ρ-Conservation: The Ineliminability Law of the Remainder across the Layer Hierarchy
ZFCρ Paper 1 established the remainder thesis: ρ(C,U)≠∅ — remainder cannot be eliminated, only displaced. Paper 2 provided the quantitative identity of the remainder: π is the first complete specimen of quantitative ρ at the L₁→L₂ transition; Euler's formula is the structural identity "one act (the exponential map) binding two remainders (i, π) to produce closure"; successor, e^x, Ω_U, and time can be interpreted as instances of the same class of self-referential generation unfolding at different layers.
This paper proposes the ρ-Conservation Principle: grounded in Paper 1's ρ-thesis (remainder is ineliminable) and the premise "extensionalization does not create information," it argues that remainder is neither created nor destroyed under any extensionalization or layer transition — it is only redistributed between remainder (extensional product) and act (operative performance).
An immediate manifestation: the same ρ presents different faces at different layers — π at L₁ is an inexhaustible infinite series; at L₂ it is the precise irrational constant 3.14159…. It is not ρ that changes, but the observational layer.
On the applied side, this paper re-reads Gödel's incompleteness theorem as ρ-overflow; proposes a conservation relationship between proof complexity and ρ-content, illustrated by Fermat's Last Theorem and the Poincaré Conjecture; and provides a ρ-diagnosis and working classification of six major open problems.
1. Introduction: From "Displacement" to "Conservation"
1.1 The Logical Chain across Three Papers
Paper 1: Remainder exists, cannot be eliminated, can only be displaced. — Qualitative.
Paper 2: The quantitative characterization of remainder requires structure from one layer above. π is the first complete specimen. The ternary structure of the exponential map / i / π (act / remainder / remainder) closes in Euler's formula. Self-referential generation unfolds across layers as successor / e^x / Ω_U / time. Remainder becomes the axiom of the next layer. — Quantitative, but without asking whether the "quantity" is conserved across transitions.
Paper 3 (this paper): ρ-conservation. — From existence to ineliminability.
1.2 Why a Conservation Law Is Needed
Paper 2 established "remainder becomes the axiom of the next layer" — 0' is a wall at L₂, a floor at L₃. But as remainder passes from wall to floor, does the total quantity of ρ change?
If it decreases — some mechanism destroys operative content, contradicting Paper 1.
If it increases — self-referential generation creates operative content ex nihilo, but self-referential generation has the compression signature "act = result"; it is not a generative mapping.
If it is conserved — we have an invariant spanning all L-layers: a meta-law about formalization itself.
1.3 Intuition from π: The Same ρ Wears Different Faces at Different Layers
At L₁ (discrete lattice perspective), π is inexhaustible — the Leibniz series π/4 = 1 - 1/3 + 1/5 - 1/7 + … never converges to a rational number. From L₁'s vantage, π is infinity: no finite number of discrete operations can reach it.
At L₂ (continuous/complex-plane perspective), π is a precise irrational constant 3.14159… — the unique self-dual fixed-point parameter of the Fourier transform, the exact value of the Gaussian integral, the half-circumference of the unit circle.
The same ρ, viewed from a different layer, shifts from "inexhaustible" to "precisely determined." ρ neither shrinks nor grows; the same structural content presents different faces at different layers. What is conserved is ρ itself; what changes is the observational layer.
2. Definitions
2.1 Operative Description Space O(Lₙ)
Mathematical objects at layer Lₙ together with their complete generative history. Not "the prime 7," but "7, together with the successor path 0→1→2→3→4→5→6→7 and the complete record of failed trial divisions by 2, 3, 4, 5, 6."
2.2 Extensional Description Space E(Lₙ)
The same mathematical objects, stripped of generative history. "7 ∈ P," where P is the set of primes.
2.3 Extensionalization Mapping f: O(Lₙ) → E(Lₙ)
An intra-layer operation: discard operative history, retain extensional result. Every definition, axiomatization, and formalization is an instance of f.
3. ρ-Conservation: Derivation
3.1 Conservation Principle I (Single Extensionalization)
Proposition: For any non-trivial operative description o ∈ O and extensionalization mapping f: O → E,
3.2 Derivation
Direction 1 (ρ is not destroyed). Assume ρ_E(f(o)) + ρ_f < ρ(o). Then some ρ has been genuinely destroyed. But destroying operative content is itself an operative act requiring ρ, so the very process of destroying ρ produces an equal amount — a self-referential contradiction. More rigorously: if some f could partially destroy ρ, then repeated application of f could drive ρ arbitrarily close to zero, contradicting Paper 1's ρ-thesis (ρ is ineliminable, can only be displaced).
Direction 2 (ρ is not created). Assume ρ_E(f(o)) + ρ_f > ρ(o). Then extensionalization has created operative content ex nihilo. But f is defined as discarding operative history and retaining extensional results — information compression, not information generation. The sole source of operative content is operative acts themselves. As a mapping, f's own operative content ρ_f derives from the act of "executing f," and this act operates on o, so ρ_f cannot exceed the operative content of o minus what is retained in the product.
Conjunction. ρ(o) = ρ_E(f(o)) + ρ_f. ∎
3.3 Conservation Principle II (Cross-Layer Transition)
Proposition: For any L-layer transition Lₙ → Lₙ₊₁, let A_n be the extensionalization lift of that layer. Then:
3.4 Recursive Conservation (Chain Propagation)
Chain expansion:
The original ρ_total(Lₙ) does not decay at any step along the layer chain. It is progressively distributed among the self-referential generation acts and the latent substrates of successive layers.
4. Gödel Incompleteness as ρ-Overflow
4.1 Axiomatization of PA as Large-Scale Extensionalization
The axiomatization of PA is F: O_math → E_PA. The conservation principle gives:
4.2 ρ_E(E_PA) = Undecidable Sentences
The latent remainder embedded in PA's formal structure exists in the form of undecidable sentences: their truth-values depend on operative content discarded during PA's axiomatization.
4.3 Gödel Diagonalization as ρ-Detector
Gödel's self-referential construction forces ρ_E from latent to manifest. Within the ρ framework, the incompleteness theorem can be re-read as a positive detection of ρ-conservation — conserved ρ necessarily overflows at the boundary of a formal system, and undecidable sentences are the specific locations of that overflow.
5. Proof Complexity as the Cost of ρ-Conservation
5.1 Core Claim
When extensional proof tools do not match the ρ-level of a proposition, ρ-conservation forces the discarded operative content to return in the form of proof complexity. The more concise the statement (the more densely ρ is compressed), the more complex the purely extensional proof. Conversely, viewing a problem at the correct ρ-level can make the proof very short.
5.2 ρ Re-reading of Fermat's Last Theorem
Fermat's Last Theorem is stated at L₁: for n ≥ 3, x^n + y^n = z^n has no positive integer solutions. Wiles's proof mobilized extremely heavy L₂ machinery (modular forms, Galois representations, elliptic curves), running to hundreds of pages.
An intuitive reading within the ρ framework: at n = 2, x² + y² = z² has infinitely many solutions (Pythagorean triples); at n ≥ 3, there are suddenly none. The rupture occurs at 2 — the "seed of all splitting" established in Paper 2, the sole remainder of L₀→L₁. Quadratic structure (squaring, two dimensions, the complex plane) is the minimal condition for closure. At cubic and above, the growth rate of multiplicative power structure exceeds that of additive linear structure, and the coupling breaks.
5.3 ρ Re-reading of the Poincaré Conjecture
The Poincaré Conjecture: every simply connected closed 3-manifold is homeomorphic to S³. Perelman proved it using Ricci flow with surgery.
Reading within the ρ framework: simply connected means all closed paths can be continuously contracted to a point — no topological remainder. In ρ language, the Poincaré criterion can be read as the extreme case "topological remainder equals zero."
Perelman's Ricci flow can be re-read as an act (continuous deformation) attempting to digest remainder (curvature non-uniformity). ρ-conservation manifests here as: curvature is smoothed at one location, remainder overflows at another as a singularity. Surgery is the technique for handling ρ-overflow — excise the singularity, re-glue, continue smoothing.
6. The ρ Reading of the Addition–Multiplication Coupling
6.1 Common Root
Both addition and multiplication derive from the successor function S. Addition is first-order iteration of S; multiplication is second-order iteration.
6.2 Directional Asymmetry
Zhang Yitang's 2013 proof of bounded gaps between primes shows L₂ tools making continuous progress without stalling. Chen Jingrun's 1973 result of 1+2 has seen zero progress toward 1+1 in five decades.
Structural reading within the ρ framework: the multiplication → addition direction (twin primes) is projection — from higher-order structure down to lower-order measure, flowing with the current. The addition → multiplication direction (Goldbach) is reverse reconstruction — from lower-order structure demanding higher-order atomic coverage, against the current. The directional asymmetry provides the structural explanation.
7. ρ-Diagnosis and Working Classification of Open Problems
7.1 Core Distinction: Missing Current-Layer ρ vs. Missing Higher-Layer ρ
Type 1: Missing current-layer ρ. Tools exist at the same layer but the correct closure path has not yet been found. ρ circulates within the layer. Working tendency: eventually proved by same-layer tools.
Type 2: Missing higher-layer ρ. Current-layer tools are in principle insufficient; a layer transition is needed. ρ circulates across layers. Working tendency: either proved independent of the current axiomatic system, or awaiting the development of higher-layer tools.
7.2 Working Pairing 1: Goldbach + Twin Primes (Static Coupling)
Goldbach: Can every even number be expressed as the sum of two primes?
Twin Primes: Can the gap between primes equal 2 infinitely often?
Within the ρ framework, both can be read as two directions of the same ρ-coupling. In number theory, both problems belong to additive number theory, employ similar sieve and circle methods, and Goldbach is widely regarded as harder (consistent with the counter-flow theory).
Working conjecture: If the ρ-coupling analysis is directionally correct, the two conjectures may admit unified treatment.
7.3 Working Pairing 2: Riemann + Collatz (Dynamic Behavior)
Riemann Hypothesis: The non-trivial zeros of the Riemann zeta function lie on the critical line Re(s) = 1/2.
Collatz Conjecture: Iteration of the map 3n+1 (if odd) and n/2 (if even) returns to 1.
Pairing logic: Riemann views L₁'s global distribution from L₂; Collatz views the local entanglement from within L₁. In the ρ framework, both can be read as different viewpoints on the same object — the distribution of successor-ρ between addition and multiplication.
7.4 Navier-Stokes (Continuous Version of Addition–Multiplication Entanglement)
The nonlinear term (u·∇)u of the NS equations is an entanglement of additive structure (vector addition) and multiplicative structure (field self-multiplication). The linear part (Stokes equations) is fully solvable within L₂; the nonlinear entanglement is not.
7.5 P ≠ NP (Computational Analogue of the ρ-Thesis)
P is deterministic polynomial-time solvable. NP is nondeterministic polynomial-time verifiable. P ≠ NP asserts: verification (extensional operation) is not reducible to search (operative act).
Reading within the ρ framework: The content of P ≠ NP can be read as an analogous manifestation of the ρ-thesis in computational theory.
7.6 Complete ρ-Diagnostic Table
| Problem | ρ type | Layer position | Working pairing | Working tendency |
|---|---|---|---|---|
| Goldbach | Static coupling (counter-flow) | L₁→L₂ boundary | Twin Primes | Spans both types; may admit unified treatment |
| Twin Primes | Static coupling (with-flow) | L₁→L₂ boundary | Goldbach | Leans Type 1; may admit unified treatment |
| Riemann | Global distribution | L₂ interior | Collatz (meta-conjecture) | Leans Type 1 |
| Collatz | Discrete entanglement iteration | L₁ (cross-layer need) | Riemann + NS (meta-conjecture) | Type 2 |
| Navier-Stokes | Continuous entanglement iteration | L₂→L₄ | Collatz (analogy) | Spans both types |
| P ≠ NP | Analogue of ρ-thesis | Meta-level | None (independent) | May consider independence from ZFC |
8. ρ-Conservation as Constraint on ρ-Arithmetic
The basic object of ρ-arithmetic is not a number n but an operation-pair (n, H_n) — a number together with its complete generative history.
ρ-Arithmetic uses conservation principles to constrain "how operative content flows." Axioms are static existence declarations; conservation principles are dynamic flow constraints.
9. Discussion
9.1 ρ-Conservation and Physical Conservation Laws
ρ-conservation runs structurally parallel to energy conservation: energy is neither created nor destroyed, only converted between kinetic and potential forms; ρ is neither created nor destroyed, only redistributed between remainder and act.
If ρ-conservation has a physical counterpart, it may be deeply connected to information conservation (the black hole information paradox, Susskind's ER=EPR).
9.2 Relation to Existing Foundations of Mathematics
ZFC: ρ-conservation is not an axiom of ZFC and does not contradict ZFC. It is a meta-law about formalization itself.
Category Theory: Tracks morphisms, closer to operative content than set theory, but morphisms are still extensionalized. ρ-conservation applies.
Intuitionism / Constructivism: Brouwer insisted on constructive proofs, attempting to preserve more operative content. ρ-conservation explains why constructivism is more restrictive — it refuses to discard as much ρ, carrying a heavier operative burden.
10. Conclusion
Paper 1 named the remainder. Paper 2 gave remainder its face. Paper 3 proposes remainder's law.
π at L₁ is infinity; at L₂ it is 3.14159…. It is not π that changed; it is the layer from which it is viewed. ρ is conserved.
If the ρ-conservation framework proves productive, then the most stubborn open problems in number theory may require not only more ingenious existing techniques but also a new language more willing to explicitly handle operative content. ρ-arithmetic is a candidate direction proposed, not a proven sole path.
← ZFCρ Paper II: The Quantitative Identity of the Remainder
ZFCρ Paper IV: A Draft Term Model for ρ-Arithmetic →
ZFCρ Series · Mathematical Foundations · Back to Papers
ZFCρ第一篇确立了余项命题：ρ(C,U)≠∅，余项不可消除，只能位移。第二篇给出余项的定量身份：π是L₁→L₂跃迁的定量样例，欧拉公式是"一个行为（指数映射）绑定两个余项（i, π）产生闭合"的结构恒等式，后继、e^x、Ω_U、时间可被解释为同一类自指生成行为在不同层的展开。
本文提出ρ守恒原理：以Paper 1的ρ命题（余项不可消除）和"外延化不增信息"前提为基础，论证余项在任何外延化和层级跃迁中既不被创造也不被消灭，只在余项（外延产物）和行为（操作性施行）之间重新分配。
守恒原理的直接表现：同一个ρ在不同层级呈现为不同的面貌——π在L₁是不可穷尽的无穷级数，在L₂是精确的无理常数3.14159...。不是ρ变了，是观测层级变了。
应用层面，本文将哥德尔不完备定理在ρ框架下重读为ρ溢出；提出证明复杂度与ρ含量之间的守恒关系，并以费马大定理和庞加莱猜想为例证；给出六大未解问题的ρ诊断与工作性分类。
1. 引言：从"位移"到"守恒"
1.1 三篇论文的逻辑链
Paper 1（ρ命题）：余项存在，不可消除，只能位移。——定性。
Paper 2（定量身份）：余项的定量刻画需要高一层结构。π是第一个完整样例。指数映射 / i / π 的三元结构（行为/余项/余项）在欧拉公式中闭合。自指生成在不同层展开为后继/ e^x / Ω_U /时间。余项是下一层的公理。——定量，但未追问"量"在跃迁中是否守恒。
Paper 3（本文）：ρ守恒。——从存在到不灭。
1.2 为什么需要守恒律
Paper 2确立了"余项是下一层的公理"——0'在L₂是墙，在L₃是地板。但余项从墙变成地板的过程中，ρ的总量变了没有？
如果衰减——存在某种机制消灭操作性内容，与Paper 1矛盾。
如果膨胀——自指生成凭空创造操作性内容，但自指生成是"行为=结果"的压缩签名，不是生成映射。
如果守恒——我们就有了贯穿所有L层级的不变量：一条关于形式化本身的元定律。
1.3 π的直觉：同一个ρ在不同层的面貌
π在L₁（离散格视角）是不可穷尽的——莱布尼茨级数 π/4 = 1 - 1/3 + 1/5 - 1/7 + … 永远不收敛到有理数。在L₁看来π就是无穷：有限步离散操作永远够不到它。
π在L₂（连续/复平面视角）是精确的无理常数3.14159...——Fourier自对偶的不动点参数，高斯积分的精确值，单位圆的半周长。
同一个ρ，换了一层来看，从"不可穷尽"变成"精确确定"。ρ不是变少了或变多了，是同一份结构性内容在不同层级呈现为不同的面貌。守恒的是ρ本身，变化的是观测层级。
2. 定义
2.1 操作性描述空间 O(Lₙ)
Lₙ层级的数学对象连同其完整生成历史。不是"素数7"，而是"7，连同后继路径0→1→2→3→4→5→6→7，以及对2,3,4,5,6的试除全部失败的完整否定记录"。
2.2 外延性描述空间 E(Lₙ)
同一层级的数学对象，剥去生成历史。"7∈P"，其中P是素数集。
2.3 外延化映射 f: O(Lₙ) → E(Lₙ)
层内操作：丢弃操作历史，保留外延结果。每一次定义、公理化、形式化都是f的实例。
3. ρ守恒：推导
3.1 守恒原理一（单次外延化守恒）
命题：对任意非平凡操作性描述 o ∈ O 和外延化映射 f: O → E，
3.2 推导
方向一（ρ不灭）。假设 ρ_E(f(o)) + ρ_f < ρ(o)。则存在ρ被真正消灭——既不在外延产物中，也不在映射过程中。但消灭操作性内容本身是一个操作行为，执行它需要ρ，因此消灭ρ的过程恰好产生等量的ρ——自指矛盾。更严格地：如果存在某个f使ρ可被部分消灭，则反复应用f可将ρ驱至任意小，极限为零，与Paper 1的ρ命题直接矛盾。
方向二（ρ不生）。假设 ρ_E(f(o)) + ρ_f > ρ(o)。则外延化凭空创造了操作性内容。但f的定义是丢弃操作历史、保留外延结果——信息压缩，非信息生成。ρ的唯一来源是操作行为本身。f作为映射，其自身的操作性内容 ρ_f 来自"执行f"这一行为，而这一行为的操作对象是o，所以 ρ_f 不能超过o所含的操作性内容减去产物中保留的部分。
合取。ρ(o) = ρ_E(f(o)) + ρ_f。 ∎
3.3 守恒原理二（跨层跃迁守恒）
命题：对任意L层级跃迁 Lₙ→Lₙ₊₁，设 A_n 为该层的外延化提升，则：
3.4 递归守恒（链式传递）
链式展开：
原始 ρ_total(Lₙ) 不在任何层级链中衰减。它被逐层分配到各层的自指生成行为和各层的潜伏基底中。
4. 哥德尔不完备性作为ρ溢出
4.1 PA的公理化作为大规模外延化
PA的公理化是 F: O_math → E_PA。守恒原理给出：
4.2 ρ_E(E_PA) = 不可判定命题
嵌入PA形式结构中的潜伏余项，以不可判定命题的形式存在：这些命题的真值依赖于PA公理化过程中被丢弃的操作性内容。
4.3 哥德尔对角化作为ρ探测器
哥德尔的自指构造迫使 ρ_E 从潜伏变为显性。在ρ框架下，不完备定理可被重读为ρ守恒的积极探测——守恒量的ρ必然从形式系统边界溢出，不可判定命题是溢出的具体位置。
5. 证明复杂度作为ρ守恒的代价
5.1 核心主张
当外延性证明工具与命题的ρ层级不匹配时，ρ守恒迫使被丢弃的操作性内容以证明复杂度的形式回归。陈述越简洁（ρ被压缩得越密），纯外延证明就越复杂。反之，如果在正确的ρ层级看问题，证明可以极其简短。
5.2 费马大定理的ρ重读
费马大定理陈述在L₁：n≥3时 x^n+y^n=z^n 无正整数解。Wiles的证明动用了极重的L₂工具（模形式、Galois表示、椭圆曲线），几百页。
ρ框架下的直觉性解读：n=2时 x²+y²=z² 有无穷多解（勾股数），n≥3突然一组都没有。断裂发生在2——Paper 2确立的"一切分裂的种子"，L₀→L₁的唯一余项。二次结构（平方、二维、复平面）是闭合的最低条件。三次及以上，乘法的幂结构增长速度超过加法的线性结构，耦合断裂。
5.3 庞加莱猜想的ρ重读
庞加莱猜想：每个单连通闭三维流形同胚于 S³。Perelman用Ricci flow加surgery证明。
ρ框架下的解读：单连通意味着所有闭合路径可连续收缩到一点——没有拓扑余项。在ρ语言中，庞加莱猜想的判别可被读作"拓扑余项为零"的极端情形。
Perelman的Ricci flow可被重读为行为（连续变形）试图消化余项（曲率不均匀）的过程。ρ守恒在此表现为：曲率在一处被磨平，余项在另一处以奇点形式溢出。Surgery是处理ρ溢出的技术手段——切掉奇点，重新黏合，继续磨平。
6. 加法-乘法耦合的ρ读法
6.1 共同根源
加法和乘法都从后继函数S派生。加法是S的一阶迭代，乘法是二阶迭代。
6.2 方向性不对称性
张益唐2013年证明素数间距有有界上界，L₂工具持续推进未卡死。陈景润1966年证明1+2后，从1+2到1+1五十年间零进展。
ρ框架下的结构性解读：乘法→加法方向（孪生素数）是投影——从高阶结构投影到低阶度量，顺流方向。加法→乘法方向（哥德巴赫）是逆向重建——从低阶结构要求高阶原子覆盖，逆流方向。方向不对称性提供了结构解释。
7. 未解问题的ρ诊断与工作性分类
7.1 核心区分：缺当前层ρ vs 缺高层ρ
第一类：缺当前层ρ。工具在同一层但还没找到正确的闭合方式。ρ在层内流转。工作性倾向：最终被同层工具证明。
第二类：缺高层ρ。当前层的工具原则上不够，需要层间跃迁。ρ跨层流转。工作性倾向：要么被证明独立于当前公理系统，要么等到高层工具发展出来。
7.2 工作性配对一：哥德巴赫 + 孪生素数（静态耦合）
哥德巴赫：每个偶数能否被两个素数之和覆盖？
孪生素数：素数间距能否无穷次等于2？
两者在ρ框架下可被读作同一个ρ耦合的两个方向。在数论中，两个问题都属于加性数论范畴，使用类似的筛法和圆法，且哥德巴赫普遍被认为更难（与逆流理论一致）。
工作性猜测：如果ρ耦合分析方向正确，两个猜想可能可以被统一处理。
7.3 工作性配对二：黎曼 + Collatz（动态行为）
黎曼猜想：黎曼zeta函数的非平凡零点的实部都等于1/2。
Collatz猜想：迭代映射3n+1（若n为奇数）和n/2（若n为偶数）回到1。
配对逻辑：黎曼从L₂看L₁的全局分布，Collatz在L₁内部看局部纠缠。在ρ框架下，两者可被读作对同一个对象——后继操作的ρ在加法和乘法之间的分配——的不同视角。
7.4 Navier-Stokes（加法-乘法纠缠的连续版）
NS方程的非线性项 (u·∇)u 是加法结构（向量加法）和乘法结构（场自乘）的纠缠。线性部分（Stokes方程）在L₂内部完全可解，非线性纠缠不可解。
7.5 P≠NP（ρ命题的计算论类比）
P是确定性多项式时间可解。NP是非确定性多项式时间可验证。P≠NP说的是：验证（外延性操作）不可还原为求解（操作性行为）。
ρ框架下的解读：P≠NP的内容可以被读作ρ命题（行为不可被还原为图）在计算理论中的类比表现。
7.6 完整ρ诊断图
| 问题 | ρ类型 | 层级位置 | 工作性配对 | 工作性倾向 |
|---|---|---|---|---|
| 哥德巴赫 | 静态耦合（逆流） | L₁→L₂边界 | 孪生素数 | 横跨两类，可能统一处理 |
| 孪生素数 | 静态耦合（顺流） | L₁→L₂边界 | 哥德巴赫 | 偏第一类，可能统一处理 |
| 黎曼 | 全局分布 | L₂内部 | Collatz（元猜测） | 偏第一类 |
| Collatz | 离散纠缠迭代 | L₁（跨层需求） | 黎曼+NS（元猜测） | 第二类 |
| Navier-Stokes | 连续纠缠迭代 | L₂→L₄ | Collatz（类比） | 横跨两类 |
| P≠NP | ρ命题的类比 | 元层级 | 无（独立） | 可考虑独立于ZFC的可能性 |
8. ρ守恒作为ρ-算术的约束条件
ρ-算术的基本对象不是数n，而是(n, H_n)——数连同其完整生成历史。
ρ-算术用守恒原理约束"操作性内容如何流动"。公理是静态的存在声明；守恒原理是动态的流量约束。
9. 讨论
9.1 ρ守恒与物理守恒律
ρ守恒与能量守恒在结构上平行：能量不灭，在动能和势能之间转换；ρ不灭，在余项和行为之间转换。
如果ρ守恒在物理层面有对应物，可能与信息守恒（黑洞信息悖论、Susskind ER=EPR）有深层关联。
9.2 与既有数学基础的关系
ZFC：ρ守恒不是ZFC公理，不与ZFC矛盾。是关于形式化本身的元定律。
范畴论：追踪态射，比集合论更接近操作性内容，但态射仍是外延化的。ρ守恒适用。
直觉主义/构造主义：Brouwer坚持构造性证明，试图保留更多操作内容。ρ守恒解释了为什么构造主义更受限——它拒绝丢弃ρ，携带更多操作负担。
10. 结论
Paper 1命名了余项。Paper 2给出了余项的面貌。Paper 3确立了余项的定律。
π在L₁是无穷，在L₂是3.14159...。不是它变了，是看它的层级变了。ρ守恒。
如果ρ守恒框架是有生产力的，那么数论中最顽固的开放问题或许不只需要更巧妙的现有技术，也需要一种更愿意显式处理操作性内容的新语言。ρ-算术是本文提出的一个候选方向，而不是已被证明的唯一出路。