Alphabeta Math
LemmaStatement: Literature-sourcedProof: AI-adaptedPipeline-generatedprecheck pass
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.

Derived comparisons give unique normalized homotopy maps

Statement

Let t and u be signed words with the same product w∈Bn, and let F(t),F(u) be the corresponding word complexes of The Rouquier complex of a braid word. Then:

  1. Hom⁡Kb(F(t),F(u))=Q⋅[γt,u] is one-dimensional over Q, its generator being represented by a morphism of internal degree zero;
  2. the canonical localization map Hom⁡Kb(F(t),F(u))⟶Hom⁡Db(F(t),F(u)) is an isomorphism of one-dimensional Q-vector spaces;
  3. the comparison element of Canonical comparisons between standard graph tensor products read through the derived graph models (Rouquier generator complexes have canonical derived graph models) is a nonzero element of the one-dimensional Hom⁡Db(F(t),F(u)), and there is a unique element γt,u∈Hom⁡Kb(F(t),F(u)) mapping to it. In particular γt,u is a homotopy equivalence with γu,tγt,u=id and γt,uγu,t=id, and its class is the unique normalized comparison between the two words.

Facts & Assumptions

[F1]

Invertibility of word complexes. Every word complex F(t) is invertible in Kb(Re-grmod): F(t)⊗RF(t−1)≃R≃F(t−1)⊗RF(t), where t−1 is the reversed word with inverted signs; this follows from the generator relations by induction on the length of the word, tensoring the identities Fi⊗RFi−1≃R, Fi⊗RFj≅Fj⊗RFi for distant i,j and the three-term relation. By the Artin presentation The braid group by Artin presentation, equal braid words differ by finitely many relation replacements and inverse-pair insertions or deletions: their quotient in the free group is a finite product of conjugates of relators. Tensoring the generator equivalences in those word contexts therefore compares any two words for the same braid. The same relations hold after localization, with tensoring by these two-sided finite-free complexes computed by ordinary totalization. (Opposite Rouquier generator complexes are homotopy inverse, Rouquier complexes satisfy far commutativity, Rouquier complexes satisfy the three-term braid relation, The Rouquier complex of a braid word)

[F2]

The unit and its endomorphisms. A degree-zero bimodule map R→R is multiplication by its value at 1, which must lie in R0=Q. The unit complexes have no possible nonzero chain homotopies, so Hom⁡Kb(R,R)=Q⋅id. Since both objects are modules in degree zero, their degree-zero derived-category Hom is the ordinary module Hom, giving Hom⁡Db(R,R)=Q⋅id as well: apply the boundary Hom formula of The canonical pair is a t structure with a=b=0 to graded Re-modules, so both cohomological and internal degrees are zero (The homotopy category of chain complexes, Derived category of an abelian category, Standard graph bimodules, support filtrations and characters). For Z≃R, an isomorphism in the homotopy category transports Hom⁡Kb(R,Z) to Hom⁡Kb(R,R); it does not assert a generic identification with all of H0(Z)0.

[F3]

Tensoring with an invertible object. Let C be a monoidal category and let X∈C admit a two-sided inverse X−1: isomorphisms X⊗X−1→1 and X−1⊗X→1. Using the associativity and unit isomorphisms of C, these exhibit natural isomorphisms (−⊗X−1)(−⊗X)≅idC and (−⊗X)(−⊗X−1)≅idC; hence −⊗X is an equivalence of categories with quasi-inverse −⊗X−1 in the sense of Equivalence, quasi-inverse, and adjoint equivalence of categories. By Every equivalence of categories can be equipped as an adjoint equivalence the pair can be equipped as an adjoint equivalence, and the adjunction then gives, by An adjoint equivalence is an adjunction whose unit and counit are natural isomorphisms together with The unit-counit, hom-set, unit-universal, and counit-universal encodings of an adjunction are equivalent, a natural bijection Hom⁡(U⊗X,W)≅Hom⁡(U,W⊗X−1). In Kb(Re-grmod) and Db(Re-grmod) the associativity and unit isomorphisms are those of Bounded bimodule tensor is associative, unital, and compatible with cones and their images under localization, so the bijection is available in both categories.

[F4]

Derived graph models. F(t)≅Rπ(w)(−e(t)) and F(u)≅Rπ(w)(−e(u)) in Db(Re-grmod), and the comparison ct,u of the two words induces an isomorphism of these models whose class in Hom⁡Db is nonzero; since the two Artin relations have equal exponent sums on both sides and inverse pairs have exponent zero, the exponent is invariant on braid words and e(t)=e(u) and the two graph models coincide. (Rouquier generator complexes have canonical derived graph models, Canonical comparisons between standard graph tensor products)

Proof

technique · direct
1.1F1F3

By [F1] the object F(t) is invertible with inverse F(t−1), so by [F3] the functor −⊗RF(t) is an equivalence with quasi-inverse −⊗RF(t−1) and the adjunction gives a natural bijection; applied with U=R and W=F(u) it identifies Hom⁡Kb(F(t),F(u)) with Hom⁡Kb(R,F(u)⊗RF(t−1)). The same argument applies in Db with the derived tensor product.

1.2F1

The complex Z:=F(u)⊗RF(t−1) is a word complex for the word ut−1, which represents the trivial braid because t and u represent the same element; by the relations of [F1] it is homotopy equivalent to the unit complex R, and likewise isomorphic to R in Db.

2.1F1F2step 1.1step 1.2

By step 1.2 choose a homotopy equivalence e:Z→R and its homotopy inverse. Composition with e gives a vector-space isomorphism Hom⁡Kb(R,Z)→Hom⁡Kb(R,R)=Q by [F2]. Combining with step 1.1 proves that Hom⁡Kb(F(t),F(u)) is one-dimensional in internal degree zero.

3.1F2F3step 1.1step 1.2step 2.1

Localizing the equivalence e and its inverse gives the same Hom transport in Db. Together with the localized tensor equivalences of step 1.1, this identifies the target Hom with Hom⁡Db(R,R)=Q. The localization square commutes with these transports, and its map on Hom⁡(R,R) sends the identity to the identity. It is therefore an isomorphism; hence so is the localization map on Hom⁡(F(t),F(u)).

4.1F4step 3.1∎

By [F4] the comparison element of the graph models is a nonzero element of the one-dimensional Hom⁡Db(F(t),F(u)) computed in step 3.1, so it has a unique preimage γt,u under the localization isomorphism. Applying step 3.1 also to the pairs (u,t) and (t,t) shows that the localization maps Hom⁡Kb(F(u),F(t))→Hom⁡Db(F(u),F(t)) and Hom⁡Kb(F(t),F(t))→Hom⁡Db(F(t),F(t)) are isomorphisms; since the comparisons satisfy cu,tct,u=ct,t=id by [F4], the unique preimages satisfy γu,tγt,u=id, and symmetrically γt,uγu,t=id. Hence γt,u is a homotopy equivalence and its class is the unique normalized comparison.

Remarks

The argument is Rouquier's §3.3.1: the invertibility of the word complexes makes −⊗RF(t) an equivalence and hence Hom⁡(F(t),F(u))≅Hom⁡(R,F(u)⊗RF(t−1)) one-dimensional, and the localization isomorphism transfers the canonical comparison from Db to a unique homotopy class. The degree-zero requirement is essential: the graded endomorphism object of the unit is the polynomial ring R, not Q, and only its internal-degree-zero part is used. The map γt,u is normalized by the derived condition of matching ct,u through the graph models, and this normalization pins it down uniquely by step 4.1. No choice principle is needed: the invertibility data are fixed by the generator relations, the equivalence-to-adjunction conversion is constructive, and γt,u is the unique preimage of ct,u.

Depends on

Used by

Dependency tree · two levels

66 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