Few major philosophers show evidence of having studied the works of their predecessors with special care, even in cases where they were subject to particular influences which they were ready to acknowledge. Hume knew that he was working in the tradition of ‘some late philosophers in England, who have begun to put the science of man on a new footing’—‘Mr Locke, my Lord Shaftsbury, Dr Mandeville, Mr Hutchinson, Dr Butler, &c.’ But there is not much sign in the Treatise or elsewhere in Hume's writings of any close acquaintance with the works of these authors; the presumption must be that he had read them at some time and extracted the main ideas, but was not in the habit of returning to their texts. He had something more important to do, namely to work at philosophical problems of his own. Similarly Kant, though he said that the Critique of Pure Reason was not meant to be ‘a critique of books and systems, but of the faculty of reason in general’, had clearly felt the impact of the thought of some important past philosophers, but equally had never spent much time in finding out just what these philosophers had to say. Plato, Aristotle, Descartes, Locke, Leibniz and Hume all get fairly frequent mention in his pages. But Kant takes his knowledge of Plato and Aristotle from J. J. Brucker's Historia critica philosophiae, a six-volume compilation which first appeared in 1742, or from doubtful sources such as Mendelssohn's doctored translation of the Phaedo, and though he doubtless knew the more recent authors at first hand clearly felt no need to study them in any depth. This was true even of writers to whom he attributed a particular importance, such as Leibniz and Hume. The references to Hume in the Critique and Prolegomena are all disappointingly general, and though the summary of Leibniz's philosophy in the section called ‘The Amphiboly of Concepts of Reflection’ has a certain force, it is not documented with references to Leibnizian texts. Kant knows that there is a difference between the views of the historical Leibniz and those which constituted the ‘Leibnizian-Wolffian system’ of his successors. But he is not very curious about the difference, or inclined to explore it.