Measuring Absolute Velocity

We argue that Roberts's argument for the thesis that absolute velocity is not measurable in a Newtonian world is unsound, because it depends on an analysis of measurement that is not extensionally adequate. We propose an alternative analysis of measurement, one that is extensionally adequate and entails that absolute velocity is measured in at least one Newtonian world. If our analysis is correct, then this Newtonian world is a counterexample to the widely endorsed thesis that if a property varies under the symmetries of a theory then, according to that theory, the property could not be measured. Thus, our paper shows that the debate over the measurability of symmetry-variant properties is more unsettled than previously supposed.