The shuffle map #
Let C be a preadditive monoidal category with w-small coproducts, in which tensoring on either
side preserves w-small coproducts (for instance ModuleCat k). For simplicial sets K and L
and objects R and S of C, the shuffle map of Eilenberg and Mac Lane is the morphism of chain
complexes SSet.shuffle K L R S from K.chainComplex R ⊗ L.chainComplex S, the tensor product of
the simplicial chains of the factors, to (K ⊗ L).chainComplex (R ⊗ S), the simplicial chains of
the product K × L. The tensor product of chain complexes is Mathlib's monoidal structure on
ChainComplex C ℕ, whose differential is d (a ⊗ b) = d a ⊗ b + (-1)^p a ⊗ d b for a of
degree p. The shuffle map is the other half, besides the Alexander–Whitney map
SSet.alexanderWhitney, of the Eilenberg–Zilber comparison between chains on a product and tensor
products of chains.
The shuffle map is first constructed on the standard simplices. The shuffle chain
SSet.shuffleChain T p q (p + q) is a (p + q)-chain of Δ[p] ⊗ Δ[q] with coefficients in T.
Unwinding its recursion, it is the signed sum of the nondegenerate (p + q)-simplices of
Δ[p] ⊗ Δ[q]: the monotone lattice paths from (0, 0) to (p, q), each with the sign
(-1)^N, where N counts the pairs of a vertical step followed, later on the path, by a
horizontal step. Here a path is built from its first step: either a horizontal step (0, 0) → (1, 0) followed by a path from (1, 0), or a vertical step (0, 0) → (0, 1) followed by a path
from (0, 1) with the sign (-1)^p. The two cases are the cone from (0, 0)
(SSet.stdSimplex.prodConeChain) on the shuffle chains of Δ[p - 1] ⊗ Δ[q] and Δ[p] ⊗ Δ[q - 1],
pushed along the zeroth face maps. The shuffle map then sends the summand of a p-simplex x of
K and a q-simplex y of L to the image of the shuffle chain under the map
Δ[p] ⊗ Δ[q] ⟶ K ⊗ L classifying (x, y) (SSet.ιChainComplex_tensorHom_ιChainComplex_shuffle_f).
This is the classical formula x ⊗ y ↦ ∑ ± (s_ν x, s_μ y) over the (p, q)-shuffles (μ, ν),
although this file works with the recursion and does not state that closed formula.
The boundary of the cone is ∂ (c σ) = σ - c (∂ σ) in positive degrees
(SSet.stdSimplex.prodConeChain_d). An induction on the degree then shows that the boundary of
the shuffle chain is the alternating sum of its faces in the first factor plus (-1)^p times the
alternating sum of its faces in the second factor, which is exactly the statement that the shuffle
map is a morphism of chain complexes.
Main definitions and results #
SSet.stdSimplex.coneChain: the cone from the vertex0on the simplicial chains ofΔ[a], withSSet.stdSimplex.coneChain_dandSSet.stdSimplex.coneChain_zero_dits boundary formulas.SSet.stdSimplex.prodConeChain: the cone from the vertex(0, 0)on the simplicial chains ofΔ[a] ⊗ Δ[b], withSSet.stdSimplex.prodConeChain_dits boundary formula.SSet.shuffleChain: the shuffle chain ofΔ[p] ⊗ Δ[q], with its defining recursionSSet.shuffleChain_zero_zero,SSet.shuffleChain_succ_zero,SSet.shuffleChain_zero_succandSSet.shuffleChain_succ_succ, its boundary formulaSSet.shuffleChain_d_succ_succ,SSet.shuffleChain_d_succ_zeroandSSet.shuffleChain_d_zero_succ, and its naturality in the coefficient object.SSet.shuffle: the shuffle map, a morphism of chain complexes.SSet.ιChainComplex_tensorHom_ιChainComplex_shuffle_f: its value on the summand of a pair of simplices, andSSet.ιChainComplex_tensorHom_ιChainComplex_shuffle_f_zeroin degree zero.SSet.shuffle_naturality: it is natural in both simplicial sets.SSet.shuffle_coefficient_naturality: it is natural in both coefficient objects.
References #
- S. Eilenberg and S. Mac Lane, On the groups
H(Π, n), I, Ann. of Math. 58 (1953). - S. Eilenberg and J. A. Zilber, On products of complexes, Amer. J. Math. 75 (1953).
- C. Weibel, An Introduction to Homological Algebra, Section 8.5.
The cone from the vertex 0 on a simplex x of the standard simplex Δ[a]: the simplex
whose vertices are 0, x 0, …, x m.
Equations
- SSet.stdSimplex.cone x = SSet.stdSimplex.objMk { toFun := Matrix.vecCons 0 ⇑x, monotone' := ⋯ }
Instances For
The zeroth face of the cone on x is x.
The (i + 1)-st face of the cone on x is the cone on the i-th face of x.
The second face (index 1) of the cone on a vertex x is the vertex 0.
A map of standard simplices which fixes the vertex 0 commutes with cones.
The cone from the vertex (0, 0) on a simplex of the product Δ[a] ⊗ Δ[b].
Equations
Instances For
The zeroth face of the cone on a simplex x of Δ[a] ⊗ Δ[b] is x.
The (i + 1)-st face of the cone on a simplex x of Δ[a] ⊗ Δ[b] is the cone on the i-th
face of x.
The cone from the vertex 0, as a map raising the degree of the simplicial chains of Δ[a]
by one. It is a contracting homotopy in positive degrees (SSet.stdSimplex.coneChain_d), and in
degree zero it contracts onto the vertex 0 (SSet.stdSimplex.coneChain_zero_d).
Equations
- One or more equations did not get rendered due to their size.
Instances For
The cone on the simplicial chains of Δ[a] is a contracting homotopy in positive degrees:
∂ (c ∘ σ) = σ - c ∘ ∂ σ for a chain σ of positive degree.
The map on the 0-chains of Δ[a] which sends the summand of every vertex to the summand of
the vertex 0.
Equations
- One or more equations did not get rendered due to their size.
Instances For
In degree zero, the boundary of the cone on a vertex x is x minus the vertex 0.
Collapsing every vertex onto the vertex 0 kills boundaries.
Collapsing every vertex onto the vertex 0 kills boundaries.
The cone from the vertex (0, 0), as a map raising the degree of the simplicial chains of
Δ[a] ⊗ Δ[b] by one. It is a contracting homotopy in positive degrees
(SSet.stdSimplex.prodConeChain_d).
Equations
- One or more equations did not get rendered due to their size.
Instances For
The cone on the simplicial chains of Δ[a] ⊗ Δ[b] is a contracting homotopy in positive
degrees: ∂ (c ∘ σ) = σ - c ∘ ∂ σ for a chain σ of positive degree.
A map of products of standard simplices which commutes with the cones commutes with the cone on simplicial chains.
A map of products of standard simplices which commutes with the cones commutes with the cone on simplicial chains.
The cone on simplicial chains of Δ[a] ⊗ Δ[b] is natural in the coefficient object.
The cone on simplicial chains of Δ[a] ⊗ Δ[b] is natural in the coefficient object.
The shuffle chain of Δ[p] ⊗ Δ[q] with coefficients in T, a chain of degree
n = p + q. It is defined by recursion on the first step of a lattice path from (0, 0) to
(p, q): the shuffle chain of Δ[0] ⊗ Δ[0] is its unique vertex (SSet.shuffleChain_zero_zero),
and otherwise it is the cone from (0, 0) on the shuffle chain of Δ[p - 1] ⊗ Δ[q] pushed along
the zeroth face Δ[p - 1] ⟶ Δ[p], plus (-1)^p times the cone on the shuffle chain of
Δ[p] ⊗ Δ[q - 1] pushed along the zeroth face Δ[q - 1] ⟶ Δ[q], where a term is absent when the
corresponding index is zero (SSet.shuffleChain_succ_zero, SSet.shuffleChain_zero_succ and
SSet.shuffleChain_succ_succ).
Equations
- One or more equations did not get rendered due to their size.
- SSet.shuffleChain T 0 0 0 x_4 = (CategoryTheory.MonoidalCategoryStruct.tensorObj (SSet.stdSimplex.obj { len := 0 }) (SSet.stdSimplex.obj { len := 0 })).ιChainComplex SSet.vertex✝
- SSet.shuffleChain T 0 0 n.succ h = absurd h ⋯
- SSet.shuffleChain T n.succ x✝ 0 h = absurd h ⋯
- SSet.shuffleChain T 0 n.succ 0 h = absurd h ⋯
Instances For
The shuffle chain is natural in the coefficient object.
The shuffle chain is natural in the coefficient object.
The shuffle chain of Δ[0] ⊗ Δ[0] is its unique vertex.
The shuffle chain of Δ[p + 1] ⊗ Δ[0] is the cone on the shuffle chain of Δ[p] ⊗ Δ[0],
pushed along the zeroth face of Δ[p + 1].
The shuffle chain of Δ[0] ⊗ Δ[q + 1] is the cone on the shuffle chain of Δ[0] ⊗ Δ[q],
pushed along the zeroth face of Δ[q + 1].
The shuffle chain of Δ[p + 1] ⊗ Δ[q + 1] is the cone on the shuffle chain of
Δ[p] ⊗ Δ[q + 1] pushed along the zeroth face of Δ[p + 1], plus (-1)^(p + 1) times the cone on
the shuffle chain of Δ[p + 1] ⊗ Δ[q] pushed along the zeroth face of Δ[q + 1].
The boundary of the shuffle chain of Δ[p + 1] ⊗ Δ[q + 1]: the alternating sum of its faces
in the first factor plus (-1)^(p + 1) times the alternating sum of its faces in the second
factor, each face being the image of a smaller shuffle chain under a face map.
The boundary of the shuffle chain of Δ[p + 1] ⊗ Δ[0]: the alternating sum of its faces in
the first factor.
The boundary of the shuffle chain of Δ[0] ⊗ Δ[q + 1]: the alternating sum of its faces in
the second factor.
The shuffle map C(K; R) ⊗ C(L; S) ⟶ C(K × L; R ⊗ S) of Eilenberg and Mac Lane. It sends
the summand of a p-simplex x of K and a q-simplex y of L to the image of the shuffle
chain SSet.shuffleChain (R ⊗ S) p q (p + q) under the map Δ[p] ⊗ Δ[q] ⟶ K ⊗ L classifying
(x, y) (SSet.ιChainComplex_tensorHom_ιChainComplex_shuffle_f). It is a morphism of chain
complexes for the Koszul sign rule on the tensor product.
Equations
- K.shuffle L R S = { f := SSet.shuffleX✝ K L R S, comm' := ⋯ }
Instances For
The shuffle map on the summand of a p-simplex x of K and a q-simplex y of L is the
image of the shuffle chain of Δ[p] ⊗ Δ[q] under the map Δ[p] ⊗ Δ[q] ⟶ K ⊗ L classifying
(x, y).
The shuffle map on the summand of a p-simplex x of K and a q-simplex y of L is the
image of the shuffle chain of Δ[p] ⊗ Δ[q] under the map Δ[p] ⊗ Δ[q] ⟶ K ⊗ L classifying
(x, y).
In degree zero, the shuffle map sends the summand of a pair of vertices (x, y) to the summand
of the vertex (x, y) of K × L.
The shuffle map is natural in both simplicial sets.
The shuffle map is natural in both simplicial sets.
The shuffle map is natural in both coefficient objects.
The shuffle map is natural in both coefficient objects.