A theme in Quine's writing is that proper languages for the austere purposes of science should be extensional. At one point, if memory serves, he says that if propositional attitude predicates must be intensional, so much the worse for them. They aren't needed in a proper science anyway. Similarly for dispositions. They are good shorthand, but they are modal notions, which are intensional, and so disposable in a proper extensional language of science. Where did this idea originate? I don't remember any arguments in Quine to the effect that scientific predicates are extensional. Is this an idea that can be traced back to the Vienna Circle or someone before then? Davidson picks it up and uses it to argue for some analysis of causation [edit: in "Causal Relations."]. Someone (name escapes me and I can't find the articl [edit: Crane and Mellor's "There is No Question of Physicalism."]) responds to Davidson by arguing that causation is actually an intensional notion [edit: and other intensional contexts appear in physics, e.g. probability]. So, why would one think that scientific notions are essentially extensional anyway? If dispositional properties and counterfactuals or other modal notions are involved in scientific theories, then one wouldn't think that scientific theories are extensional. No idea really.