On the periods of automorphic forms on special orthogonal groups and the Gross-Prasad conjecture (Q2269897): Difference between revisions

Let \(F\) be a number field, \(A\) its adele ring. Let \(G\) be a reductive group over \(F\); \(H\) a closed subgroup of \(G\). For \(\phi\) a cuspidal automorphic form on \(G\) associates a period: \[ P_H(\phi)=\int_{H(F)\backslash H(A)} \phi(h)\,dh. \] There is much evidence that the values of \(P_H(\phi)\) are related to the \(L\)-values of \(\pi\) if \(\phi\) is a vector in the space of automorphic representations \(\pi\). An example of such relation is the global Gross-Prasad conjecture. Here the group \(G\) is \(\text{SO}_{Q_1}\times \text{SO}_{Q_0}\) and \(H\) is \(\text{SO}_{Q_0}\) sitting diagonally in \(G\), where \((V_0,Q_0)\subset (V_1,Q_1)\) are quadratic forms over \(F\) with rank \(n\) and \(n+1\). The conjecture states that the nonvanishing of the periods \(P_H(\phi)\) for all \(\phi\in \pi\) is equivalent to the local condition that \(\pi_v\) is \(H_v\) distinguished along with the global condition that \(L(\pi,\frac12)\) is nonzero. The paper explicates the global Gross-Prasad conjecture. For \(\pi\) an irreducible tempered cuspidal automorphic representation of \(G\), the authors conjecture the following explicit identity: with suitable choice of measures, for \(\phi\in \pi\), there is an integer \(\beta\) depending on \(\pi\) with \[ |P_H(\phi)|^2=2^{\beta}\int_{H(A)}\langle \pi(h)\phi,\phi\rangle\,dh. \] Here \(\langle,\rangle\) is inner product, and the integration is regularized. The paper also gives a relation between \(\beta\) and the other parameter of \(\pi\). The authors also comment on the situation when \(\pi\) is not tempered. The authors compute the local factors of the above integral, and show the factors are indeed the local factors of \(L(\pi,\frac12)\) (up to some fixed constant). The authors' conjecture is a remarkable one. Many low rank examples are given in the paper where the conjecture is known. These examples are significant results when they were published, now the Ichino-Ikeda conjecture fits all these results into a very simple framework.