Alphabeta Math
LemmaStatement: AI-adaptedProof: AI-adaptedPipeline-generatedjudge pass (gpt-6-luna)
How statement and proof provenance work

The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.

  • Literature-sourced: the exact statement appears in a cited source; only wording and notation differ.
  • AI-adapted: a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
  • AI-generated: a genuinely novel statement formulated by AI, with no source for the claim itself.

These labels describe origin, not correctness: citations and verification chips remain separate evidence.

Partial-trace closure of negligible morphisms

Statement

In the pivotal setting of Negligible morphisms for a pivotal trace, negligible morphisms are closed under addition, scalar multiplication, arbitrary composition on either side, and tensor product on the right by arbitrary morphisms. They are also closed under tensor product on the left, hence form a two-sided tensor ideal, under either of these sufficient hypotheses: the two pivotal traces agree on every endomorphism, or the category admits a braiding. In particular the conclusion applies to the planar Temperley–Lieb category and to the braided full tilting category used below. No factorization of an arbitrary tensor-product intertwiner is assumed.

Facts & Assumptions

Given: A C-linear rigid monoidal category with pivotal structure j, and a negligible morphism f:X→Y.

[F1]

Negligibility means Tr⁡L(jXgf)=0 for all g:Y→X (Negligible morphisms for a pivotal trace).

[F2]

The pivotal structure is a monoidal natural isomorphism, and the categorical trace is cyclic, including the typed different-object identity (Pivotal structure, Basic properties of the categorical trace).

[F3]

Evaluation and coevaluation satisfy the two zig-zag identities (Rigid object and rigid monoidal category). Write tX(u)=Tr⁡L(jXu) and rX(u)=Tr⁡R(ujX−1) for the two pivotal traces.

[F4]

When the braided alternative is used, Braiding supplies natural isomorphisms cZ,X:Z⊗X→X⊗Z. In the other alternative assume explicitly tX(u)=rX(u) for every object and every endomorphism.

[F5]

The corrected finite rank-one R series gives the full tilting ribbon structure constructs a natural invertible braiding on the full prequotient rank-one tilting category for every ℓ≥3, including even ℓ, by the finite integral R series; its Proofs 3.1 and 4.1 extend the construction from tensor powers to all tilting summands.

[L1]

For the Temperley–Lieb specialization, the partial trace is the cap/cup closure of Unnormalized diagram closure and partial trace, and the component-counting identity is proved in Diagram closure trace is cyclic and satisfies the tower rules.

Proof

technique · direct
1.1F1F2algebra

For any a:Y→Z and any h:Z→X, the trace of h(af) is Tr⁡L(jXhaf). By the different-object cyclicity in [F2] and naturality jYf=f∨∨jX, this equals Tr⁡L(jYfha). Since ha:Y→X, [F1] makes this zero. The same argument with a map b:W→X and a test map g:Y→W gives Tr⁡L(jXbgf)=0; hence post- and pre-composition preserve negligibility. Closure under sums and scalar multiples follows from linearity of trace.

2.1F1F2F3step 1.1construct

Let Z be any object and h:Y⊗Z→X⊗Z. Write ev⁡~Z=ev⁡Z∨(jZ⊗1Z∨):Z⊗Z∨→1, the right evaluation determined by the pivotal structure, and define ptr⁡Z(h)=(1X⊗ev⁡~Z)(h⊗1Z∨)(1Y⊗coev⁡Z):Y→X. Expanding the trace of jX⊗Zh(f⊗1Z) gives ev⁡Z∨⊗X∨(((jX⊗jZ)h(f⊗1Z))⊗1Z∨⊗X∨)(1X⊗coev⁡Z⊗1X∨)coev⁡X. The tensor-evaluation formula factors ev⁡Z∨⊗X∨ by first evaluating the middle Z∨∨⊗Z∨ pair and then the outer X∨∨⊗X∨ pair; substituting ev⁡~Z leaves exactly Tr⁡L(jXptr⁡Z(h)f). This regrouping uses monoidality of j and the dual zig-zags [F2,F3]. Since ptr⁡Z(h):Y→X is an allowed test map in [F1], the value is zero, so f⊗1Z is negligible.

3.1F1F2F3F4step 2.1construct

Suppose first that the two pivotal traces agree on all endomorphisms. For h:Z⊗Y→Z⊗X set coev⁡~Z=(1Z∨⊗jZ−1)coev⁡Z∨ and q=(ev⁡Z⊗1X)(1Z∨⊗h)(coev⁡~Z⊗1Y):Y→X. Expand the right pivotal trace rZ⊗X(h(1Z⊗f)) using the tensor evaluation and coevaluation formulas. Its inner closure is the Z-closure defining q; the remaining outer closure is rX(qf). This is the mirror of the nested calculation in step 2.1 and uses only the two snakes and monoidality of j. Thus tZ⊗X(h(1Z⊗f))=rZ⊗X(h(1Z⊗f))=rX(qf)=tX(qf)=0, by the trace-agreement hypothesis and negligibility of f. This proves left tensor closure in this alternative; the two changes of trace are essential.

3.2F1F4step 1.1step 2.1algebra

Suppose instead that a braiding is given. Step 2.1 makes f⊗1Z negligible. Naturality gives 1Z⊗f=cZ,Y−1(f⊗1Z)cZ,X. The two outside factors have the displayed compatible types, and composition closure from step 1.1 therefore proves 1Z⊗f negligible. This alternative does not require equality of the two pivotal traces.

4.1F1F2F5L1step 1.1step 2.1step 3.1step 3.2algebra∎

For any g:Z→W, f⊗g=(1Y⊗g)(f⊗1Z) and g⊗f=(g⊗1Y)(1Z⊗f). Composition closure from step 1.1 and step 2.1 give right tensor closure for every pivotal category. Under either stated additional hypothesis, step 3.1 or step 3.2 also gives left tensor closure. In the Temperley–Lieb category, [L1] also verifies the partial-trace identities by counting components of the closed finite stacked graph. For the full rank-one tilting category, [F5] supplies the prequotient braiding required in step 3.2, so its negligible morphisms also form a two-sided tensor ideal for every ℓ≥3. Together with linearity this proves every stated closure assertion.

Remarks

The extra hypothesis for left tensor closure cannot simply be omitted. Consider the C-linear category of orientation-compatible noncrossing pairings of boundary words in X,X∨. Composition joins pairings; each resulting oriented circle contributes 0 in the orientation defining tX(1X) and 1 in the opposite orientation. A closed component remains closed in every further composition, so the total component decomposition proves associativity and tensor interchange. Cups and caps satisfy the snakes, rotation gives duals, and two rotations give the identity pivotal comparison. Taking the additive envelope gives the setting of the Definition, with End⁡(1)=C and End⁡(X)=C1X. Thus tX(1X)=0 makes 1X negligible. However, for a=coev⁡X∨:1→X∨⊗X and b=ev⁡X:X∨⊗X→1, cyclicity gives tX∨⊗X(ab)=ba=1. Hence 1X∨⊗1X is not negligible. This is a counterexample to the unqualified generality, not an open problem.

Depends on

Used by

Dependency tree · two levels

26 results within two dependency steps of this one, each drawn at its shortest distance from it. An arrow runs from a result to what uses it, so the chart reads left to right and ends at this result, which carries a heavier outline. Every node is a link to that result. Click elsewhere on the chart to enlarge it.

Sources