On the boundary and intersection motives of genus 2 Hilbert-Siegel varieties (Q2312694)

The aim of this paper is to construct a Chow motive for genus \(2\) Hilbert-Siegel varieties. The main technical input is to show how a general result of Wildeshaus can be applied in this context. The construction of Chow motives for Shimura varieties goes back to \textit{A. J. Scholl} [Invent. Math. 100, No. 2, 419--430 (1990; Zbl 0760.14002)], who proved the existence of such a motive realising Deligne representations attached to elliptic cuspforms. This paper extends such a result to the case of genus \(2\) Hilbert-Siegel varieties, namely, Shimura varieties \(S_K\) attached to the group \(G=\mathrm{Res}_{F/\mathbb{Q}}(\mathrm{GSp}_{4,F})\), where \(F\) is a totally real number field of degree \(d\) over \(\mathbb{Q}\), and a neat open compact subgroup \(K\) of \(G(\mathbb{A}_f)\), where \(\mathbb{A}_f\) are the finite ideles of \(\mathbb{Q}\). The main result can be stated as follows. Let \(V_\lambda\) be an irreducible representation of \(G\) of highest weight \(\lambda\). The highest weight \(\lambda\) corresponds to a choice of a couple of vectors of non-negative integers \((k_{1,\sigma})_{\sigma\in I_F}\) and \((k_{2,\sigma})_{\sigma\in I_F}\) where \(I_F\) are the embeddings of \(F\) into \(\mathbb{R}\). We say that \(\lambda\) is regular if \(k_{1,\sigma}>k_{2,\sigma}\) for all \(\sigma\in I_F\). Let \(\mu(V_\lambda)\) denote the local system on \(S_K(\mathbb{C})\) and \(\mu_\ell(V_\lambda)\) be the \(\ell\)-adic sheaf on \(S_K\) attached to \(V_\lambda\). The author shows that if \(\lambda\) is regular there exists a Chow motive over \(\mathbb{Q}\) whose Hodge realisation is equal to \(H^{3d}_!(S_K(\mathbb{C}),\mu(V_\lambda))\) and whose \(\ell\)-adic realisation is equal to \(H^{3d}_!(S_K\otimes_\mathbb{Q}\bar{\mathbb{Q}},\mu_\ell(V_\lambda))\). He also shoes that every element of the Hecke algebra attached to \((K,G(\mathbb{A}_f)\) acts naturally on this motive. The main technical input to prove such a result is to show that a criterion by \textit{J. Wildeshaus} is satisfied in this context (see in particular the papers [Compos. Math. 145, No. 5, 1196--1226 (2009; Zbl 1179.14021); Adv. Math. 305, 515--600 (2017; Zbl 1360.14078); ``On the intersertion motive of certain Shimura varieties: the case of Siegel threefolds'', to appear in Annals of K-Theory (2019)]). Basically, one needs to shows that the boundary of the Baily-Borel compactification of \(S_K\) satisfies a certain cohomological condition. More precisely, let \(S_K^*\) denote the Baily-Borel compactification of \(S_K\), \(j:S_K\hookrightarrow S_K^*\) the canonical map, \(i:\partial S_K^*\hookrightarrow S_K^*\) the closed immersion of the boundary of the compactification. Consider the complex of \(\ell\)-adic sheaves \(i^*j_*\mu_\ell(V_\lambda)\). The condition to check is that there are no weights \(0\) and \(1\) in these sheaves, where the weights are those defined by the Galois action. This is checked by analysing the perverse cohomology of \(i^*j_*\mu_\ell(V_\lambda)\) stratum by stratum on the boundary. Even if the paper is long and technical, the style in which results and proofs are presented is very clear and provides numerous references and background which can be used to the reader to orient her/himself. A long and detailed introduction is especially useful to fully understand the main result and the technical tools used to prove it.