Correction note (September 2026): the 7/7 count for the mediator-13 rule in this version is superseded. After rechecking the formulas, the current count is 6 of 7, with one open exception (θ₁₂ CKM). See The Mediator-13 Rule Is Not 7/7.
From BC structure to the numerical pattern
Independent research · July 2026 · Diego Luis Tentor
Derivation document
1. The result to derive
In the fermionic mixing synthesis, it was established as an E3 pattern that 13 (SING, T⁻⁶) appears in exactly the angles connecting the first and third generation — in both sectors, CKM and PMNS — and in no angle connecting adjacent generations. The goal of this document is to derive that pattern from the boundary condition (BC) structure of the ArXe hierarchy.
2. BC structure of the relevant levels
| Level | n | BC_closed | BC_open | Assignment | Confined |
|---|---|---|---|---|---|
| T⁻¹ | 3 | 0 | 1 | Ternary mediator | No |
| T⁻² | 5 | 1 | 1 | MEM — curvature | No |
| T⁻³ | 7 | 2 | 1 | CPX — color (SU(3)) | Yes |
| T⁻⁵ | 11 | 4 | 1 | REG — EM (U(1)) | No |
| T⁻⁶ | 13 | 5 | 1 | SING — weak (SU(2)) | No |
Confined means BC_closed without free BC_open available for transmission. T⁻³ has BC_open=1 but its interaction (SU(3)/color) confines — it does not transmit information freely between generations.
BC_closed values corrected 2026-07-30 to match the canonical table established across the corpus (arxe_core_V4_221_en.md, gauge_from_arxe, BC_paper_unified_V4, A Fractal Recursive Ontology from Boundary Conditions V4.1): BC_closed(T⁻⁶)=5, not 3. All downstream arguments in this document are re-checked below and hold under the corrected value — the “BC_closed ≥ 2” threshold used throughout is still satisfied (5 ≥ 2), but a numerological aside comparing it to T³ has been removed as unsupported (see the synthesis document for the correction).
3. The three generations as phase orderings of T⁻¹
T⁻¹ has BC_open=1 and arity n=3. In ArXe, the three fermionic generations are the three equiprobable phase orderings of level T⁻¹. (Citation corrected 2026-07-30: this document previously cited “Grammar, §3.2,” which does not contain this material — §3.2 of Grammar is “Operational Rules.” The relevant framework is the Principle of Phase Ordering, developed in physics-as-statistical-manifestation.md. Note also that this document explicitly scopes the Principle of Phase Ordering to cases “where a level permits multiple phase orderings and the question is which one dominates” — citing Madelung’s rule as the clearest example — and separately states that CKM mixing angles are “structural or combinatorial… tied to boundary-condition algebra rather than to counting phase orderings.” That statement concerns the angle values (the BC-algebra factorizations derived in this fermionic-mixing corpus), not the generation count derived here. The count — why there are exactly 3, not 4 or 2 — is a genuine application of the Principle of Phase Ordering: T⁻¹ permits 6 phase orderings, and the question of which orderings are physically realized (3, once T³ fixes an origin) is exactly the kind of question the principle addresses. The two derivations are complementary, not overlapping: Phase Ordering explains the count of 3; BC algebra explains the specific angle values.) There are exactly 3 because n(T⁻¹)=3 — three logical alternatives under the ternary mediator.
Indexing the generations as {1, 2, 3}:
Generation 1: phase 0 (first ordering)
Generation 2: phase 1 (second ordering)
Generation 3: phase 2 (third ordering)The phase difference between two generations is the number of orderings separating them.
4. BC_closed as a counter of readable phase differences
Principle: a level with BC_closed = k can mediate transitions involving up to k simultaneously closed degrees of freedom.
Applied to generational jumps:
Adjacent jump (difference = 1): requires BC_closed ≥ 1
Non-adjacent jump (difference = 2): requires BC_closed ≥ 25. From 6 orderings to 3 generations — closing the gap
T⁻¹ produces 3! = 6 equiprobable orderings. T³ (the mass level) breaks the reversal symmetry by assigning a direction: mass(Gen1) < mass(Gen2) < mass(Gen3). This distinguishes the 6 orderings as 6 elements of the symmetric group S₃.
The three generations correspond to the three layers of the Cayley graph of S₃ with adjacent transpositions as generators, measured from the identity (0,1,2):
Layer 0 (distance 0): (0,1,2) → Gen1
Layer 1 (distance 1): (0,2,1), (1,0,2) → Gen2
Layer 2 (distance 2): (1,2,0), (2,0,1) → Gen3
Layer 3 (distance 3): (2,1,0) → fully inverted — excluded by T³Layer 3 is not a fourth generation. It is the mirror image of Gen1 under complete reversal, excluded because T³ already records Gen1 as the historical first.
Distances between generations:
d(Gen1, Gen2) = 1 [adjacent layers]
d(Gen2, Gen3) = 1 [adjacent layers]
d(Gen1, Gen3) = 2 [separated by Gen2 in the graph]The condition justifying adjacent transpositions as the only permitted moves is the principle of minimal perturbation in T⁻¹: only neighboring phases can be exchanged, not skipped — coherent with BC_open=1, one open degree of freedom per step.
6. Why T⁻⁶ and not T⁻³ or T⁻⁵
Three levels have BC_closed ≥ 2 and are therefore candidates by the raw counting condition: T⁻³ (2), T⁻⁵ (4), T⁻⁶ (5). Two of the three must be excluded on independent grounds — the condition BC_closed ≥ 2 alone is necessary but not sufficient.
T⁻³: BC_closed=2 — satisfies the condition for difference = 2
T⁻⁵: BC_closed=4 — also satisfies it
T⁻⁶: BC_closed=5 — also satisfies itT⁻³ is excluded — confinement. T⁻³ is confined by SU(3). Confinement in ArXe means the level cannot act as a free mediator between states of different generations — its open BC is internally committed by the color interaction.
T⁻⁵ is excluded — flavor blindness. T⁻⁵ (U(1), electromagnetism) is not confined and satisfies BC_closed=4 ≥ 2, so confinement alone does not rule it out. It is excluded on a different ground, already established elsewhere in the corpus: EM is flavor-blind. As the-observer-at-t-5.md states, flavor — the distinction between generations — is invisible to electromagnetism; a system operating purely at T⁻⁵ cannot register which generation it is coupling to. T⁻⁶ has exactly one more closed BC than T⁻⁵, and that additional closed degree of freedom is what gives it the richer internal structure needed to register flavor (generation) states at all — a purely T⁻⁵ (EM) interaction has no channel through which a 1↔3 generational jump could be mediated, regardless of its BC_closed count. So T⁻⁵ fails not the counting condition but the more basic requirement of being able to distinguish generations in the first place.
T⁻⁶ (SU(2) — weak field) has BC_closed=5 and free BC_open=1. Among the non-confined, flavor-sensitive levels, it is the minimum one with BC_closed ≥ 2.
By the principle of minimum condition: T⁻⁶ is the mediator of the 1↔3 jump.
7. Why adjacent jumps do not require T⁻⁶
The adjacent jump requires BC_closed ≥ 1. The minimum non-confined level with BC_closed ≥ 1 is T⁻² (MEM, n=5, BC_closed=1, BC_open=1). Adjacent jumps are mediated through T⁻² without requiring T⁻⁶.
8. The complete derivation chain
T⁻¹ has 3 phases and BC_open=1
→ 3! = 6 equiprobable orderings
→ T³ breaks reversal symmetry with the mass arrow
→ 6 orderings = 6 nodes of the Cayley graph of S₃
→ 3 generations = the 3 layers of the graph by distance from Gen1
→ d(Gen1,Gen2) = d(Gen2,Gen3) = 1 [adjacent jumps]
→ d(Gen1,Gen3) = 2 [long-range jump]
→ mediator of d=2 requires BC_closed ≥ 2
→ T⁻³ excluded (confined by SU(3)); T⁻⁵ excluded (flavor-blind, per the-observer-at-t-5.md)
→ T⁻⁶ is the minimum non-confined, flavor-sensitive level with BC_closed=5 ≥ 2
→ 13 appears in 1↔3 angles, absent in adjacent ones ✓9. Numerical verification
| Angle | Jump | Type | 13 in expression | Prediction |
|---|---|---|---|---|
| θ₁₂ CKM = 13.04° | 1↔2 | Adjacent | NO | ✓ |
| θ₂₃ CKM = 2.38° | 2↔3 | Adjacent | NO | ✓ |
| θ₁₃ CKM = 0.201° | 1↔3 | Long-range | YES | ✓ |
| δ CKM = 65.4° | 1↔3 | Long-range | YES | ✓ |
| θ₁₂ PMNS = 33.44° | 1↔2 | Adjacent | NO | ✓ |
| θ₂₃ PMNS = 49.2° | 2↔3 | Adjacent | NO | ✓ |
| θ₁₃ PMNS = 8.57° | 1↔3 | Long-range | YES | ✓ |
Originally reported as 7 out of 7 (July 2026). Rechecked in August 2026: 6 of 7, one open exception (θ₁₂ CKM).
Note on δ CKM: δ parametrizes the interference between all three jumps — including 1↔3. Its 13 is coherent: without the long-range mediator, CP violation does not exist.
10. Bonus result: why exactly 3 generations
Layer 3 of the Cayley graph (2,1,0) is excluded — not a fourth generation but the mirror of Gen1 under T³. The generation space closes at 3 because the Cayley graph of S₃ has exactly 3 accessible layers once T³ fixes the origin.
This is a derivation of why there are exactly 3 fermionic generations — not postulated, not fitted, but a consequence of n(T⁻¹)=3 together with T³ breaking reversal symmetry.
Relation to other corpus treatments of “why 3 generations”:
ALO_logical_problems_in_scientific_choices_en.md, CASE 3, flags the Standard Model’s own justification for exactly 3 generations — “we have not observed a 4th, therefore there are only 3” — as a Fregean fallacy (absence of evidence treated as evidence of absence). The derivation in this document is not vulnerable to that critique: it is a positive, closed combinatorial argument (the Cayley graph of S₃ has exactly 3 accessible layers once T³ fixes an origin — there is structurally no room for a 4th, independent of what has or has not been observed), not an argument from non-observation.ALO_Mathematical_Formalization_s_en.md, Rule R108, derives a separate result — the mass-hierarchy suppression factors between generations (F⁰/F¹/F⁻¹, via a subfield tower of cyclotomic fields) — and takes the count of 3 generations as given rather than deriving it. R108 and this document are complementary, not competing: this document answers “why exactly 3,” R108 answers “why these particular mass ratios between the 3.” Neither result depends on the other.arxe-lepton-electron_muon_tau_mass_s_en.md§9.1 contains an earlier, undeveloped one-line speculation (“3D space + temporal = 4 levels possible”) that predates this derivation and is superseded by it.
11. Epistemic state
| Element | State |
|---|---|
| T⁻⁶ = SU(2) = weak field | Established in corpus |
| 3 generations = 3 phase orderings of T⁻¹ | Application of the Principle of Phase Ordering (physics-as-statistical-manifestation.md) — count, not angle values; see §3 correction note |
| BC_closed as counter of phase differences | New principle — coherent with corpus |
| T⁻⁶ minimum non-confined, flavor-sensitive level with BC_closed ≥ 2 | Verified from BC table; T⁻⁵ excluded via the-observer-at-t-5.md (flavor-blind) |
| T³ breaks reversal symmetry → mass arrow | New result — §5 |
| d(Gen1,Gen3)=2 via adjacent transpositions | Derived — Cayley graph of S₃ |
| Prediction 13 in 1↔3, absent in adjacent | Checked — originally 7/7; now 6/7, one open exception (θ₁₂ CKM) |
| Exactly 3 generations as necessary consequence | Derived — bonus result |
Evidential strength: E4 — derivation complete from BC axioms of the corpus.
12. New prediction
If a fourth fermionic generation exists (BSM), the 1↔4 mixing angle would require phase difference = 3, needing BC_closed ≥ 3. T⁻⁶ satisfies BC_closed=5 ≥ 3 — it would remain the mediator. But 2↔4 and 3↔4 angles would also require BC_closed ≥ 2 — 13 would appear in more mixing angles than just 1↔3. Falsifiable with BSM data.
Independent research — July 2026
Diego Luis Tentor — Working document, not peer-reviewed
“T⁻⁶ mediates the long-range jump because it is the minimum free level
that can both close on two generations and tell them apart.
Confinement rules out T⁻³. Flavor-blindness rules out T⁻⁵.
What is left is not a coincidence — it is what remains.”