Monday, 28 May 2012

rt.representation theory - Invariant quadratic forms of irreducible representations

Let G be a finite group, and k be a field of characteristic zero (not necessarily algebraically closed!). Let rho:GtomathrmEndkleft(knright) be a irreducible representation of G over k. Consider the vector space



S=leftlbraceHinmathrmEndkleft(knright)midrholeft(gright)THrholeft(gright)=HtextforanyginGrightrbrace



=leftlbracesumlimitsginGrholeft(gright)THrholeft(gright)midHinmathrmEndkleft(knright)rightrbrace



and its subspace



T=leftlbraceHinSmidHtextisasymmetricmatrixrightrbrace.



It is easy to show that, if we denote our representation of G on kn by V, then the elements of S uniquely correspond to homomorphisms of representations VtoVast (namely, HinS corresponds to the homomorphism vmapstoleft(wmapstovTHwright)), while the elements of T uniquely correspond to G-invariant quadratic forms on V (namely, HinT corresponds to the quadratic form vmapstovTHv).



(1) In the case when k=mathbbC, Schur's lemma yields dimSleq1, with equality if and only if VcongVast (which holds if and only if V is a real or quaternionic representation). Thus, dimTleq1, and it is known that this is an equality if and only if V is a real representation. (Except of the equality parts, this all pertains to the more general case when k is algebraically closed of zero characteristic).



(2) In the case when k=mathbbR, it is easily seen that Tneq0 (that's the famous nondegenerate unitary form, which in the case k=mathbbR is a quadratic form), and I think I can show (using the spectral theorem) that dimT=1. As for S, it can have dimension >1.



(3) I am wondering what can be said about other fields k; for instance, k=mathbbQ. If ksubseteqmathbbR, do we still have dimT=1 as in the mathbbR case? In fact, Tneq0 can be shown in the same way.

No comments:

Post a Comment