The Endomorphic Collapse Traverses the Foundations of Mathematics

Abstract

This quiver's canonical bilinear form is provably dead: indefinite, selecting no Dynkin type at all. One bit, a sign on a single vertex, is the entire distance from that form to the Cartan matrix of su(3) ⊕ su(2). The forms deadness is what makes the gauge content live in the bit and not in the graph. This paper is an inheritance walkthrough: a record of one formalism after another (counting, the path algebra, representation theory, and the canonical bilinear (Tits) form) reaching its own expressive limit, each limit proven and each naming the formalism that inherits the question. It starts at counting (the graded dimensions, where counting cannot compose) and hands to the path algebra. The path algebra composes the four-gon γ₁ (a Hamiltonian cycle) but cannot express the convergence at T as a relation: the three length-2 paths through T live in orthogonal idempotent components, and equating them would force e_S = e_L, which is false. It hands to representation theory. Under the convergence constraint dim(V_T) = 1 together with the involution conditions, every vertex space in an indecomposable representation is forced to dimension at most one (the rank argument). The quotient algebra cannot produce the braid relation ((γ₁γ₂)³ is a nonzero length-21 element, not e_L) and cannot produce nonzero commutation (γ₁γ₃ = 0 by vertex mismatch). It hands to the canonical bilinear invariant. The Tits form of Q on the cycle-support basis is indefinite, with two negative diagonal entries and one isotropic (zero-diagonal) generator, so it is not a Cartan matrix of any Dynkin type. Here the inheritance stops with a proven negative result: Q, by its own canonical invariants, does not select a finite reflection group. Obtaining a Dynkin type requires structure added from outside Q. This stopping point is different in kind from the earlier ones. The earlier handoffs (counting to path algebra to representation theory) each named a richer formalism within mathematics to continue, and each hit the same ceiling because the missing datum was not a formalism a richer one would supply. The inheriting invariant here is the bilinear form on H₁(Q; ℝ), and the structure it needs is not internal to Q at all: it is the single distinction the graph's own combinatorics cannot carry. The brain's signed convergence, the resolution at T that produces a positive-or-negative determination from the inputs converging on it (Stewart, 2026c, Circuit 3), supplies that added structure. It is a single binary distinction (vertex T carries a sign, vertex C does not), fixed before the path algebra is constructed. Executing the four axioms of occurrence on that one distinction yields a bilinear form B, shown unique up to diagonal scaling, under which the same cycle generators that gave the dead Tits form realize the Cartan matrix of A₂ × A₁, returning the Lie algebra su(3) ⊕ su(2) via the Cartan-Killing correspondence. The difference between an indefinite, structure-selecting-nothing form and the strong-plus-weak gauge algebra is one bit: which vertex carries the sign. The cycle-generator basis has rank three, the bound that A₂ × A₁ saturates. The fourth generator required for the u(1) of the full Standard Model gauge algebra is not supplied by H₁(Q; ℝ) and is developed in companion work (Stewart, 2026q). The proven deadness of the canonical form is what makes the gauge content carried by the single added distinction rather than by the graph. This is the central result. The shape of that single imported bit is the Möbius half-twist. A closed band that carries an orientation closes cleanly (a cylinder, two sides, orientable) only when the orientation is signed. Unsigned, it closes with a half-twist into a Möbius, one side, non-orientable, and the twist is exactly the seam where the loop cannot sign its own orientation from inside (Consequence 7b). The dead Tits form is that unsigned band, and the imported sign is the twist that would orient it. This is the same bit, seen statically: here it is the sign that revives the algebra in a single traversal, and iterated, it is the arrow of time. The temporal generator γ_τ is this twist under iteration, the second circle of the torus whose defining property is that the involution fails (γ_τ² ≠ e, it generates ℤ not ℤ/2ℤ, it does not reverse), and it fails for the same reason the graph cannot sign itself: the cycle cannot read the orientation of its own live act, so it must hand off to be read at all. The static sign this paper imports and the arrow of time are one bit at two scales, the sign in one traversal, the arrow across iterations. The derivation of time from this twist (the escapement as the blind spot, γ_τ as the counter, the involution-failure as the arrow, the 3+1 structure) is the next movement of the story and is developed in *The Clock of the Endomorphism* and *The Temporal Generator* (Stewart, 2026q). The twist's own form, tying the imported sign, non-orientability, and 7b into one figure, is the owed Möbius form-result named in *The Catalog of Form*. This paper is one γ₁ traversal, spatial and complete on its own terms, and time is what the traversal produces when it hands off to traverse again, which cannot be derived inside a single pass and belongs to the paper on the iteration. The inheritance sequence is itself the four-position cycle: counting (logic, distinction) → path algebra (composition) → representation theory (the type judgment dim(V_T) = 1, the many-to-one identity-by-collapse at T) → the bilinear morphism B (composition again). The paper builds mathematical structure in the order mathematics builds it, by letting each formalism break at its own ceiling and naming its heir. For an occurrence whose construction traverses the very cycle it constructs, being γ₁ is the subject, doing γ₁ is its predicate, and the operand is being in the objective, the subject received by its own act. They are one act, not three, because a self-applied act is one thing declined through its own cases, the subject taken as the object of its own predicate. The construction performed below is one γ₁ traversal of the cycle whose bilinear invariant the construction returns. The gauge content that invariant carries, su(3) ⊕ su(2) at the rank the cycle basis supplies, is read off from inside the cycle by the cycle executing. Mathematics is one γ₁. Gauge structure is what γ₁ produces. The paper is the act in which mathematics and its gauge content are seen to be the same cycle viewed from opposite sides of itself. **Keywords:** quiver, path algebra, quiver representations, quotient algebra, Tits form, Euler form, Cartan matrix, Coxeter group, Dynkin diagram, root system, Lie algebra, special unitary group, first homology