1 Transcendental idealism and metaphysics: Kant’s commitment to things as they are in themselves 1 One of Kant’s central central claims in the Critique of Pure Reason is that we cannot have knowledge of things as they are in themselves. He says: our cognition reaches appearances only, leaving the thing in itself as something actual for itself, but uncognized by us (Bxx). The status of things in themselves in Kant’s system has been regarded as problematic in a number of ways, for example, whether Kant is entitled to assert both that there are things in themselves and that we cannot have knowledge of them, and, more generally, what Kant’s commitment to things in themselves amounts to. A number of commentators deny that Kant is committed to there actually being an aspect of reality which we cannot cognise; they argue that he is committed merely to the idea that we cannot avoid the concept of things as they are in themselves. 1 Perhaps one of the most important proponents of this view is Henry Allison (1983, 2004), who argues that transcendental idealism is not a metaphysical position, but a methodological standpoint. Allison famously argues that Kant’s distinction between things in themselves and appearances should be understood in terms of the idea of considering things in terms of epistemic conditions, and the idea of things considered apart from such conditions, together with an emphasis on the idea that cognition requires both concepts and intuitions. Apparent support for anti-metaphysical, or deflationary, readings is provided by the implausibility of extreme metaphysical interpretations. 2 Deflationary interpreters (correctly, it seems to me) reject the idea that Kantian appearances should be understood as ideal in the Berkelean sense, and reject the idea that Kant’s notion of things in themselves should be understood as a commitment to there being supersensible, non-spatio-temporal ‘objects,’ of which we cannot have knowledgeintelligibilia. However, from the facts that Kant is not committed to the existence of intelligibilia and that he is not a Berkelean idealist, it does not follow that he is not committed to there being a way things are in themselves, which we cannot cognise, or that he is not committed to appearances being genuinely dependent on our minds in some (non-Berkelean) sense. And while the claim that we cannot know things in themselves is of course an epistemic claim, this does not mean that it involves no metaphysical commitmentsuch as a commitment to the existence of an aspect of reality which we cannot cognise. I will argue in this paper that while transcendental idealism is partly an epistemological position, it is also partly a metaphysical position, and in specific, that Kant is committed to the claim that the things we cognise have, in addition to the way they appear to us, a nature that is independent of us, which we cannot cognise. 3 Kant starts out with an epistemic 1 See Bird, who says of the division between phenomena and noumena in the negative sense that “such a division is not, like an empirical distinction, between two genuine kinds of object, but only between phenomena, which are the things we ordinarily perceive and know about, and the empty (but not inconsistent) concept of a non-phenomenon” (Bird 1962:74), and that there “are two ways of looking at the same thing only because, on Kant’s view, there is only one thing at which to look, namely appearances” (Bird 1962:29). See also Bird 2006: 553, 579; Grier 2001; Hanna 2006; Schrader 1968: 173, 181; Senderowicz 2005. In contrast, for authors who take it as entirely obvious that Kant does have a commitment to things in themselves, see Adams 1997; Adickes 1924; Walker (forthcoming). 2 Bird seems to tie a commitment to there being a way things are in themselves to what he calls the traditional idealist interpretation, which sees Kant as a Berkelean idealist about appearances (2006:566). 3 My concern here is to argue that Kant is committed to things in themselves in the sense of there being a way in which the things we cognise are in themselves, which we cannot cognise. It is of course