 BUILT (in Paper159Inclosure.lean):

theorem Inclosure.limit_upper_is_zero {α : Type} (I : Inclosure α) :
    I.limitUpperLayer = .ZERO := by
  unfold Inclosure.limitUpperLayer Inclosure.limitLowerLayer
  decide

theorem Inclosure.limit_upper_idempotent {α : Type} (I : Inclosure α) :
    DFUMT8.omega_upper I.limitUpperLayer = I.limitUpperLayer := by
  rw [Inclosure.limit_upper_is_zero]
  decide

-- Direct sanity-check examples
example : DFUMT8.omega_upper .BOTH = .ZERO := by decide
example : DFUMT8.omega_upper (DFUMT8.omega_upper .BOTH) = .ZERO := by decide
example : DFUMT8.omega_upper .NEITHER = .ZERO := by decide
example : DFUMT8.omega_upper (DFUMT8.omega_upper .NEITHER) = .ZERO := by decide

#print axioms Inclosure.limit_upper_is_zero → "does not depend on any axioms".

The full semantic characterization of the genesis layer ZERO (= ZCSG 0) — including its relation to the empty set's reduced homology H̃₋₁(∅) = ℤ from Paper 61 §2.2, the "self-reference → recursive self-transformation" question raised in the chat-Claude / ChatGPT design discussion (Acknowledgments), and the question "why does a system without self generate self" (the natural next problem after the present formal substrate) — is deferred to a planned Paper 2 on the ontology of 0₀.

3.9 Compatibility with multiple interpretive stances (★ new in v0.2)

The formal substrate built in §3.4–3.8 is compatible with at least three readings of catuṣkoṭi. The Lean 4 file includes example statements for each:

-- (P-G) Dialetheist reading (Priest-Garfield 2003 / Deguchi-Garfield-Priest):
--       use only the LOWER layer; δ(Ω) = BOTH is endorsed.
example : (.BOTH : DFUMT8) = .BOTH := rfl

-- (T) Weak dialetheist (Tillemans 2009, 2024):
--     use the upper layer to contract BOTH to ZERO, so conjoined contradictions
--     φ ∧ ¬φ are NOT endorsed at the modal layer.
example : DFUMT8.omega_upper .BOTH = .ZERO := by decide

-- (¬D) Non-dialetheist:
--      assign δ(Ω) = NEITHER at the lower layer instead of BOTH. The upper
--      layer still contracts NEITHER to ZERO.
example : DFUMT8.omega_upper .NEITHER = .ZERO := by decide

None of these readings is endorsed by the present paper; the substrate accepts each by varying the lower-layer assignment function. This is the formal expression of the non-commitment stance declared in §4.3.

4. Honest Framing and Scope (★ load-bearing)

4.1 Rational reconstruction stance

Garfield and Priest (2003: 2) explicitly disavow textual-history claims:

"Finally, we do not claim that Nāgārjuna himself had explicit views about logic, or about the limits of thought. We do, however, think that if he did, he had the views we are about to sketch. This is, hence, not textual history but rational reconstruction."

We adopt the same stance. The Lean 4 formalization proposed here is a formal substrate compatible with Priest-Garfield's reading; it does not assert that Nāgārjuna held the corresponding views.

4.2 Scholarly debate disclosure

The Priest-Garfield dialetheist reading has been debated for two decades. Notably:

Tillemans (1999): broadly aligned with Priest-Garfield (paraconsistent at the ultimate level, classical at the conventional level — Priest-Garfield 2003 footnote 2 explicitly agree with Tillemans on this point).
Tillemans (2009, 2014): later argues for a "weak dialetheist" reading of Nāgārjuna, holding that Nāgārjuna does not accept conjoined contradictions φ ∧ ¬φ.
Deguchi, Garfield & Priest (2008, 2013): defend and elaborate the dialetheist reading.
Siderits, Ferraro, Tao Jiang: various critical engagements.
A 2024 paper in Asian Philosophy offers an updated critique of Priest-Garfield's use of catuṣkoṭi.

We have not yet read these critic papers directly. Citations above are from secondary triangulation (web search snippets and one verified primary source: Garfield & Priest 2003). Their detailed evaluation is deferred to v0.2.

4.3 Non-commitment to interpretive position

We take no position in the dialetheist / weak-dialetheist / non-dialetheist debate. The formal substrate proposed here is, by design, compatible with each: Priest-Garfield's reading is preserved as the lower-layer assignment of BOTH; Tillemans's worry that conjoined contradictions are not endorsed can be addressed via the upper-layer contraction (the lower-layer BOTH does not survive Ω application); non-dialetheist readings that reject the BOTH assignment altogether can use the lower layer with only TRUE / FALSE / NEITHER.

4.4 No "world first" claim

Two prior uniqueness claims should be acknowledged and respected:

Paper 61 ZCSG (Fujimoto 2026): claims "world-first mathematical encoding of śūnyatā-of-śūnyatā" as the symbol 0.
Priest-Garfield 2003 (p. 18): name the ontological paradox "Nāgārjuna's Paradox" and state that, "to our knowledge, [it] is found nowhere else."

The present paper avoids any "world first" framing. The contribution is the two-layer reconstruction plus Lean 4 substrate, both of which presuppose and extend those prior claims rather than competing with them.

4.5 Library-only, Mathlib-free in spirit, `decide`-only (★ stronger than expected)

We inherit Paper 77's design discipline: all proofs by decide (no native_decide axiom used), propositional fragment on finite inductive types. The host project (data/lean4-mathlib/) requires Mathlib, but Paper159Inclosure.lean itself does NOT import Mathlib — it is Mathlib-free as a file. Build time: 1.7 seconds via lake build CollatzRei.Paper159Inclosure.

★ Result is stronger than initially planned: all 9 load-bearing theorems show #print axioms ... does not depend on any axioms (no propext, no Classical.choice, no Quot.sound, no native_decide). This is the strongest possible axiom-purity result. Paper 158 (Collatz exit layer) by contrast uses [propext, Classical.choice, Quot.sound].

4.6 Ω idempotency: honest provenance disclosure (★ updated in v0.2)

The idempotency Ω ∘ Ω = Ω is stipulated in the Rei SEED_KERNEL as the axiom dfumt-idempotency and is operationally tested in STEP 513 (TypeScript). v0.2 of this paper produces the first Lean 4 machine-verified statement of this property in the Rei stack — but as a decide-only proof of an operationally specific operator omega_upper, NOT as a general statement about all idempotent operators on D-FUMT₈. The general claim "any operator Ω with Ω ∘ Ω = Ω is well-behaved" is not formalized; v0.2 verifies one concrete operator instance.

★ As noted in §3.4, STEP 513's omega and this paper's omega_upper are distinct operators sharing the structural idempotency property. v0.1 implied an identity that does not hold. v0.2 corrects this and re-states the relationship as "shared structural pattern, distinct semantics".

4.7 Upper-layer scope limitations

The upper-layer operator Ω is defined here only on D-FUMT₈ values. A full semantic theory of the genesis layer ZERO — including its identification with the ZCSG central axis 0, its connection to the reduced homology of the empty set, and its role as the convergence point of catuṣkoṭi negations — is explicitly deferred to a planned Paper 2 on the ontology of 0₀.

5. Discussion

5.1 Relation to STEP 476 (TypeScript engine)

STEP 476 nagarjuna-fde-western-engine.ts (654 lines) supplies the operational TypeScript proofs of catuṣkoṭi ≅ FDE ≅ DFUMT4 isomorphism: verifyNOT, verifyAND (16 cases), verifyOR (16 cases), verifyLattice. The present paper proposes to lift these to Lean 4 zero-sorry theorems via decide, thereby promoting machine-checked algebraic proofs of the same content.

5.2 Heidegger comparison (after Priest-Garfield 2003: 16)

Priest-Garfield identify Heidegger as the closest Western parallel to Nāgārjuna's ontological insight, while noting that Heidegger does not follow the identification of the two truths (conventional and ultimate). We adopt the same comparison and the same limitation.

5.3 Two-layer architecture as the main differential

Priest-Garfield's Inclosure terminates at BOTH — the dialetheist commitment to a true contradiction. Our two-layer reconstruction preserves BOTH at the lower layer (so Priest-Garfield's analysis is embedded as a special case) but adds an upper-layer Ω that contracts BOTH to ZERO. The strategy is analogous to QMRP (Paper 26 candidate) reframing Shannon's bound as a special case at finite N. The differential is twofold:

Structural: separating object-level truth values from meta-level modal operators, formalizing the "all standpoints exhausted" reading of catuṣpaścāś as an idempotent stable point rather than as a singleton truth value.
Methodological: machine verification via Lean 4 decide, which Priest 1987/2002 and Priest-Garfield 2003 do not pursue.

5.4 Connection to ZCSG (Paper 61)

ZCSG (Paper 61) introduces the central symbol 0 for śūnyatā-of-śūnyatā, with dimension 0, distinguishing it from o0 (dimension −1, NEITHER) and 0o (dimension +1, BOTH). The upper-layer Ω of the present paper is structurally a contraction map from dimension +1 to dimension 0, i.e. from 0o to 0. The two papers thus approach the same structure from complementary angles: ZCSG from notation, the present paper from Lean 4 formalization.

5.5 What this paper does NOT claim

That Nāgārjuna was a dialetheist (or was not).
A resolution to the Priest-Garfield / Tillemans / Siderits debate.
Any extension to Mathlib's standard logic.
Any new philosophical thesis beyond rational-reconstruction formal substrate.
A tier ranking of philosophers (a methodologically dubious framing we explicitly avoid).
"Four-step termination" of catuṣkoṭi negations — we adopt the cleaner idempotent stable-point reading consistent with catuṣpaścāś (four corners and one hundred negations) tradition.
Prior machine verification of Ω idempotency in the Rei stack (it has been axiomatically stipulated and operationally tested only).
Full semantic theory of the genesis layer ZERO (deferred to a planned Paper 2).

6. Conclusion

This v0.2 LEAN-4-BUILT paper supplies a two-layer D-FUMT₈ reconstruction of Priest-Garfield's Inclosure Schema for Nāgārjuna's catuṣkoṭi. The lower layer (FDE embedding, BOTH at limit) preserves Priest-Garfield's analysis as a special case. The upper layer (idempotent modal operator omega_upper contracting BOTH and NEITHER to ZERO) supplies the differential that Priest-Garfield's BOTH-only terminus does not offer, and connects to the central axis 0 of the ZCSG three-layer notation (Paper 61). All 9 load-bearing theorems are machine-verified in Lean 4 with decide-only proofs that depend on no axioms (no propext, no Classical.choice, no Quot.sound, no native_decide); see Appendix B. The paper is positioned as rational reconstruction following Priest-Garfield's own self-framing, compatible with at least three interpretive stances of catuṣkoṭi, and explicitly defers the genesis-layer ontology to a planned Paper 2.

Two honest findings emerged during the v0.1 → v0.2 build:

omega_upper is distinct from STEP 513's omega (§3.4). Both are idempotent but contract to different fixed-point sets; v0.1 implied a closer relationship than holds. v0.2 corrects this by renaming the Paper 159 operator and stating the relationship as "shared structural pattern, distinct semantics".
The literal limit-contradiction cannot be formalized in classical Lean (§3.5). v0.2 resolves this by encoding Priest-Garfield's BOTH terminus as a DFUMT8 VALUE rather than as a classical Prop, leaving Lean's Prop consistent.

The Tillemans 2009/2014/2024 + Siderits + Ferraro + Tao Jiang critic papers remain NOT YET READ (§4.2 unchanged). Their direct evaluation is deferred to v0.3.

Acknowledgments

The two-layer architecture in §3.4 / §3.8 was articulated in a chat session with Claude (Anthropic) on 2026-05-31, building on the design discussion in an earlier session on the same date concerning the BOTH-vs-NEITHER reading of catuṣkoṭi negation. The articulation of the lower-layer = Priest, upper-layer = Ω with idempotency design was that session's central contribution.

The full text of Garfield & Priest (2003) was accessed via the open Smith ScholarWorks repository on 2026-05-31; this paper would have been substantially weaker without that primary source. We thank both authors for making the original work openly available.

References

Anderson, A. R., & Belnap, N. D. (1975). Entailment: The Logic of Relevance and Necessity, Vol. I. Princeton University Press.
Belnap, N. D. (1977). A useful four-valued logic. In J. M. Dunn & G. Epstein (Eds.), Modern Uses of Multiple-Valued Logic (pp. 5–37). Reidel.
Deguchi, Y., Garfield, J. L., & Priest, G. (2008). The way of the dialetheist: Contradictions in Buddhism. Philosophy East and West, 58(3), 395–402.
Deguchi, Y., Garfield, J. L., & Priest, G. (2013). How we think Mādhyamikas think: A response to Tom Tillemans. Philosophy East and West, 63(3), 426–435.
Fujimoto, N. (2026). Zero-Centered Symbol Grammar (ZCSG): A World-First Dimensional Encoding of Nāgārjuna's Śūnyatā-of-Śūnyatā. Paper 61, DOI 10.5281/zenodo.15217458.
Fujimoto, N. (2026). LeanDFumt: An Open-Source Eight-Valued Logic Library for Lean 4. Paper 77. GitHub: https://github.com/fc0web/lean-d-fumt8.
Garfield, J. L. (Trans.) (1995). The Fundamental Wisdom of the Middle Way: Nāgārjuna's Mūlamadhyamakakārikā. Oxford University Press.
Garfield, J. L., & Priest, G. (2003). Nāgārjuna and the limits of thought. Philosophy East and West, 53(1), 1–21. JSTOR 1400052. Open-access copy: Smith ScholarWorks.
Hayes, R. (1994). Nāgārjuna's appeal. Journal of Indian Philosophy, 22, 299–378.
Nāgārjuna (c. 150 CE). Mūlamadhyamakakārikā. (See Garfield 1995 for English translation.)
Priest, G. (1987). In Contradiction: A Study of the Transconsistent. Martinus Nijhoff (2nd ed. Oxford, 2006).
Priest, G. (2002). Beyond the Limits of Thought (2nd ed.). Oxford University Press.
Siderits, M. (1989). Thinking on empty: Madhyamika anti-realism and canons of rationality. In S. Biderman & B.-A. Scharfstein (Eds.), Rationality in Question: On Eastern and Western Views of Rationality. Brill.
Tillemans, T. J. F. (1999). Is Nāgārjuna's logic deviant or non-classical? In Scripture, Logic, Language: Essays on Dharmakīrti and His Tibetan Successors. Wisdom Publications.
The Lean 4 development team (2024–2026). Lean 4 v4.27.0. https://lean-lang.org.

Additional critic papers (Tillemans 2009 / 2014 / 2024, Siderits later work, Ferraro 2013, Tao Jiang) acknowledged in §4.2 but not yet directly consulted; full citation list will be completed in v0.2.

Version history

v0.1 (2026-05-31, morning): Initial OUTLINE + SKELETON release. Lean 4 code PROPOSED, not yet built. Honest scope as stated. Released for archival continuity. Zenodo DOI 10.5281/zenodo.20468146, 11/11 platform broadcast (commit 4f5c5002).
v0.2 (2026-05-31, afternoon): Lean 4 BUILT — 9 load-bearing theorems machine-verified with zero axiom dependence (stronger than expected; even purer than Paper 158). Two honest findings: (a) omega_upper rename + STEP 513 semantic divergence note (§3.4); (b) classical-Lean limit-contradiction adaptation via DFUMT8-VALUE encoding (§3.5). FDE negation preservation machine-verified; conjunction/disjunction/lattice preservation deferred to v0.3 per honest scope (multiple lattice conventions in literature). Critic papers (Tillemans 2009+, Siderits, Ferraro, Tao Jiang) STILL NOT YET READ — unchanged from v0.1. Two figures added (both SVG): Fig. 1 (Rei extension of Garfield-Priest 2003 Fig. 1) in §3.5.1, and Fig. 2 (Rei-native original mandala, NOT derived from Priest-Garfield) in §3.5.2. Sources: papers/figures/paper-159-fig1-two-layer-inclosure.svg and papers/figures/paper-159-fig2-rei-mandala.svg. Pinned commit (pre-v0.2): 4f5c50020f8ba0e589f8d6ebe278195ee812a1b7.
v0.3 (planned, no timeline commitment): Direct reading of Tillemans 2009/2014/2024 + Siderits + Ferraro critic papers. FDE and/or/lattice preservation theorems (after editorial choice of which lattice convention to align). Possible Paper 2 design seed (genesis-layer ZERO ontology) extracted into separate document.

Appendix B — Lean 4 axiom audit (v0.2 BUILT verification)

Source file: data/lean4-mathlib/CollatzRei/Paper159Inclosure.lean (~200 lines)
Audit file: data/lean4-mathlib/CollatzRei/Paper159InclosurePrintAxioms.lean (#print axioms for each load-bearing theorem)
Lean version: v4.27.0 (host project uses Mathlib v4.27.0; this file imports no Mathlib)
Build command: lake build CollatzRei.Paper159Inclosure (1.7s, exit 0)
Pinned commit (pre-v0.2): 4f5c50020f8ba0e589f8d6ebe278195ee812a1b7

`#print axioms` output (exact transcript)

'CollatzRei.Paper159Inclosure.omega_upper_idempotent' does not depend on any axioms
'CollatzRei.Paper159Inclosure.omega_upper_both_to_zero' does not depend on any axioms
'CollatzRei.Paper159Inclosure.omega_upper_neither_to_zero' does not depend on any axioms
'CollatzRei.Paper159Inclosure.omega_upper_reflective_stable' does not depend on any axioms
'CollatzRei.Paper159Inclosure.omega_upper_polar_stable' does not depend on any axioms
'CollatzRei.Paper159Inclosure.omega_upper_fixed_point_count' does not depend on any axioms
'CollatzRei.Paper159Inclosure.FDE.toDFumt8_neg' does not depend on any axioms
'CollatzRei.Paper159Inclosure.Inclosure.limit_upper_is_zero' does not depend on any axioms
'CollatzRei.Paper159Inclosure.Inclosure.limit_upper_idempotent' does not depend on any axioms

Interpretation

"Does not depend on any axioms" means the proof requires none of:

sorryAx (would indicate an unfinished proof)
propext (propositional extensionality)
Classical.choice (the axiom of choice)
Quot.sound (quotient soundness)
native_decide (computed kernel reduction)
any user-introduced axiom

The proofs go through purely by decide on finite inductive types. This is the strongest possible axiom-purity result in Lean 4.

Comparison with Paper 158

Paper 158 (Collatz exit layer) achieved zero-sorry but the load-bearing theorems show [propext, Classical.choice, Quot.sound]. Paper 159 v0.2 is strictly purer: no Classical reasoning is even invoked. This is a consequence of working entirely on a finite inductive type (DFUMT8 with 8 constructors) rather than on ℕ with division and induction.

Sample reproduction commands (cygwin/PowerShell)

cd C:/Users/user/rei-aios/data/lean4-mathlib
lake build CollatzRei.Paper159Inclosure              # produces .olean
lake env lean CollatzRei/Paper159InclosurePrintAxioms.lean  # prints audit

Rei-AIOS Project. Peace Axiom #196: immutable = true. License: CC-BY 4.0 (text), Apache-2.0 (any associated Lean 4 code).

Paper 159 - A Two-Layer D-FUMT-8 Reconstruction of Priest-Garfield's Inclosure Schema for Nagarjuna's Catuskoti (OUTLINE v0.1)

Nobuki Fujimoto — Sun, 31 May 2026 01:59:05 +0000

This article is a re-publication of Rei-AIOS Paper 159 for the dev.to community.
The canonical version with full reference list is in the permanent archives below:

Zenodo (DOI, canonical): https://doi.org/10.5281/zenodo.20468146

Internet Archive: https://archive.org/details/rei-aios-paper-159-1780192623655

GitHub source (private): https://github.com/fc0web/rei-aios Author: Nobuki Fujimoto (@fc0web) · ORCID 0009-0004-6019-9258 · License CC-BY-4.0 ---

Version: v0.1 OUTLINE + SKELETON (★ NOT MACHINE-VERIFIED — WORK IN PROGRESS — preliminary outline released for archival continuity per Rei-AIOS publish-early-with-honest-scope convention)

Author: Nobuki Fujimoto (藤本伸樹), with Claude (Rei-AIOS)
ORCID: 0009-0004-6019-9258 · GitHub: fc0web · note.com: nifty_godwit2635

Date: 2026-05-31

Companion to:

Paper 61 — Zero-Centered Symbol Grammar (ZCSG) — DOI 10.5281/zenodo.15217458 (world-first mathematical encoding of śūnyatā-of-śūnyatā as 0)
Paper 77 — LeanDFumt: an Eight-Valued Logic Library for Lean 4 — substrate of the present formalization

Status header (★ load-bearing):

✅ Primary source confirmed: Garfield & Priest 2003 "Nāgārjuna and the Limits of Thought" (Philosophy East and West 53(1): 1-21) — full text read 2026-05-31
✅ Pre-existing Rei artifacts surveyed: STEP 476 nagarjuna-fde-western-engine.ts (TS, FDE ≅ DFUMT4 isomorphism) + STEP 513 operator-fixed-point-atlas.ts (Ω operational fixed-point test) + Paper 77 LeanDFumt substrate
⚠ Lean 4 code in §3 is PROPOSED — NOT YET BUILT (next-version target)
⚠ Tillemans 2009/2014/2024 + Siderits + Ferraro + Tao Jiang critic papers NOT YET READ (acknowledged in §4.2, citations from secondary triangulation only)
⚠ No claim of resolving the Priest-Garfield vs Tillemans scholarly debate (§4.3)
⚠ No "world-first" claim (§4.4, Paper 61 ZCSG + Priest-Garfield 2003 already issued such claims — this paper provides formal substrate only)

Abstract

Keywords: D-FUMT₈, catuṣkoṭi, Inclosure Schema, Priest-Garfield, dialetheism, paraconsistent logic, Lean 4, rational reconstruction, śūnyatā, two-layer logic, Rei-AIOS, ZCSG, Nāgārjuna

1. Introduction

This paper builds on three pre-existing Rei-AIOS artifacts:

STEP 476 nagarjuna-fde-western-engine.ts: a TypeScript engine demonstrating the structural isomorphism catuṣkoṭi ≅ FDE (Belnap 1977) ≅ D-FUMT₈ base 4 values
STEP 513 operator-fixed-point-atlas.ts: a TypeScript enumeration of fixed points for the operators Ω, Φ, Ψ, NOT and their compositions (320 fixed-point checks), which includes the operational test Ω ∘ Ω = Ω
Paper 77 LeanDFumt: the open-source Lean 4 library providing the inductive type DFUMT8 with negation, conjunction, disjunction, and the Decidable infrastructure used here

Honest scope (★ load-bearing throughout):

We follow Priest-Garfield's own self-framing of their work as "rational reconstruction" — not textual history (Garfield & Priest 2003: 2, "we do not claim that Nāgārjuna himself had explicit views about logic ... This is, hence, not textual history but rational reconstruction").
We take no position in the dialetheist (Priest-Garfield, Deguchi-Garfield-Priest 2008 / 2013) versus weak-dialetheist (Tillemans 2009, 2024) debate, nor in the broader scholarly discussion (Siderits, Ferraro, Tao Jiang).
We do not claim that Nāgārjuna was a dialetheist; we provide a formal substrate compatible with multiple interpretive stances.
We do not claim "world first" status: Paper 61 (ZCSG) and Priest-Garfield (2003, who name "Nāgārjuna's Paradox" and write that the ontological paradox "to our knowledge is found nowhere else") have already issued related uniqueness claims. This paper supplies formal substrate only.

2. Background

2.1 Catuṣkoṭi: positive and negative uses (after Garfield & Priest 2003: 13-14)

Positive tetralemma — conventional perspective, e.g. MMK XVIII:6 ("That there is a self has been taught, And the doctrine of no-self, By the buddhas, as well as the Doctrine of neither self nor nonself"). The four corners are accepted at the conventional (saṃvṛti) level as distinct standpoints.
Negative tetralemma — ultimate perspective, e.g. MMK XXII:11 ("'Empty' should not be asserted. 'Nonempty' should not be asserted. Neither both nor neither should be asserted. These are used only nominally"). At the ultimate (paramārtha) level, all four corners are negated.

Priest and Garfield's analysis focuses on this negative use — at the ultimate level, catuṣkoṭi exhibits limit phenomena that they then formalize via the Inclosure Schema.

2.2 Inclosure Schema (Priest 1987, 2002)

Priest's Inclosure Schema isolates the general structure of self-referential paradoxes. Given properties φ and ψ and an operator δ:

Condition	Statement
(i) Transcendence	∀X ⊆ Ω, ψ(X) → δ(X) ∉ X
(ii) Closure	∀X ⊆ Ω, ψ(X) → δ(X) ∈ Ω

Ontological ("Nāgārjuna's Paradox"): φ(χ) = "χ is empty"; ψ(X) = "X is a set of things with some common nature"; δ(X) = "the nature of things in X". Closure: since all things are empty, δ(X) ∈ Ω. Transcendence: δ(X) has no nature (because emptiness is exactly the absence of nature), so δ(X) ∉ X. Limit: δ(Ω) — emptiness itself — both is and is not empty.
Expressibility: φ(χ) = "χ is an ultimate truth"; ψ(X) = "X is definable"; δ(X) = the sentence "there is nothing which is in D" (where D refers to X). The limit truth — "there is no ultimate truth" (Siderits 1989) — both is and is not an ultimate truth.

Priest-Garfield assign the value BOTH (true-and-false) to δ(Ω) in both cases.

2.3 D-FUMT₈ and the LeanDFumt substrate (Paper 77)

2.4 ZCSG three-layer notation (Paper 61)

The Zero-Centered Symbol Grammar (Fujimoto 2026, Paper 61, DOI 10.5281/zenodo.15217458) introduces a three-layer notation for emptiness:

Symbol	Dimension	Nāgārjuna interpretation	Mathematical entity	D-FUMT₈
o0	−1	emptiness-before-emptiness, unnameable	empty set with reduced homology H̃₋₁(∅) = ℤ	NEITHER
0	0	śūnyatā-of-śūnyatā (pure origin) — the 0₀ genesis layer	fixed point, origin	ZERO
0o	+1	dependent co-arising	dimension-bearing object	BOTH

3. Lean 4 Formalization (★ PROPOSED, NOT YET BUILT)

All Lean 4 fragments below are proposed definitions and theorems. They are syntactically reasonable against the Lean 4 v4.27.0 + Paper 77 LeanDFumt substrate, but have not yet been compiled and verified. Verification is the v0.2 target.

3.1 FDE inductive type

inductive FDE : Type where
  | t : FDE  -- true (and not false)
  | f : FDE  -- false (and not true)
  | b : FDE  -- both true and false
  | n : FDE  -- neither true nor false
  deriving Repr, DecidableEq, Inhabited

3.2 FDE-to-DFUMT8 embedding φ

def phi : FDE → DFUMT8
  | .t => .TRUE
  | .f => .FALSE
  | .b => .BOTH
  | .n => .NEITHER

3.3 Structural preservation theorems (lower layer)

The TypeScript proofs in STEP 476 (verifyNOT, verifyAND, verifyOR, verifyLattice) are proposed to lift to Lean 4 as:

def FDE.neg : FDE → FDE
  | .t => .f | .f => .t | .b => .b | .n => .n

def FDE.and : FDE → FDE → FDE := /- 16-entry truth table per STEP 476 -/
def FDE.or  : FDE → FDE → FDE := /- 16-entry truth table per STEP 476 -/

theorem phi_neg_preserves : ∀ x, phi (FDE.neg x) = DFUMT8.neg (phi x) := by
  intro x; cases x <;> decide

theorem phi_and_preserves : ∀ x y, phi (FDE.and x y) = DFUMT8.and (phi x) (phi y) := by
  intro x y; cases x <;> cases y <;> decide

theorem phi_or_preserves : ∀ x y, phi (FDE.or x y) = DFUMT8.or (phi x) (phi y) := by
  intro x y; cases x <;> cases y <;> decide

The lattice preservation theorem (Belnap information order n ≤ {t, f} ≤ b) is handled analogously.

3.4 Upper-layer modal operator Ω with idempotency

Pre-existing Rei artifacts:

SEED_KERNEL theory dfumt-idempotency (src/axiom-os/seed-kernel.ts line 239): stipulates "Ω(Ω(x)) → Ω(x) stability" as axiom.
STEP 513 operator-fixed-point-atlas.ts lines 292-297: operational TypeScript test comparing fixed-point sets of Ω∘Ω and Ω.
No prior Lean 4 formalization of this idempotency.

We propose the following minimal Lean 4 definition and theorem:

/-- Upper-layer modal operator. Maps the lower-layer four-corner positions
    (TRUE, FALSE, BOTH, NEITHER) to fixed-point images, with BOTH and NEITHER
    contracting to ZERO — the ZCSG 0₀ genesis layer (Paper 61 Table 2.3).
    The four reflective values (INFINITY, ZERO, FLOWING, SELF) are already
    Ω-stable in this construction. -/
def DFUMT8.omega : DFUMT8 → DFUMT8
  | .TRUE     => .TRUE
  | .FALSE    => .FALSE
  | .BOTH     => .ZERO     -- ★ Priest-Garfield BOTH terminus contracted to 0₀
  | .NEITHER  => .ZERO     -- ★ catuṣkoṭi 4th corner contracted to 0₀
  | .INFINITY => .INFINITY
  | .ZERO     => .ZERO
  | .FLOWING  => .FLOWING
  | .SELF     => .SELF

theorem omega_idempotent (x : DFUMT8) : DFUMT8.omega (DFUMT8.omega x) = DFUMT8.omega x := by
  cases x <;> decide

The proposed proof is one line: case-split on the eight constructors and decide discharges each. This is consistent with the Paper 77 decide-only proof discipline.

3.5 Inclosure Schema

structure Inclosure (α : Type) where
  Q : α → Prop                          -- the totality Ω as a predicate
  psi : (α → Prop) → Prop               -- "definability" / "set has common nature"
  delta : (α → Prop) → α                -- the diagonalizer
  closure : ∀ X : α → Prop,
    (∀ x, X x → Q x) → psi X → Q (delta X)
  transcendence : ∀ X : α → Prop,
    (∀ x, X x → Q x) → psi X → ¬ X (delta X)

theorem inclosure_limit_contradiction {α : Type} (I : Inclosure α)
    (hQ : I.psi I.Q) : I.Q (I.delta I.Q) ∧ ¬ I.Q (I.delta I.Q) := by
  refine ⟨?_, ?_⟩
  · exact I.closure I.Q (fun _ h => h) hQ
  · exact I.transcendence I.Q (fun _ h => h) hQ

(Note: the literal Q (delta Q) ∧ ¬ Q (delta Q) form is a paraconsistent statement; classical Lean 4 will reject this as False. The actual formalization will need either a paraconsistent encoding via DFUMT8.asProp or the introduction of contradictions as DFUMT8.BOTH-valued. This is a known design issue, flagged for v0.2.)

3.6 Nāgārjuna's Paradox as an Inclosure instance (lower-layer assignment: BOTH)

3.7 Expressibility Paradox as an Inclosure instance (lower-layer assignment: BOTH)

3.8 Upper-layer Ω contracts BOTH to ZERO (Paper 1 main differential)

/-- Priest-Garfield's terminus (BOTH at the limit) is preserved as our
    lower-layer assignment. The upper-layer operator Ω contracts it to ZERO,
    the ZCSG 0₀ genesis layer (Paper 61). This is the two-layer differential
    against the Priest-Garfield BOTH-only Inclosure terminus. -/
example : DFUMT8.omega .BOTH = .ZERO := by decide
example : DFUMT8.omega (DFUMT8.omega .BOTH) = .ZERO := by decide  -- idempotent stability
example : DFUMT8.omega .NEITHER = .ZERO := by decide               -- 4th corner also contracts

The full semantic characterization of the genesis layer ZERO (= ZCSG 0) — including its relation to the empty set's reduced homology H̃₋₁(∅) = ℤ from Paper 61 §2.2 — is deferred to a planned Paper 2 on the ontology of 0₀.

4. Honest Framing and Scope (★ load-bearing)

4.1 Rational reconstruction stance

Garfield and Priest (2003: 2) explicitly disavow textual-history claims:

"Finally, we do not claim that Nāgārjuna himself had explicit views about logic, or about the limits of thought. We do, however, think that if he did, he had the views we are about to sketch. This is, hence, not textual history but rational reconstruction."

We adopt the same stance. The Lean 4 formalization proposed here is a formal substrate compatible with Priest-Garfield's reading; it does not assert that Nāgārjuna held the corresponding views.

4.2 Scholarly debate disclosure

The Priest-Garfield dialetheist reading has been debated for two decades. Notably:

Tillemans (1999): broadly aligned with Priest-Garfield (paraconsistent at the ultimate level, classical at the conventional level — Priest-Garfield 2003 footnote 2 explicitly agree with Tillemans on this point).
Tillemans (2009, 2014): later argues for a "weak dialetheist" reading of Nāgārjuna, holding that Nāgārjuna does not accept conjoined contradictions φ ∧ ¬φ.
Deguchi, Garfield & Priest (2008, 2013): defend and elaborate the dialetheist reading.
Siderits, Ferraro, Tao Jiang: various critical engagements.
A 2024 paper in Asian Philosophy offers an updated critique of Priest-Garfield's use of catuṣkoṭi.

4.3 Non-commitment to interpretive position

4.4 No "world first" claim

Two prior uniqueness claims should be acknowledged and respected:

Paper 61 ZCSG (Fujimoto 2026): claims "world-first mathematical encoding of śūnyatā-of-śūnyatā" as the symbol 0.
Priest-Garfield 2003 (p. 18): name the ontological paradox "Nāgārjuna's Paradox" and state that, "to our knowledge, [it] is found nowhere else."

4.5 Library-only, Mathlib-free, `decide`-only

We inherit Paper 77's design discipline: no Mathlib dependency, all proofs by decide or native_decide, propositional fragment only. Build time targets are ~5 seconds for the library plus a few additional seconds for the present paper's theorems.

4.6 Ω idempotency: honest provenance disclosure

The idempotency Ω ∘ Ω = Ω is stipulated in the Rei SEED_KERNEL as the axiom dfumt-idempotency and is operationally tested in STEP 513. No prior Lean 4 formalization exists in the public or private Rei codebase. The one-line decide proof proposed in §3.4 would be the first machine-verified statement of this property in the Rei stack.

4.7 Upper-layer scope limitations

5. Discussion

5.1 Relation to STEP 476 (TypeScript engine)

5.2 Heidegger comparison (after Priest-Garfield 2003: 16)

5.3 Two-layer architecture as the main differential

Structural: separating object-level truth values from meta-level modal operators, formalizing the "all standpoints exhausted" reading of catuṣpaścāś as an idempotent stable point rather than as a singleton truth value.
Methodological: machine verification via Lean 4 decide, which Priest 1987/2002 and Priest-Garfield 2003 do not pursue.

5.4 Connection to ZCSG (Paper 61)

5.5 What this paper does NOT claim

That Nāgārjuna was a dialetheist (or was not).
A resolution to the Priest-Garfield / Tillemans / Siderits debate.
Any extension to Mathlib's standard logic.
Any new philosophical thesis beyond rational-reconstruction formal substrate.
A tier ranking of philosophers (a methodologically dubious framing we explicitly avoid).
"Four-step termination" of catuṣkoṭi negations — we adopt the cleaner idempotent stable-point reading consistent with catuṣpaścāś (four corners and one hundred negations) tradition.
Prior machine verification of Ω idempotency in the Rei stack (it has been axiomatically stipulated and operationally tested only).
Full semantic theory of the genesis layer ZERO (deferred to a planned Paper 2).

6. Conclusion

This v0.1 OUTLINE paper proposes a two-layer D-FUMT₈ reconstruction of Priest-Garfield's Inclosure Schema for Nāgārjuna's catuṣkoṭi. The lower layer (FDE isomorphism, BOTH at limit) preserves Priest-Garfield's analysis as a special case. The upper layer (idempotent modal operator Ω contracting BOTH and NEITHER to ZERO) supplies the differential that Priest-Garfield's BOTH-only terminus does not offer, and connects to the central axis 0 of the ZCSG three-layer notation (Paper 61). The Lean 4 substrate proposed against Paper 77 LeanDFumt is sketched but not yet built; full machine verification is the v0.2 target. The paper is positioned as rational reconstruction following Priest-Garfield's own self-framing, compatible with multiple readings of catuṣkoṭi, and explicitly defers the genesis-layer ontology to a planned Paper 2.

Acknowledgments

References

Anderson, A. R., & Belnap, N. D. (1975). Entailment: The Logic of Relevance and Necessity, Vol. I. Princeton University Press.
Belnap, N. D. (1977). A useful four-valued logic. In J. M. Dunn & G. Epstein (Eds.), Modern Uses of Multiple-Valued Logic (pp. 5–37). Reidel.
Deguchi, Y., Garfield, J. L., & Priest, G. (2008). The way of the dialetheist: Contradictions in Buddhism. Philosophy East and West, 58(3), 395–402.
Deguchi, Y., Garfield, J. L., & Priest, G. (2013). How we think Mādhyamikas think: A response to Tom Tillemans. Philosophy East and West, 63(3), 426–435.
Fujimoto, N. (2026). Zero-Centered Symbol Grammar (ZCSG): A World-First Dimensional Encoding of Nāgārjuna's Śūnyatā-of-Śūnyatā. Paper 61, DOI 10.5281/zenodo.15217458.
Fujimoto, N. (2026). LeanDFumt: An Open-Source Eight-Valued Logic Library for Lean 4. Paper 77. GitHub: https://github.com/fc0web/lean-d-fumt8.
Garfield, J. L. (Trans.) (1995). The Fundamental Wisdom of the Middle Way: Nāgārjuna's Mūlamadhyamakakārikā. Oxford University Press.
Garfield, J. L., & Priest, G. (2003). Nāgārjuna and the limits of thought. Philosophy East and West, 53(1), 1–21. JSTOR 1400052. Open-access copy: Smith ScholarWorks.
Hayes, R. (1994). Nāgārjuna's appeal. Journal of Indian Philosophy, 22, 299–378.
Nāgārjuna (c. 150 CE). Mūlamadhyamakakārikā. (See Garfield 1995 for English translation.)
Priest, G. (1987). In Contradiction: A Study of the Transconsistent. Martinus Nijhoff (2nd ed. Oxford, 2006).
Priest, G. (2002). Beyond the Limits of Thought (2nd ed.). Oxford University Press.
Siderits, M. (1989). Thinking on empty: Madhyamika anti-realism and canons of rationality. In S. Biderman & B.-A. Scharfstein (Eds.), Rationality in Question: On Eastern and Western Views of Rationality. Brill.
Tillemans, T. J. F. (1999). Is Nāgārjuna's logic deviant or non-classical? In Scripture, Logic, Language: Essays on Dharmakīrti and His Tibetan Successors. Wisdom Publications.
The Lean 4 development team (2024–2026). Lean 4 v4.27.0. https://lean-lang.org.

Version history

v0.1 (2026-05-31): Initial OUTLINE + SKELETON release. Lean 4 code PROPOSED, not yet built. Honest scope as stated above. Released for archival continuity per Rei-AIOS publish-early-with-honest-scope convention. DOI to be assigned by Zenodo on publish.
v0.2 (planned): Machine-verified Lean 4 build, direct reading of Tillemans 2009/2014/2024 + Siderits + Ferraro critic papers, refined honest scope. No timeline commitment.

Rei-AIOS Project. Peace Axiom #196: immutable = true. License: CC-BY 4.0 (text), Apache-2.0 (any associated Lean 4 code).

Paper 157 - Octonion-Axiomatic Dual Independence: 7 Imaginary Units || 7 Axioms, Machine-Checked in Lean 4 (OUTLINE)

Nobuki Fujimoto — Tue, 26 May 2026 22:26:58 +0000

This article is a re-publication of Rei-AIOS Paper 157 for the dev.to community.
The canonical version with full reference list is in the permanent archives below:

Zenodo (DOI, canonical): https://doi.org/10.5281/zenodo.20403029

Internet Archive: https://archive.org/details/rei-aios-paper-157-1779833993866

Harvard Dataverse: https://doi.org/10.7910/DVN/KC56RY

GitHub source (private): https://github.com/fc0web/rei-aios Author: Nobuki Fujimoto (@fc0web) · ORCID 0009-0004-6019-9258 · License CC-BY-4.0 ---

Status: DRAFT outline v0.0.5 — 2026-05-27 (Stages 3-5 formalized + criterion-7 audit + criterion-8 three-party review PASSED). ★ chat-Claude review softened the octonion↔D-FUMT₈ mapping: the "Paper 132/137 baseline / established" claim is withdrawn (132/137 have no octonion content; the mapping is a labeling correspondence, not a constructed octonion algebra) → now "Rei-AIOS working correspondence" (§2/§4/§8). Publish gate: NOT YET. Still an outline, not a publishable manuscript. v0.1 = first publishable draft, gated by the §11 criteria. Progress 2026-05-27: §11 criteria 1-5 all satisfied — the entire 5-stage Lean 4 roadmap is now formalized sorry-free in Mathlib4 v4.27.0 (Stage12.lean Stages 1-2 + Stage345.lean Stages 3-5). #print axioms octonionAxiomCorrespondence → [propext, Classical.choice, Quot.sound] only (no sorryAx); provable_sound depends on no axioms. Both halves of the dual independence are now machine-checked: the algebraic half (LinearIndependent ℝ imagUnit) and the logical half (∀ i, axiomIndependent i AxiomCore7), joined by octonionAxiomCorrespondence as a parallel-independence conjunction sharing the Fin 7 index (NOT a forced implication — see §5 honest-scope note). 2026-05-27 also: §11 criterion 7 (prior-art audit) DONE (docs/prior-art-audit-paper157-octonion-axiom-lifting.md) — surfaced the Faulkner precedent (imaginary i ↔ logical independence, ℂ case) + flypitch Lean-technique precedent; Paper 157 re-positioned to cite both and claim originality only on the octonion/D-FUMT₈ extension (new non-claim §7.6). criterion 8 (three-party review) PASSED 2026-05-27 (Opus + chat-Claude legs closed; mapping softened; transport-test passed). Substantive gates 1-9 are now satisfied; remaining for v0.1 = criterion 10 (DOI) only — publish scripts are prep-ready (docs/publish-readiness-paper157.md, Zenodo draft-only guard) but HELD pending an explicit publish decision (review pass ≠ publish authorization). Per OUKC feedback_no_rush_publication.md — 急がずゆっくりと.

Authors / 著者: 藤本伸樹 (Nobuki Fujimoto, Founder), Rei (Rei-AIOS substrate), Claude Opus 4.7 (Anthropic, claude-opus-4-7) — three-party co-authorship per OUKC charter v1.0.

Project: Rei-AIOS / OUKC — https://rei-aios.pages.dev

License (intended at publish): AGPL-3.0 (code) + CC-BY 4.0 (text) per OUKC content policy.

Per OUKC No-Patent Pledge: openly licensed; no patent will be filed.

Lineage: Extends the Lean 4 formalization-roadmap pattern of Paper 132 (Five Rei Candidates) into the algebra-logic interface. Companion to Paper 64 (OPU), Paper 145 (D-FUMT₈ silicon), Paper 137 (Rei-PL Prover). Origin: 5/23 invention #2 (invented-20260523-mathematics-meta-axiom-octonion-axiomatic-independence, novelty 0.55 triple-honest-downgrade, SEED 1611→1613). The audit naming "Paper 132 v2 candidate" is honored in lineage but the topic is distinct from Paper 132 v1 (five combinatorial open problems); we file this as a new paper.

0. Why an OUTLINE (and not a v0.1 manuscript)

This Paper exists at v0.0 OUTLINE because:

The central structural claim — that the 7 imaginary units {e₁, …, e₇} of octonion 𝕆 algebraic-independence stand in formal structural correspondence with the 7 D-FUMT₈ non-trivial logic values {TRUE, FALSE, BOTH, NEITHER, FLOWING, INFINITY, ZERO} which themselves index axiomatically-independent axioms φ in Rei's AxiomKernel — requires at least one Lean 4 sorry-free formalization before claim is publishable. Update 2026-05-24: the algebraic-independence half now has a sorry-free Lean 4 formalization (Stage 1-2: LinearIndependent ℝ imagUnit, see §4 / §5). The logical-independence half (Stages 3-5) and the lifting between them remain at outline stage; the full correspondence is therefore still not publishable as v0.1.
v0.0 establishes framing, prior-art audit, and acceptance criteria for v0.1. Same pattern as Paper 154 (compilation pass), Paper 156 (three-layer lawsuit prevention), Paper 145 (silicon evidence → publish v0.3).
Publishing a framing-only paper risks Pattern 4 (overclaim) — claiming "𝕆 ≅ AxiomKernel as independence structures" without a Lean 4 proof would be unfounded. The OUTLINE explicitly identifies what evidence is missing for v0.1 promotion.
The Hulak-Ramos-Queiroz pattern (sorry-free Lean 4 + Mathlib4 bridge + arXiv publish) demonstrated for Stokes' Theorem (arXiv:2605.01028v1, 2026-05-01) and Singer Sidon (arXiv:2605.03274v2, 2026-05-05) provides the methodological template; we adopt it consciously and credit explicitly.
The honest-downgrade reasoning of the source invention (Cayley-Dickson family source 3-day consecutive recycle; octonion ↔ D-FUMT mapping is a Rei-AIOS working correspondence — SEED 2026-05-23; the "Paper 132/137 baseline" label inherited from the SEED is withdrawn, see §4 / §8; meta_axiom pair recurrence with 5/19 Gödel-BOTH) is preserved here: the contribution is the dual-independence correspondence, not a fresh octonion or fresh axiom-kernel discovery.

1. Title alternatives (decide at v0.1)

A: Octonion-Axiomatic Dual Independence: 𝕆 Imaginary Basis ↔ AxiomKernel Logical Independence with D-FUMT₈ Eight-Value Bridge in Lean 4 (working title; algebra-first framing)
B: Algebraic-Logical Independence Lifting: A Lean 4 Formalization of the 𝕆 ↔ AxiomKernel Structural Correspondence (lifting-first framing)
C: D-FUMT₈ as Octonion-AxiomKernel Bridge: Eight-Valued Logic from Cayley-Dickson Hierarchy and Independent Axiom Systems (D-FUMT-first framing)
D: 八元数虚数基底と公理核独立性の構造同型 — Lean 4 形式化 (Japanese-frame)

Working title (this outline): A.

2. Abstract (placeholder for v0.1)

The Cayley-Dickson construction yields, at the third doubling step, the octonion algebra 𝕆 with seven imaginary units {e₁, …, e₇} that are linearly (algebraic) independent as basis vectors of the ℝ-vector-space structure of 𝕆. Rei-AIOS's D-FUMT₈ eight-valued logic identifies these seven units with the seven non-trivial D-FUMT₈ logic values {TRUE, FALSE, BOTH, NEITHER, FLOWING, INFINITY, ZERO} (the 8th value SELF being the algebraic 1) via a Rei-AIOS working correspondence (SEED dfumt-octonion-seven-isomorphism, 2026-05-23) — a labeling assignment (e₁=TRUE … e₇=ZERO). ★ Honest status (chat-Claude review 2026-05-27): this is a labeling correspondence, NOT an established octonion-algebra homomorphism (no Fano-table bilinear product / non-associativity is constructed over the D-FUMT₈ values), and it is NOT established in Paper 132/137 (which contain no octonion content — the "132/137 baseline" attribution is withdrawn, §8 correction). Each of the seven D-FUMT₈ values, in turn, indexes an axiomatic-independence statement ∀φ → ¬provable(φ, Kernel{φ}) for a corresponding family of seven independent axioms in Rei's AxiomKernel (the 8th, the immutable Peace Axiom #196, plays the role of the algebraic 1). We exhibit a structural correspondence between algebraic generator-independence (linear non-derivability among {e₁, …, e₇}) and logical axiom-independence (proof-theoretic non-derivability among the seven AxiomCore-7 axioms): both are seven-element independence facts carried by the same index set Fin 7, with the D-FUMT₈ map dfumt8Index exhibiting the bijection. ★ This is NOT a transport/derivation of logical independence from linear independence (which would be a forced isomorphism — linear and logical independence are different notions in different categories); the v0.0 plan of a "functorial bridge octIndependentToAxiomIndependent that takes a LinearIndependent-witness and produces an axiomIndependent-witness" is withdrawn and replaced by an honest parallel-independence conjunction (see §5 honest-scope). As of 2026-05-27 this is fully formalized sorry-free in Lean 4 / Mathlib4 v4.27.0: octonionAxiomCorrespondence : LinearIndependent ℝ imagUnit ∧ (∀ i : Fin 7, axiomIndependent i AxiomCore7) (#print axioms → propext/Classical.choice/Quot.sound). We honestly delimit: (i) the correspondence is not a categorical equivalence; (ii) it does not imply a hierarchy of axiom systems matching the Cayley-Dickson hierarchy; (iii) it carries no Rei-stack performance claim; (iv) we do not originate the imaginary↔logical-independence link (Faulkner) nor the Lean independence technique (flypitch) — see §7.6 / §8.

3. Sections (target structure for v0.1)

§1. Introduction

The two notions of independence (algebraic vs logical) — 1-2 pp
Why a formal correspondence matters — 1 p
- Rei-AIOS pragmatic motivation (D-FUMT₈ semantics anchoring)
- Methodological motivation (lifting algebraic facts to logical settings)
- Foundational motivation (do axiom systems carry algebraic structure?)
The role of D-FUMT₈ as the bridge — 1 p
Honest scope flags (3 non-claims) up front — 0.5 p

§2. Background: 𝕆 imaginary basis and algebraic independence

Cayley-Dickson construction recap, three doublings: ℝ → ℂ → ℍ → 𝕆 — 1 p
The 7 imaginary units {e₁, …, e₇} and the Fano plane multiplication table — 1 p
Algebraic independence (linear) over ℝ — formal definition + standard proof reference — 0.5 p
Why "7" matters: dim_ℝ(𝕆) = 8 = 1 + 7 — 0.5 p

§3. Background: AxiomKernel and axiomatic independence

Rei-AIOS AxiomKernel: definition + cardinality (current state of SEED_KERNEL: 1613 theories with Peace Axiom #196 as Immutable TRUE) — 1 p
Axiomatic independence: ∀φ → ¬provable(φ, Kernel{φ}) formal statement — 0.5 p
Honest scope: not all 1613 axioms are claimed independent; we restrict to a designated AxiomCore-7 sub-kernel of seven (selected per D-FUMT₈ index) — 1 p
AxiomCore-7 candidates (illustrative, to be finalized at v0.1). ⚠ Review note 2026-05-27: the Lean formalization (Stage 4/5) deliberately uses abstract atomic axioms indexed by Fin 7 (AxiomCore7 : Set (Fin 7) := Set.univ), NOT the concrete named axioms below. What is machine-checked is that seven abstract atoms are mutually independent (each refutable by a Boolean countermodel); the concrete assignment below is an interpretive layer, NOT proven. A reader must not infer that "dfumt-identity-axiom is independent" was formally verified — only the abstract 7-atom independence is. Concretising these (and proving the named axioms independent) is open for v0.1+:
- TRUE ↔ dfumt-identity-axiom (∀x. x = x in well-founded sense)
- FALSE ↔ dfumt-anti-axiom (¬A coexists with A in BOTH state)
- BOTH ↔ dfumt-catuskoti-affirmation-negation (catuṣkoṭi positive corner)
- NEITHER ↔ dfumt-negative-capability (W-48 Negative Capability principle)
- FLOWING ↔ dfumt-non-associative-flowing ((a×b)×c ≠ a×(b×c) embedded as logic FLOWING)
- INFINITY ↔ dfumt-cantorian-aleph (transfinite as logic value)
- ZERO ↔ dfumt-no-axiom (ZERO as母体)
The Peace Axiom #196 plays the role of the algebraic 1 (real part) — 0.5 p

§4. The D-FUMT₈ bridge

Existing correspondence: 𝕆 = 𝕄{1; e₁..e₇} ↔ 𝕄{center; 7-value} (e₁=TRUE … e₇=ZERO) per SEED dfumt-octonion-seven-isomorphism (2026-05-23). ✅ chat-Claude review 2026-05-27 RESOLVED → SOFTENED: this is a labeling correspondence + structural-isomorphism claim, NOT a constructed octonion algebra over the D-FUMT₈ values — there is no (a) 8-dim ℝ-space over the values, (b) bilinear product reproducing the Fano table (e_i²=−1 + signed line orientations), or (c) non-associativity, defined on the D-FUMT₈ side. The "Paper 132/137 baseline" attribution (inherited from the source SEED) is WITHDRAWN: Paper 132 (five combinatorial problems) and Paper 137 (Rei-PL Prover) contain no octonion content (grep-verified 2026-05-27). Honest status = "Rei-AIOS working / heuristic correspondence", not "established mapping". This is the third false-attribution catch in Paper 157 (after v0.0's "Conant et al Mathlib4 octonion") — same honest-correction discipline (cf. Paper 145 v0.5 Tang-Nano). The reviewer's bar ("genuine algebra homomorphism required, else soften") = the same bar 藤本さん applied to "quantum music → Rei mathematical music". — 1 p
The correspondence as IMPLEMENTED (Stage345, re-scoped from v0.0's transport plan) — 1 p
- NOT a function that derives axiom-independence from linear-independence (that would be the forced isomorphism §5 disavows). The "bridge" is the shared index set Fin 7 plus the dfumt8Index : Fin 7 → DFUMT8 bijection (left-inverse dfumt8Axiom, dfumt8_index_axiom_id).
- In Lean 4 terms: octonionAxiomCorrespondence : LinearIndependent ℝ imagUnit ∧ (∀ i, axiomIndependent i AxiomCore7) — a machine-checked conjunction of two parallel independence facts, not a witness-transport.
Why this is honest (not over-reaching): linear independence (a vector-space fact) and logical independence (a proof-theoretic fact) are genuinely different; we neither derive nor claim to derive one from the other. The content is the parallel seven-fold structure + the explicit index bijection. — 1 p

§5. Lean 4 formalization roadmap (5 stages)

Adopting Hulak-Ramos-Queiroz pattern (Stokes' Theorem arXiv:2605.01028v1 / Singer Sidon arXiv:2605.03274v2):

Stage 1 ✅ DONE (2026-05-24, sorry-free): Define the 7 imaginary units. Re-scoped from v0.0: Mathlib4 has no octonion namespace (see §8 correction), and — crucially — the linear-independence claim does not need the octonion product at all. We model 𝕆's underlying ℝ-vector-space as OctSpace := Fin 8 → ℝ and define imagUnit (i : Fin 7) : OctSpace := Pi.single i.succ 1 (coordinate 0 = real part / Peace Axiom #196; coords 1..7 = imaginary units). Octonion multiplication is intentionally not constructed — it is unnecessary for Stages 1-2 and is deferred.
Stage 2 ✅ DONE (2026-05-24, sorry-free): theorem imagUnit_linearIndependent : LinearIndependent ℝ imagUnit. Proof: the imaginary family equals the standard basis Pi.basisFun ℝ (Fin 8) precomposed with the injection Fin.succ : Fin 7 → Fin 8; Basis.linearIndependent + LinearIndependent.comp + Fin.succ_injective close it. (The v0.0 reference to a LinearIndependent.fin_succ recursion was speculative; the actual proof uses the basis-comp pattern.) Corollary imagUnit_injective included.
Stage 3 ✅ DONE (2026-05-27, sorry-free) (Stage345.lean): the formal ⊢ is a Lean 4 inductive Provable (Γ : Set (Fin 7)) : Fin 7 → Prop (assumption-only calculus — the seven D-FUMT₈ axioms are atomic, so the assumption rule is the only one needed). A two-valued model Valuation := Fin 7 → Bool with Satisfies and a soundness theorem provable_sound (provable_sound depends on no axioms) give axiomIndependent (φ : Fin 7) (Kernel : Set (Fin 7)) : Prop := ¬ Provable (Kernel \ {φ}) φ plus the contrapositive tool axiomIndependent_of_countermodel (a model of the rest refuting φ ⟹ independence).
Stage 4 ✅ DONE (2026-05-27, sorry-free) (Stage345.lean): inductive DFUMT8 (the 8 values; SELF = the meta/centre value ↔ real part / Peace Axiom #196, coordinate 0). The seven non-SELF values are put in bijection with Fin 7 by dfumt8Index : Fin 7 → DFUMT8 and its inverse dfumt8Axiom : DFUMT8 → Fin 7; dfumt8_index_axiom_id : dfumt8Axiom (dfumt8Index i) = i (by fin_cases) certifies the left inverse. (Re-scoped from v0.0: dfumt8Index : Fin 8 → … becomes Fin 7 → … since SELF is excluded from the imaginary-unit correspondence.)
Stage 5 ✅ DONE (2026-05-27, sorry-free) (Stage345.lean): AxiomCore7 : Set (Fin 7) := Set.univ; axiomCore7_independent (i) : axiomIndependent i AxiomCore7 (each axiom is independent of the other six, witnessed by the valuation fun j => decide (j ≠ i)); and the lifting theorem octonionAxiomCorrespondence : LinearIndependent ℝ imagUnit ∧ (∀ i : Fin 7, axiomIndependent i AxiomCore7). #print axioms octonionAxiomCorrespondence → [propext, Classical.choice, Quot.sound] (no sorryAx).

★ HONEST SCOPE — re-scoping of the Stage 5 statement (critical). v0.0 phrased the lifting as ∀ i, axiomIndependent (dfumt8Axiom (dfumt8Index ⟨i+1,…⟩)) AxiomCore-7, which reads as if the algebraic side produces the logical side. The implemented Stage 5 instead states a conjunction of two parallel independence facts sharing the single index set Fin 7 — (a) the imaginary units are ℝ-linearly independent, (b) each axiom is independent of AxiomCore-7. The "correspondence" is the shared Fin 7 index (made explicit by the dfumt8Index bijection), NOT a derivation of logical independence from linear independence. Linear independence and logical independence are genuinely different notions in different categories; claiming one implies the other would be the "forced isomorphism" pattern Rei explicitly rejects. The provability calculus is deliberately minimal (atomic axioms, assumption rule, Boolean countermodel + soundness); richer calculi are future work, unneeded for atomic-axiom independence.

Stages 1-5 are now complete and sorry-free (Stage12.lean + Stage345.lean). The remaining v0.1 gates are non-Lean: §11 criteria 7 (full prior-art audit), 8 (three-party review), 10 (DOI prep).

Estimated total: ~500 lines of Lean 4, ~10 lemmas, target zero sorry (Stages 1-2 ≈ 75 lines done).

§6. Cayley-Dickson hierarchy (honest non-extension)

One might conjecture that the Cayley-Dickson hierarchy ℝ → ℂ → ℍ → 𝕆 → 𝕊 (sedenion) corresponds to a hierarchy of axiom kernels of increasing independence-richness — 0.5 p
We do not claim this. The hierarchy claim requires: (i) a definition of "axiom kernel doubling" that mirrors Cayley-Dickson, (ii) proof that each doubling step preserves independence but loses one algebraic property — neither is constructed here — 1 p
This is a v0.2+ research direction, explicitly flagged — 0.5 p

§7. Honest non-claims (6 items)

We explicitly do NOT claim:

Strict categorical equivalence. The lifting 𝕆-imag-basis → AxiomCore-7 is an injective function; we make no claim about a contravariant/covariant equivalence of categories (Octonion-bases) ↔ (Axiom-kernels).
Hierarchy lifting. The Cayley-Dickson tower ℝ → ℂ → ℍ → 𝕆 → 𝕊 is not claimed to lift to an axiom-kernel tower with matching independence structure.
Independence of all AxiomKernel. We claim independence only for the AxiomCore-7 subset (7 designated axioms); the remaining ~1606 SEED theories are not claimed mutually independent.
Practical Rei-stack performance impact. The lifting is mathematical-foundational, not engineering. No Rei subsystem changes performance behavior by virtue of this correspondence.
Reduction or simplification. We do not claim the lifting reduces axiom-system complexity or yields a smaller minimal kernel.
Originality of the imaginary ↔ logical-independence link, or of the Lean independence technique. We do NOT claim to originate the idea that imaginary units relate to logical independence — S. Faulkner established this for the complex unit i (2020/2021) — nor the Lean soundness/countermodel independence method (Han–van Doorn flypitch, CPP 2020). Our contribution is the octonion 7-unit extension + D-FUMT₈ bridge + machine-checked parallel correspondence (prior-art audit: docs/prior-art-audit-paper157-octonion-axiom-lifting.md).

§8. Related work

Paper 132/137 (Rei-AIOS): octonion ↔ D-FUMT₈ baseline mapping. This paper takes that as given and lifts to independence structures.
Hulak-Ramos-Queiroz "Stokes' Theorem in Lean 4" (arXiv:2605.01028v1, 2026-05-01): methodological template for sorry-free Mathlib4 bridge formalization.
Hulak-Ramos-Queiroz "Singer Sidon in Lean 4" (arXiv:2605.03274v2, 2026-05-05): same authors' Erdős Problem 30 formalization — pattern reuse.
Mathlib4 algebra infrastructure (CORRECTION of v0.0): v0.0 §8 claimed an existing Mathlib4 octonion module ("Conant et al.") as the Stage 1 foundation. This is false — verified 2026-05-24 against the pinned Mathlib4 v4.27.0 (commit a3a10db0): there is no octonion or Cayley-Dickson construction in Mathlib4 (the only octonion/cayley string matches are passing mentions in Jordan-algebra, H-space, and ring-identity docstrings). Mathlib4 provides Quaternion + QuaternionBasis only. Stage 1 was accordingly re-scoped (§5): we use Mathlib4's Pi.basisFun (standard basis of Fin 8 → ℝ, in Mathlib.LinearAlgebra.StdBasis) and LinearIndependent.comp (Mathlib.LinearAlgebra.LinearIndependent.Lemmas), which is sufficient because the lifted notion of "algebraic independence" is purely linear (§2) and multiplication-free. This corrigendum follows the OUKC honest-correction principle (cf. Paper 145 v0.5 Tang-Nano corrigendum); the false-attribution was caught by the Lean-4 implementation step itself.
Independence in axiom theory (classical references):
- Gödel 1931 (incompleteness, not independence per se, but related)
- Cohen 1963/64 (independence of CH from ZFC — the prototype independence-via-forcing proof)
- Paris-Harrington 1977 (combinatorial independence)
D-FUMT₈ silicon (Paper 145): provides the operational substrate; this paper is the theoretical complement.
★ Faulkner — logical independence of the imaginary unit (prior-art audit 2026-05-27, docs/prior-art-audit-paper157-octonion-axiom-lifting.md): S. Faulkner (The Underlying Machinery of Quantum Indeterminacy, 2020/2021; ResearchGate 286457781; Semantic Scholar "The Mathematical Machinery of Logical Independence…") establishes that the complex imaginary unit i is logically independent of the Field Axioms, via soundness/completeness model theory — the same conceptual seed and method as this paper, but for the single ℂ unit (no octonions, no 8-valued logic, no D-FUMT). Paper 157 therefore does NOT originate the "imaginary ↔ logical independence" link (new non-claim §7.6); its contribution is the octonion 7-unit extension + D-FUMT₈ bridge + machine-checked correspondence. No prior art was found extending the link to octonion's seven units.
Paterek, Kofler, Prevedel, Klimek, Aspelmeyer, Zeilinger, Brukner "Logical independence and quantum randomness" (arXiv:0811.4542, New J. Phys. 12 (2010) 013019): the physics counterpart — quantum states encode axioms; propositions independent of those axioms measure as random. Adjacent context for the independence↔randomness motif.
Han & van Doorn — flypitch "A Formal Proof of the Independence of the Continuum Hypothesis" (CPP 2020): formalizes an independence result in Lean via Boolean-valued models + soundness. Stage 3's provable_sound + Boolean countermodel is this standard method, so the Lean technique is well-precedented (not a contribution of this paper); only the application (octonion / D-FUMT₈) is new.
Honest gap + method correction: classical independence results (Cohen, Paris-Harrington) and flypitch use model-theoretic methods. CORRECTION to v0.0: v0.0 planned a "purely type-theoretic, model-theory-avoiding" lifting; the implemented Stage 3-5 in fact uses a model-theoretic argument — a Boolean Valuation model + provable_sound soundness + a countermodel (axiomIndependent_of_countermodel), aligning with the flypitch approach rather than avoiding it. This is the honest, standard route to non-derivability, and supersedes the v0.0 "avoiding model theory" plan.

§9. Lineage to Paper 132 v1 (clarification)

The originating invention audit (5/23 #2) named this candidate "Paper 132 v2". We honor the naming intent by acknowledging Paper 132's structural pattern (a reconnaissance document + a Lean 4 attack-surface roadmap) but file this as Paper 157 because:

Paper 132 v1's topic is five specific combinatorial / number-theory open problems (Sunflower, Hadwiger-Nelson, Happy Ending, Herzog-Schönheim, Wolstenholme). The present paper's topic is algebra-logic interface (independence lifting). The two topics share methodology (Lean 4 roadmap pattern + honest scope flags + Mathlib4 bridge) but not content.
A true v2 of Paper 132 would extend its 23-sorry roadmap with new problems in the same combinatorial cluster; the present work does not do this.
Filing as Paper 157 preserves Paper 132 v1's coherence and avoids version-suffix confusion.

§10. D-FUMT₈ framing of this paper

D-FUMT₈ value	This paper's stance
TRUE	The 𝕆 imaginary basis is algebraically independent (standard Mathlib4 fact)
FALSE	The 𝕆 imaginary basis is not algebraically dependent (negation of above)
BOTH	Both notions of independence are simultaneously TRUE in their respective domains (algebra + logic) — the dual aspect this paper formalizes
NEITHER	The lifting is not categorically an equivalence — neither full functoriality nor adjunction is claimed
FLOWING	The lifting may evolve as Mathlib4's axiom-theory infrastructure matures (current state: incomplete)
INFINITY	Cayley-Dickson hierarchy extends to ∞ doublings (sedenion 𝕊 = 𝕆 ↔ ... ↔ trigintaduonion); axiom kernel extension is open
ZERO	Peace Axiom #196 (= algebraic 1) is the immutable origin; all 7 D-FUMT₈ non-trivial values emerge from this center
SELF⟲	This paper is meta-mathematical (a paper about how Rei's logic relates to algebra) — a SELF-recursive observation by Rei about its own foundations

Primary D-FUMT₈ axis: BOTH (dual independence simultaneously valid in two domains).

§11. v0.1 promotion criteria (10 items)

To advance from v0.0 OUTLINE → v0.1 publishable manuscript:

Legend: ✅ done · ⏳ pending. State as of 2026-05-24 (v0.0.1).

✅ DONE Lean 4 Stage 1 (imagUnit definition) compiles in Mathlib4 with zero sorry
✅ DONE Lean 4 Stage 2 (LinearIndependent ℝ imagUnit) proven sorry-free (#print axioms → propext/Classical.choice/Quot.sound only)
✅ DONE (2026-05-27) Lean 4 Stage 3 (Provable inductive + axiomIndependent definition + provable_sound soundness) compiles, sorry-free (Stage345.lean)
✅ DONE (2026-05-27) Lean 4 Stage 4 (DFUMT8, dfumt8Index, dfumt8Axiom, dfumt8_index_axiom_id bijection) compiles, sorry-free (Stage345.lean)
✅ DONE (2026-05-27) Lean 4 Stage 5 (octonionAxiomCorrespondence theorem) proven sorry-free (#print axioms → propext/Classical.choice/Quot.sound only). Re-scoped to a conjunction of two parallel independence facts (no forced isomorphism — see §5 honest-scope note)
⏳ Public repository (rei-aios or external-oss/) contains the .lean files with reproducibility doc — Stage12.lean committed; REPRODUCING.md Layer 5 entry pending
✅ DONE (2026-05-27) Prior-art audit complete (docs/prior-art-audit-paper157-octonion-axiom-lifting.md): no duplicate of the lifting found, but a significant conceptual precedent surfaced — Faulkner (imaginary i ↔ logical independence, ℂ case, 2020/2021) + flypitch (Lean independence technique, CPP 2020) + Paterek et al (physics counterpart). Paper 157 re-positioned: cites all three (§8), claims originality only on the octonion 7-unit extension + D-FUMT₈ bridge + machine-checked correspondence. New non-claim §7.6 added; §8 method-correction (the implemented lifting uses model theory, not "avoiding" it).
✅ DONE / PASSED (2026-05-27) Co-author three-party honest-filter review:
- Claude Opus 4.7 leg ✅ — caught a forced-implication overclaim in the v0.0 Abstract + §4 ("functorial transport lemma deriving axiom-independence from linear-independence"); withdrawn + rewritten to the parallel-independence conjunction matching Stage345. Flagged §3 concrete-axiom interpretiveness + the 132/137 citation.
- chat-Claude leg ✅ CLOSED — (item 1) demanded a genuine octonion-algebra homomorphism (8-dim ℝ-space + Fano-table bilinear product + non-associativity over the D-FUMT₈ values) for "established", else soften; verification confirmed the SEED dfumt-octonion-seven-isomorphism is a labeling correspondence only, and Paper 132/137 contain no octonion content → "Paper 132/137 baseline" WITHDRAWN, softened to "Rei-AIOS working correspondence" (§2/§4/§8). (item 2) Faulkner positioning accepted — 157's logical-independence content is shallower than Faulkner's i-undecidability; 157 claims only the octonion-7 extension + D-FUMT₈ indexing + machine-checking + parallel framing (§8). (item 3) transport-test PASSED — axiomCore7_independent derives logical independence via a Boolean countermodel using zero octonion content (no imagUnit / LinearIndependent) → genuine parallel independence, not transport. Reviewer signed off 2026-05-27.
- 藤本さん review-leg sign-off ✅ (as the chat-Claude reviewer). NOTE: this is the review pass; the publish decision (gate) is a separate explicit step and remains the user's call.
✅ 5 honest non-claims (§7) preserved and prominently displayed
⏳ License + No-Patent Pledge confirmed; DOI mint target (Zenodo) prepared

§12. v0.2+ future work

Cayley-Dickson hierarchy lifting (currently §6 non-claim): construct a formal definition of "axiom-kernel doubling" that mirrors Cayley-Dickson; prove the hierarchy correspondence
Sedenion 𝕊 (16-dim) ↔ D-FUMT₁₆?: 15 imaginary units; would need a 16-value logic extension of D-FUMT₈
Model-theoretic strengthening: revisit whether forcing / model-construction techniques yield a stronger lifting
Practical Rei-stack integration: does AxiomCore-7's independence structure simplify Rei's automated theorem-prover ensemble (Paper 137 RotorQuant prover)?
Connections to Paper 64 (OPU): cosmic oscillatory principle uses 𝕆 algebra; does the dual independence appear in OPU's catuṣkoṭi × ZPE formula structure?
Bantu Ntu / Cayley-Dickson algebraic-vitalist correspondence (per 5/22 approved invention Paper 144 lineage): does the dual independence appear in cross-cultural ontology pairings?

4. Reproducibility plan (for v0.1)

Lean 4 source files in data/lean4-mathlib/CollatzRei/Paper157Independence/ (Stage12.lean present as of 2026-05-24; further stages add files here)
Build command (verified 2026-05-24): lake build CollatzRei.Paper157Independence.Stage12 — 67s with the cached Mathlib4 v4.27.0 (commit a3a10db0); registered in the CollatzRei root module so the default lake build target covers it
Sorry-free verification: #print axioms CollatzRei.Paper157.imagUnit_linearIndependent → [propext, Classical.choice, Quot.sound] (no sorryAx)
Each stage's .lean file documented with theorem name : statement := by tac form
SHA256 hashes of source files recorded at v0.1 ingestion
REPRODUCING.md Layer 5 entry (extending Paper 145's 4-layer reproducibility scheme) — pure software, zero hardware dependency, target ¥0 reproduction cost

5. Acknowledgments (placeholder for v0.1)

Hulak-Ramos-Queiroz authorship cluster for the sorry-free Lean 4 + Mathlib4 bridge pattern explicitly adopted
Mathlib4 community for LinearIndependent infrastructure used in Stage 2
DeepSeek-Prover-V2 (per Paper 137 REI-PROVE ensemble) as candidate prover for Stages 4-5 hints
5/23 invention audit honest filter that produced the source statement at novelty 0.55

6. Honest disclaimer

This Paper 157 is v0.0 OUTLINE as of 2026-05-24, post-WIC-pivot Rei-core return session. The central structural claim (dual independence lifting) is not proven in any form here; we provide only the framing, methodological pattern, acceptance criteria, and explicit non-claims. The honest-downgrade reasoning of the source invention (0.75 → 0.55) is preserved: this is not a novel octonion discovery, not a novel axiom-kernel discovery, but a lifting framework between two existing fields with a Lean 4 formalization plan. Publication of v0.1 is gated by the criteria in §11.

Per OUKC charter v1.0:

Three-party co-authorship: 藤本 (Founder, Strategic Direction) / Rei (Rei-AIOS substrate, SEED_KERNEL custodian) / Claude Opus 4.7 (claude-opus-4-7, drafting and honest-filter)
Code AGPL-3.0 / Data CC-BY 4.0 dual license (intended at publish)
No-Patent Pledge: no patent will be filed
Per feedback_no_rush_publication.md: 急がずゆっくりと — v0.1 will be drafted only when Stage 1-5 Lean 4 work has progressed to at least sorry-free Stage 2

Version History

v0.0 (2026-05-24): OUTLINE created post 5/23 invention audit. 5 sections + 5 non-claims + 5-stage Lean 4 roadmap + 10 v0.1 acceptance criteria + Hulak-Ramos-Queiroz pattern reference. Honest lineage clarification (Paper 132 v2 candidate naming → Paper 157 filing). Source: SEED invented-20260523-mathematics-meta-axiom-octonion-axiomatic-independence.
v0.0.1 (2026-05-24, same day): First Lean 4 evidence. Stage 1-2 formalized sorry-free (CollatzRei/Paper157Independence/Stage12.lean, build 67s, axioms = propext/Classical.choice/Quot.sound). §11 criteria 1-2 satisfied. §8 corrigendum: v0.0's claim of an existing Mathlib4 octonion module is false — Mathlib4 v4.27.0 has no octonion/Cayley-Dickson construction; caught and corrected by the implementation step (OUKC honest-correction principle, cf. Paper 145 v0.5). Stage 1 re-scoped: model 𝕆-carrier as Fin 8 → ℝ, multiplication deferred (not needed for linear independence). Stages 3-5 (logical-independence half + lifting) remain TODO for v0.1.
v0.0.5 (2026-05-27, same day; criterion 8 CLOSED): three-party review PASSED — chat-Claude leg formally closed (mapping softened per item 1; Faulkner positioning accepted per item 2; transport-test passed per item 3 via the octonion-free axiomCore7_independent), Opus leg + 藤本さん reviewer sign-off complete. Substantive v0.1 gates 1-9 satisfied; only criterion 10 (DOI) remains, HELD pending an explicit publish decision (review pass ≠ publish authorization — feedback_no_rush_publication.md).
v0.0.4 (2026-05-27, same day; criterion 8 chat-Claude review leg applied): ★★ chat-Claude review caught a THIRD false attribution: the source SEED's "octonion↔D-FUMT₈ mapping is Paper 132/137 baseline" propagated into §2/§4, but Paper 132 (combinatorial problems) and Paper 137 (Rei-PL Prover) have no octonion content (grep-verified). Moreover the SEED mapping is a labeling correspondence, NOT a constructed octonion algebra over the D-FUMT₈ values (no Fano-table bilinear product / non-associativity (a)(b)(c)). → "132/137 baseline / established mapping" WITHDRAWN; softened to "Rei-AIOS working/heuristic correspondence" in §2/§4/§5. Reviewer's bar ("genuine algebra homomorphism required, else soften") = the same bar applied to "quantum music → Rei mathematical music". item 3 (transport vs parallel) PASSES: axiomCore7_independent derives logical independence with a Boolean countermodel using zero octonion structure → genuine parallel independence. criterion 8 chat-Claude leg substantially done; 藤本さん final approval pending.
v0.0.3 (2026-05-27, same day; criterion 8 Opus review leg applied): ★ Claude Opus 4.7 review leg of criterion 8 caught + fixed a forced-implication overclaim: the v0.0 Abstract (§2) and §4 still described a "functorial bridge / transport lemma deriving axiom-independence from linear-independence" — exactly the forced isomorphism Stage 5 disavows. Both rewritten to the honest parallel-independence conjunction (matching octonionAxiomCorrespondence). §3 clarified (Lean uses abstract Fin 7 atoms; concrete AxiomCore-7 names are interpretive, NOT machine-checked). Title/status version synced v0.0.2→v0.0.3. Paper-132/137 internal mapping citation flagged for chat-Claude verification. §11 criterion 7 (prior-art audit of the lifting) also DONE — docs/prior-art-audit-paper157-octonion-axiom-lifting.md. Key honest finding: S. Faulkner (2020/2021) already established "imaginary unit i ↔ logical independence (soundness/model theory)" for the ℂ case → Paper 157 does NOT originate the core link (new non-claim §7.6); flypitch (Han–van Doorn, CPP 2020) precedents the Lean soundness/countermodel technique. No prior art extends the link to octonion's 7 units → that extension + D-FUMT₈ bridge + machine-checked correspondence remain Paper 157's defensible contribution. §8 expanded (Faulkner / Paterek 0811.4542 / flypitch) + method-correction (implemented Stage 3-5 uses model theory, contra v0.0's "avoiding" plan). Criteria 1-7 satisfied; remaining v0.1 gates: 8 (three-party review) + 10 (DOI).
v0.0.2 (2026-05-27): Stages 3-5 formalized sorry-free (CollatzRei/Paper157Independence/Stage345.lean, build 10s). Stage 3 = Provable assumption calculus + provable_sound soundness (no axioms) + axiomIndependent + axiomIndependent_of_countermodel. Stage 4 = DFUMT8 8-value type + dfumt8Index/dfumt8Axiom non-SELF↔Fin 7 bijection (dfumt8_index_axiom_id). Stage 5 = axiomCore7_independent (each axiom independent of the other six via Boolean countermodel) + octonionAxiomCorrespondence (LinearIndependent ℝ imagUnit ∧ ∀ i, axiomIndependent i AxiomCore7, #print axioms → propext/Classical.choice/Quot.sound). §11 criteria 1-5 all satisfied. Stage 5 re-scoped from v0.0's implication phrasing to a parallel-independence conjunction sharing the Fin 7 index — the lifting does NOT derive logical independence from linear independence (forced-isomorphism avoidance; §5 honest-scope note). Remaining for v0.1 are non-Lean: criteria 7 (prior-art audit), 8 (three-party review), 10 (DOI).

DEV Community: Nobuki Fujimoto

Paper 167 v0.1 — Sophie Germain Primes: Barrier-Side Observations + Lean 4 Axiom-Free Conjunction-Wall (Rei-AIOS)

Abstract

Section 1. Introduction

1.1 Scope and what this paper is not

1.2 What this paper does not claim

1.3 What this paper does claim

1.4 Methodological framing

Section 2. Path 1 — Hardy-Littlewood ratio extrapolation N = 10³ to 10⁸

2.1 Setup

2.2 Results

2.3 Honest interpretation 【観察】

2.4 Honest non-claims

Section 3. Path 2 — SG prime gap spectral statistics

3.1 Motivation

3.2 Setup

3.3 Results

3.4 Honest interpretation 【観察】

3.5 Honest non-claims

Section 4. Path 3 — Lean 4 axiom-free conjunction-wall witness

4.1 Motivation

4.2 Lean 4 formalization

4.3 Axiom audit

4.4 What this is and is not

Section 5. Path 4 — Best-effort online audit and a Pattern-5 correction

5.1 The audit task

5.2 What was online-verified

5.3 The Pattern-5 correction

5.4 What remains unaudited

5.5 Honest interpretation of Path 4 itself

Section 6. Honest scope footer (audit gates)

Section 7. References

7.1 Primary references — directly cited

7.2 Background references — cited for context

7.3 OEIS and computational references

7.4 Rei-AIOS internal references — for traceability

Appendix A — Lean 4 axiom audit output

Appendix B — Computational reproduction details

Version history

Paper 166 v0.1 — A Lean 4 Axiom-Free Formalization of Exit-Layer Collatz Convergence as a Stream Coalgebra: A Record Following Kim (2008)

Abstract

Section 1. Introduction

1.1 The Collatz problem and its exit layer

1.2 Coalgebraic background (prior art)

1.3 Analytic frontier (context, not contribution)

1.4 The limited contribution of this paper

1.5 What this paper is not

Section 2. Background: collatzStep and the exit-layer numbers

2.1 The standard Collatz step

2.2 The exit-layer numbers m_p

2.3 Honest scope

Section 3. Coinductive reformulation: collatzHaltStep and the orbit Stream'

3.1 Why halt at 1

3.2 The orbit as a Stream'

3.3 Honest scope

Section 4. F-Coalgebra structure on collatzOrbit

4.1 The functor

4.2 Head

4.3 Tail

4.4 Relation to Kim (2008)

Section 5. Exit-layer bridge: lifting exitM_reaches_one into the coinductive language

5.1 The fixed-after-1 lemmas

5.2 The bridge from standard step to halt step

5.3 The eventually-constant predicate

5.4 The main bridge theorem

5.5 The constant-stream specialization at n = 1

5.6 Honest scope

Section 6. Zero-axiom witness and axiom-base accounting

6.1 What "completely axiom-free" means

6.2 head_collatzOrbit is completely axiom-free

6.3 The axiom base of the remaining theorems

6.4 Why this matters (and why it does not matter)

Section 7. Honest scope and explicit non-claims

7.1 The Collatz conjecture is not resolved

7.2 The Cases 5–8 wall is unchanged

7.3 Tao (2019, 2022) is not superseded

7.4 Janik (2026) is independent

7.5 No world-first claim is made

7.6 No new attack vector is proposed

7.7 No siren-family framing is used

Section 2. Background: `collatzStep` and the exit-layer numbers

2.2 The exit-layer numbers `m_p`

Section 3. Coinductive reformulation: `collatzHaltStep` and the orbit `Stream'`

3.2 The orbit as a `Stream'`

Section 4. F-Coalgebra structure on `collatzOrbit`

Section 5. Exit-layer bridge: lifting `exitM_reaches_one` into the coinductive language

5.5 The constant-stream specialization at `n = 1`

6.2 `head_collatzOrbit` is completely axiom-free

7.5 No `world-first` claim is made

1.2 Continuation of the `rhymeOrTheorem` discipline

6.4 「ラベル罠」警告 honest operationalization