Abstract
So-called average subsystem entropies are defined by first taking partial traces over some pure state to define density matrices, then calculating the subsystem entropies, and finally averaging over the pure states to define the average subsystem entropies. These quantities are standard tools in quantum information theory, most typically applied in bipartite systems.
We shall first present some extensions to the usual bipartite analysis (including a calculation of the average tangle and a bound on the average concurrence), follow this with some useful results for tripartite systems, and finally extend the discussion to arbitrary multipartite systems. A particularly nice feature of tripartite and multipartite analyses is that this framework allows one to introduce an "environment" to which small subsystems can couple.