Proponents of a recent interpretation of Ernst Mach's views on thought experiments argue that for Mach thought experiments must be continuous with and return to cached experiences. These criteria, the proponents hold, explain the tension which has been noted in Mach's views on thought experiments: on the one hand, Mach reprimands Newton in “extending principles beyond the boundaries of experience” when critiquing Newton's bucket argument, while on the other, Mach himself engages in speculative reasoning when presenting ‘Mach's principle’. Against this interpretation, I argue that Mach's critique of Newton's argument does not turn on Newton’s illegitimate use of thought experiments, but instead turns on Newton's illegitimate use of thoughts in thought experiments. Recognizing this leads to a simple and more compelling interpretation of Mach's views. I close the paper by gesturing towards such an interpretation and noting that taking Mach's views on thought experiments seriously implies that Mach is much less of a strict empiricist than he is usually understood to be.
