The Shape Minimality Challenge

We turned an impossible target — "prove the shape is the simplest" — into a reviewable contest: a declared game, a declared cost, a competitor ledger, and target-blind failure tests. Here is the whole scorecard, with nothing hidden.

The question

The 13-dimensional shape at the center of this program is

$$ M_4 \times K_6 \times S^2 \times S^1_Y/\mathbb{Z}_2, \qquad K_6 = SU(3)/T^2, $$

and from it, once it is selected and frozen, a striking amount of Standard-Model structure reads off: the gauge algebra $su(3)\oplus su(2)\oplus u(1)$, three chiral generations, charge quantization $Q = T_3 + Y$ with $Y \in \tfrac{1}{6}\mathbb{Z}$. The obvious challenge from any honest skeptic is: why this shape and not a simpler one? Is it the minimal geometry that does the job, or just a geometry that does the job?

This page answers that challenge head-on. It does not claim the shape is the simplest possible — that claim cannot be made by anyone, for any theory, as we will show. Instead it states exactly which minimality game the shape is winning, which game it is not yet winning, and what theorem would let a reviewer compare every contender fairly. The honesty here is not a hedge; it is the methodology. A theory that tells you precisely where its own floor is gives you something far more useful than a theory that claims to have no floor.

Three discipline lines run underneath everything below, and we never drop them:

• selection $\ne$ derivation — the shape is chosen and frozen, not deduced from nothing;
• given-$E$ $\ne$ derivation-of-$E$ — downstream results follow given the observed matter content $E$, which the framework does not itself force;
• frozen/reproducible $\ne$ proven-unique — that a construction is hash-frozen and reproduces blind does not make it the unique answer.

Why "simplest" is the wrong target — and what replaces it

The naive minimality claim is "no simpler shape, anywhere, reproduces the Standard Model." That sentence is a universal negative over all conceivable architectures, and it is not merely hard — it is uncomputable. The exact description-length floor of a target $T$ is its Kolmogorov complexity $K(T)$, and $K$ is not a computable function. There is a schema $B_P$ — any finite program $P$ that outputs the target $T$ to operational resolution $\Delta_0$ — that already produces infinitely many distinct generators. You cannot enumerate them, and you cannot prove none is shorter. "Prove the shape is THE simplest" is a unicorn: every theory that has ever existed faces the same wall, whether or not it admits it.

The move that makes the question answerable is to stop chasing the unicorn and declare the game. We fix a finite role-mechanism grammar $\mathcal{G}$ — the finite list of known mechanisms by which any architecture can fill each functional role the Standard Model requires. Inside $\mathcal{G}$, "is there a shorter recipe?" becomes a finite, decidable problem: a lower-bound matrix over finitely many normal-form classes. Converting an uncomputable universal negative into a finite, checkable contest is the real achievement. It does not by itself close the contest — but it makes "open" mean a specific named theorem is not yet proven, rather than a unicorn was not caught.

The grammar's five axes — each a declared posit, not a fact of nature — are: gauge-origin, chirality-origin, family-count, flavor-origin, scale-origin. A "normal-form class" is one chosen mechanism per axis. The 13D shape is itself just one tuple in this grammar (gauge $=$ isometry/coset; chirality $=$ index theorem; family-count $=$ topological index; flavor $=$ geometric overlap plus a fitted normalization; scale $=$ posited anchors). It gets no privileged row: it is charged by the same codebook as every rival. If a competitor uses a mechanism not on a list, it has not refuted the theorem — it has found a new mechanism, which extends the grammar. The contest is therefore living: broadening the axes is the legitimate way to attack it.

The three claim levels

We do not say "the shape is simplest." We say something sharper and three-tiered: selected, frozen, adversarially benchmarked.

Level 1 — Frozen candidate. ACHIEVED.

The shape is fully specified, hash-frozen on branch $\mathtt{dcc66f1b2685}$ / $\mathtt{a5b1e6f9d951}$, reproduced blind, with no silent retuning. This is the freeze-before-compare discipline: a generator must be locked before any downstream comparison, so that nothing can be quietly adjusted after seeing the data. If it changes after the comparison, it is a new branch, scored fresh. Level 1 is the precondition for every honest claim that follows — and it is met.

Level 2 — In-grammar minimality. PARTIALLY ACHIEVED + hardening.

Within the grammar where "forces are geometry" — gauge symmetry as the isometry of a compact internal space — cheaper shelves have been eliminated and the internal carriers are forced by architecture-neutral arguments:

• Weak carrier $S^2$ (fact F1): no abelian or torus carrier, of any dimension, has non-abelian $SU(2)$ among its isometries. This closes the cheaper direction not for one named rival but for an entire shelf of carriers.
• Hypercharge carrier $S^1_Y/\mathbb{Z}_2$ (fact F2): a bare closed odd-dimensional factor keeps both handednesses, producing mirror fermions, which are excluded by the measured LEP $Z$-width. The $\mathbb{Z}_2$ fold is what supplies chirality. This too closes a whole shelf.
• Color carrier $K_6 = SU(3)/T^2$ (abelian-isotropy uniqueness): the maximal torus $T^2$ is the unique purely abelian isotropy subgroup of $SU(3)$, with centralizer $C_{SU(3)}(T^2)=T^2$ (Cartan only). It injects no spurious non-abelian factor, keeping color, weak, and hypercharge separable.

The genuinely cheaper rival on the color rung is $CP^2 = SU(3)/U(2)$, the minimal-dimension (4D) $SU(3)$ carrier. We did not wave it away — we built it end-to-end, adversarially, as an 11D variant. It breaks at the gauge gate: $U(2) = (SU(2)\times U(1))/\mathbb{Z}_2$ is non-abelian and sits inside color $SU(3)$, so by the coset-reduction centralizer rule it is gauge-active. That forces a lose-lose fork — either keep $S^2, S^1$ and over-produce an extra $SU(2)+U(1)$ (the gauge-group equality fails), or drop them and isotropy-lock the electroweak factors inside color ($C_{SU(3)}(U(2)) = U(1)$ only, by Schur). Four independently built sectors converged on this same break. This is the cleanest real elimination in the program: $CP^2$ is excluded for a structural reason, from representation theory and the centralizer rule alone — not reverse-engineered — and it is not cheaper once forced to reproduce the full target, because it over-produces and fails.

So inside the isometry grammar,

$$ G_{\text{isometry}} + T + C_{\text{MDL}} \;\Rightarrow\; K_6 \times S^2 \times S^1_Y/\mathbb{Z}_2 . $$

Why "partial," not "done"? Because Level 2 rests on two conditional pillars, and a frank account names both. First, "shortest inside $\mathcal{G}$" is scored under a description-length (MDL) metric; the architecture-neutral theorem that selects MDL over a dimension-first metric is conditional on a bridge axiom (see the next section). Second, the in-grammar exhaustion is relative to the declared taxonomy — proven over the considered class, not yet a closed classification over all admissible carriers. And the flavor rulebook $F^+$ — which still uses fitted normalizations $N_d, N_e, N_\nu$ for the down-type, charged-lepton, and neutrino sectors — is the weakest link: a fitted normalization is a charged injection, scored like an anchor, not a derivation (charged openly, the sector still delivers: the flavor gate closes on the live board with its hardest output, the up-quark mass, landing at +0.058σ — SG-8). Level 2 is real, it is hardening, and its open sub-lemmas are named.

Level 3 — Architecture-neutral minimality. NOT OWED — dissolves as a universal negative no theory can prove.

"No simpler competitor across all grammars" is the universal-negative wall again. This is the unicorn, and it is honest to call it that: no theory ever written can prove it, because the exact description-length floor is a Kolmogorov quantity and uncomputable. On the live gate board the demand therefore dissolves rather than being owed — the Shape deep-root gate closes at DERIVED-GIVEN-anchor, RESOLVED at +0, with the dissolution stated on the row itself (dossier). What is reachable — and what the whole program aims at — is the bounded claim: shortest recipe within a declared, extensible grammar, given $E$. That bounded claim is the ceiling, and the ceiling is worth far more than a false absolute.

Public claim, in one line: Level 1 achieved · Level 2 partial + hardening · Level 3 not owed — dissolved as a universal negative every theory faces, and tracked here as the bounded research frontier.

The competitor ledger

This is the heart of the challenge: one row per rival, scored by a common codebook, with each candidate's failure mode named. The Standard-Model spectrum $E$ is presupposed on both sides and cancels; $E$-automatic facts (anomaly cancellation, written-spectrum chirality, accidental proton stability) are free to everyone and not charged. What is counted is the independent measured reals each side must inject, plus charged structural and rulebook bits. The honest charged cost of the 13D branch is not the "4-in" headline: it is roughly 4 measured anchors plus 9–10 fitted reals, so $n_{13} \approx 13\text{–}14$.

Two entries deserve a closer look, because they are where the contest is most often misread.

The 4D effective field theory is the honest cheap rival. Under a dimension-first metric it simply wins: $4 < 13$, full stop. It is allowed to postulate $su(3)\oplus su(2)\oplus u(1)$, the representation table, the family count, the hypercharges, the masses and mixings. But every one of those is a primitive input unless a mechanism generates it. Scored by description length, the EFT injects the full $\sim 25$-real ledger directly — it wins the weak-target game only because it does not generate the quantitative target at all; it leaves the constants as free parameters. The honest summary is: lower metric dimension, higher primitive-label cost. This is not "the EFT loses, full stop" — it is "the EFT wins one game and the 13D shape wins another, and the two games are decided by which metric is correct."

String / F / M-theory ties on the gauge outcome and is not refuted by anything in the isometry grammar. These frameworks have ample expressive power to host the Standard Model. After unfolding into stage, rulebook, and actors, they typically carry large charged burdens — compactification choice, flux choice, brane localization, moduli stabilization, exotics removal, and a landscape with no compact selector. They are not a preferred competitor today because none has been supplied as a frozen, reproducible certificate with fewer effective primitives than the active branch and a mechanism that actually selects its vacuum. That is an honest "not currently beaten," not a claim that they are impossible. They remain serious future candidates, and the moment one supplies a lower-burden frozen package, the contest reopens. String theory is not ruled out by the 13D branch — it plays a different game, and the bridge between games is exactly what is open.

The same posture holds for SO(10) and exceptional GUTs (elegant actor compression that reappears as breaking-chain, doublet-triplet, and proton-decay burden), and for finite-state / discrete geometry (a possible deeper compression route, currently underdeveloped against the full gate burden). Each is eliminated as a current preferred competitor, not as a future one — a current-record statement under a declared audit protocol, never a universal nonexistence proof.

The closure path — three named, finite targets

What stands between today's honest standing and the strongest claim that is reachable is not a search for a unicorn. It is three specific theorems, each finite and each falsifiable:

(1) The metric bridge. Show that finite operational reality — granularity, the cost of finite records — implies that description length is the correct simplicity order. Concretely: a measured real costs $\Theta\!\big(\log(1/\Delta_0)\big)$ bits to specify to resolution $\Delta_0$, while a discrete structural choice costs $O(1)$ bits. If that bridge holds, MDL is the right ruler and dimension-first is inadmissible (it calls "simpler" the architecture that needs more finite records). This currently rests on a natural bridge axiom — "simplicity $=$ minimal injected record cost" — and is therefore conditional, not proven outright. A refuted-economy outcome, where the 4D EFT is genuinely cheaper under the correct metric, is an equally valuable result of this same theorem.

(2) The role-mechanism normal-form / exhaustion theorem. Show that every admissible architecture that reproduces the target reduces, without getting cheaper (the no-smuggling rule), to one of the finitely many normal-form tuples over the five axes, then lower-bound each class. This is what turns the dimension-ladder survey — which today reports, under MDL, zero refutations, one structural failure-to-generate, and ten rivals losing to the 13D recipe across the considered class — into a classification. It also subsumes the color-carrier sub-shelf: the $SU(3)$-carrier shelf must be exhausted, not merely enumerated. The analogy is the classification of simple Lie groups: finite and checkable precisely because the invariants cut the space into finitely many classes.

(3) The bundle-uniqueness theorem (the given-$E$ wall). The whole stack bottoms on $E$. The Euler characteristic computation $\chi(K_6, E) = -3$ delivers exactly three families — but it does so as a rigid integer given a chosen bundle on $E$, not as a selection of $E$. (On $CP^2$ the same count appears as a discrete $\text{Spin}_c$ index, $\mathrm{ind} = \tfrac{r(r+1)}{2} = 3$ at $r=2$ — discrete on both carriers, which is exactly why the family count traces to $E$, not to geometry alone.) The only route that could remove "given-$E$" is to prove that the SM charges canonically force the bundle via the isotropy embedding, without smuggling "three generations" in as input. The honest prior is that this is hard and may not hold; the realistic ceiling may remain "rigid integer given a selected bundle." For now, "$E$ is forced" is refuted, because anomaly cancellation is a filter with infinitely many solutions, not a selector — passing $A(E_{\text{SM}}) = 0$ does not single out $E_{\text{SM}}$.

None of these three is a universal negative. Each is a bounded object that a reviewer can attack, prove, or refute. That is what it means to replace an impossible target with a reviewable one.

Status — reached vs open

Reached. Level 1 is achieved: the shape is frozen, hashed, and reproduced blind. Within the isometry grammar, two of the three internal carriers (weak $S^2$, hypercharge $S^1_Y/\mathbb{Z}_2$) are forced by general theorems that close entire shelves, and the one cheaper color rival, $CP^2$, was built end-to-end and breaks at the gauge gate for a structural reason — the strongest current result. Across the considered competitor class, under the description-length metric, no rival exhibits a strictly shorter complete recipe; the eliminations are themselves description-length eliminations (a tunable match $=$ an injected real $=$ a longer recipe), which is what keeps the win non-circular.

Open, and named. Level 2's two conditional pillars — the metric bridge and full in-grammar exhaustion — plus the weakest link, the fitted flavor normalizations in $F^+$. Level 3 — architecture-neutral minimality — is the universal-negative wall, out of reach in full for every theory ever written; on the live board that demand is dissolved as not owed (the Shape deep-root row), and it is tracked here as the bounded research frontier. The given-$E$ wall is the central open question: the framework reads off enormous structure given the matter content, and forcing that content is exactly what no known principle yet does.

Where this sits on the gate board. The three levels above are this page’s own, stricter axis. On the live gate board — 33 requirement-gates, all 33 RESOLVED at +0 · 0 ANCHORED at +1 · 0 OPEN (ratified 2026-07-08) — the Shape deep-root gate closes at DERIVED-GIVEN-anchor, its one named axiom (the common-currency / description-length rule) declared openly and the universal negative dissolved as not owed. The same board states the complementary honest axis just as plainly: 0 of 33 gates are physics-closed — every closure rests on declared measured anchors. The two statements are not in tension; they are the two halves of the one discipline this page practices.

The strongest honest sentence the whole challenge supports:

Given the frozen 13D shape (with $K_6 = SU(3)/T^2$) and given the observed matter content $E$, the framework reads off the Standard-Model gauge algebra, three chiral generations, and charge quantization from geometry; within the internal-isometry grammar the 13D branch is the strongest known survivor, with two carriers forced outright and the cheaper color rival refuted by construction. The shape itself remains a declared, frozen anchor — not a uniqueness theorem.

That standing — unrefuted but not certified — is not a weakness dressed up as a virtue. It is exactly what a skeptical external reviewer should conclude, stated before they conclude it. The honesty is the methodology: we say precisely which game the shape is winning, which it is not yet winning, and what theorem would let everyone compare every game fairly.

Where to go next

This page is the synthesis. The pieces it ties together each have their own page:

• A0 — The Master Anchor — the single audit rule that grades the shape and every competitor by the same standard.
• Layer 1 — Why "simplest" means description-length — the metric bridge, the decisive open seam.
• Layer 2 — No unpaid exact labels — the anti-smuggling rule, and why anomaly cancellation is a filter, not a selector.
• Layer 3 — The allowed search grammars — the four games and the no-cross-grammar rule.
• Layer 4 — Carrier-forcing and the given-$E$ wall — where the isometry grammar's dominance is argued in full.
• The Anchors — how we judge the theory honestly — the navigational overview.

Supporting gate dossiers:

• The shape, in depth — the deep-root shape dossier.
• SG-2 — gauge recovery · SG-3 — three generations · SG-4 — hypercharge & anomaly.