Alphabeta Math
LemmaStatement: Literature-sourcedProof: AI-adaptedPipeline-generatedprecheck passverified 2026-11-05 (gpt-6.1-sol)
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.

The separability projector realizes the relative tensor product

Statement

Let C and the special Frobenius algebra data be as in MTC-valued Frobenius algebra and defect data for oriented rational CFT. Let A,B be such algebras, X an A-B-bimodule, and M a left B-module. Put eB=ΔB∘ηB:1→B⊗B. With canonical associators and unitors inserted, define PX,M=(ρXr⊗ρM)∘(id⁡X⊗eB⊗id⁡M):X⊗M→X⊗M. Then PX,M is an idempotent, and its split image is the coequalizer X⊗BM of the right B-action on X and left B-action on M. The left A-action on X descends to this image.

Facts & Assumptions

Given: The finite semisimple complex ribbon category, the special Frobenius algebra B, the A-B-bimodule X, and the left B-module M with the action and associator conventions of the linked definition.

[F1]

MTC-valued Frobenius algebra and defect data for oriented rational CFT: mB∘ΔB=id⁡B, the Frobenius identities hold, and the left and right actions on X commute. The relative tensor product is the coequalizer of the two stated action maps.

Proof

1.1givenF1algebra

Define eB=ΔBηB. The Frobenius identities and the two unit identities give the centrality equality (mB⊗id⁡B)∘(id⁡B⊗eB)=ΔB=(id⁡B⊗mB)∘(eB⊗id⁡B), as maps B→B⊗B. Specialness gives mB∘eB=ηB. These are equalities of morphisms; in Sweedler shorthand they say beB(1)⊗eB(2)=eB(1)⊗eB(2)b and eB(1)eB(2)=1B.

2.1givenF1step 1.1algebra

In the same shorthand, interpreted as composition with the morphism eB and the indicated module actions rather than as elements of the objects, PX,M(x⊗m)=xeB(1)⊗eB(2)m. Applying P twice gives xeB(1)eB′(1)⊗eB′(2)eB(2)m, where the primed copy denotes the second occurrence of eB. Apply centrality to the primed copairing with the input slot eB(2): eB′(1)⊗eB′(2)eB(2)=eB(2)eB′(1)⊗eB′(2). The double application therefore equals xeB(1)eB(2)eB′(1)⊗eB′(2)m. Now mBeB=ηB and the unital right action on X reduce it to xeB′(1)⊗eB′(2)m=PX,M(x⊗m). This substitution composes the centrality morphism with the second output of the unprimed copairing; no interchange of the two outputs or braiding is used. Thus P2=P as endomorphisms of X⊗M.

3.1givenF1step 1.1step 2.1algebra

Centrality also gives P∘(ρXr⊗id⁡M)=P∘(id⁡X⊗ρM). If q:X⊗M→Z equalizes those action maps, then qP(x⊗m)=q(xeB(1)⊗eB(2)m)=q(x⊗eB(1)eB(2)m)=q(x⊗m), so qP=q. Conversely, if qP=q, the displayed equality for P implies that q equalizes the two action maps. Hence the coequalizing maps are exactly the maps fixed by precomposition with P.

4.1givenF1step 2.1step 3.1algebra∎

Finite semisimplicity splits P as X⊗BM→iX⊗M→rX⊗BM with ir=P and ri=id⁡. For every coequalizer q, step 3.1 gives qP=q, so q=(qi)r; if q′=q′′r, then q′′=q′i by ri=id⁡. Thus r has the coequalizer universal property. Since the left A-action on X commutes with its right B-action, its tensor with id⁡M commutes with P and preserves the image; restricting along i,r defines the descended left A-action.

Depends on

Dependency tree · 2 levels

Sources