Summary Canfield and Lehrer argued in [1] that it is impossible to predict an event deductively from given laws and initial conditions. In [2] I showed (1) that there are counterexamples to C&L's claim, (2) that their argument was based on an assumption which they failed to recognize, let alone support (the assumption that there are ‘rejectible’ facts), (3) that if certain simple conditions on predictability were granted, it could be shown that C&L's tacit assumption is false. In a recent (...) issue of this Journal Dr. Harnatt has criticized (3) (ignoring (1) and (2)), and he has concluded that we should still wonder whether there are any deductive predictions. Dr. Harnatt's arguments are shown to be defective. (shrink)