Relativistic Algebra over Finite Ring Continuum

We present a formal reconstruction of the conventional number systems, including integers, rationals, reals, and complex numbers, based on the principle of relational finitude over a finite field \( \Fp \). Rather than assuming actual infinity, we define arithmetic and algebra as observer-dependent constructs grounded in finite field symmetries. Consequently, we formulate relational analogues of the conventional number classes, expressed relationally with respect to a chosen reference frame.
The rows state the paper’s claims in their current form. The source column names the body statement that carries each row and links its witnesses as tokens — py, lean and leanML, each followed by the row’s label: the python check that decides the row, its declaration in the core (no axioms) and its declaration on Mathlib.
#DescriptionStatus Sources
A. Inputs
A1p01001Finite-field algebra: $\Fpx$ is cyclic of order $\p-1$ with primitive roots; $\gen^{\p-1}=1$; $-1$ is a square in $\Fp$ iff $\p\equiv1\pmod4$ (Euler’s criterion); the affine maps $x\mapsto a+bx$, $b\neq0$, are the bijections preserving the field structure up to relabelling.I
[Burton 2010; Dummit & Foote 2004]
A2p01002The sign of the quarter-turn exponent: the oriented quarter-turn is $\im=\gen^{-\kap}=-\gen^{\kap}$, the exponent negative by pullback covariance (00:C7); the positive-exponent reading is the conjugate chart. Declared here, derived in the corpus.I
master: 00:C7
A3p01025The finite subgroups of $SO(3)$: cyclic, dihedral, tetrahedral ($12$), octahedral ($24$), icosahedral ($60$) — Klein’s classification; the binary icosahedral group as the $600$-cell’s vertices. Imported for E5.I
[Klein 1884; Coxeter 1973]
B. The shell and its frame
B1p01003The shell, capacity-first: $\kap$ primary, $\p=4\kap+1$ prime; the Subject’s frame $(\tau;0,1,\gen)$ with $\gen$ a primitive root and the drive $x\mapsto\gen x$; the cycle $\Phi_\p\cong C_{\p-1}$; $\im:=-\gen^{\kap}$ (A2), $-1=\gen^{2\kap}$.D
Def. 1, Section 2
master: 00:C1
B2p01004Symmetry completeness (Thm. 2.1): the fourth roots of unity form the unique order-four subgroup $Q_4=\{1,\im,-1,-\im\}$; under the Klein four-group $\langle x\mapsto-x,\,x\mapsto x^{-1}\rangle$, $Q_4$ is the union of the two size-two orbits $\{\pm1\}$, $\{\pm\im\}$, and $\Fpx\setminus Q_4$ splits into exactly $\kap-1$ orbits of size four. For $\p\equiv3\pmod4$ no $\im$ exists.T
python sourceLean, the coreLean, Mathlib
B3p01005The oriented quarter-turn: $\im=-\gen^{\kap}$ satisfies $\im^{2}=-1$ for every primitive root $\gen$; the two square roots of $-1$ are $\{-\gen^{\kap},\gen^{\kap}\}$, so the drive fixes which is $\im$. On $\F_{13}$, $\gen=2$: $\im=5$.T
python sourceLean, the coreLean, Mathlib
master: 00:C7
B4p01006Affine reframing (Def. 2, Lemma 2.2): $\varphi_{a,b}(x)=a+bx$, $b\neq0$, is a ring isomorphism $(\Fp,+,\cdot)\to(\Fp,\oplus,\otimes)$ with origin $a$ and unit $a+b$; every polynomial identity is frame-covariant.T
python sourceLean, the coreLean, Mathlib
master: 00:A6
B5p01007Origin and unit are observer data (affine gauge); the generator $\gen$ is the Subject’s drive, not gauge; the observer’s window is the naturals below $\sqrt\p$ (00:B8).D
Section 5; 00:A8, 00:C1
master: 00:B12, 00:C1
C. Arithmetic symmetries and the orbital complex
C1p01008The four symmetries as maps of the shell: the successor $C_1:k\mapsto k+1$; translations $T_a$, $a\in\Fp$, the cycle $C_\p$; scalings $S_m$, $m\in\Fpx$, the cycle $C_{\p-1}$; power maps $P_\varepsilon:x\mapsto x^{\varepsilon}$, $\varepsilon\in C_{\p-1}$, bijective iff $\gcd(\varepsilon,\p-1)=1$.D
Section 3, (3.1)–(3.4)
C2p01009The frame group (Prop. 1): $\langle T_a,S_m\rangle$ is the affine group of order $\p(\p-1)$, acting simply transitively on the frames $(a,b)$, $b\neq0$; the set of frames is a torsor under it (00:C5, 00:D13).T
python sourceLean, the core
manuscript
C3p01010Meridians and latitudes (Def. 3): $M_n(a)=a\gen^{n}$, $a\in\Fp$, the ordered list of the shell with step $\gen^{n}$; $L_a(m)=a\gen^{m}$, $m\in C_{\p-1}$, $a\neq0$, the multiplicative cycle from $a$; the orbital complex $\Sp$ with vertex set $\{(n,\pm a)\}\cup\{0\}$. The $2$-sphere reading is 2-geometry’s [Akhtman 2025].D
Def. 3
C4p01011Involutions and counts (Def. 3 (a)–(c)): $M_n(a)=M_{n+2\kap}(-a)$ and $L_a(m)=L_{-a}(m+2\kap)$ since $\gen^{2\kap}=-1$; $\p-1$ distinct meridian lists in $2\kap$ great circles; $2\kap$ latitude pairs; one-sided vertex count $(\p-1)^{2}/2+1$, the double cover $(\p-1)^{2}$ (00:B11).T
python sourceLean, the coreLean, Mathlib
D. Framed numbers
D1p01012Framed integers (Def. 4): the window $W_H=\{z\in\Z:|z|\le H\}$, $2H<\p$, embedded by $z\mapsto z\bmod\p$; sign, order and magnitude are the window’s; sums read back for $4H<\p$, products for $2H^{2}<\p$ — the corpus window $\sqrt\p$ (00:B1).D
Def. 4; 00:B1
master: 00:B1
D2p01013The window law: $z\mapsto z\bmod\p$ is injective on $W_H$ when $2H<\p$; for $x,y\in W_H$ the residue of $x\pm y$ reads back the integer when $4H<\p$, and that of $xy$ when $2H^{2}<\p$ (the same injectivity on the windows $2H$ and $H^{2}$).T
python sourceLean, the coreLean, Mathlib
D3p01014Framed rationals (Def. 5): $a/b$ with $a,b\in W_H$, $b\not\equiv0$, of value $ab^{-1}\in\Fp$; the scale grid $[x,n]:=x\gen^{-n}$, $x\in\{0,\dots,\p-1\}$, $n\in C_{\p-1}$; $G_n$ the ordered list $(x\gen^{-n})_x$, the meridian $M_{-n}$ of C3.D
Def. 5; 6:B1
master: 00:B1
D4p01015Scale-periodicity (Lemma 4.1, in the field): $S_n(x)=x\gen^{-n}$ satisfies $S_{n+(\p-1)}=S_n$; the grids $G_n$ are $(\p-1)$-periodic as ordered lists; the zoom $[x,n]\mapsto[x,n+1]$ has order $\p-1$. As lists of rationals $x/\gen^{n}$ ($\gen$ lifted to $\Z$) the grids are not periodic: the periodicity is a relation of the field.T
python sourceLean, the coreLean, Mathlib
D6p01017The continuum as chart (Def. 6, Props. 3–4, Cor. 1): the framed real line is the reading of the nested grids against $\R$, a labelled idealisation entered nowhere as an object (00:B1); no completion, compactness or “complete continuum” is claimed.D
Section 4.4; 00:B1
master: 00:B1
D7p01027Limit-like approximation (second clause), across the tower: for every real $r$ and $\varepsilon>0$ some shell $\p=4\kap+1$ carries a framed rational $x/2^{n}$, $|x|\le2\kap$, within $\varepsilon$ of $r$ (primes $\equiv1\pmod4$ beyond every bound). One shell trades range for resolution; the limit is the tower, not a completion.T
python sourceLean, Mathlib
E. The complex chart, polynomial equations and continuous symmetry
E1p01018The framed complex chart (Def. 7): the pair $(a,b)\in W_H^{2}$ read as $a+b\im\in\Fp$; the quarter-turn acts on the shell as the rotation $x\mapsto\im x$ of order four; the chart is the plane picture of $Q_4$ acting on the shell, not an extension of $\Fp$.D
Def. 7
E2p01019The extension by $X^{2}+1$ is trivial (Prop. 5 reversed): on the shell $\Fp[X]/(X^{2}+1)\cong\Fp\times\Fp$ has zero divisors, $(X+\im)(X-\im)=0$, so it is not a field; $\Fp[X]/(X^{2}+1)$ is a field exactly when $\p\equiv3\pmod4$; the quadratic extension of the shell is $\Fp[\eta]$, $\eta^{2}=\nu$ a non-square (8-dirac B6).T
python sourceLean, the coreLean, Mathlib
E3p01026Solving in the substrate (first clause): $f\in\Fp[X]$ has a root in $\Fp$ iff $f$ and $X^{\p}-X=\prod_{a}(X-a)$ are not coprime; the gcd counts the distinct roots and Cantor–Zassenhaus splitting finds them, exactly. An equation without a root in $\Fp$ is recognised, not extended (E2).T
python sourceLean, the coreLean, Mathlib
E4p01028Continuous symmetry, the abelian case (third clause): $\Fpx\simeq C_{\p-1}\simeq SO(2,\Fp)$ (6:E2) is a $\pi/(\p-1)$-net of $U(1)$ under $k\mapsto e^{2\pi\im k/(\p-1)}$; the rounding $k(\theta)=\lfloor(\p-1)\theta/2\pi+\tfrac12\rfloor$ has angle and chord error $\le\pi/(\p-1)$ and group-law defect at most one step.T
python sourceLean, Mathlib
E5p01029Continuous symmetry, the non-abelian obstruction: no finite subgroup of $SO(3)$ is an $\varepsilon$-net for $\varepsilon<\varepsilon_0=\arccos\big((3\sqrt5-1)/8\big)\approx44.5^\circ$, the covering radius of the icosahedral group (A3); the tetrahedral, cyclic and dihedral groups have radius $\ge\pi/2$, the octahedral $\approx62.8^\circ$. Finite subgroups cannot approximate $SO(3)$; the window does (E6).T
python source
E6p01030Continuous symmetry, the non-abelian case resolved by the window: the framed quaternions $W_H^{4}$, read in a shell $\p>8H^{2}$, compose exactly (products land in $W_{4H^{2}}$); normalised, they are an $\varepsilon$-net of $SO(3)$ with $\varepsilon\le2\arcsin(1/H)$ (round $Hs$ to the lattice). At $H=3$, $\p=73$: $38.9^\circ<\varepsilon_0$, finer than any finite subgroup; measured $30.0^\circ$.T
python sourceLean, the core
Section 6
E7p01032The Lie-algebra layer, one axis: on a shell $\p>8H^{2}$, for $q=a+v\in W_H^{4}$ with $v\ne0$, the rotation $\rho_q\colon x\mapsto qxq^{-1}$ is the Cayley step (8:C3) of the self-adjoint $\eta S_v$ (E2’s $\eta$, $S_vx=v\times x$, $N(v)=v\cdot v$) at $\alpha=\eta/(\nu a)$, and the half-turn $U_\infty=I+2N(v)^{-1}S_v^{2}$ at $a=0$. With $U_\infty$ these steps form a cyclic group of order $\p-\bigl(\tfrac{N(v)}{\p}\bigr)$ containing every iterate; at most $2H^{2}+2H$ of its elements are window rotations.T
python sourceLean, the coreLean, Mathlib
Section 6; 8:C3, 8:C5
E8p01033The Lie-algebra layer, two axes and range: for pure quaternions $u,w$ with $1+N(u)$, $1+N(w)$ and $1-u\cdot w$ nonzero, $\rho_{1+u}\rho_{1+w}=\rho_{1+u\circ w}$ with $u\circ w=(u+w+u\times w)/(1-u\cdot w)$, the bracket being $[u,w]=2u\times w$; at $u\cdot w=1$ the product is the half-turn about $u+w+u\times w$. The entries of $q^{m}$, $q\in W_H^{4}$, are at most $(2H)^{m}$, and $q^{m}$ reads back from the shell iff $2\|q^{m}\|_\infty<\p$ (D2).T
python sourceLean, the coreLean, Mathlib
Section 6; D2
Z. The horizon
Z1p01020No element of additive order two (Thm. 5.1): $2s=0$ forces $s=0$; the antipode of the origin on the additive cycle is not a residue; it sits between $2\kap=(\p-1)/2$ and $2\kap+1=(\p+1)/2=2^{-1}$ (20-rh B8).T
python sourceLean, the coreLean, Mathlib
Z2p01021The antipode named $\infty$: the horizon of the observer at $0$ (00:C25, 00:D13), the compactification point of the framed line [chart]; its reading as classical infinity is the chart, not a residue.D
Section 5.1; 00:C25, 00:D13
master: 00:C25, 00:D13

Status

I
Import. A standard result used without reproof.
D
Definition. A naming or declared convention, carrying no empirical content.
T
Theorem. Derived, and machine-verified in exact arithmetic by the named witness.

Sources

python source
Gold: the python check that decides the predicate, at its line in the script (the source, rendered on this site); on hover the deciding family and the families that corroborate it, and the verdict of the last run (red when it fails). The script page's bar runs the predicate's cell of the package's notebook on Google Colab (one cell per predicate; the whole notebook from the header badge).
Lean, the coreLean, Mathlib
The predicate as one Lean declaration, named by its accession key (p01004) — the conjunction of the theorems that prove it — at its line in the rendered module, one mark per library that proves it, the core's first. Green: the core (lean/FrcCore) — the statement on the shell ring, by computation and induction, no Mathlib, no axioms: the finitist witness. Violet: Mathlib (lean/FrcLedger) — the classical statement for every shell on Mathlib's hierarchy (its axioms, propext, Classical.choice, Quot.sound, on hover). Red: a proof the gate rejects (an unfinished proof, native_decide) — not a witness. A predicate proved in both libraries shows both marks; one mark, one proof. The module page's bar opens the module as one executable file in Lean Live (the core on plain Lean, no Mathlib; the file elaborates in the browser and prints every predicate's axioms at its end); the header's two Lean badges count the marks of each colour and open the two executables.
section
The statement in the paper's body that carries the predicate: a proposition, a remark or a section.
rests on
The predicates it rests on, by label: realisations, definitions and imports of this ledger, predicates of other ledgers (32:B3) and of the master (00:D15).
What the reader must grant. Three imports: the finite-field algebra of A1, the sign convention of A2 (derived in the corpus, declared here) and Klein’s list of A3. Nine declarations (B1, B5, C1, C3, D1, D3, D6, E1, Z2): the frame, the window, the grids, the charts. Sixteen theorem rows (B2–B4, C2, C4, D2, D4, D7, E2–E8, Z1): nine are proved in Lean in full (B3, B4, D4, D7, E3, E4, E7, E8, Z1), four in their main clause with a counting or converse clause outstanding (B2, C4, D2, E2), E5 is numerical on the imported list, E6 is proved by an elementary argument (Section 6), and C2 is settled in the manuscript. Eleven of them (B2, B3, B4, C2, C4, D2, D4, E2, Z1, E3’s criterion with the root bound, E6’s finitary clause) are proved in the core with no axioms at all, every clause, for every shell. E7 and E8 have their identities proved there: the Cayley form, the half-turn, the kernel, the composition law and the bracket. Their counts are proved on Mathlib: the order and cyclicity of the axis group, the window count, the heights and the read-back. The Mathlib proofs are classical. What no Lean row states: E3’s gcd count and Cantor–Zassenhaus (the package’s check of E3), and E5’s radii and E6’s net bound (the charts of E5 and E6). No realisation row: the paper makes no physical identification of its own; the readings it anticipates are owned by the master rows it cites (00:A8, B1, B8, B11, C1, C5, C7, C25, D13). No row is open: E3 (solving), D7 (approximation) and E4–E8 (continuous symmetry) decide the conjecture of the conclusion clause by clause. The title claim — the classical number classes reconstructed as frame-relative structures of one finite field — is B4 with D2, D4 and E2, read through the charts D6 and E1 (the approximation theorem and its obstruction, Thm. 4.2, stay in the body); the “complete continuum” of Corollary 1 is not a row.