It is well known that universally composable multiparty computation cannot, in general, be achieved in the standard model without setup assumptions when the adversary can corrupt an arbitrary number of players. One way to get around this problem is by having a trusted third party generate some global setup such as a common reference string (CRS) or a public key infrastructure (PKI). The recent work of Katz shows that we may instead rely on physical assumptions, and in particular tamper-proof hardware tokens. In this paper, we consider a similar but strictly weaker physical assumption. We assume that a player (Alice) can partially isolate another player (Bob) for a brief portion of the computation and prevent Bob from communicating more than some limited number of bits with the environment. For example, isolation might be achieved by asking Bob to put his functionality on a tamper-proof hardware token and assuming that Alice can prevent this token from communicating to the outside world...