Foulkes modules and decomposition numbers of the symmetric group. (Q404367)

This paper connects two important themes in the representation theory of symmetric groups: on-going attempts at the very difficult problem of describing explicitly the decomposition matrices of the symmetric groups in prime characteristic, and the study of Foulkes modules, which has seen a resurgence of interest in recent years. This recent interest in Foulkes modules has concentrated on the case of characteristic zero (and in particular on the Foulkes Conjecture), but the paper under review uses the Foulkes module \(H^{(2^m)}\) in odd characteristic \(p\) to give information about decomposition numbers; in particular, the authors describe explicit columns of the decomposition matrix for blocks of arbitrary \(p\)-weight. To state the main theorem, suppose \(p\) is an odd prime, \(\gamma\) is a \(p\)-core and \(k\) is a non-negative integer. Let \(w_k(\gamma)\) denote the minimum number of \(p\)-hooks that need to be added to \(\gamma\) to give a partition with exactly \(k\) odd parts, and let \(\mathcal E_k(\gamma)\) denote the set of partitions that can be obtained in this way. Then (provided \(w_k(\gamma)\leqslant w_{k-p}(\gamma)\)) the set \(\mathcal E_k(\gamma)\) can be written as a disjoint union \(\mathcal X_1\sqcup\cdots\sqcup\mathcal X_c\) in such a way that for each \(j\) there is a column of the decomposition matrix of the corresponding symmetric group in characteristic \(p\) which has \(1\)s in the rows labelled by the partitions on \(\mathcal X_j\), and \(0\)s in all other rows. The proof involves twisted Foulkes modules. Recall that the Foulkes module \(H^{(2^m)}\) is the permutation module for the symmetric group \(S_{2m}\) on the set of partitions of \(\{1,\ldots,2m\}\) into pairs. The twisted Foulkes module \(H^{(2^m;k)}\) is obtained by taking the outer tensor product with the sign representation for \(S_k\) and inducing to \(S_{2m+k}\). The important ingredient in the proof is the identification of the vertices and Green correspondents of the summands of \(H^{(2^m;k)}\). Knowledge of the ordinary character of \(H^{(2^m)}\), together with the machinery of modular representation theory, gives the main result. The paper is meticulously written and very nice to read. The authors collect a wide variety of background material (and take the opportunity to give a survey of results on decomposition numbers) in an accessible way, and organise their proof very neatly. The paper concludes with some applications and examples.