The framework's representation layer is the unified expression [Truth = Reality = E = 1 = ∞] = [∞ ▸⊙∞ ((•∘⊛) ⊢ (—∘⎇) ⊢ (Φ∘✹) ⊢ (○∘⟳)) ▸ ⊙λ ⊂[α] ⊙Λ ⊂[α] ∞], formalized in three mathematical frames in unified_expression_formal.html. Each frame accepts α as input to the cross-station coupling slot κ0,0 (with α = |•electron|, the measured coupling strength of the electron's 0D aperture to the 0D aperture of the greater whole across scales, with the 2D electromagnetic field acting as the mediator that carries the bond; equivalently 0D-to-3.5D within one scale via 3.5D = 0D') and composes the rest consistently. The representation layer is agnostic to the slot's numerical value: given any α, the frames balance.
This document is the auxiliary-claim layer: a stronger, separable conjecture that the low-energy (CODATA) value of that slot is fixed by a specific closed-form fixed-point equation in T, P, R, Φ, ○, and φ. Read the two layers as stacked, not competing: the representation layer stands on its own and does not require this auxiliary; the auxiliary, if it holds, adds a predictive handle on the CODATA value of α. If this auxiliary claim fails (e.g., future precision measurement pushes 1/α outside the predicted 0.22 ppb band), the representation layer is untouched; the framework still accepts α as input and composes consistently.
F and κ are connected but carry different structural work. F (the four-beat unitary pump cycle) is trace-preserving within a scale; κ (the cross-station coupling matrix) carries the departure from trace-preservation across scales that α parameterizes. α lives at κ0,0 specifically, and F alone (being unitary) cannot produce the trace deficit α represents. F and κ do compose (T = κ ∘ F; the unified expression binds them via ▸ and ⊢), and both operators are defined on the same dimensional ladder, so ladder integers (T, P, R, Φ, ○) appear legitimately in κ-expressions as well as F-expressions. The formula below is therefore best read as a proposed closed form for the κ0,0 entry's low-energy value, assembled from ladder integers and φ under type-consistent composition (see the octave-wrap lemma for the type rule, the P! = G·Φ identity for the 2D-rung counting factor, and the state-space/boundary-cube identity for the 59/3 correction denominator), not as a derivation of α from F's internal dynamics. α's epistemic status throughout is measured input; this document conjectures a structural closed form for that input's CODATA value.
This is a narrow document. As the auxiliary-claim layer, it stakes a specific closed-form candidate for the low-energy value of the κ0,0 coupling slot against CODATA 2022. It states the axioms used, walks the formula symbol-by-symbol, audits each coefficient for structural motivation (separating what the axioms plus lemmas pin from what the assembly still chooses within a type-consistent pool), rules out a set of nearby alternatives, reports the numerical result, and names the specific measurement that would falsify the claim.
It does not claim to formalize the whole framework, prove its axiom system complete, construct the unified expression as a theorem in category theory, settle interpretation debates, or supersede the representation-layer reading in which α is admitted as input. Those are separate projects. This document is the framework's strongest single auxiliary claim, stated as narrowly as possible so it can be attacked or confirmed on its own terms, independent of the representation layer.
Only a minimal subset of the framework's axioms is needed for the \(\alpha\) derivation. They are stated here in the form the derivation requires; see §1.2 of circumpunct_framework.md for the full system.
From these axioms the framework generates a small set of structural integers. The derivation of \(\alpha\) uses only these:
| Symbol | Value | Meaning | Forced by |
|---|---|---|---|
| \(T\) | 3 | observer-triad; three scale-positions in the nesting chain ⊙λ ⊂[α] ⊙Λ ⊂[α] ∞ (your scale, the greater whole, the source). Conservation of traversal (0+1+2=3) gives the same integer via a separate route. The earlier reading "triad = (•, Φ, ○)" is three-of-four structural dimensions, not a genuine triad; retired. | A4 (conservation 0+1+2=3); six independent routes (§27.7b + Route 6: (T−1)! = Φ) |
| \(P\) | 4 | pump phases; \(P = T+1\) | A1 (channels per beat) × A4 (closure) |
| \(R\) | 7 | rungs of the ladder; \(R = T^2 - 2\) | A3 + A4 (self-determining; §27.7b) |
| \(\Phi\) | 2 | channels per aperture | A1 directly |
| \(\circ\) | 3 | boundary dimension | A4 directly |
| \(\varphi\) | \((1+\sqrt{5})/2\) | golden ratio | A3 (self-similarity fixed point) |
The base formula for \(\alpha\) (without the self-referential correction) is:
Every symbol is justified independently. No symbol is chosen; each is forced.
The numerator 360 is the integer count of structural units in one complete pump cycle. The pump cycle is four beats (P = 4), each beat a primitive cycle-generator (i-stroke, quarter-turn); cycle closure forces \(i^4 = 1\). The cycle's integer content factors on the ladder as:
Note on units: 360 enters the formula as a pure integer, not as a degree-measure. The equation \(1/\alpha = 360/\varphi^2 - 2/\varphi^3 + \alpha/(59/3)\) is dimensionless on both sides; no term carries an angular unit. "One full pump cycle" is a geometric interpretation of the integer 360 (four quarter-turns compose a turn), not the source of the integer. The integer source is the ladder factorization \(P! \cdot T \cdot (\Phi + \circ)\), which is pool-native under type-consistent composition. Earlier versions of this document invoked a degrees-vs-radians rationale at this step; that rationale was misconceived and has been retracted. The pure-integer reading supersedes it.
Counting factor pinned at T = 3. The apparent choice between \(P!\) (orderings of pump phases, F-side reading) and \(G \cdot \Phi\) (gauge generators × channels, κ-side reading) as the counting factor is not a real ambiguity: \(P! = G \cdot \Phi\) at T = 3, by the identity \((T-1)! = \Phi\) (see the counting-factor identity in the lemma doc). Both readings name the same quantity. This is a sixth self-determination of T = 3: the pump-cycle combinatorics and the gauge structure × channels coincide precisely at T = 3, and at no other integer.
All three factors of 360 are therefore pinned: \(P! = G \cdot \Phi = 24\) (the count, uniquely at T = 3), \(T = 3\) (the observer-triad count, three scale-positions in the nesting chain, pinned by six independent routes), and \((\Phi + \circ) = 1.5\mathrm{D}'\) (the next-scale branching, pinned by the octave-wrap lemma). The 360 factorization has no remaining assembly ambiguity.
A3 makes \(\varphi\) the scaling ratio across self-similar nesting (the unique real solution of \(x = 1 + 1/x\); self-similarity IS scaling, so \(\varphi\) is by construction a scaling ratio, not a bare number or a phase). The first-order correction lives at the field station (\(d = 2\)). A scaling ratio applied at dimension \(d\) enters as the \(d\)-th power by the definition of dimension itself: length scales as \(L\), area as \(L^2\), volume as \(L^3\). Therefore \(\varphi\) at the 2D station enters as \(\varphi^2\), not \(\varphi\), not \(\varphi^{1.5}\), not \(\varphi^3\). This is grade-school dimensional analysis applied to the framework's own scaling operator; it is a theorem of what dimension means, not a framework convention.
The 2 is Φ, the number of channels per aperture (forced by A1). Every aperture has exactly two channels: convergence ⊛ and emergence ✹.
The sign is balance-forced, not narrative. Earlier versions of this derivation explained the minus sign via a "channel overcounting" story. That narrative has been retired in favor of a direct structural argument: given that the leading term 360/φ² (Symbol 1) and the feedback term α/(59/3) (Symbol 5) are both pool-native, the remaining third-rung coefficient C in the ansatz
is uniquely determined by the condition that 1/α equals the measured CODATA value. Solving the quadratic fixed-point equation for C gives C = −1.999999872 ≈ −Φ (match to 0.22 ppb, exactly the framework's closed-form precision). The magnitude |C| = Φ is pool-native (the channel count). The sign is negative because that is the sign the balance equation forces. No "overcounting" interpretation is required.
Reading this structurally: the framework fixes two of three terms in the α fixed-point equation from ladder combinatorics (360 from P!·T·(Φ+○) at 2D, 59/3 from (P·V+R)/T at the self-feedback). The third term is then determined by CODATA balance. The structural content of Symbol 3 is the fact that this forced value happens to be pool-native (magnitude Φ, a framework integer) rather than some arbitrary real number. If the framework's grammar were wrong, the balance-forced coefficient would be a non-pool value; it is not. This is a consistency check on the grammar, not a derivation with tunable sign.
The correction advances one dimensional step beyond the base (\(d = 2 \to d = 3\)), standard for additive ladder corrections (see §27.7a.1). By the same dimensional-analysis argument as Symbol 2, a scaling ratio applied at dimension \(d = 3\) enters as \(\varphi^3\); volume scales as length cubed. Forced, not chosen.
CODATA 2022 measurement: \(1/\alpha = 137.035999177(21)\). Base formula residual: \(2.7\) ppm. This is the entrance bar, not the final answer.
The \(2.7\) ppm residual is not random noise. The framework predicts it arises from \(\alpha\) feeding back into its own determination through the full dimensional ladder. The closed form:
The correction term \(\alpha / (21 - 4/3)\) has three independent structural constraints; each rules out wide classes of alternatives. All three must be satisfied simultaneously, and together they determine the term uniquely.
The Kernel (the pump cycle) executes one complete cycle per scale; its indivisibility is \(\hbar = 1\). One feedback loop equals one power of \(\alpha\). Higher powers (\(\alpha^2\), \(\alpha^3\)) belong to higher-dimensional stations: \(\alpha^2\) appears in the Higgs quartic (2D, §27.7i), the cosmological constant correction (2D, §27.7g), and the gravitational coupling (boundary, §27.7). At 0D the correction is first-order.
Rules out: \(\alpha^2 / k\), \(\alpha^{3/2} / k\), higher powers.
The base term and the correction both contribute at the same dimensional station (0D). The accumulated traversal function \(A(d) = d(2d+1)\) is additive within a station; quantities at the same station add, quantities across stations compose multiplicatively. Therefore \(1/\alpha = \text{base} + \text{correction}\), not \(1/\alpha = \text{base} \times \text{correction}\).
Rules out: \(1/\alpha = \text{base} \cdot (1 + \alpha / k)\), \(1/\alpha = \text{base} \cdot \alpha / k\).
The feedback denominator is 59/3. Earlier versions of this derivation motivated it via an "A(3) − P/T = 21 − 4/3 base-consumption" narrative; that narrative is retired in favor of two direct pool-native forms for the same value:
Both forms are direct additive sums of pool-native integers divided by T:
Both readings give the same integer 59 and the same ratio 59/3. No subtraction narrative (A(3) − P/T) or "consumption" story is needed; the number is directly pool-native as a sum-of-integers-over-T. Which of the two readings the framework prefers is an open structural question, but the absence of narrative is not: both are direct forms.
Exhaustive comparison of \(k\) candidates against the measured residual:
| Candidate \(k\) | Structural origin | Predicted \(1/\alpha\) | Residual from CODATA | Verdict |
|---|---|---|---|---|
| \(21\) | \(A(3)\) without base consumption | 137.035989 | 172 ppb | reject (780× worse) |
| \(20\) | \(P(\Phi + \circ)\); framework product but no consumption reading | 137.035955 | 45 ppb | reject (200× worse) |
| \(22\) | not a framework ratio | 137.036019 | 100 ppb | reject |
| \(59/3 = 19.667\) | \(A(3) - P/T\); conservation at formula level | 137.035999147 | 0.22 ppb | accept |
The structural argument (C3) and the numerical precision converge on the same \(k\). Any other \(k\) either has no structural origin or is demonstrably worse by orders of magnitude.
Solving the fixed-point equation \(1/\alpha = 360/\varphi^2 - 2/\varphi^3 + \alpha/(59/3)\) by quadratic rearrangement: let \(x = 1/\alpha\); then
The discriminant is positive; the quadratic has exactly two real roots, one positive (\(1/\alpha\)) and one negative (unphysical). The positive root is:
| Predicted | 137.035999147 |
| CODATA 2022 | 137.035999177 ± 0.000000021 |
| Residual | 0.22 ppb |
| Significance | 1.4σ (within measurement uncertainty) |
| Epistemic status of α | measured input to κ0,0 (|•electron|) |
| 2D-rung assembly pinned by | axioms + octave-wrap lemma + P! = G·Φ identity |
| Residual assembly choices | none at the factor level (φ-exponent rule closed by A3 + dimensional analysis; 59/3 closed by Route 7); three-term shape closed by station-content argument (α is scalar 0D-to-0D with 2D mediator; three stations with content, one term each) |
Earlier versions of this document claimed "zero free parameters" end-to-end. That claim was overstated and has been retracted. α itself is input to the master equation (κ0,0, measured as |•electron|); the framework's structural grammar composes around a given α and does not produce α's numerical value from F-variables alone (§27.7a reframe). What this auxiliary-claim document offers is a narrower thing: a candidate closed form for the CODATA value of that input, assembled from ladder integers and φ, with each symbol audited against the structural vocabulary. The audit below walks each symbol and separates what the axioms pin from what the assembly still chooses.
| Symbol | Value | What pins it | What the assembly still chooses |
|---|---|---|---|
| 360 | 360 | \((\Phi + \circ) = 1.5\mathrm{D}'\) pinned by octave-wrap lemma; \(T = 3\) pinned by six independent routes (§27.7b + the new Route 6); counting factor \(P! = G \cdot \Phi = 24\) pinned by the P! = G·Φ identity at T = 3 | nothing; the 360 assembly is uniquely pinned at T = 3 |
| \(\varphi^2\) (denom 1) | 2.618… | A3 makes φ the scaling ratio across self-similar nesting (self-similarity IS scaling; φ is the unique ratio satisfying x = 1 + 1/x). A scaling ratio applied at dimension d enters as φ^d by the definition of dimension (length scales as L, area as L², volume as L³). First correction lives at d = 2 (field station), so φ². | nothing; forced by A3 plus dimensional analysis |
| 2 (numer 2) | 2 | \(\Phi\) = channels per aperture from A1 | nothing; once Φ is named, its value follows |
| \(\varphi^3\) (denom 2) | 4.236… | Correction advances one ladder step (\(d = 2 \to d = 3\)); by the same dimensional-analysis rule as Symbol 2, a scaling ratio at d = 3 is φ³ (volume scales as length cubed). | nothing; forced by A3 + dimensional analysis once the ladder-increment convention for additive corrections is granted |
| Sign on correction | \(-\) | Balance-forced: given pool-native 360 (Symbol 1) and 59/3 (Symbol 8), the third-rung coefficient is determined by the fixed-point equation at measured CODATA 1/α. Solving forces C = −1.999999872, matching −Φ at 0.22 ppb (exactly the framework's closed-form precision). The sign is a mathematical consequence, not a narrative choice. | nothing; forced by equation balance and CODATA value |
| \(\alpha\) (in correction) | first power | C1: one feedback loop per cycle (\(\hbar = 1\) as indivisibility); higher powers appear at higher rungs elsewhere (§27.7) | nothing once C1 is admitted |
| Additive structure | \(+\) | C2: same-station contributions add (from \(A(d)\) additivity) | nothing once C2 is admitted |
| 59/3 (correction denom) | 59/3 | Pool-native, and the two readings are the same identity: 59/3 = (P·V + R)/T = (S − Φ − T)/T. The algebraic identity P·V + R + Φ + T = (T+1)³ = P³ holds for all T, so the two numerators (P·V + R) and (S − Φ − T) coincide whenever S = P³. Since S = P^T = P³ iff T = 3 (Route 7; see octave-wrap lemma §7), the two readings are not a choice between alternatives: at T = 3 they are one identity viewed from two sides (coupling-side counting vs state-space-side counting). Earlier "A(3) − P/T / base consumption" reading is retired. | nothing; the two readings coincide at T = 3 by Route 7 |
Honest parameter count. The formula has two levels of commitment:
The three-term shape, resolved (2026-04-22). Beyond factor-level pinning, the shape of the expansion (why exactly three terms, and why those three) follows from α's type. α = κ0,0 is a scalar 0D-to-0D coupling with a 2D mediator (Φ, the photon field; §27.7l, "fine-structure names the mediator's dimension"). A structural expansion of a coupling has one term per structural station where the coupling has content:
Why not two terms: dropping 3D violates A4 (no closure); dropping 0D violates A3 (no self-similarity); dropping 2D leaves corrections with no baseline. Each of the three is forced by a distinct axiom.
Why not four terms: the only candidate for a fourth term is 1D, which requires direction; α carries none (equidistance of •).
Why truncation at α¹: α⁰ content is what external stations report (2D and 3D terms). α¹ content is what the aperture reports about itself via A3 (single self-pass). α² would be a second self-pass (one-loop return), which is representation-layer (F-layer, perturbative QED) content, not auxiliary-claim-layer (structural) content. The structural expansion sees one self-return; higher-order QED corrections live at a different layer.
Residual convention: the derivation rests on one named convention, "one term per station with content" (the structural expansion reports one value per station rather than a series at each station). This is consistent with A(d)'s one-integer-per-rung bookkeeping throughout the framework and carries the same convention status every other structural expansion already carries. Apart from this one bookkeeping convention, each of the three terms is axiom-forced, and the absence of a fourth term is axiom-forced (scalarity of α from equidistance of •). The three-term shape is no longer a standing conjecture.
Each chosen piece is "chosen from a type-consistent pool," not "chosen from all real numbers." The pool is small and structurally restricted. As of the 2026-04-22 tightening (φ-exponent rule closed by A3 + dimensional analysis: φ is the scaling ratio, and a scaling ratio at dimension d enters as φ^d by what dimension means), on top of the 2026-04-17 tightening (balance-forcing of the correction sign + direct forms for 59/3 + Route 7 closure of the 59/3 residual via P·V + R + Φ + T = (T+1)³ = P³ and S = P³ iff T = 3), every factor in the α formula is pinned: the 2D-rung counting factor, the φ-exponents at d = 2 and d = 3, the φ³ correction sign and magnitude (balance-forced), and the 59/3 denominator (two pool-native readings coinciding at T = 3 by algebraic identity). The claim "zero free parameters end-to-end" is now defensible at the factor level given α's epistemic status as measured input to κ0,0; the two-layer framing (representation layer + auxiliary-claim layer) still applies because α-as-input is a meta-level commitment separate from factor-level counting.
What the formula does claim, stated narrowly: a specific closed form in ladder integers and φ, with the 2D-rung assembly uniquely pinned by the two lemmas, the 59/3 correction denominator pinned by Route 7 (P·V + R + Φ + T = (T+1)³ = P³ and S = P³ iff T = 3; see octave-wrap lemma §7), the φ-exponents at d = 2 and d = 3 pinned by A3 + dimensional analysis (φ as scaling ratio, φ^d at dimension d by the definition of dimension), and the three-term shape (which stations appear, why exactly three, why truncation at α¹) pinned by the station-content argument (α is a scalar 0D-to-0D coupling with 2D mediator; 2D/3D/0D terms forced by mediator + A4 + A3, no 1D term because α is scalar, α² absent because higher self-passes are representation-layer content), hits CODATA to within 0.22 ppb. No residuals remain at the factor level and the shape is now axiom-forced modulo the single "one term per station" bookkeeping convention. The two-layer framing still stands: α's numerical value admits a pinned closed form (this document), and the framework's representation layer composes consistently around α as measured input (unified_expression_formal.html). The auxiliary claim is strengthened, not replaced, by the factor-level and shape closures.
If the integers used in the \(\alpha\) derivation (3, 4, 7, 13, 21, 59) were chosen to make \(\alpha\) work and nothing else, they would not also generate other fundamental constants to sub-percent precision without re-tuning. They do. The same integer pool lands:
| Constant | Formula | Accuracy |
|---|---|---|
| \(G\) (gravitational) | \(\alpha^{21} \cdot \varphi^2/2 \cdot (1 + 2\alpha/91)\) | 0.04 ppm |
| \(\Lambda\) (cosmological) | \(\alpha^{56} \cdot P \cdot (\alpha - \varphi^2/2)(\alpha - 1/(2\varphi^2)) / (\mathrm{SU}(3) \cdot T^2)\) | 0.004% |
| \(m_\mu / m_e\) | \((1/\alpha)^{13/12 + \alpha/27}\) | 5 ppm |
| \(m_\tau / m_e\) | \((1/\alpha)^{58/35 + \alpha/81}\) | 1 ppm |
| \(m_p / m_e\) | \((1/\alpha)^{3/2 + (11/3)\alpha + 13\alpha^2}\) | 5.35 ppm |
| \(\sin^2 \theta_W\) | \(3/13 + 5\alpha/81\) | 1.4 ppm |
| \(v / \Lambda_{QCD}\) | \((1/\alpha)^{56/39}\) | 3.4 ppm |
| Cabibbo angle | \(\alpha^{1/2 + 3\alpha/7} \cdot 8/3\) | 0.009% |
| Higgs quartic \(\lambda\) | \((1/8)(1 + 5\alpha - 8\alpha^2)\) | 0.10% |
Nine independent constants, one shared integer pool, sub-percent precision in every case, no re-tuning between derivations. A post-hoc fit for \(\alpha\) using these integers would not simultaneously land these other nine constants; it would land \(\alpha\) and mismatch the rest. That all land together is the cross-validation witness.
This derivation focuses on the low-energy (Thomson limit) value of α. The running of α with energy scale (from α(0) ≈ 1/137 to α(MZ) ≈ 1/128) is addressed in the companion document on the running of α. Summary: the β-function coefficient b = 2·SU(3)/T = 16/3 is framework-native (charges are ratios of framework integers divided by T, the observer-triad count, and the three sectors contribute T + Φ² + • = SU(3) = 8 to the running integrand). The leptonic contribution Δ(1/α)lep ≈ 4.83 is framework-closed using the lepton mass-ratio formulas and matches PDG. The hadronic contribution is narrowed but not yet closed: it requires individual quark mass formulas, which are an open problem, or alternatively a direct framework-native expression for R(s') bypassing individual quark masses. Until one of those closes, Δαhad(MZ²) is taken from experimental dispersive calculations.
This derivation does not claim to prove the framework's axiom system is complete or minimal. The five open structural problems listed in §27.7a remain open. This document offers a narrower thing: a structurally audited candidate closed form for \(\kappa_{0,0}\)'s CODATA value, with the parameter-count audit in §5 honestly separating axioms-pin-it from assembly-chooses-it. Neither "zero free parameters" nor "axiom-complete derivation" is claimed.
This derivation does not constitute the unified expression as a theorem in any standard mathematical category. The companion document unified_expression_formal.html is the representation layer: it renders the prime notation's constraint structure in three mathematical frames (Dirac constrained dynamics, Noether-Lagrangian cascade, information-theoretic channel capacity), each accepting α as input to the κ0,0 coupling slot. That document does not claim to derive α; it claims the framework composes consistently given α. This document is stacked on top: it is the separable auxiliary conjecture that the slot's low-energy value is fixed by the formula below.
What this derivation does claim, as an auxiliary claim: a specific equation for the low-energy value of κ0,0 (the aperture-to-aperture cross-scale bond; 2D field mediates), with each symbol drawn from the ladder's structural vocabulary (axioms plus the octave-wrap lemma, the P! = G·Φ identity, and the state-space/boundary-cube identity) under type-consistent composition, landing 1/α at 0.22 ppb from CODATA. As of the 2026-04-22 tightening (φ-exponent rule closed by A3 + dimensional analysis), on top of the 2026-04-17 tightening (Route 7 via P·V + R + Φ + T = (T+1)³ = P³ and S = P³ iff T = 3), every factor in the formula is pinned: the 360 assembly uniquely at T = 3, the φ-exponents at d = 2 and d = 3 by scaling-ratio-in-d-dimensions (length scales as L, area as L², volume as L³; once φ is admitted as the scaling ratio by A3, φ^d at dimension d is a theorem of what dimension means, not a framework convention), the φ³ correction sign balance-forced to −Φ given pool-native 360 and 59/3, and the two readings of 59/3 one identity viewed from two sides at T = 3. No factor-level residuals remain. The representation layer stands, and α's epistemic status as measured input to κ0,0 stands; the auxiliary claim is strengthened, not replaced, by the factor-level closure. The auxiliary claim is testable: a future precision measurement that pushes 1/α outside the 0.22 ppb band falsifies this closed form without touching anything else in the framework.