ScalingStacks

8.3.5. Action of τ\tau

The action of τ\tau on E⁡(T,−)⊗E⁡(−,S)E(T,-)\otimes E(-,S) corresponds to an endomorphism of Rξ2−2⊕Rξ2−​Rξ1−⊕Rξ1−​Rξ2−⊕Rξ1−2R_{\xi_{2}^{-}}^{2}\oplus R_{\xi_{2}^{-}}R_{\xi_{1}^{-}}\oplus R_{\xi_{1}^{-}}R_{\xi_{2}^{-}}\oplus R_{\xi_{1}^{-}}^{2} given in (5.3.4).

0PE2

Lemma 8.3.6. We have τ∘((f2,f1)⊗(f2,f1))=((f2,f1)⊗(f2,f1))∘τ\tau\circ((f_{2},f_{1})\otimes(f_{2},f_{1}))=((f_{2},f_{1})\otimes(f_{2},f_{1}))\circ\tau.

0PE3

Proof. Consider αi∈Rξi−∙​(T,U)\alpha_{i}\in R_{\xi_{i}^{-}}^{\bullet}(T,U) and βi∈Rξi−∙​(U,S)\beta_{i}\in R_{\xi_{i}^{-}}^{\bullet}(U,S). In order to prove that the equality of the lemma holds when applied to ((α2,α1)⊗(β2,β1))((\alpha_{2},\alpha_{1})\otimes(\beta_{2},\beta_{1})), we can assume that (αi)|T=id(\alpha_{i})_{|T}=\operatorname{id}\nolimits and (βi)|S=id(\beta_{i})_{|S}=\operatorname{id}\nolimits, since the morphisms involved in the equality commute with the right action of 𝒮M​(Z){\mathcal{S}}_{M}(Z).

We have

τ∘(fi⊗fj)​(αi⊗βj)=\tau\circ(f_{i}\otimes f_{j})(\alpha_{i}\otimes\beta_{j})=
=τ(((αi)ξi−​(−1)⋅[ξ1−(−1)→ξi−(−1)]⊠id)⊗(id⊠(βj)ξj−​(−1)⋅[ξ1−(−1)→ξj−(−1)]))\displaystyle=\tau\Bigl(\bigl((\alpha_{i})_{\xi_{i}^{-}(-1)}\cdot[\xi_{1}^{-}(-1)\to\xi_{i}^{-}(-1)]\boxtimes\operatorname{id}\nolimits\bigr)\otimes\bigl(\operatorname{id}\nolimits\boxtimes(\beta_{j})_{\xi_{j}^{-}(-1)}\cdot[\xi_{1}^{-}(-1)\to\xi_{j}^{-}(-1)]\bigr)\Bigr)
=di,j((βj)ξj−​(−1)⋅[ξ1−(−1)→ξj−(−1)]⊠id)⊗(id⊠(αi)ξi−​(−1)⋅[ξ1−(−1)→ξi−(−1)])\displaystyle=d_{i,j}\bigl((\beta_{j})_{\xi_{j}^{-}(-1)}\cdot[\xi_{1}^{-}(-1)\to\xi_{j}^{-}(-1)]\boxtimes\operatorname{id}\nolimits\bigr)\otimes\bigl(\operatorname{id}\nolimits\boxtimes(\alpha_{i})_{\xi_{i}^{-}(-1)}\cdot[\xi_{1}^{-}(-1)\to\xi_{i}^{-}(-1)]\bigr)

where d2,1=0d_{2,1}=0, d1,2=1d_{1,2}=1 and di,i=1d_{i,i}=1 if ((αi)ξi−​(−1)⋅[ξi−(−2)→ξi−(−1)]⊠(βi)ξi−​(−1))⋅τ≠0\bigl((\alpha_{i})_{\xi_{i}^{-}(-1)}\cdot[\xi_{i}^{-}(-2)\to\xi_{i}^{-}(-1)]\boxtimes(\beta_{i})_{\xi_{i}^{-}(-1)}\bigr)\cdot\tau\neq 0 and di,i=0d_{i,i}=0 otherwise. We deduce that the lemma holds when applied to ((α2,α1)⊗(β2,β1))((\alpha_{2},\alpha_{1})\otimes(\beta_{2},\beta_{1})), hence it holds in general. ∎

Original mathematics by the credited authors. Source collection and HTML conversion remain in progress.

Andrew Manion, Raphael Rouquier

Original source: arXiv:2009.09627v2