In his book Metaphysics, George Schlesinger claims that the so-called infinite regress argument is “one of the most versatile weapons in the armoury of the metaphysician” (1983: 218). He is right, first, because the argument is truly one that is frequently used; several of the most notoriously well-known and discussed arguments in philosophy have been (and are) infinite regress arguments. He is right, also, in that not only are infinite regress arguments frequently used, they are also arguments attributed with great force. Often it is claimed that a particular philosophical proposition, or even the entire theory of which the proposition in question is only one part, should be rejected on the grounds that it generates an infinite regress. The infinite regress argument is indeed a “weapon”. Yet, for all its frequency and force, the argument itself (apart from its many instantiations) is surprisingly little discussed and there is, as I will try to show, reason to doubt whether it is always as effective as it is claimed to be. The key to understanding the infinite regress argument lies in our understanding the distinction between a regress that is vicious, and one that is benign. In my talk I discuss this distinction and argue that although the “standard view” is that a regress may be vicious for several different reasons, there is in fact but one (substantive) sense in which a regress can be vicious. Whether a particular regress is vicious (i.e. whether it exemplifies this “vicious-making feature”), moreover, is something that can only be decided once the theoretical framework in which the position from which the regress has been generated has been identified. The viciousness of a regress, I will argue, is, in this sense, “context-sensitive”.