ZFCρ Paper LXVIII: The {2,3}-Skeleton and the Minimum Crossing Mechanism — From 18 Kills to the Irreducibility of the Discrete
ZFCρ 第LXVIII篇：{2,3}-骨架与最小跨越机制——从18次否决到离散不可替代性

§1. Status After Paper 67

Paper 67 (DOI: 10.5281/zenodo.19705713) established four facts:

1. C_eff(W) starts near 0.3 and eventually increases, crossing the critical threshold 1.0 at all j = 22–32.
2. When plotted against W/m_j, the C_eff trajectory collapses across scales, indicating a block-relative mechanism.
3. Proposition 67.A provides an exact log-budget identity: log Ψ(m_j) = log Ψ(W₀) + Σ(1 − C_r)·Δ_r.
4. Corollary 67.B: if the cumulative budget is O(1) and Ψ(W₀) = O(j), then Ψ_full = O(j), yielding 59.1.

Paper 67 identified the empirical signature. Paper 68 asks: what microscopic mechanism within the IC recursion produces this signature?

---

§2. Too Complex: Kill 16

2.1. Multiplicity Is Not the Driver

The simplified recursion ρ_simple(n) = min(ρ(n−1)+1, ρ(n/p_min)+1), where p_min is the smallest prime factor of n, allows exactly one factorization per composite (M_ε = 1 by construction).

Result: ρ_simple exhibits C_eff crossing 1.0 at all j = 22–26, with the same two-regime structure as full IC.

Kill 16: near-optimal factorization multiplicity is not necessary for C_eff crossing. The mechanism survives radical simplification.

2.2. The Coarseness Distribution

For each multiplicative event in full IC, the winning factorization a·b = p−1 was identified and its coarseness χ = log(min(a,b))/log(p−1) computed.

The distribution is remarkably concentrated: 50.3% of winning factorizations use factor 2 (χ < 0.05), 45.6% use factor 3 or a small prime (χ in [0.05, 0.10)), 4.0% use intermediate factors, and 0.0% use balanced splits (χ near 0.5). This distribution is stable across j = 25, 27, 29 (mean χ = 0.064, 0.060, 0.056).

Balanced factorizations never win because their smooth baseline cost exceeds the add path by approximately 1 full unit, and O(1) fluctuations cannot compensate.

Full IC's winning factorizations fall in the very-small-prime range (χ < 0.10) in 96% of cases, dominated by factor 2 (50.3%) and factor 3 (9.1%), with remaining small primes making up the rest.

2.3. Per-Event Decorrelation Is Gentle

The conditional lag-1 correlation is 0.191 following an add event and 0.176 following a mult event (ratio 0.92). Each individual mult event reduces correlation by only 8%, with no significant scale dependence within blocks.

Naive compounding (0.92^{0.74W}) would predict exponential cutoff, which Paper 66 killed. Successive mult events are correlated through shared small-factor structure, producing power-law rather than exponential decorrelation.

---

§3. Too Simple: Kills 17–18

3.1. Kill 17: Pure Halving Is Not Sufficient

The recursion ρ_halve(n) = min(ρ(n−1)+1, ρ(n/2)+1) for even n, with ρ(n−1)+1 for odd n, was computed to N = 3×10⁸.

Result: C_eff never crosses 1.0 at j = 25–27 (maximum reliable C_eff ≈ 0.77). The cumulative budget grows systematically from 5.1 to 6.3. Ψ/j at the largest windows reaches 337–1016, far exceeding O(j) behavior.

The lag-1 autocorrelation γ(1) = 0.71, compared with 0.18 for full IC, indicates that halving provides dramatically insufficient decorrelation. The map n → n/2 sends the system to a structurally nearby region (ρ gap ≈ λ_halve · ln 2 ≈ 1.5).

Kill 17: factor 2 alone does not provide sufficient decorrelation for C_eff crossing.

λ_halve = 2.154, and ρ_halve(2^k) = k+1 exactly, confirming the connection to binary digit structure.

3.2. Kill 18: {2,5} Is Not Sufficient

The recursion ρ_{2,5} adds factor-5 access (n/5 path when 5|n) but omits factor 3.

Result: C_eff does not cross at j = 25–27 (maximum ≈ 0.77). Budget grows. Factor 5 cannot substitute for factor 3.

Factor 3 covers 1/3 of odd integers; factor 5 covers only 1/5 — a 67% density disadvantage. Moreover, factor 3 has a smooth baseline 0.24 units more favorable than add, while factor 5's baseline is less favorable. Factor 3 wins at more positions with higher probability.

Kill 18: among tested two-prime skeletons, the crossing mechanism requires factor 3 specifically; factor 5 cannot substitute.

---

§4. The Goldilocks Skeleton: {2,3}

4.1. ρ_{2,3} Crosses

The recursion ρ_{2,3}(n) = min(ρ(n−1)+1, ρ(2)+ρ(n/2)+2 if 2|n, ρ(3)+ρ(n/3)+2 if 3|n) was computed to N = 3×10⁸.

Result: C_eff crosses 1.0 at all j = 22–27. λ_{2,3} = 5.25.

Representative trajectories (j = 26): 0.180, 0.196, 0.257, 0.256, 0.305, 0.365, 0.495, 0.701, 0.988, 1.646, 1.600. Clear eventual increase through crossing.

4.2. The Mod-6 Residue Automaton

The {2,3}-skeleton induces a natural partition of integers by residue mod 6:

• n ≡ 0 (mod 6): maximum choice (add, /2, /3) — the "flood gate"
• n ≡ 2, 4 (mod 6): two options (add, /2)
• n ≡ 3 (mod 6): two options (add, /3) — odd-sector renormalization
• n ≡ 1, 5 (mod 6): add only — forced inheritance

Two-thirds of integers possess at least one multiplicative channel. The remaining third (residues 1 and 5) must use the add path exclusively. All primes greater than 3 reside in these forced-inheritance classes.

4.3. The Goldilocks Table

Skeleton	Factor 3?	Crosses?	Assessment
ρ_halve ({2} only)	No	No	Too weak
ρ_{2,5} ({2,5})	No	No	Wrong odd-sector
ρ_{2,3} ({2,3})	Yes	Yes	Goldilocks minimum
ρ_{2,3,5}	Yes	Yes	Enriched, same mechanism
ρ_simple	Yes	Yes	Same mechanism, more factors
Full IC	Yes (96% small-prime)	Yes	≈ {2,3}-skeleton governed

4.4. Why Factor 3 Is Essential

Factor 3 provides three capabilities that halving alone cannot:

First, an odd-sector renormalization channel. Integers n ≡ 3 (mod 6) — i.e., 9, 15, 21, 27, ... — are forced to use the add path under ρ_halve. Factor 3 opens a /3 channel at these positions, providing renormalization access in the odd sector.

Second, deeper structural distance. The map n → n/3 produces a ρ gap of approximately λ · ln 3, compared with λ · ln 2 for halving. For the {2,3}-skeleton (λ ≈ 5.25), these gaps are 5.8 and 3.6 respectively — factor 3 jumps 61% deeper in ρ-space.

Third, a favorable smooth baseline. Factor 3's baseline cost is approximately 0.24 units below the add path (for full IC with λ ≈ 3.86), while factor 2 is 0.33 units above add. Factor 3 wins more readily, which accounts for its 45.6% share of winning factorizations despite being available at fewer positions than factor 2.

---

§5. Full IC as {2,3}-Skeleton Plus Perturbation

5.1. The 96% Equivalence

Full IC's winning factorizations: exact factor identification shows factor 2 wins 50.3%, factor 3 wins 9.1%, factor 5 wins 1.7%, with remaining small primes contributing approximately 5%. By coarseness bins, χ < 0.10 (dominated by factor 2 and factor 3 but also including other small primes) accounts for 96%. Mean coarseness χ = 0.06. Balanced splits never win. This distribution is stable across j = 25–29.

Full IC uses very small factors in the overwhelming majority of its multiplicative events, making it mechanistically close to the {2,3}-skeleton.

5.2. The 4% Perturbation

The remaining 4% of multiplicative events use larger factors (χ = 0.10–0.50), representing deeper jumps into IC structure with stronger decorrelation. This perturbation strengthens, rather than weakens, the decorrelation mechanism.

The extension from {2,3}-skeleton to full IC is best framed as a perturbative budget comparison: full IC equals the {2,3}-skeleton plus a sparse correction to the shell budget, concentrated at 4% of positions where the correction only accelerates erosion.

5.3. Conditional Theorem and Extension Principle

Theorem 68.A ({2,3}-Skeleton Criterion). If the {2,3}-skeleton's C_eff eventually crosses 1.0 with bounded cumulative budget, then the {2,3}-skeleton satisfies Ψ = O(j). Combined with Paper 65's conditional theorem, this yields the L² analogue of 59.1 for the {2,3}-skeleton.

Extension Principle 68.B (Empirical/Conjectural). Full IC can be modeled as the {2,3}-skeleton plus a sparse shell-budget correction concentrated at approximately 4% of multiplicative positions where larger factors win. These positions involve deeper structural jumps, which on physical grounds should strengthen rather than weaken decorrelation. If this perturbative budget comparison holds, full IC also satisfies Ψ = O(j). Rigorous justification of this extension is part of the Paper 69 target.

Status: Theorem 68.A is conditional on formal proof of C_eff crossing for the {2,3}-skeleton. Crossing is empirically confirmed in this paper. Extension 68.B is empirically motivated and formally conjectural.

---

§6. Why Factor 2 Wins Half the Time

For even n = p−1, factor 2 costs ρ(2) + ρ(n/2) + 2 = ρ(n/2) + 4. The add path costs ρ(n−1) + 1.

On the smooth baseline: factor 2 costs approximately λ·ln n − λ·ln 2 + 4, while add costs approximately λ·ln n + 1. The gap is 3 − λ·ln 2.

For full IC (λ ≈ 3.86): gap ≈ +0.33. Factor 2 is typically 0.33 units more expensive than add on the smooth baseline.

Factor 2 wins when ρ(n/2) experiences a negative fluctuation exceeding 0.33. Since ρ-fluctuations are O(1) with standard deviation approximately 0.75, this occurs in roughly 50% of cases — matching the observed 50.3% win rate.

The mean ρ-drop conditioned on mult winning is 1.86, consistent with conditional selection: mult wins only when fluctuations are sufficiently favorable, and the conditional mean reflects the tail of the fluctuation distribution.

For factor 3 (full IC): the smooth gap is 4 − λ·ln 3 ≈ −0.24. Factor 3 is baseline-favorable — it wins even without fluctuation assistance, explaining its substantial 45.6% share.

For balanced splits: the gap is approximately +1.0 — always worse than add by a full unit, unrescuable by O(1) fluctuations.

---

§7. Toward the Formal Proof: Dual-Channel Renormalization

7.1. Single-Channel Failure

The pure-halving RG recurrence γ_j(k) ≈ p_add^k + (1−p_add^k)·γ_{j-1}(k/2) has a fixed point g* < 1, consistent with ρ_halve's failure to cross.

7.2. Dual-Channel Extension

Adding the factor-3 channel modifies the recurrence:

γ_j(k) ≈ p_A · γ_j(k−1) + p_2 · γ_{j-1}(k/2) + p_3 · γ_{j-1}(k/3) + ε

where p_A + p_2 + p_3 = 1, with weights depending on the mod-6 class distribution.

Factor 3 provides a faster descent channel: log_3(k) < log_2(k). This additional pathway may shift the RG fixed point from g < 1 to g > 1.

7.3. Shellwise Renormalization Inequality

The natural theorem target for Paper 69 is a shellwise inequality:

Ψ(2W) ≤ 2^{1−c_{2,3}(W)} · Ψ(W) + error

where c_{2,3}(W) = c_2(W) + δ_3(W), and δ_3(W) ≥ 0 represents factor 3's additional erosion contribution.

Under pure halving, c_2(W) alone is insufficient (g* < 1). Adding δ_3(W) may push c_{2,3}(W) past the crossing threshold.

7.4. Why the 8% Effect Does Not Compound Independently

If individual mult decorrelations (8% per event) were independent, compounding would yield 0.92^{0.74W} — exponential cutoff. Paper 66 killed exponential cutoff.

Resolution: successive mult events are correlated. Halving n and n+2 sends both to n/2 and (n+2)/2, which are structurally adjacent in IC space. Correlated resets produce power-law decay (γ ~ k^{−C}), not exponential. C_eff crossing emerges from cumulative shell budget structure, not per-event independence.

---

§8. Philosophical Insights: The Irreducibility of the Discrete

This section records philosophical observations prompted by the mathematical discoveries of §1–§7. These observations grow naturally from the mathematical skeleton and resonate structurally with the SAE framework's core concepts. They are interpretive extensions of the mathematical results — not theorem-level conclusions and not premises for the preceding mathematical arguments. The scope and philosophical status of each observation require further development in independent SAE texts.

8.1. The Discrete Precedes the Continuous

The following is a methodological observation specific to the ZFCρ domain.

The conventional view in computer science holds that computers use discrete approximation to simulate a fundamentally continuous world. Within the IC recursion domain, ZFCρ suggests a reversal of this priority.

The IC recursion ρ(n) = min(ρ(n−1)+1, min_{ab=n}[ρ(a)+ρ(b)+2]) is defined on the integers. The min operation is non-smooth — it has a kink that no differentiable function can reproduce. The {2,3}-skeleton's mod-6 structure is an arithmetic fact. Within the specific domain of IC recursion, no continuous PDE has been found that reproduces these structural properties without losing the min-kink and mod-6 skeleton.

The non-smoothness of min is not an analytical artifact but a fundamental feature of IC recursion. This suggests that in integer complexity and similar discrete recursive systems, continuous approximation cannot completely capture structural information.

Continuity can be understood as an emergent property: when discrete structure is sufficiently dense, cognition perceives it as continuous. The calibration constant λ ≈ 3.86 appears continuous but emerges from discrete recursion. The residual η = ρ − λ·ln n is what remains after continuous compression — the irreducibly discrete remainder.

8.2. Time Is Hopping Stones, Not a River

The following records a structural resonance between IC recursion and the phenomenology of consciousness — analogous to the Kingdom of Ends discussion in Thermo X. This is structural analogy, not derivation.

The min(add, mult) binary structure of IC recursion resonates suggestively with the continuity-versus-restructuring alternation in conscious experience:

• The add path (顺, inheritance): continue from the previous moment's context — the substrate of experiential continuity.
• The mult path (凿, restructuring): jump to a structurally different state — the substrate of insight, surprise, and creative thought.
• The min operation (不得不, forced choice): the system must select between continuity and restructuring at each step.

In IC, most updates preserve locality — even when entering a multiplicative channel, 96% of events fall in the very-small-prime range (gentle, nearby renormalization). If this resonance reflects deeper structure (an open question), the feeling of temporal continuity corresponds to the dominance of locality-preserving updates, while rare deep structural jumps correspond to creative moments.

Whether this resonance reflects a deeper connection or is merely a structural analogy remains an open question. This paper records it as a suggestive observation.

8.3. The Remainder and the Thing-in-Itself: A Structural Analogy

ρ = λ·ln n + η. Within the SAE framework, this decomposition bears significant structural analogy to Kant's distinction between phenomenon and thing-in-itself.

λ-calibration functions as a narrative faculty: it takes the wild variation of discrete ρ-values and compresses them into the smooth trend λ·ln n. The residual η is what remains after this narrative smoothing — the irreducibly discrete remainder after continuous compression.

In the SAE vocabulary, the remainder (余项) and the thing-in-itself (物自体) share structural features: both represent what is irreducibly left after cognitive or analytical smoothing. This is a structural analogy within the SAE framework, not a claim of philosophical equivalence.

η is precisely the object that determines 59.1 — what continuous approximation cannot absorb.

8.4. The Applicable Domain of ZFCρ

ZFCρ provides irreplaceable insight precisely at the boundary where continuous approximation breaks down:

• Number theory: native domain (discrete, min-plus)
• Thermodynamics: structural isomorphism (discrete states, Boltzmann selection; established in Thermo series I–X)
• Phase transitions: C_eff crossing 1.0 bears structural analogy to critical phenomena (formal mapping is an open question)

Where continuous approximation is already highly effective — classical mechanics, smooth cosmology — ZFCρ's discrete tools may not provide additional first-principles structural insight.

The power boundary of ZFCρ's tools correlates with the discrete-versus-continuous character of each domain — a methodological observation whose exact boundaries require case-by-case verification.

8.5. The {2,3}-Skeleton and Twin Primes: A Structural Alignment

The following is a preliminary arithmetic observation requiring rigorous analysis in Paper 69 and beyond.

All primes greater than 3 occupy residues 1 or 5 modulo 6. Twin prime pairs (p, p+2) correspond to coordinates (5, 1) in the mod-6 automaton.

Between every twin prime pair lies the integer 6k, which occupies residue 0 — the position of maximum renormalization capacity (add + /2 + /3). This "flood gate" provides a natural decorrelation opportunity between twin prime pairs.

The {2,3}-skeleton's mod-6 structure exhibits structural alignment with prime gap distribution. The IC approach has the potential to provide new analytical tools for prime distribution — though this potential requires concrete verification in future work and is not an established result of this paper.

8.6. From Proving One Conjecture to Creating a Class of Tools

H' is stronger than the twin prime conjecture: it asserts not merely that gap 2 occurs infinitely often, but that all gaps distribute exactly as the random model predicts, asymptotically.

The {2,3}-skeleton offers a possible explanation for why H' is natural from the IC perspective: the mod-6 decorrelation mechanism does not discriminate between specific gaps but acts uniformly across all correlations at all scales. This makes H' as a uniform statement more natural than individual gap statements within this framework.

The series' potential contribution extends beyond 59.1/H': the {2,3}-skeleton's mod-6 decorrelation structure, the C_eff scale-dependent erosion framework, and the log-budget identity all possess applicability that may extend to other discrete recursive systems — though full application to other problems is future work.

8.7. Boundary of This Section

This section does not claim that "ZFCρ proves any philosophical proposition" or that "mathematics derives ontology."

Each observation in §8.1–§8.6 is a structural resonance triggered by the mathematical results of §1–§7: the specific behavior of the IC recursion's {2,3}-skeleton on the mod-6 automaton exhibits identifiable structural analogy with core concepts of the SAE framework (remainder, chisel-construct cycle, Via Negativa/Via Rho). These analogies are suggestive, not demonstrative.

Specific boundaries:

First, "the discrete precedes the continuous" is a methodological observation within the IC recursion domain. Whether it extends to general ontology is an open question — this paper makes no general ontological claim.

Second, "time is hopping stones" records a structural resonance between IC recursion's add/mult structure and the phenomenology of consciousness. Following the principle established in Thermo X: this is philosophical resonance, not derivation. Humans are not merely recursive systems.

Third, "the remainder and the thing-in-itself" is a structural analogy within the SAE framework, not a claim of philosophical equivalence.

Fourth, the connection between the {2,3}-skeleton and prime gap distribution is a preliminary arithmetic observation requiring future rigorous analysis.

The technical contribution of Paper 68 is confined to §1–§7. Section 8 records philosophical readings prompted by these technical findings; the philosophical status and scope of each reading require further development in independent SAE texts.

---

§9. Paper 69 Target

9.1. What Paper 68 Accomplished

1. Identified {2,3}-skeleton as the minimum crossing mechanism (Goldilocks)
2. Produced 18 hypothesis kills (Kills 14–18 in this paper)
3. Established that full IC is governed by the {2,3}-skeleton (96% of winning factorizations in very-small-prime range)
4. Characterized the mod-6 residue automaton as the natural proof target
5. Demonstrated that factor 3 is essential among tested two-prime skeletons

9.2. What Remains Open

1. Formal proof of C_eff crossing for the {2,3}-skeleton
2. Analytic derivation of factor 2/3 win rates from IC axioms
3. Unconditional 59.1

9.3. Paper 69 Objectives

Derive, from the mod-6 recursion structure of the {2,3}-skeleton:

(a) That C_eff is eventually increasing (via dual-channel RG analysis)

(b) That the cumulative budget is bounded (via shellwise renormalization inequality)

(c) Extension to full IC via the 4% perturbation argument

Candidate tools: spectral analysis of the mod-6 residue automaton; dual-channel renormalization inequalities; Dirichlet characters mod 6 for residue-class decomposition; Gelfond-type digit sum estimates for ρ_{2,3} fluctuation distributions.

---

§10. Via Negativa to Via Rho: The Methodology of 18 Kills

10.1. Complete Kill Table

#	Hypothesis	How Killed	Paper
1–5	Coordinate and low-rank illusions	Surrogate matching	64–65
6–7	State-space shortcuts	Instant mixing of transfer operator	66
8–10	Screening models	Thermodynamic divergence, tail artifacts	66
11–13	Ancestry shortcuts	Chain, GCD, and attenuation failures	66
14	Simple bridge projection	Corrects only 10–30%	67
15	Dyadic geometry essential	Non-dyadic blocks show identical crash	67
16	Multiplicity drives C_eff	ρ_simple (M_ε = 1) still crosses	68
17	Pure halving sufficient	ρ_halve fails to cross	68
18	{2,5} sufficient	ρ_{2,5} fails to cross	68

10.2. Via Negativa in the Theoretical Phase

Kill 17 prevented Paper 68 from committing to an impossible proof (C_eff crossing for ρ_halve). Kill 18 prevented targeting the wrong minimum skeleton. The principle generalizes: never commit theoretical resources to proving a statement without first empirically confirming it holds.

10.3. The Goldilocks Principle as 凿構循環

Kill 16 removed excessive complexity. Kills 17–18 removed insufficient simplicity. What remains is the {2,3}-skeleton: neither too complex nor too simple, but precisely the minimum structure that produces C_eff crossing.

This is the SAE methodology at atomic resolution: chisel away (凿) excess complexity and insufficient simplicity until the precise skeleton (構) is revealed. Each kill is not a failure but a structural constraint that narrows the remaining space. Eighteen constraints collectively point to a unique mechanism.

---

§11. Status Map

Content	Level
Theorem 68.A: {2,3}-skeleton crossing + bounded budget → Ψ=O(j)	Conditional
{2,3}-skeleton crosses C_eff = 1.0	Empirically confirmed (j = 22–27)
Full IC governed by {2,3}-skeleton (96% small-prime)	Empirically confirmed
ρ_halve does NOT cross	Empirically confirmed (Kill 17)
ρ_{2,5} does NOT cross	Empirically confirmed (Kill 18)
Factor 3 essential among tested two-prime skeletons	Empirically confirmed
18 hypotheses killed	Empirical/analytic
{2,3} crossing formal proof	Open (Paper 69)
59.1 unconditional	Open (Paper 69)

---

Methods

A. IC Variants

All variants computed from scratch on the integers up to N = 3×10⁸: ρ_halve (add + /2 for even); ρ_{2,3} (add + /2 if 2|n + /3 if 3|n); ρ_{2,5} (add + /2 if 2|n + /5 if 5|n); ρ_{2,3,5} (add + /2, /3, /5); ρ_simple (add + smallest prime factor); ρ_largest (add + largest prime factor). Full IC ρ(n) precomputed to N = 10¹⁰ (455 million primes).

B. Sub-Block Variance and C_eff

Following Paper 67 Methods C–D. Half-decade W resolution with K ≥ 4 minimum block count.

C. Coarseness Measurement

For full IC multiplicative events, the winning factorization a·b = p−1 with ρ(a)+ρ(b)+2 = ρ(p−1) was identified via trial division. Coarseness χ = log(min(a,b))/log(p−1).

D. Corrector Spectrum

Raw corrector h(p) = ρ(p) − λ·ln p − class_mean. Centered residual η via Paper 65 adjacent gauge decomposition. Direct Fourier transform at selected low frequencies.

---

References

[1] H. Qin, "ZFCρ Paper LXVII," Zenodo (2026). DOI: 10.5281/zenodo.19705713.

[2] H. Qin, "ZFCρ Paper LXVI," Zenodo (2026). DOI: 10.5281/zenodo.19701479.

[3] H. Qin, "ZFCρ Paper LXV," Zenodo (2026). DOI: 10.5281/zenodo.19687343.

[4] H. Qin, "ZFCρ Paper LX," Zenodo (2026). DOI: 10.5281/zenodo.19480563.

[5] H. Qin, "SAE Methodology Paper 00 (Via Rho)," Zenodo (2026). DOI: 10.5281/zenodo.19657439.

§1. Paper 67后的状态

Paper 67（DOI: 10.5281/zenodo.19705713）建立了：

1. C_eff(W)从~0.3起步eventually increasing，在所有j=22-32穿过critical threshold 1.0
2. 以W/m_j为横轴时C_eff轨迹collapse（block-relative mechanism）
3. Proposition 67.A（exact log-budget identity）：log Ψ(m_j) = log Ψ(W₀) + Σ(1-C_r)·Δ_r
4. Corollary 67.B：若budget = O(1)且Ψ(W₀) = O(j)，则Ψ_full = O(j) → 59.1

Paper 68的target：识别C_eff crossing的微观机制。

---

§2. Too Complex Killed（Kill 16）

2.1. Near-optimal multiplicity不是driver

定义ρ_simple(n) = min(ρ(n-1)+1, ρ(n/p_min)+1)，其中p_min是n的最小素因子。每个composite只有ONE factorization option（M_ε = 1）。

Result： ρ_simple的C_eff仍然crossing 1.0，at all j=22-26。Same pattern as full IC。

Kill 16： 近最优乘法multiplicity is not necessary for C_eff crossing。

2.2. Full IC的coarseness distribution

对full IC中每个mult event，找到胜出factorization a*b = p-1的coarseness χ = log(min(a,b))/log(n)。

χ range	Meaning	Fraction
[0.00, 0.05)	Factor 2	50.3%
[0.05, 0.10)	Factor 3 or small prime	45.6%
[0.10, 0.50)	Intermediate	4.0%
[0.50, 1.00)	Balanced a≈√n	0.0%

96%的winning factorization落在χ < 0.10的very-small-prime区间，以factor 2（50.3%）和factor 3（9.1%）为主导。 Mean χ = 0.06。Balanced splits never win。Stable across j=25-29。

Full IC ≈ {2,3}-skeleton at mechanism level。

2.3. Per-event decorrelation tiny

Event type	Corr(η_i, η_{i+1})
Add (inherit)	0.191
Mult (reset)	0.176
Ratio mult/add	0.92

Each individual mult event reduces lag-1 correlation by only ~8%。Not dramatic reset——gentle decorrelation。C_eff crossing comes from cumulative effect of millions of tiny differences。

8%不能独立相乘（would give exponential cutoff, killed in Paper 66）。Successive mult events are correlated（共享small factors），producing power-law not exponential decorrelation。

---

§3. Too Simple Killed（Kills 17-18）

3.1. Kill 17：Pure halving不sufficient

定义ρ_halve(n) = min(ρ(n-1)+1, ρ(n/2)+1) for even n，ρ(n-1)+1 for odd n。

Result： C_eff NEVER crosses 1.0 at j=25-27。Budget systematically growing（5.1 → 6.3）。Ψ/j at largest W达337-1016。

	γ(1)	C_eff max (reliable)	Budget trend
ρ_halve	0.71	0.77	Growing
ρ_simple	~0.5	>1.0	Bounded
Full IC	0.18	>1.0	Bounded

λ_halve = 2.15。ρ(2^k) = k+1（purely additive along powers of 2）。

Kill 17： Factor 2 alone不provide sufficient decorrelation。Halving sends n → n/2——structurally too close（ρ gap ≈ 1.5）。

3.2. Kill 18：{2,5}也不sufficient

定义ρ_{2,5}(n) = min(ρ(n-1)+1, ρ(n/2)+ρ(2)+2 if 2|n, ρ(n/5)+ρ(5)+2 if 5|n)。

Result： C_eff does NOT cross at j=25-27。Max C_eff = 0.77 at j=25。Budget growing。

Kill 18： Factor 5不能替代factor 3。Not about "any second prime"——specifically factor 3。

3.3. 为什么factor 5不能替代factor 3

Factor 3在odd integers中的density：1/3（每3个odd integer有一个是3的倍数）。

Factor 5在odd integers中的density：1/5（每5个odd integer有一个是5的倍数）。

Factor 3 density比factor 5高67%。

且factor 3的smooth baseline = 4 - λ·ln 3 ≈ -0.24（比add更favorable）。Factor 5的baseline更差。Factor 3在更多位置以更高概率胜出。

---

§4. Just Right：{2,3}-Skeleton

4.1. ρ_{2,3} CROSSES

定义ρ_{2,3}(n) = min(ρ(n-1)+1, ρ(n/2)+ρ(2)+2 if 2|n, ρ(n/3)+ρ(3)+2 if 3|n)。

Result： C_eff crosses 1.0 at ALL j=22-27。

j	C_eff trajectory	Crosses?
22	...0.94, 1.20, 0.59, 4.15	✅ (K小，有noise)
24	...0.58, 0.75, 1.31, 1.14	✅
26	...0.50, 0.70, 0.99, 1.65, 1.60	✅

j=22-23处K_eff较小，trajectory有numerical noise；crossing pattern在j=24-27（K_eff更大）处robust。

λ_{2,3} = 5.25。

4.2. Mod-6 Residue Automaton

{2,3}-skeleton在mod 6下有natural结构：

n mod 6	Options	Description
0	Add, /2, /3	Maximum choice（both channels）
2	Add, /2	Even-only channel
4	Add, /2	Even-only channel
3	Add, /3	Odd-factor-3 channel
1	Add only	Forced inheritance
5	Add only	Forced inheritance

2/3 of integers have mult option。1/3 (residues 1, 5) must add only。

所有primes > 3在residues 1或5。所以primes themselves always forced to add（ρ(p) = ρ(p-1)+1 for all primes in {2,3}-skeleton）。但p-1（prime predecessor）falls in residues 0或4——always has mult option。

4.3. Goldilocks框架

Skeleton	Factor 3?	Crosses?	Mechanism
ρ_halve ({2})	No	NO	Too weak decorrelation
ρ_{2,5} ({2,5})	No (has 5)	NO	Wrong odd-sector coverage
ρ_{2,3} ({2,3})	Yes	YES	Goldilocks minimum
ρ_{2,3,5} ({2,3,5})	Yes	YES	Enriched but same mechanism
ρ_simple (smallest)	Yes	YES	Same mechanism + more factors
Full IC (all)	Yes (96% small-prime)	YES	≈ {2,3}-skeleton主导

{2,3}是minimum crossing skeleton。 Neither too complex（multiplicity unnecessary）nor too simple（halving insufficient）。

4.4. 为什么是factor 3

Factor 3提供了halving不能给的三件事：

一、Odd-sector renormalization channel。 Residue 3 mod 6（即9, 15, 21, 27...）在ρ_halve下forced to add。Factor 3在these positions opens /3 path。

二、Deeper structural distance。 n → n/3的ρ gap ≈ λ·ln 3 ≈ 5.8 for ρ_{2,3}。比n → n/2的gap ≈ λ·ln 2 ≈ 3.6更大。Stronger per-event decorrelation。

三、Favorable baseline。 Factor 3的smooth baseline比add better by 0.24 units。Factor 3 wins MORE easily than factor 2（which is typically 0.33 WORSE than add）。

在已测试的二素数骨架中，这不是density alone。 Factor 5 has comparable coverage but doesn't cross。Factor 3的specific combination of density（1/3 of odds）+ structural distance（deeper jump）+ favorable baseline在已测试范围内是unique。是否存在其他未测试的二素数组合也能crossing是open question。

---

§5. Full IC as {2,3}-Skeleton + Perturbation

5.1. 96% equivalence

Full IC的winning factorization中，exact factor identification shows：factor 2胜出50.3%，factor 3胜出9.1%，factor 5胜出1.7%，其余小素因子共占约5%。按coarseness χ分布，χ < 0.10区间（dominated by factor 2和factor 3但也包含其他小素因子）占96%。Mean coarseness χ = 0.06。Balanced splits（χ ≈ 0.5）never win。

这意味着full IC在绝大多数mult events处使用very small factors，与{2,3}-skeleton的机制essentially equivalent。

5.2. 4% perturbation

Remaining 4%使用larger factors（χ = 0.10-0.50）。These are DEEPER jumps（更强decorrelation）。

4% perturbation对C_eff的contribution：deeper jumps只strengthens decorrelation。Full IC在perturbation positions decorrelates MORE than {2,3}-skeleton。

更precise的extension principle（per ChatGPT review）：full IC = {2,3}-skeleton + sparse correction to shell budget。不claim γ_full(k) ≤ γ_simple(k) as strong monotonicity，而是perturbative budget comparison。

5.3. Conditional theorem and extension principle

Theorem 68.A（{2,3}-Skeleton Criterion）。 若ρ_{2,3}的C_eff eventually crosses 1.0并且cumulative budget bounded，则ρ_{2,3}的Ψ = O(j)。结合Paper 65条件定理，this gives 59.1 for the {2,3}-skeleton。

Extension Principle 68.B（Empirical/Conjectural）。 Full IC can be modeled as {2,3}-skeleton plus a sparse shell-budget correction concentrated at ~4% of mult positions where larger factors win。These positions involve deeper structural jumps，which on physical grounds should strengthen rather than weaken decorrelation。若this perturbative budget comparison成立，则full IC的Ψ also O(j)。Full IC extension的rigorous论证是Paper 69 target的一部分。

状态： Theorem 68.A conditional on {2,3}-skeleton C_eff crossing的formal proof。Crossing empirically confirmed（this paper）。Extension 68.B empirically motivated, formally conjectural。

---

§6. 为什么Factor 2 Win 50%？

每个even n = p-1，factor 2的cost = ρ(2) + ρ(n/2) + 2 = ρ(n/2) + 4。Add cost = ρ(n-1) + 1。

Smooth baseline比较：

factor 2: ρ(n/2) + 4 ≈ λ·ln(n/2) + 4 = λ·ln n - λ·ln 2 + 4

add: ρ(n-1) + 1 ≈ λ·ln n + 1

Gap = (4 - λ·ln 2 - 1) = 3 - λ·ln 2。

对full IC（λ = 3.86）：gap ≈ +0.33。Factor 2在smooth baseline上略差于add。

但ρ-fluctuation是O(1)量级。当ρ(n/2)有negative fluctuation ≥ 0.33时，factor 2翻盘。这发生在approximately 50%的cases。

Mean ρ-drop when mult wins = 1.86——consistent with conditional selection bias：mult只在fluctuation足够negative时win。

对factor 3（λ = 3.86）：smooth gap = 4 - λ·ln 3 ≈ -0.24。Factor 3在baseline上比add更好。所以factor 3 wins even more easily，explaining its 45.6% share。

对balanced split（a ≈ √n）：cost ≈ 2·λ·ln √n + 2 = λ·ln n + 2。Gap = +1。Always worse than add by 1 full unit。 O(1) fluctuations cannot rescue——balanced splits NEVER win。

---

§7. 双通道重整化方程

7.1. 从单通道到双通道

Gemini在brainstorm中提出的RG recurrence（pure halving）：

γ_j(k) ≈ p_add^k + (1-p_add^k)·γ_{j-1}(k/2)

这个方程的fixed point g* < 1——consistent with ρ_halve NOT crossing。

7.2. {2,3} dual-channel extension

加入factor 3通道后：

γ_j(k) ≈ p_A·γ_j(k-1) + p_2·γ_{j-1}(k/2) + p_3·γ_{j-1}(k/3) + ε

其中p_A + p_2 + p_3 = 1（取决于position的mod-6 class）。

Factor 3提供k → k/3的"faster descent"——log_3(k) < log_2(k)。这改变了fixed point，可能pushing g* > 1。

7.3. Shellwise renormalization inequality

ChatGPT提出的theorem target：

Ψ(2W) ≤ 2^{1-c_{2,3}(W)}·Ψ(W) + error

其中c_{2,3}(W) = c_2(W) + δ_3(W)，δ_3(W) ≥ 0是factor-3 channel的额外erosion contribution。

Pure halving：c_2(W) alone不够大。加上δ_3(W)后c_{2,3}(W) pushes past crossing。

7.4. 为什么8%的per-event effect不独立相乘

如果independent：0.92^{0.74W} → exp(-0.062W) → exponential cutoff。但Paper 66 killed exponential cutoff。

Resolution：successive mult events correlated。n/2和(n+2)/2在IC space中structurally close——send to related regions。Correlated resets give power-law decay（γ ~ k^{-C}），not exponential。

C_eff crossing comes from cumulative shell budget structure，not from per-event independence。

---

§8. 哲学洞察：离散不可替代性

本节记录由§1-§7数学发现引发的若干哲学读法。这些读法从数学骨架中自然生长，与SAE框架的核心概念形成结构共振。它们是数学结果的解释性延伸——不是定理推论，不作为前文数学论证的前提。每个观察的适用范围和哲学地位需在SAE独立文本中进一步论证。

8.1. 离散先于连续

以下是ZFCρ domain内的方法论观察。

传统CS认知：计算机用离散模拟连续世界。

ZFCρ揭示了IC recursion domain内相反的priority：ρ(n) = min(...)定义在ℤ上。Min operation有kink（不可微）。{2,3}-skeleton的mod-6 structure是arithmetic fact。在当前IC recursion这一特定域内，尚未看到任何连续PDE在不丢失min-kink与mod-6骨架的前提下重现这些性质。

Min的non-smoothness不是approximation artifact，是IC recursion的fundamental feature。这提示：在integer complexity以及类似discrete recursive systems里，continuous approximation无法completely capture structural information。

连续性可以被理解为emergent property——当离散结构足够dense，cognition将其"感知"为continuous。

λ·ln n是discrete ρ(n)的continuous compression。η = ρ - λ·ln n是compression residual——continuous approximation之后的irreducibly discrete remainder。

8.2. 时间是跳格子

以下是IC recursion结构与意识现象学之间的structural resonance观察，类比Thermo X的Kingdom of Ends discussion——structural analogy, not derivation。

IC recursion的min(add, mult)二元性与意识经验中continuity-vs-restructuring的交替有suggestive structural resonance：

• Add path = 顺（inherit）：继承上一刻的context，对应意识经验中连续性的基底。
• Mult path = 凿（restructure）：跳到structurally different state，对应insight、surprise、创造性重组的moments。
• Min = 不得不（forced choice）：system在continuity和restructuring之间做forced selection。

在IC中，多数更新保持局部性——即便进入mult通道，96%仍落在very-small-prime的近域重整化（gentle, nearby transitions）。若这个resonance反映deeper structure（open question），"时间感觉连续"可理解为局部性主导的phenomenological counterpart，而少数deep structural jumps对应意识的creative moments。

是否这个resonance反映deeper connection或仅是mathematical analogy是open question——本文将其作为suggestive observation记录。

8.3. 余项与物自体的结构类比

ρ = λ·ln n + η

在SAE框架内，这个分解与Kant的现象/物自体区分有significant structural analogy：

λ-calibration作为narrative function：takes wild discrete ρ values, compresses into smooth λ·ln n trend。The residual η is what remains after narrative smoothing——continuous approximation之后的irreducibly discrete remainder。

η is precisely the object that matters for 59.1：它是连续性压缩无法消化的余项。

在SAE的vocabulary中：余项（remainder）与物自体（thing-in-itself after cognitive smoothing）在structural level相通——两者都represent cognitive/analytical smoothing之后irreducibly remaining的部分。这是SAE框架内的structural analogy，不是哲学等价性的claim。

8.4. ZFCρ的applicable domain

ZFCρ provides irreplaceable insight precisely where continuous approximation breaks：

• Prime distribution（discrete, min-plus native）：Direct application
• Thermodynamics（discrete states, Boltzmann selection）：Structural isomorphism（Thermo series I-X已建立）
• Phase transitions：C_eff crossing 1.0与critical phenomena有structural analogy（formal mapping是open question）

在continuous approximation已高度有效的领域（classical mechanics, electrodynamics, smooth cosmology），ZFCρ型discrete tools未必提供额外的第一性结构增益。

ZFCρ tools的power boundary与问题的discrete-vs-continuous性质相关——这是methodological observation, 具体边界需case-by-case验证。

8.5. {2,3}-skeleton与孪生素数的结构对齐

以下是初步arithmetic observation，需Paper 69+的rigorous analysis。

所有大于3的primes在residues 1或5 (mod 6)。Twin prime pairs (p, p+2) 对应coordinates (5, 1) mod 6。

每个twin prime pair之间的6k处在residue 0 (mod 6)——拥有maximum重整化能力（add + /2 + /3三个通道）。这个"泄洪闸"提供了twin prime对之间的natural decorrelation机会。

{2,3}-skeleton的mod-6 structure与prime gap distribution之间存在structural alignment。IC approach有潜力为prime distribution提供新的分析工具——though this potential需要在未来work中具体验证，不是本文established result。

8.6. 从"证一个猜想"到"创造一类工具"

H'比twin prime更强：不是说"gap 2 exists infinitely"而是"ALL gaps distribute exactly as random model predicts"。

{2,3}-skeleton提供了一个可能的explanation：mod-6的decorrelation不discriminate between specific gaps，而是uniformly作用于所有correlations at所有scales。这使得H'作为uniform statement比individual gap statements更natural from IC perspective。

Series的potential contribution超出59.1/H'本身：{2,3}-skeleton的mod-6 decorrelation structure, C_eff scale-dependent erosion framework, log-budget identity都具备向其他discrete recursive系统延展的applicability——尽管full application to other problems是future work。

8.7. 本节的边界

本节不claim"ZFCρ证明了任何哲学命题"或"数学推导出了本体论"。

§8.1-§8.6的每条观察都是由§1-§7数学结果触发的结构共振：IC递推的{2,3}-skeleton在mod-6 automaton上的具体behavior，与SAE框架的核心概念（余项、凿构循环、Via Negativa/Via Rho）之间存在可识别的structural analogy。这些analogy是suggestive的，不是demonstrative的。

具体边界：

一、"离散先于连续"是IC recursion domain内的methodological observation。它是否extend到general ontology是open question——本文不做general ontological claim。

二、"时间是跳格子"是IC recursion的add/mult structure与意识phenomenology之间的structural resonance。如同Thermo X所确立的原则：这是philosophical resonance, not derivation。人不仅仅是递推系统。

三、"余项与物自体"是SAE框架内的structural analogy，不是哲学等价性的claim。

四、{2,3}-skeleton与prime gap distribution的关联是preliminary arithmetic observation，需要future rigorous analysis。

Paper 68的technical contribution限定在§1-§7。§8记录了这些技术发现引发的哲学读法，每条读法的哲学地位和适用范围需在SAE独立文本中进一步论证。

---

§9. Paper 69 Target

9.1. Paper 68完成了什么

1. 识别{2,3}-skeleton as minimum crossing mechanism（Goldilocks）
2. 18次假说否决（Kills 14-18 in this paper, 1-13 in Papers 62-66）
3. Full IC由{2,3}-skeleton主导（96% coarseness落在very-small-prime区间）
4. Mod-6 residue automaton as proof target
5. 在已测试的二素数骨架中，Factor 3是essential的第二通道

9.2. Paper 68没有完成什么

1. 没有formal proof of C_eff crossing for {2,3}-skeleton
2. 没有从IC axiom推导factor 2/3 win rates
3. 没有unconditional 59.1

9.3. Paper 69 target

从{2,3}-skeleton的mod-6 recursion structure推导：

(a) C_eff eventually increasing（from dual-channel RG analysis）

(b) Cumulative budget bounded（shellwise renormalization inequality）

(c) Extension to full IC via 4% perturbation argument

候选工具：

• Mod-6 residue automaton的spectral analysis
• Dual-channel renormalization inequality（factor 2 + factor 3）
• Dirichlet characters mod 6 for residue-class decomposition
• Gelfond-type digit sum estimates for ρ_{2,3} fluctuation distribution

---

§10. Via Negativa → Via Rho：18 Kills的方法论

10.1. Kill表

#	Hypothesis	Result	Paper
1-5	Coordinate/low-rank illusions	Surrogates match	64-65
6-7	State-space shortcuts	Transfer instant mixing	66
8-10	Screening models	Thermo diverges, tail artifact	66
11-13	Ancestry shortcuts	Chain/GCD/attenuation fail	66
14	Simple bridge projection	Too weak (10-30%)	67
15	Dyadic geometry essential	Non-dyadic same crash	67
16	Multiplicity drives C_eff	ρ_simple M_ε=1 crosses	68
17	Pure halving sufficient	ρ_halve doesn't cross	68
18	{2,5} sufficient	ρ_{2,5} doesn't cross	68

10.2. Via Negativa在理论阶段的价值

Kill 17 prevented Paper 68 from committing to proving wrong theorem（C_eff crossing for ρ_halve——impossible）。Kill 18 prevented proving wrong minimum skeleton（{2,5} doesn't cross）。

Principle： never commit to proving something without first empirically confirming it's true。This applies even when the target seems "obvious"。

10.3. Goldilocks as 凿構循環

Kill 16 removed "too complex"。Kills 17-18 removed "too simple"。Remaining：{2,3}-skeleton = just right。

Neither too complex（full IC with arbitrary factorizations）nor too simple（pure halving or {2,5}）。The minimum mechanism that still produces C_eff crossing。

这是SAE methodology at atomic level：凿去excess complexity（Kill 16）和insufficient simplicity（Kills 17-18），构出the precise skeleton。

---

§11. Status Map

内容	层级
Theorem 68.A: {2,3}-skeleton crossing + bounded budget → Ψ=O(j)	Conditional theorem
{2,3}-skeleton crosses C_eff = 1.0	Empirically confirmed (j=22-27)
Full IC由{2,3}-skeleton主导（96% small-prime）	Empirically confirmed
ρ_halve does NOT cross	Empirically confirmed (Kill 17)
ρ_{2,5} does NOT cross	Empirically confirmed (Kill 18)
在已测试二素数骨架中Factor 3 essential	Empirically confirmed
Per-event decorrelation ~8%	Empirically confirmed
Centering reduces absolute but not normalized power	Empirically confirmed
18 hypotheses killed total	Empirical/analytic
{2,3} C_eff crossing formal proof	Open (Paper 69)
59.1 unconditional	Open (Paper 69)

---

Methods

A. IC Variants

All variants computed from scratch on ℤ up to N = 3×10^8:

• ρ_halve: min(ρ(n-1)+1, ρ(n/2)+1) for even n; ρ(n-1)+1 for odd
• ρ_{2,3}: min(ρ(n-1)+1, ρ(2)+ρ(n/2)+2 if 2|n, ρ(3)+ρ(n/3)+2 if 3|n)
• ρ_{2,5}: min(ρ(n-1)+1, ρ(2)+ρ(n/2)+2 if 2|n, ρ(5)+ρ(n/5)+2 if 5|n)
• ρ_{2,3,5}: min over add + factor 2, 3, 5
• ρ_simple: min(ρ(n-1)+1, ρ(p_min)+ρ(n/p_min)+2) for smallest prime p_min dividing n
• ρ_largest: min(ρ(n-1)+1, ρ(p_max)+ρ(n/p_max)+2) for largest prime factor

Full IC ρ(n) precomputed to N = 10^10 (455M primes).

B. Sub-block Ψ and C_eff

As Paper 67 Methods C-D. Half-decade W resolution. K ≥ 4 minimum.

C. Coarseness measurement

For full IC mult events (ρ(p-1) ≠ ρ(p-2)+1): find winning factorization a*b = p-1 with ρ(a)+ρ(b)+2 = ρ(p-1). Coarseness χ = log(min(a,b))/log(p-1). Factor identification via trial division with precomputed small primes.

D. Corrector spectrum

h(p) = ρ(p) - λ·ln p - class_mean. η = h/p - branch_mean. Direct Fourier transform at selected low frequencies.

---

参考文献

[1] H. Qin, "ZFCρ Paper LXVII," Zenodo (2026). DOI: 10.5281/zenodo.19705713.

[2] H. Qin, "ZFCρ Paper LXVI," Zenodo (2026). DOI: 10.5281/zenodo.19701479.

[3] H. Qin, "ZFCρ Paper LXV," Zenodo (2026). DOI: 10.5281/zenodo.19687343.

[4] H. Qin, "ZFCρ Paper LX," Zenodo (2026). DOI: 10.5281/zenodo.19480563.

[5] H. Qin, "SAE Methodology Paper 00 (Via Rho)," Zenodo (2026). DOI: 10.5281/zenodo.19657439.