Alphabeta Math
TheoremStatement: Literature-sourcedProof: AI-adaptedPipeline-generatedprecheck passaudited 2026-09-27
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.

Double leaves form graded R-bases of type-A diagrammatic Hom spaces

Statement

Let n≥2, let D be the type-A diagrammatic Soergel category of The type-A diagrammatic Soergel category and its candidate bimodule functor and let x‾=(x1,…,xp) and y‾=(y1,…,yq) be two words, with products x,y∈Sn in the Coxeter presentation of Type-A reduced words and the Coxeter presentation. Write LL for the recursively constructed light leaves of Elias–Williamson, Construction 6.1. For each w∈Sn, fix a reduced word w‾ (with e‾=∅), used as the common target for every leaf expressing w, as in Remark 6.3 of that source. Then for every word x‾, every element w∈Sn and every subexpression e of x‾ expressing w one has a light leaf LLx‾,e∈Hom⁡D(x‾,w‾),deg⁡LLx‾,e=d(e), where d(e)=#U0−#D0∈Z is the defect of the subexpression, and a double leaf is the composite LL‾y‾,f∘LLx‾,e, first the light leaf from x‾ and then the flipped light leaf towards y‾, with e,f subexpressions of x‾,y‾ that express the same element w∈Sn and  ‾ denoting the vertical flip, so that the composite is an element of Hom⁡D(x‾,y‾). Then:

  1. the set of double leaves is a homogeneous free left R-basis of the graded left R-module Hom⁡D(x‾,y‾);
  2. when y‾=∅, the subfamily with w=e (the identity of Sn) is a homogeneous free left R-basis of Hom⁡D(x‾,∅);
  3. consequently all hom spaces of D are free graded left R-modules, and their graded ranks are ∑w∑e,f: we=wf=wvd(e)+d(f), over the indexing pairs of subexpressions.

Facts & Assumptions

Given: The category D with its presentation (The type-A diagrammatic Soergel category and its candidate bimodule functor), whose candidate assignment on generators respects the relations by (The type-A diagrammatic relations hold for Soergel bimodules), and words x‾,y‾ as above.

[F1]

D is the Q-linear monoidal category generated by colored strands, polynomial boxes, dots, trivalent vertices and the 2mst-valent vertices, with the degrees of Elias–Williamson, Definition 5.1 and the relations of §§5.1–5.6, so that D is graded and its generators have those degrees (The type-A diagrammatic Soergel category and its candidate bimodule functor).

[F2]

The elements of Sn and the subexpression relation come from the Coxeter presentation of Type-A reduced words and the Coxeter presentation: a subexpression of x‾ is a word obtained by replacing each letter xa by either xa (a step of type 1) or 1 (a step of type 0), and each step is moreover of "up" or "down" type according to whether the length of the partial product increases or decreases; writing U0,U1,D0,D1 for the four possibilities, the defect is d(e):=#U0−#D0, an integer, following Elias–Williamson §2.4.

[F3]

Imported construction (Elias–Williamson, §6.1, Construction 6.1 with Figure 2): for every (x‾,e) expressing w and every k≤ℓ(x‾) the truncated light leaf on the first k letters is obtained from the truncated light leaf on the first k−1 letters by a rex move β when the k-th letter is a down step, followed by exactly one of four local moves, namely the dot U0 of degree +1, the identity U1 of degree 0, the merging trivalent vertex D0 of degree −1 and the cap D1 of degree 0; the recursion is thus on the prefixes of x‾, each local move has degree equal to the change in the defect, so the resulting light leaf has degree d(e) (The type-A diagrammatic Soergel category and its candidate bimodule functor).

[F4]

Imported statement (Elias–Williamson, Theorem 6.11, with Proposition 6.12 and Corollary 6.13): the double leaves form a free R-basis of Hom⁡D(Bx‾,By‾), the light leaves LLx‾,e with w=e form a free R-basis of Hom⁡D(Bx‾,1), and all hom spaces of D are free graded R-modules (The type-A diagrammatic Soergel category and its candidate bimodule functor).

[F5]

Imported input (Elias–Williamson, §7, Proposition 7.6 with the coupled maxwidth induction of §7 applied to the lower-term ideals): with the lower-term ideals Iw generated by the maps killing the top summands, every morphism of D between Bott–Samelson objects lies in the span of the double leaves modulo the ideal of strictly lower terms, which supplies the spanning half of [F4]; the proof uses only the relations of D and the induction on the maximum width of a graph (The type-A diagrammatic Soergel category and its candidate bimodule functor).

[F6]

Imported input (Elias–Williamson, Proposition 6.9, with the localization argument of §6.3): after localization the space of maps between Bott–Samelson objects is a direct sum of terms indexed by pairs of subsequences, these terms carry a partial order, and the double-leaf maps are upper triangular with respect to it with an invertible diagonal; hence the double leaves are linearly independent over the fraction field of R (The type-A diagrammatic Soergel category and its candidate bimodule functor).

Proof

1.1

The indexing set: by [F2], subexpressions of x‾ describing a fixed w are finite in number, each has a well-defined defect d(e)∈Z, and the target of LLx‾,e is the chosen reduced-word object w‾ for the element w that e expresses; consequently the double leaves are indexed by the pairs (e,f) of subexpressions of x‾,y‾ with the same product w, and each such pair carries the degree d(e)+d(f).

F2F3
1.2

Homogeneity: by [F3] each local move in the recursion for LLx‾,e has degree equal to the change of the defect, so LLx‾,e is homogeneous of degree d(e) in the grading of D fixed by [F1]; the vertical flip preserves degrees, so the double leaves are homogeneous of the displayed degrees.

F1F3
1.3

The unit case: when y‾ is empty the only admissible product is w=e and the flipped light leaf LL‾∅,∅ is the empty diagram, so the double leaves LLx‾,e with w=e are exactly the light leaves to the unit.

F3F4
1.4

Spanning: this is the imported content of [F5]: every morphism between Bott–Samelson objects is congruent modulo the lower-term ideal to an R-linear combination of double leaves, and induction on the maximum width of a graph descends through the lower terms, so the double leaves span Hom⁡D(x‾,y‾) over R.

F5
1.5

Linear independence: the independent-family clause of the imported double-leaves theorem [F4] applies to the exact category D of [F1]. Its localization proof uses the upper-triangular double-leaf calculation of [F6] on pairs of source and target subsequences; a calculation for source light leaves alone would not establish independence of their composites. Thus the double leaves indexed in step 1.1 are linearly independent over R.

F1F4F6step 1.1
2.1

Freeness: by step 1.4 the finite family of double leaves spans the graded left R-module Hom⁡D(x‾,y‾), and by step 1.5 it is linearly independent. It is therefore a homogeneous free basis. The degree of each basis element is d(e)+d(f) by step 1.2, so its graded rank is the stated finite sum of Laurent monomials.

step 1.2step 1.4step 1.5
3.1

Conclusion: by steps 1.5, 2.1 and 1.4 the double leaves form a homogeneous free left R-basis of Hom⁡D(x‾,y‾), which is claim (1); claim (2) is step 1.3 together with the same argument in the case y‾=∅, which is the flipped-leaves clause of the imported basis theorem [F4]; claim (3) is the existence of a finite free basis in each target degree, the graded rank being the sum of the monomials vd over the indexing pairs. ∎

F4step 1.3step 2.1step 1.4

Depends on

Used by

Dependency tree · two levels

15 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