The Quantitative Identity of the Remainder: From ρ≠∅ to Euler's Formula
余项的定量身份：从 ρ≠∅ 到欧拉公式

1. The ρ-Proposition: Recap and Boundaries

Paper I established the following:

> The ρ-proposition. For any domain formalized extensionally within ZFC, the formalization structurally reduces operational performance to extensional output. Merely enriching the formal language extensionally cannot absorb that performance; it can only relocate the remainder.

The ρ-proposition distinguishes two things: the remainder (the extensional product of closure, a structural trace that has been set aside) and the act (the operational performance that makes closure happen, not exhausted by its graph). The regress argument of Paper I shows that the act cannot be reduced to a graph: every attempt to formalize the act produces a new remainder.

The ρ-proposition has a clear boundary: it is qualitative. $\rho(C,U)\ne\varnothing$ says that a remainder exists. It does not say what the remainder equals, what the act looks like, or what precise relation holds between them.

Clarification on standard closure operators. In the standard closure-operator framework, if one defines the remainder as $\mathrm{cl}(A)\setminus A$, then already-closed objects have empty remainder. The ρ of ZFCρ is not a set-difference. It is a meta-theoretic marker — marking the structural asymmetry between extensional representation and operational performance. The regress argument applies to any extensional formalization, including the formalization of the state "already closed".

---

2. π: First Complete Instance of Quantitative Remainder

2.1 Where π comes from

Consider the simplest discrete domain: the integer lattice $\mathbb{Z}^d$. Applying the closure operation of Poisson summation exchanges the lattice with its dual:

$$\theta_3(e^{-\pi t}) = t^{-1/2}\cdot\theta_3(e^{-\pi/t})$$

The Fourier self-duality of the Gaussian $e^{-\pi x^2}$ forces π into the exchange. π is not imported from outside; it is squeezed out by the closure operation itself.

2.2 Why ρ can only qualify π

The ρ-proposition tells us that closure necessarily leaves a remainder ($\rho\ne\varnothing$) and that the remainder cannot be absorbed by finitely many operations. It cannot tell us that the remainder equals $3.14159\ldots$ rather than some other transcendental. Quantitative determination requires the structure of the layer above.

2.3 The input from $L_2$: the Fourier self-dual fixed-point condition

To evaluate the integral of a one-dimensional linearly decaying function over the whole real axis, one must raise the dimension to two and introduce polar coordinates — that is, introduce the geometry of the circle — in order to close:

$$\left(\int e^{-x^2}dx\right)^2 = \int_0^{2\pi}\!\!\int_0^\infty e^{-r^2}r\,dr\,d\theta = \pi$$

The linear space is forced to close into circular geometry, and π is the measure of that forced closure.

2.4 The dual identity of π

π therefore carries a double identity: as remainder it marks the failure of closure; as geometric constant it measures the success of closure. The polar-coordinate proof of the Gaussian integral reveals the common root of the two identities: the structural tension between infinite linear space and closed circular geometry.

2.5 General principle

The existence of $\rho(L_n)$ can be read inside $L_n$, but its quantitative determination requires the structure of $L_{n+1}$. Computing $\rho(L_n)$ with the tools of $L_{n+k}$ is more efficient — Ramanujan computed π with $L_3$ tools (modular equations, singular values) at eight digits per term, far outstripping the $L_2$ tool of the Leibniz series.

---

3. The $L_1\to L_2$ Transition: Two Remainders and One Act

3.1 $i$ and π: two remainders

π = the lattice/dual-lattice facet (harmonic analysis). The discrete lattice ($L_1$) attempts to exchange with its dual by Fourier transform. The exchange is not clean, and π is squeezed out. π is the extensional product of a closure operation — a definite constant computable to arbitrary precision.

$i$ = the linear/planar facet (algebra). The real line ($L_1$: ordered, linear) attempts algebraic closure. $x^2+1=0$ has no solution on the one-dimensional real line; one must jump to the two-dimensional complex plane. $i$ is the literal remainder of the quotient ring $\mathbb{R}[x]/\langle x^2+1\rangle$ — among all mathematical constants, the ρ-identity of $i$ is the most literal and the least in need of interpretation.

3.2 The exponential map: the act

Strictly, the act is not the constant $e$ but the exponential map $x\mapsto e^x$. This paper uses $e$ to mark the normalized signature of that act.

$e^x$ is the unique function satisfying $f'=f$ with $f(0)=1$. That is: differentiation (the operation) applied to $e^x$ returns $e^x$ itself. The act equals the result. Operation and output are indistinguishable — precisely the formal signature of "the act is irreducible to its graph".

Without the exponential map, $i$ and π are two static constants with no relation between them. The exponential map $t\mapsto e^{it}$ is the action that puts them in relation — winding the real line (linear, infinite) onto the unit circle (compact, closed).

3.3 Structural summary

Role	Symbol	Mathematical identity	ZFCρ correspondent
Remainder (algebraic facet)	$i$	remainder of the quotient ring $\mathbb{R}[x]/\langle x^2+1\rangle$	ρ (extensional product)
Remainder (harmonic-analytic facet)	$\pi$	the unique constant of the Fourier self-dual fixed point	ρ (extensional product)
Act	$x\mapsto e^x$	unique solution of $f'=f$ (act = result)	act (operational performance)

3.4 Independence of $i$ and π

The definition of $i$ is purely algebraic (root of $x^2+1=0$); it requires no analysis. The definition of π is purely geometric (ratio of circumference to diameter); it requires no algebra. Within elementary algebra and geometry there is no direct bridge between the two definitions.

The weaker mathematical statement that may be retained is: π is transcendental over $\mathbb{Q}(i)$. Equivalently, if $P\in\mathbb{Q}[X,Y]$ and $P(i,\pi)=0$, then $P$ lies in the ideal $(X^2+1)$ — there is no nontrivial additional algebraic relation involving π.

〔Note that one may not write "there is no nonzero $P$ with $P(i,\pi)=0$": $P=X^2+1$ is a counterexample, since $i$ of course satisfies $x^2+1=0$.〕

The exponential map is the canonical binding source, within this trajectory, of the relation between $i$ and π.

> 〔Revised in v2〕 Version 1 wrote that "the exponential map is the sole source of the relation between the two remainders", which is too strong. There are at least two bridges that do not pass through the complex exponential: > > 1. Rotational geometry. Once the complex plane is in place, multiplication by $i$ is a quarter turn, and a quarter turn in radian measure is $\pi/2$. This bridge uses only complex multiplication and Euclidean circle geometry. > 2. Contour integration. Take $z(t)=\cos t+i\sin t$; then $dz=i\,z(t)\,dt$, so $\oint_{|z|=1}dz/z=\int_0^{2\pi}i\,dt=2\pi i$. The computation uses no complex exponential. > > Moreover SAE Mathematics Paper 4 already registers the deck-generator identity and the residue-normalisation identity of $2\pi i$ as two presentations of a single statement, so "sole source" is unsupported from within the series as well. "No algebraic relation" does not amount to "no geometric, analytic, or topological bridge". See SAE Mathematics Paper 5 §§ 9.9 and 11.4, Erratum 11.

---

4. Euler's Formula: Structural Identity of Act Binding Remainders

4.1 Precise mechanism

The exponential map $t\mapsto e^{it}$ is a covering map from $\mathbb{R}$ to $S^1$ with kernel $2\pi\mathbb{Z}$. The exact sequence:

$$0\to\mathbb{Z}\xrightarrow{\ \times 2\pi\ }\mathbb{R}\xrightarrow{\ e^{i(\cdot)}\ }S^1\to 1$$

What $e^{i\pi}$ does is this: the exponential map runs in the orthogonal dimension opened by the algebraic remainder ($i$), traverses the half-turn distance measured by the harmonic-analytic remainder (π), and arrives at the antipode of the starting point, $-1$.

$$e^{i\pi}+1=0$$

4.2 ρ re-reading

Euler's formula can be re-read as: one act (the exponential map) applied to two remainders, producing closure. The exponential map binds two static failures of closure — a line cannot become a plane ($i$), a lattice cannot dualise cleanly (π) — into one dynamic success of closure (a half-turn to the antipode), cancelling against the unit to zero.

4.3 Remainders are not defects

The production of $i$ and π is not a collapse of mathematics but a generative product of a domain extending toward its natural completion.

A remainder is not a crack but the seed of the next dimension. The act is not repair but growth.

---

5. Cross-Layer Remainders and Layer-Internal Remainders

Cross-layer remainders: produced by $L_n\to L_{n+1}$ transitions. Examples: $i$, π ($L_1\to L_2$). The irreducibility of cross-layer remainders is easier to prove (transcendence of π, Lindemann 1882).

Layer-internal remainders: residues left when a closure operation within the same layer fails to fully close. Example: Catalan's constant $G=\beta(2)$. The Dirichlet beta function at odd arguments yields clean multiples of π (self-duality fully closes), but at even arguments self-duality is insufficient, leaving new periods. $G$ is an $L_2$-internal remainder. The irreducibility of layer-internal remainders is harder to prove (irrationality of $G$ remains unproven).

Working conjecture: the irreducibility of layer-internal remainders may be harder to establish than that of cross-layer remainders. $G$ is the first test case.

---

6. Structural Re-reading of Known Results: $L_2\to L_3$

6.1 Two remainders and one act

Remainder 1: formal/completeness facet (the boundary of proof). Gödel's incompleteness theorem (1931): any sufficiently strong consistent formal system contains undecidable propositions. Remainder = the class of undecidable sentences.

〔v2 cross-reference: this class is made precise in SAE Mathematics Paper 5 as $\mathrm{Ind}_T=\{\varphi: T\nvdash\varphi\text{ and }T\nvdash\neg\varphi\}$, of complexity $\Pi^0_1$, degree $0'$, and $\Pi^0_1$-many-one complete. The set contains both sides of each independent pair, and is not the same object as "the true but unprovable", $U_T=\mathrm{Th}(\mathbb{N})\setminus\mathrm{Th}(T)$, whose degree is $0^{(\omega)}$. The present paper's "class of undecidable sentences" means the former. See Paper 5 § 12.3. This item does not alter the original text; it supplies a cross-reference.〕

Remainder 2: linguistic/semantic facet (the boundary of expression). Tarski's undefinability theorem (1936): a sufficiently strong consistent formal language cannot define its own complete truth predicate within itself. Remainder = there is no single complete truth predicate within the base language realising all levels uniformly — the family of stratified truth predicates has no unifying member.

Note that in an external metatheory there is an effective procedure producing, for each $n$, the code of the $n$-th partial truth predicate $\mathrm{Tr}_{\Sigma_n}$; what does not exist is a unifier for all levels inside the base language. Its witness is complete arithmetical truth $\mathrm{Th}(\mathbb{N})$, of degree $0^{(\omega)}$.

> 〔Revised in v2〕 Version 1 wrote "Remainder = the class of $\Sigma_1$ truth predicates". That formulation requires disambiguation, and the outcome of disambiguating it is replacement rather than clarification: > > - Read as the stratified family having no unifying member, it is Tarski's theorem itself, and the reading is correct; > - Read as the set of true $\Sigma_1$ sentences $\mathrm{Th}_{\Sigma_1}(\mathbb{N})$, it fails: for each $n\ge 1$ there is a partial truth predicate $\mathrm{Tr}_{\Sigma_n}$ of complexity $\Sigma_n$, with PA proving the corresponding T-biconditionals formula by formula, so $\mathrm{Th}_{\Sigma_1}(\mathbb{N})$ is definable and effectively enumerable, of degree $0'$ rather than $0^{(\omega)}$, and cannot carry undefinability. > > Version 1 explicitly writes $\Sigma_1$, and construing that as the full stratified family is too charitable; the present item therefore replaces the original object with the correct one. See Paper 5 § 11.4, Erratum 8.

Act: algorithmic/computational facet. The diagonalization / universal evaluation operation (a program running its own encoding as input). Chaitin's constant $\Omega_U$ (halting probability relative to a prefix-free universal machine $U$) is the holographic measure of this act over the entire program domain — just as $e$ is the normalized signature of the exponential map, $\Omega_U$ is the normalized condensation of the diagonalization operation.

6.2 Coincidence in Turing-degree space

In the Turing-degree space $\mathcal{D}$:

$$[\Omega_U]_T = [K]_T = [\mathrm{Th}_{\Sigma_1}]_T = 0'$$

Here $K$ is the halting set, $\mathrm{Th}_{\Sigma_1}$ the set of Gödel numbers of true $\Sigma_1$ sentences, and $\Omega_U$ Chaitin's constant. These three degree equivalences are known results in recursion theory (Chaitin 1975, Post 1944).

The re-reading offered here is: two products of the universal-evaluation act ($K$ and $\Omega_U$) coincide in the degree quotient with the first non-trivial fragment of the truth hierarchy ($\mathrm{Th}_{\Sigma_1}$).

Two misreadings must be guarded against:

1. The equation does not contain the complete remainder 2. $\mathrm{Th}(\mathbb{N})$ has degree $0^{(\omega)}$ and cannot enter this single point; and $\mathrm{Th}_{\Sigma_1}$ is a successfully captured fragment (definable, effectively enumerable), not a member of any remainder.
2. "Syntactic side" cannot be read off $0'$. The degree quotient forgets provenance — the same $0'$ is realised jointly by $K$, $\Omega_U$, $\mathrm{Th}_{\Sigma_1}$, $\mathrm{Th}(T)$, $\mathrm{Ind}_T$ and other heterogeneous objects.

> 〔Revised in v2〕 Version 1 read the equation as "two remainders and one act, departing from different directions, collapse to the same boundary point". That reading does not stand, for the two reasons above. The equation itself is unchanged, letter for letter; what changes is its interpretation — and the item is therefore registered as a substantive revision rather than a disambiguation. See Paper 5 § 8.2 and § 11.4, Erratum 9.

6.3 The return-to-zero mechanism: relativization

$0'$ is irreversible within $L_2$: there is no $A$-algorithm that uniformly decides, for all $A$-relative programs, whether they halt, and that uniform self-interrogation produces the strictly higher $A'$. The only way to return to zero is relativization: absorbing $0'$ itself as an oracle:

$$\deg^{0'}(0')=0$$

> 〔Revised in v2〕 Version 1 wrote "the Turing-degree space is an upper semilattice; $0'$ cannot be 'subtracted' within $L_2$". The semilattice property (joins always exist, meets need not — there are pairs of degrees with no infimum) bears no generative relation to the irreversibility of $0'$, which comes from self-application; and "meets need not exist" does not yield "cannot be subtracted" (a meet is a common lower bound, not a subtraction, and a pair of degrees lacking a meet produces no uniform, canonical remainder). Juxtaposing the two invites a misreading, so they are stated separately here. > > This item falls outside the four registered in Paper 5 § 11.4; it is a clarification added by v2 on its own account.

	$L_1\to L_2$	$L_2\to L_3$
Quotient space	$S^1$ ($\mathbb{R}$ mod $2\pi\mathbb{Z}$)	$\mathcal{D}$ (sets mod Turing reducibility)
Collapse point	$-1\in S^1$	$0'\in\mathcal{D}$
Return-to-zero	algebraic inverse: $+1$	oracle relativization: $(\bmod\ 0')$
Closure equation	$e^{i\pi}+1=0$	$\deg^{0'}(0')=0$
Nature of return	symmetric (algebraic cancellation)	asymmetric (transition to higher layer)

The return-to-zero at $L_1\to L_2$ is symmetric: $-1$ and $+1$ are algebraic objects on a par. At $L_2\to L_3$ it is asymmetric: $0'$ cannot be reversed within $L_2$ but can only be absorbed by $L_3$ from outside. The system is forced to transition upward, hard-coding the insurmountable wall as the floor.

> Two notes (added in v2) > > Note 1 · $0'$ is the collapse point, not a boundary of the degree profile. The $L_3$ degree profile given in SAE Mathematics Paper 5 § 5.2 is the ideal of arithmetical degrees $\mathrm{AR}=\{d:\exists n\ d\le_T 0^{(n)}\}$, downward closed and closed under the jump, so $0'$ is merely one of its members. Moreover, by Spector's exact pair theorem, $\mathrm{AR}$ has no least upper bound in $\mathcal{D}$: a countable degree ideal $I$ has an exact pair $(a,b)$ with $I=\{d: d\le a\text{ and }d\le b\}$; if $I$ had a least upper bound $c$ then $c\le a,b$, whence $c\in I$, so $c$ would be the greatest element of $I$, contradicting $0^{(n)}<_T 0^{(n+1)}$. Hence one may not say "the supremum of the profile is $0^{(\omega)}$" — $0^{(\omega)}$ is the least uniform upper bound of the finite jump tower (relative to a fixed effective enumeration), not the supremum of $\mathrm{AR}$. > > Note 2 · This return-to-zero equation does not exhibit the complete remainder 2 as a non-deletable input. $\mathrm{Th}(\mathbb{N})$ (degree $0^{(\omega)}$) does not appear in the equation above, which therefore has not been shown to instantiate "closure of two independent remainders". This is a verdict of insufficient evidence, not a theorem that the equation is one-sided — the latter would require a binary closure map already defined, and none is. See Paper 5 §§ 10.3 and 15.2.

6.4 Remainders become axioms of the next layer

What a lower-dimensional system cannot close becomes the ground zero of the higher-dimensional one. Development is not the elimination of remainders but their conversion into the floor of the next layer.

---

7. A Research Programme: From Self-Referential Generation to a Unified Inter-Layer Act

What follows no longer states known mathematical results but advances an interpretive programme. The following is structural conjecture, not completed proof.

7.1 $L_0\to L_1$: why closure requires two remainders

$L_0$ is pure binary distinction: presence/absence, 0 and 1. $L_0$ has only one distinction; closure can fail in only one direction, producing a single remainder: 2 — the first object squeezed out of the distinction between 0 and 1.

$L_0\to L_1$ has an act: the successor operation $S(n)=n\cup\{n\}$. But $L_0\to L_1$ has no closure equation. The successor can only push forward: $0\to1\to2\to3\to\cdots$, never returning.

Closure requires two independent failure directions so that the act can bind them into a loop returning to zero. With only one direction, there is nothing to cancel against. A single remainder can only unfold, not close. The non-closability of a single remainder is the source of infinity.

7.2 Self-referential generation: candidate definition for the unified act

The core research programme proposed by this paper: the act driving all layer transitions may be the same class of operation — self-referential generation, an operation that incorporates itself into its own output.

Layer	Guise	Signature
$L_0\to L_1$	successor $S(n)=n\cup\{n\}$	incorporates self as element
$L_1\to L_2$	exponential map ($f'=f$)	operation outputs itself
$L_2\to L_3$	diagonalization (Ω as condensation)	program judges itself
$L_3\to L_4$	time (causal succession)	each moment contains the previous (irreversible)

7.3 The full picture

	$L_0\to L_1$	$L_1\to L_2$	$L_2\to L_3$	$L_3\to L_4$
Act	successor $S$	exponential map	diagonalization (Ω)	time
Remainders	1	2	2	2 (predicted)
Remainder 1	2 (distinction)	$i$ (algebra)	Gödel sentence (proof)	? (observation)
Remainder 2	—	π (harmonic analysis)	Tarski truth predicate (semantics)	? (spacetime geometry)
Result	unfolding ($\infty$)	$e^{i\pi}+1=0$	$\deg^{0'}(0')=0$	closure (predicted)

〔v2 note: the entry "Remainders = 2" in the $L_2\to L_3$ column is a structural claim of this paper. By Note 2 of § 6.3, the existing closure equation has not been shown to take both remainders as non-deletable inputs; should that argument never be supplied, the transition should be called a coupled re-zeroing of two boundaries rather than a closure of two independent remainders. See Paper 5 §§ 9.7 and 15.2.〕

---

8. Division of Labour

Completed in this paper

1. Established that the ρ-proposition is qualitative; quantitative determination of the remainder requires a higher layer.
2. Presented π as the first complete instance of remainder quantification.
3. Repositioned the exponential map (marked by $e$) as "act", distinguished from $i$ and π (remainders). $f'=f$ is the formal expression of act-irreducible-to-graph.
4. Confirmed that $i$ and π have no direct bridge within elementary algebra and geometry; the exponential map is the canonical binding source, within this trajectory, of the relation between them (revised in v2; version 1 wrote "sole source" — see § 3.4).
5. Re-read Euler's formula as the structural identity "one act binding two remainders to produce closure".
6. Juxtaposed the residue theorem, Gauss–Bonnet, and Atiyah–Singer as isomorphic patterns (juxtaposition, not derivation).
7. Distinguished cross-layer from layer-internal remainders; proposed the working conjecture that layer-internal remainders may be harder to prove irreducible.
8. Identified the $L_2\to L_3$ transition's "two remainders + one act" structure; re-read the coincidence in Turing-degree space and the relativization return-to-zero (revised in v2; version 1 wrote "collapse" — see § 6.2).
9. Through analysis of $L_0\to L_1$ (single remainder, only unfolding), proposed the structural argument that closure requires two remainders.

Research programme proposed

Core programme: the successor, the exponential map, the diagonalization operation, and time may be interpretable as unfoldings of the same class of self-referential generation across layers. For this to become a rigorous mathematical proposition, additional formalization is required.

Prediction: the $L_3\to L_4$ transition involves an act (time) binding two remainders to produce closure; the closure equation requires the $L_4$ perspective.

Registered as incomplete in v2

• Whether the closure equation of $L_2\to L_3$ satisfies "both remainders non-deletable" has not been shown (§ 6.3, Note 2);
• Should that argument never be supplied, the wording for the transition must change (§ 7.3, note).

---

Related Papers

• ZFCρ Paper I: On the Remainder of Choice: A Meta-Theoretic Proposition Concerning ZFC. DOI: 10.5281/zenodo.18914682
• SAE Mathematics Paper 4 ($L_2$). DOI: 10.5281/zenodo.21466722
• SAE Mathematics Paper 5 ($L_3$). DOI: 10.5281/zenodo.21538494 — the source of the present v2 revisions

1. ρ 命题回顾与边界

第一篇论文确立了以下结论:

> ρ 命题. 对于任何在 ZFC 内部外延式形式化的域, 操作的形式化在结构上将操作性行为还原为其外延输出. 仅仅外延式地丰富形式语言不能完全吸收那个行为, 只能重新安置余项.

ρ 命题区分了两样东西: 余项 (外延闭合的产物, 被搁置的结构性痕迹) 和行为 (使得闭合得以发生的操作性施行, 不被图所穷尽). 第一篇论文的回退论证表明行为不可被还原为图: 每一次试图形式化行为的努力都产生一个新的余项.

但 ρ 命题有一个明确的边界: 它是定性的. $\rho(C,U)\ne\varnothing$ 说余项存在. 它不说余项等于什么, 不说行为长什么样, 更不说余项和行为之间有什么精确关系.

关于标准闭包算子的澄清. 在标准闭包算子框架中, 若把余项定义为 $\mathrm{cl}(A)\setminus A$, 则已闭合的对象余项为空. ZFCρ 的 ρ 不是集合差, 而是一个元理论标记 —— 标记外延表示与操作性施行之间的结构不对称. 回退论证适用于任何外延式形式化, 包括对「已经闭合」这一状态的形式化.

---

2. π: 余项定量化的第一个完整样例

2.1 π 从哪来

考虑最简单的离散域: 整数格 $\mathbb{Z}^d$. 对这个 $L_1$ 对象执行闭合操作 —— Poisson 求和 —— 把格与对偶格交换. Gaussian $e^{-\pi x^2}$ 的 Fourier 自对偶性在交换过程中把 π 带进来. π 不是从外部加入的, 是闭合操作本身挤出来的不可还原量.

$$\theta_3(e^{-\pi t}) = t^{-1/2}\cdot\theta_3(e^{-\pi/t})$$

2.2 ρ 为什么只能给 π 定性

ρ 命题告诉我们: 闭合必然留下余项 ($\rho\ne\varnothing$), 余项不可被有限操作吸收. 但它不能告诉我们余项等于 $3.14159\ldots$ 而非其他超越数. 定量刻画需要高一层的结构.

2.3 $L_2$ 的输入: Fourier 自对偶不动点条件

计算一维线性衰减函数的全实轴积分, 必须升维到二维并引入极坐标 —— 即引入圆的几何 —— 才能闭合:

$$\left(\int e^{-x^2}dx\right)^2 = \int_0^{2\pi}\!\!\int_0^\infty e^{-r^2}r\,dr\,d\theta = \pi$$

线性空间被迫闭合到圆形几何, π 就是这个强制闭合的度量.

2.4 π 的双重身份

π 因此具有双重身份: 作为余项, 它标记闭合的失败; 作为几何常数, 它度量闭合的成功. Gaussian 积分的极坐标证明揭示了两个身份的共同根源: 无穷线性空间与闭合圆形几何之间的结构性张力.

2.5 一般原则

$\rho(L_n)$ 的存在性可以在 $L_n$ 内部读出, 但其定量刻画需要 $L_{n+1}$ 的结构. 用 $L_{n+k}$ 的工具去计算 $\rho(L_n)$, 效率更高 —— Ramanujan 用 $L_3$ 工具 (模方程、奇异值) 算 π, 每项 8 位数字, 远超 $L_2$ 工具 (莱布尼茨级数).

---

3. $L_1\to L_2$ 跃迁: 两个余项与一个行为

3.1 $i$ 和 π: 两个余项

π = 格／对偶格面向 (调和分析). 离散格 ($L_1$) 试图通过 Fourier 变换与对偶格交换. 交换不干净, 挤出 π. π 是闭合操作的外延产物 —— 一个确定的常数, 可以被计算到任意精度.

$i$ = 线性／平面面向 (代数). 实数轴 ($L_1$, 有序、线性) 试图代数闭合. $x^2+1=0$ 在一维实轴上无解, 需要跳到二维复平面. $i$ 是商环 $\mathbb{R}[x]/\langle x^2+1\rangle$ 的字面余项 —— 在所有数学常数中, $i$ 的 ρ 身份是最字面、最不需要解释的.

3.2 指数映射: 行为

严格说, 行为不是常数 $e$ 本身, 而是指数映射 $x\mapsto e^x$. 本文用 $e$ 标记这一行为的规范化签名.

$e^x$ 是唯一满足 $f'=f$ ($f(0)=1$) 的函数. 这意味着: 微分 (操作) 作用于 $e^x$, 得到的是 $e^x$ 本身. 行为等于结果. 操作和它的输出不可区分 —— 这正是「行为不可被还原为图」的形式化签名.

没有指数映射, $i$ 和 π 只是两个静态常数, 之间没有关系. 指数映射 $t\mapsto e^{it}$ 是让它们发生关系的动作 —— 把实数轴 (线性、无穷) 卷到单位圆 (紧致、闭合).

3.3 结构总结

角色	记号	数学身份	ZFCρ 对应
余项 (代数面向)	$i$	$\mathbb{R}[x]/\langle x^2+1\rangle$ 的商环余项	ρ (外延产物)
余项 (调和分析面向)	$\pi$	Fourier 自对偶不动点唯一常数	ρ (外延产物)
行为	$x\mapsto e^x$	$f'=f$ 唯一解 (行为 = 结果)	act (操作性施行)

3.4 $i$ 和 π 的独立性

$i$ 的定义是纯代数的 ($x^2+1=0$ 的解), 不需要分析. π 的定义是纯几何的 (圆周与直径之比), 不需要代数. 在基本代数与初等几何的范畴内, 两个定义之间没有直接桥梁.

可保留的较弱数学陈述是: π 在 $\mathbb{Q}(i)$ 上超越. 等价地, 若 $P\in\mathbb{Q}[X,Y]$ 且 $P(i,\pi)=0$, 则 $P$ 必属于理想 $(X^2+1)$ —— 即不存在涉及 π 的非平凡额外代数关系.

〔须注意: 不可写「不存在非零 $P$ 使 $P(i,\pi)=0$」—— $P=X^2+1$ 即是反例, 因 $i$ 自身当然满足 $x^2+1=0$.〕

指数映射是 $i$ 与 π 关系在本 trajectory 中的规范绑定来源.

> 〔v2 修订〕 v1 写作「指数映射是使另外两个余项发生关系的唯一原因」, 该措辞过强. 至少有两座不经复指数的桥: > > 1. 旋转几何: 复平面构造完成后, 乘以 $i$ 是四分之一转, 而弧度制中四分之一转是 $\pi/2$. 该桥只用复乘法与欧氏圆周几何; > 2. 围道积分: 取 $z(t)=\cos t+i\sin t$, 则 $dz=i\,z(t)\,dt$, 故 $\oint_{|z|=1}dz/z=\int_0^{2\pi}i\,dt=2\pi i$. 该计算不用复指数. > > 且 SAE 数学 Paper 4 已把 $2\pi i$ 的覆叠生成元身份与留数归一身份登记为同一陈述的两种呈现 —— 从系列内部亦不支持「唯一」. 「没有代数关系」不等于「没有任何几何、解析或拓扑桥」. 详 SAE 数学 Paper 5 § 9.9 与 § 11.4 勘误 11.

---

4. 欧拉公式: 行为绑定余项的结构恒等式

4.1 精确机制

指数映射 $t\mapsto e^{it}$ 是从 $\mathbb{R}$ 到 $S^1$ 的覆盖映射, 核是 $2\pi\mathbb{Z}$. 正合列:

$$0\to\mathbb{Z}\xrightarrow{\ \times 2\pi\ }\mathbb{R}\xrightarrow{\ e^{i(\cdot)}\ }S^1\to 1$$

$e^{i\pi}$ 做的事情是: 指数映射在代数余项 ($i$) 打开的正交维度中运行, 走过调和分析余项 (π) 度量的半圈距离, 到达起点的对径点 $-1$.

$$e^{i\pi}+1=0$$

4.2 ρ 重读

欧拉公式可被重读为: 一个行为 (指数映射) 作用于两个余项, 产生闭合. 指数映射把两个静态的闭合失败 —— 线不能变面 ($i$)、格不能干净对偶 (π) —— 绑定为一次动态的成功闭合 (半圈旋转到对径点), 与单位元抵消为零.

4.3 余项不是缺陷

$i$ 和 π 的产生不是数学的崩溃, 而是域向其自然完成形式扩展时的生成性产物.

余项不是裂缝, 是生成下一个维度的种子. 行为不是修补, 是生长.

---

5. 跨层余项与层内余项

跨层余项: $L_n\to L_{n+1}$ 跃迁产生的余项. 如 $i$, π ($L_1\to L_2$). 跨层余项的不可还原性更容易证明 (π 的超越性, Lindemann 1882).

层内余项: 同一层闭合操作未能完全闭合时留下的残余. 如 Catalan 常数 $G=\beta(2)$. $G$ 是 $L_2$ 内部的余项, 没有跨出形式连续统的边界. 层内余项的不可还原性更难证明 ($G$ 的无理性至今未证).

工作性猜想: 层内余项的不可还原性证明可能比跨层余项更棘手. Catalan 常数 $G$ 只是这一猜想的首个测试点.

---

6. 已知数学结果的结构重读: $L_2\to L_3$

6.1 两个余项与一个行为

余项一: 形式／完备面向 (证明的边界). 哥德尔不完备定理 (1931): 任何足够强的一致形式系统包含在其内部不可判定的命题. 余项 = 不可判定命题类.

〔v2 交叉引用: 该类在 SAE 数学 Paper 5 中被精确化为 $\mathrm{Ind}_T=\{\varphi: T\nvdash\varphi\ \text{且}\ T\nvdash\neg\varphi\}$, 复杂度 $\Pi^0_1$, 度 $0'$, 且 $\Pi^0_1$-many-one 完全. 该集含每对独立句的两侧, 与「真而不可证」$U_T=\mathrm{Th}(\mathbb{N})\setminus\mathrm{Th}(T)$ 不是同一对象 ($U_T$ 的度为 $0^{(\omega)}$). 本文的「不可判定命题类」指前者. 详 Paper 5 § 12.3. 本条不改原文, 只作交叉引用.〕

余项二: 语言／语义面向 (表达的边界). Tarski 不可定义性定理 (1936): 足够强的一致形式语言不能在自身内部定义自身的完整真值谓词. 余项 = 不存在原语言内的单一完整真谓词统一实现所有层级 —— 即分层真谓词族无统一成员.

须注意: 在外部元理论中存在有效过程, 给定 $n$ 产生第 $n$ 层部分真谓词 $\mathrm{Tr}_{\Sigma_n}$ 的代码; 不存在的是原语言内的全层统一器. 其见证是完整算术真理 $\mathrm{Th}(\mathbb{N})$, 度为 $0^{(\omega)}$.

> 〔v2 修订〕 v1 写作「余项 = $\Sigma_1$ 真值谓词类」. 该表述须消歧, 而消歧的结果是替换而非澄清: > > - 读作分层真谓词族无统一成员, 即是 Tarski 定理本身, 该读法正确; > - 读作真 $\Sigma_1$ 句的集合 $\mathrm{Th}_{\Sigma_1}(\mathbb{N})$, 则不成立 —— 对每个 $n\ge 1$ 存在复杂度为 $\Sigma_n$ 的部分真谓词 $\mathrm{Tr}_{\Sigma_n}$, PA 逐公式证其 T-biconditional; 故 $\mathrm{Th}_{\Sigma_1}(\mathbb{N})$ 是可定义且可有效枚举的, 其度为 $0'$ 而非 $0^{(\omega)}$, 不能承担不可定义性. > > v1 明确写 $\Sigma_1$, 把它解释为全层分层族过于善意, 故本次是以正确对象替换原对象. 详 Paper 5 § 11.4 勘误 8.

行为: 算法／计算面向. 对角化／通用求值操作 (程序将自身编码作为输入并运行). Chaitin 常数 $\Omega_U$ (相对于某个 prefix-free 通用机 $U$ 的停机概率) 是这一行为在整个程序域上的全息测度 —— 正如 $e$ 是指数映射的规范化签名, $\Omega_U$ 是对角化操作的规范化凝结.

6.2 图灵度空间中的坍缩

在图灵度空间 $\mathcal{D}$ 中:

$$[\Omega_U]_T = [K]_T = [\mathrm{Th}_{\Sigma_1}]_T = 0'$$

其中 $K$ 是停机集, $\mathrm{Th}_{\Sigma_1}$ 是真 $\Sigma_1$ 句的 Gödel 数集, $\Omega_U$ 是 Chaitin 常数. 这三个图灵度等价是递归论的已知结果 (Chaitin 1975, Post 1944).

本文的重读是: 通用求值行为的两个产物 ($K$ 与 $\Omega_U$) 与真理分层的第一个非平凡片段 ($\mathrm{Th}_{\Sigma_1}$) 在图灵度商空间中重合于同一点.

须防两处误读:

1. 该式不含完整的余项二. $\mathrm{Th}(\mathbb{N})$ 的度是 $0^{(\omega)}$, 进不了这个单点; 而 $\mathrm{Th}_{\Sigma_1}$ 是被成功捕获的片段 (可定义、可有效枚举), 不是余项的成员;
2. 不可由 $0'$ 读出「语法侧」. 度商遗忘来源 —— 同一个 $0'$ 由 $K$、$\Omega_U$、$\mathrm{Th}_{\Sigma_1}$、$\mathrm{Th}(T)$、$\mathrm{Ind}_T$ 等异质对象共同实现.

> 〔v2 修订〕 v1 读作「两个余项和一个行为从不同方向出发, 在图灵度商空间中坍缩为同一个边界点」. 该读法不成立 (理由见上两条). 等式本身一字未改, 改的是它的解释 —— 故本条登记为实质修订而非消歧. 详 Paper 5 § 8.2 与 § 11.4 勘误 9.

6.3 归零机制: 相对化

$0'$ 在 $L_2$ 内部不可逆: 不存在由 $A$ 自身统一决定所有 $A$-相对程序是否停机的 $A$-算法, 该统一自询问产生严格更高的 $A'$. 归零的唯一方式是相对化: 把 $0'$ 本身作为预言机吞入系统:

$$\deg^{0'}(0')=0$$

> 〔v2 修订〕 v1 写作「图灵度空间是上半格, $0'$ 不能在 $L_2$ 内部被『减去』」. 上半格性质 (join 总存在而 meet 未必, 即存在无下确界的度对) 与 $0'$ 的不可逆性无生成关系 —— 后者来自自应用, 而「meet 不存在」推不出「不能相减」(meet 是共同下界, 不是减法; 缺 meet 的度对不产生任何统一的、规范的余项). 二者并置易致误读, 故分开陈述. > > 该条超出 Paper 5 § 11.4 已登记的四条, 是 v2 自行追加的澄清.

	$L_1\to L_2$	$L_2\to L_3$
商空间	$S^1$ ($\mathbb{R}$ mod $2\pi\mathbb{Z}$)	$\mathcal{D}$ (集合 mod 图灵可归约)
坍缩点	$-1\in S^1$	$0'\in\mathcal{D}$
归零方式	代数逆: $+1$	预言机相对化: $(\bmod\ 0')$
闭合方程	$e^{i\pi}+1=0$	$\deg^{0'}(0')=0$
归零性质	对称 (代数抵消)	不对称 (向高层过渡)

$L_1\to L_2$ 的归零是对称的 —— $-1$ 和 $+1$ 是对等的代数对象. $L_2\to L_3$ 的归零是不对称的 —— $0'$ 不能在 $L_2$ 内部被逆转, 只能被 $L_3$ 从外部吞噬. 系统被迫跃迁到更高层, 把不可逾越的墙硬编码为地板.

> 两条注 (v2 增补) > > 注一 · $0'$ 是坍缩点, 不是度剖面的边界. SAE 数学 Paper 5 § 5.2 给出的 $L_3$ 度剖面是算术度理想 $\mathrm{AR}=\{d:\exists n\ d\le_T 0^{(n)}\}$, 它下闭且对 jump 封闭, 故 $0'$ 只是其中一个度. 且由 Spector 的 exact pair 定理, $\mathrm{AR}$ 在 $\mathcal{D}$ 中无最小上界 —— 可数度理想 $I$ 有 exact pair $(a,b)$ 使 $I=\{d:d\le a\ \text{且}\ d\le b\}$; 若 $I$ 有最小上界 $c$ 则 $c\le a,b$ 故 $c\in I$, 即 $c$ 为 $I$ 的最大元, 与 $0^{(n)}<_T 0^{(n+1)}$ 矛盾. 故不可说「剖面的上确界是 $0^{(\omega)}$」 —— $0^{(\omega)}$ 是有限 jump 塔的最小一致上界 (相对于固定的有效枚举), 不是 $\mathrm{AR}$ 的上确界. > > 注二 · 该归零式未把完整余项二展示为不可删除的输入. 上式中不出现 $\mathrm{Th}(\mathbb{N})$ (度 $0^{(\omega)}$), 故该式尚未被证明构成「两独立余项闭合」的实例. 这是证据不足之判定, 不是「已证明该式单侧」的定理 —— 后者须先有一个已定义的二元闭合映射, 而尚无. 详 Paper 5 § 10.3 与 § 15.2.

6.4 余项是下一层的公理

低维系统闭合不了的东西, 成为高维系统的基底零点. 发展不是消除余项, 而是把余项变成下一层的地面.

---

7. 研究纲领性猜想: 从自指生成到层间行为的统一

下文不再陈述已知数学结果, 而提出一条解释性纲领. 以下论述是结构猜想, 而非已完成证明.

7.1 $L_0\to L_1$: 为什么闭合需要两个余项

$L_0$ 是纯粹的二元区分: 有／无, 0 和 1. $L_0$ 只有一个区分, 因此闭合操作只能沿一个方向失败, 只能产生一个余项: 2 —— 第一个从「有／无」的区分中被挤出来的东西.

$L_0\to L_1$ 有行为: 后继操作 $S(n)=n\cup\{n\}$. 但 $L_0\to L_1$ 没有闭合方程. 后继只能向前推: $0\to1\to2\to3\to\cdots$, 永远不回头.

闭合需要两个独立的失败方向, 行为才能把它们绑成一个环回到零. 只有一个方向, 没有东西可以跟它对消. 一个余项只能展开, 不能闭合. 单一余项的不可闭合性就是无穷的来源.

7.2 自指生成: 统一行为的候选定义

本文提出的核心研究纲领: 驱动所有层间跃迁的可能是同一类行为 —— 自指生成 (self-referential generation): 操作将自身纳入自身的输出.

层	自指生成的面目	签名
$L_0\to L_1$	后继 $S(n)=n\cup\{n\}$	把自身包含为元素
$L_1\to L_2$	指数映射 ($f'=f$)	操作输出自身
$L_2\to L_3$	对角化 (Ω 为其凝结)	程序判定自身
$L_3\to L_4$	时间 (因果后继)	时刻包含前一时刻 (不可逆)

7.3 纲领的完整图景

	$L_0\to L_1$	$L_1\to L_2$	$L_2\to L_3$	$L_3\to L_4$
行为	后继 $S$	指数映射	对角化 (Ω)	时间
余项数	1	2	2	2 (预测)
余项一	2 (区分)	$i$ (代数)	哥德尔句 (证明)	? (观测)
余项二	—	π (调和分析)	Tarski 真谓词 (语义)	? (时空几何)
结果	展开 ($\infty$)	$e^{i\pi}+1=0$	$\deg^{0'}(0')=0$	闭合 (预测)

〔v2 注: $L_2\to L_3$ 一列的「余项数 = 2」是本文的结构主张. 按 § 6.3 注二, 现有的闭合方程尚未被证明同时以两余项为不可删除输入; 若该论证始终不能给出, 该跃迁应改称「两边界耦合归零」而非「两独立余项闭合」. 详 Paper 5 § 9.7 与 § 15.2.〕

---

8. 分工

本文完成

1. 确立 ρ 命题是定性的, 余项的定量刻画需要高一层结构.
2. 给出 π 作为余项定量化的第一个完整样例.
3. 将指数映射 (以 $e$ 标记) 定位为「行为」, 与 $i$、π (余项) 区分. $f'=f$ 是行为不可还原为图的形式化表达.
4. 确认 $i$ 和 π 在初等代数与几何的范畴内没有直接桥梁; 指数映射是二者关系在本 trajectory 中的规范绑定来源 (v2 修订: v1 写作「唯一来源」, 见 § 3.4).
5. 重读欧拉公式为「一个行为绑定两个余项产生闭合」的结构恒等式.
6. 并置留数定理、Gauss–Bonnet、Atiyah–Singer 为同构模式 (并置, 非递推).
7. 区分跨层余项与层内余项, 提出层内余项可能更难证明不可还原性的工作性猜想.
8. 识别 $L_2\to L_3$ 跃迁的「两个余项 + 一个行为」结构, 重读其在图灵度空间中的重合与相对化归零 (v2 修订: v1 写作「坍缩」, 见 § 6.2).
9. 通过分析 $L_0\to L_1$ (只有一个余项, 只能展开不能闭合), 提出闭合需要两个余项的结构性论证.

本文提出的研究纲领

核心研究纲领: 后继、指数映射、对角化操作、时间, 可以被解释为同一类自指生成行为在不同层的展开. 这一解释若要成为严格数学命题, 仍需额外的形式化工作.

预测: $L_3\to L_4$ 跃迁中, 行为 (时间) 绑定两个余项产生闭合, 闭合方程需要 $L_4$ 视角.

v2 未完成而登记的

• $L_2\to L_3$ 的闭合方程是否满足「两余项皆不可删除」, 尚未被证明 (§ 6.3 注二);
• 若该论证始终不能给出, 该跃迁的措辞须改 (§ 7.3 注).

---

关联论文

• ZFCρ Paper I: 《论选择的余项: 关于 ZFC 的一个元理论命题》 DOI: 10.5281/zenodo.18914682
• SAE 数学 Paper 4 ($L_2$): DOI: 10.5281/zenodo.21466722
• SAE 数学 Paper 5 ($L_3$): DOI: 10.5281/zenodo.21538494 — 本次 v2 修订的来源