Nonequilibrium work theorem for a system strongly coupled to a thermal environment

In a recent paper (2004 J. Stat. Mech.: Theor. Exp. P07006), Cohen and Mauzerall (CM) have argued that the derivation of the nonequilibrium work relation given in a previous work of Jarzynski (1997 Phys. Rev. Lett. 78 2690) is flawed. Here I attempt to answer their criticisms, both by presenting a detailed version of that derivation and by addressing specific objections raised by CM. The derivation presented here is in fact somewhat stronger than the one I gave in 1997, as it does not rely on the assumption of a weak coupling term connecting the system of interest and its thermal environment.