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 and be signed words with the same product , and let be the corresponding word complexes of The Rouquier complex of a braid word. Then:
- is one-dimensional over , its generator being represented by a morphism of internal degree zero;
- the canonical localization map is an isomorphism of one-dimensional -vector spaces;
- 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 , and there is a unique element mapping to it. In particular is a homotopy equivalence with and , and its class is the unique normalized comparison between the two words.
Facts & Assumptions
Given: Signed words with product , the word complexes of The Rouquier complex of a braid word, and the generator relations of Opposite Rouquier generator complexes are homotopy inverse, Rouquier complexes satisfy far commutativity, Rouquier complexes satisfy the three-term braid relation.
Invertibility of word complexes. Every word complex is invertible in : , where is the reversed word with inverted signs; this follows from the generator relations by induction on the length of the word, tensoring the identities , for distant 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)
The unit and its endomorphisms. A degree-zero bimodule map is multiplication by its value at , which must lie in . The unit complexes have no possible nonzero chain homotopies, so . Since both objects are modules in degree zero, their degree-zero derived-category Hom is the ordinary module Hom, giving as well: apply the boundary Hom formula of The canonical pair is a t structure with to graded -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 , an isomorphism in the homotopy category transports to ; it does not assert a generic identification with all of .
Tensoring with an invertible object. Let be a monoidal category and let admit a two-sided inverse : isomorphisms and . Using the associativity and unit isomorphisms of , these exhibit natural isomorphisms and ; hence is an equivalence of categories with quasi-inverse 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 In and 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.
Derived graph models. and in , and the comparison of the two words induces an isomorphism of these models whose class in 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 and the two graph models coincide. (Rouquier generator complexes have canonical derived graph models, Canonical comparisons between standard graph tensor products)
Proof
By [F1] the object is invertible with inverse , so by [F3] the functor is an equivalence with quasi-inverse and the adjunction gives a natural bijection; applied with and it identifies with . The same argument applies in with the derived tensor product.
The complex is a word complex for the word , which represents the trivial braid because and represent the same element; by the relations of [F1] it is homotopy equivalent to the unit complex , and likewise isomorphic to in .
By step 1.2 choose a homotopy equivalence and its homotopy inverse. Composition with gives a vector-space isomorphism by [F2]. Combining with step 1.1 proves that is one-dimensional in internal degree zero.
Localizing the equivalence and its inverse gives the same Hom transport in . Together with the localized tensor equivalences of step 1.1, this identifies the target Hom with . The localization square commutes with these transports, and its map on sends the identity to the identity. It is therefore an isomorphism; hence so is the localization map on .
By [F4] the comparison element of the graph models is a nonzero element of the one-dimensional computed in step 3.1, so it has a unique preimage under the localization isomorphism. Applying step 3.1 also to the pairs and shows that the localization maps and are isomorphisms; since the comparisons satisfy by [F4], the unique preimages satisfy , and symmetrically . Hence 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 an equivalence and hence one-dimensional, and the localization isomorphism transfers the canonical comparison from to a unique homotopy class. The degree-zero requirement is essential: the graded endomorphism object of the unit is the polynomial ring , not , and only its internal-degree-zero part is used. The map is normalized by the derived condition of matching 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 is the unique preimage of .
Depends on
- The Rouquier complex of a braid word
- Rouquier complexes satisfy far commutativity
- Rouquier complexes satisfy the three-term braid relation
- Opposite Rouquier generator complexes are homotopy inverse
- Canonical comparisons between standard graph tensor products
- Rouquier generator complexes have canonical derived graph models
- Derived category of an abelian category
- The localization functor sends quasi isomorphisms to isomorphisms
- The homotopy category of chain complexes
- Standard graph bimodules, support filtrations and characters
- Equivalence, quasi-inverse, and adjoint equivalence of categories
- Every equivalence of categories can be equipped as an adjoint equivalence
- An adjoint equivalence is an adjunction whose unit and counit are natural isomorphisms
- The unit-counit, hom-set, unit-universal, and counit-universal encodings of an adjunction are equivalent
- Bounded bimodule tensor is associative, unital, and compatible with cones
- Homology factors uniquely through the homotopy category
- The canonical pair is a t structure
- The braid group by Artin presentation
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
- Raphaël Rouquier, Categorification of the braid groups, arXiv:math/0409593v1 (30 September 2004), §3 "The 2-braid group" (standard reference, not scraped)