Kant was never satisfied with the version of his “Refutation” published in 1787 (KrV, B 275-279). His dissatisfaction is already evident in the footnote added to the preface of the second edition of the Critique in 1787. As a matter of fact, Kant continued to rework his argument for at least six years after 1787. The main exegetical problem is to figure out who is the target of the “Refutation”: a non-sceptic idealist or a global sceptic of Cartesian provenance or (...) both. In this last case, a related problem is to know whether either of them is the Cartesian sceptic of the first Meditation, the idealist sceptic of the second Meditation and first part of the third Meditation, or some other non-sceptic idealist. I present and defend a new reconstruction of Kant’s “Refu­tation” as a successful argument against Mendelssohnian idealism of Cartesian provenance. This defence is based on a simple logical sketch of the proof provided by Dicker, but essentially modified in the light of Dyck’s insight about Kant’s opponent. How shall I support my reading? First, by appealing to overwhelming textual evidence according to which the proof is of the existence of mind-independent things, showing that Kant’s main opponent is Mendelssohn’s idealism of Cartesian provenance. Finally, I support my reading by showing that Kant’s “Refutation” is doomed to fail against all forms of global scepticism, but is quite successful against Mendelssohn’s idealism. (shrink)