Kant's views on the epistemological status of physical science provide an important example of how a philosophical system can be applied to understand the foundation of scientific theories. Michael Friedman has made considerable progress towards elucidating Kant's philosophy of science; in particular, he has argued that Kant viewed Newton's law of universal gravitation as necessary for the possibility of experiencing what Kant called true motion, which is more than the mere relative motion of appearances but is different from Newton's concept of absolute motion. In this context, Friedman has provided an account of how Kant must have viewed Newton's supposed derivation of universal gravitation from Kepler's laws, based on, among other things, Kant's claim that Newton really needed to make extra assumptions in order to derive universal gravitation. In this paper, I argue that Friedman's account is incomplete for three reasons. First, Friedman has overlooked an important aspect of how Newton's third law is applied in the relevant sections of the Principia; as a result, Friedman's account partially misconstrues the relation between the planetary phenomena and the theory of universal gravitation. Second, his account fails to account for Kant's apparent belief that Kepler's laws are only empirically-based rules, even though they seem to be necessary for the derivation of universal gravitation and hence also necessary for Kant's own definition of true motion. Third, Friedman has overlooked some remarks by Kant that indicate that Kant thought the crucial properties of universal gravitation could be known without reference to the empirically determined motions of the planets and hence seemingly without any help from Newton.