On the decomposition of the Foulkes module. (Q1945791): Difference between revisions

Suppose \(a\) and \(b\) are positive integers. The `Foulkes module' \(H^{(a^b)}\) is the permutation module over \(\mathbb C\) for the symmetric group \(\mathfrak S_{ab}\) with basis the set of partitions of \(\{1,\dots,ab\}\) into \(b\) sets of size \(a\). The `Foulkes conjecture' asserts that when \(a\geqslant b\) there is an injective homomorphism from \(H^{(a^b)}\) to \(H^{(b^a)}\); in other words, the multiplicity of any irreducible character for \(\mathbb C\mathfrak S_{ab}\) as a constituent of the character \(\varphi^{(a^b)}\) of \(H^{(a^b)}\) is less than or equal to its multiplicity as a constituent of \(\varphi^{(b^a)}\). One approach to this conjecture is to calculate (or at least estimate) these composition multiplicities in certain cases. In the paper under review, a sufficient condition is given on the partition \(\lambda\) for the irreducible character \(\chi^\lambda\) not to appear at all in \(\varphi^{(a^b)}\); computer experiments show that, excluding characters dealt with by more elementary considerations, this accounts for a large proportion of the irreducible characters not appearing in the Foulkes character. A special case of the main result is that if \(\lambda\) is a hook partition \((ab-k,1^k)\) then the character \(\chi^\lambda\) does not appear in the Foulkes character. The methods used are very classical, working with permutation modules and elementary character theory, and a few well-known results on the ordinary representation theory of the symmetric group, such as the Littlewood-Richardson rule. The main method of proof is induction, using generalised Foulkes characters. The paper is very carefully written, and the author's exposition is excellent; there are only a couple of places where he reveals that English is not his first language. In addition to proving the main theorem, the paper provides a basic introduction to the Foulkes conjecture for the reader unfamiliar with the subject.