Covariant representations of Hecke algebras and imprimitivity for crossed products by homogeneous spaces (Q932306)

The motivation for the questions addressed in this article comes from earlier work by the authors on formulating and solving an extension problem, namely, that of extending a unitary representation of a subgroup \(H\) to its ambient group \(G\), by using the theory of non-Abelian duality for crossed products. More precisely, the extension problem is reduced to that of imprimitivity, meaning, an equivalence between an induced representation and a certain type of regular representation. In the case that \(H\) is a normal subgroup of \(G\), the problem comes down to deciding when, for a given representation \(\nu\) of \(c_0(G/H)\), there is a unitary representation \(U\) of \(G/H\) such that the pair \((\nu, U)\) is covariant for the action of \(G/H\) on \(C_0(G/H)\); if this happens, \(\nu\) is equivalent to a multiple of the representation \(M\) by multiplication on \(L^2(G/H)\) (via the Stone-von Neumann theorem). In the present paper, the authors take one step further by considering non-normal subgroups. More precisely, they assume that \((G,H)\) forms a (discrete) Hecke pair, in the sense that the orbits of the left action of \(H\) on the coset space \(G/H\) are finite, and they view the Hecke algebra \(\mathcal{H}(G, H)\) of \((G, H)\) as the appropriate analogue of the quotient group from the normal case. The first step is to formulate a suitable covariance condition for pairs \((\nu, V)\) of \(*\)-representations of \(c_0(G/H)\) and \(\mathcal{H}(G, H)\), respectively. Then the first main result (Theorem 1.6) characterises such pairs \((\nu, V)\) as those which are unitarily equivalent to \((1\otimes M, 1\otimes \rho)\), where \(M\) is multiplication and \(\rho\) is the right-regular representation of the Hecke algebra. In particular, a representation \(\nu\) is equivalent to a multiple of \(M\) if and only if there is a representation \(V\) of the Hecke algebra such that \((\nu, V)\) forms a covariant pair. As a first application, they obtain an imprimitivity theorem (Theorem 3.2) which says the following: for a Hecke pair \((G, H)\) and a maximal coaction of \(G\) on a \(C^*\)-algebra \(B\), a representation \((\pi, \nu)\) of the Echterhoff-Quigg crossed product \(B\times_\delta G/H\) has the property that \(\pi\) is equivalent to a regular representation if and only if there is a representation \(V\) of the Hecke algebra such that \((\nu, V)\) forms a covariant pair and \(\pi\) and \(V\) have commuting ranges. For Hecke algebras which are semigroup crossed products in the sense of Brenken and Laca-Larsen, the authors also obtain a characterisation of covariance for pairs \((\nu, V)\) in terms of Murphy-covariance for a semigroup dynamical system arising from the Hecke pair and of ordinary covariance for a group dynamical system also arising from the particular type of Hecke pair. The final application, Theorem 4.3, employs the imprimitivity result to give an answer to the extension problem in the case of a subgroup \(H\) of \(G\) from a Hecke pair \((G, H)\). For the class of Hecke pairs that give rise to semigroup crossed products, this answer takes a particularly elegant and accessible form.