Some remarks on Gill's theorems on Young modules. (Q2339481): Difference between revisions

A paper of Gill considers the multiplicities obtained by expressing the tensor product of two Young modules for a symmetric group as a direct sum of Young modules, namely \[ Y(\lambda)\otimes Y(\mu)=\bigoplus_\nu y^\nu_{\lambda,\mu}Y(\nu). \] Gill works over a field of characteristic \(p>0\) and deploys methods from the modular representation theory of finite groups to show, \textit{inter alia}, that \(y_{p\lambda,p\mu}^{p\nu}=y^\nu_{\lambda,\mu}\). He also gives a lower bound for the Cartan invariant \(c_{\lambda,\mu}\) of the Schur algebra in terms of base \(p\) expansions of \(\lambda\) and \(\mu\). The author provides an alternative approach to this problem by using polynomial representation theory of general linear groups and then applying the Schur functor. The main result is a formula for the Young module multiplicities in a tensor product decomposition stated in terms of the characters of the simple polynomial modules for general linear groups. This approach yields a different proof of the result of Gill above; the lower bound for the Cartan numbers is also obtained. A feature of this methodology is that it is no more difficult to work with Young modules for Hecke algebras of type \(A\) at roots of unity. In this setting the formula involves the characters of simple polynomial modules for quantised general linear groups. However the author points out one striking difference in the quantum case: namely whereas \(Y(\lambda)\otimes Y(\mu)\) has a natural module structure in the classical situation, one only assigns the tensor product of Young modules for the Hecke algebra a meaning as a virtual module. An example is given to demonstrate that this is not in general represented by a direct sum of Young modules in the quantised case.