Don Garrett's strategy, in Cognition and Commitment in Hume's Philosophy, is to reconcile apparent contradictions in Hume's theory. Two examples occur early in Part I of Book I of Treatise. Hume argues that all ideas are derivable impressions (what Garrett calls the Copy Principle). Garrett defends Hume's rigorous attempt to found this principle empirically, taking Hume at his word when he describes it as the first principle I establish in science of human nature. (T 7)1 Notoriously, Hume almost immediately provides a counter-example: suppose a person had experienced all shades of blue except one, and further suppose they were presented with a table of all these shades, with a blank where missing shade ought to be. Clearly, Hume says, such a person could supply from his own relevant idea, without having first experienced corresponding impression. This presents an apparent counter-example to Copy Principle. Another example occurs a little later. Hume claims that Where-ever imagination perceives a difference among ideas, it can easily produce a separation. (T 10) Distinct ideas are separable. Garrett calls this the Separability Principle, and it is almost as important for Hume as Copy Principle. For one thing, it shows that Hume's theory of ideas, though heavily influenced by Locke, is importantly different earlier theory: Locke thought there were some necessary connections between distinct ideas, connections that could not be reduced to identities or partial identities.2 More importantly, separability of distinct ideas is primary ground of Hume's conceivability criterion of demonstration, crucially employed in arguments concerning causation, necessary connection and induction. But Hume seems to present an apparent counter-example to Separability Principle as well. When discussing distinctions of reason, Hume admits that we can distinguish between the