A Philosophical Model of the Relation between Things in Themselves and Appearances

The question of how best to understand Kant’s distinction between things in themselves and appearances is one of the central interpretive questions concerning Kant’s theoretical philosophy, as well as one of the thorniest. By and large, philosophers have approached this distinction by recasting it in terms for which they antecedently believe they have a clearer understanding. In short, these philosophers provide what we can think of as translations of Kant’s technical jargon into the jargon of contemporary philosophy. In what follows, I will propose what I will call, for want of a better name, a non-translationist approach to this distinction, which stresses a methodology I will call philosophical modeling. I will then apply this methodology to one of the thorniest questions in Kant interpretation, specifically, to the question of what the distinction between things in themselves and appearances consists in. As the name suggests, a non-translationalist approach does not attempt to recast, reconstruct, explicate, or explicitly define the Kantian technical terminology of things in themselves and appearances in terms of something we are antecedently more comfortable with. Instead, a non-translationalist approach takes these technical expressions on their own terms as it were, and attempts to provide a coherent account of how they are used in the Kantian philosophy. But of course it is hard to see how else one provides such an account without offering explicit definitions of the Kantian jargon. This is where the methodology of philosophical modeling will prove important. So let me turn to a brief discussion of the idea of philosophical modeling. Let me start with something relatively obvious. Whenever a philosopher offers a definition of a bit of jargon employed by her historical predecessor, she runs a risk of attributing to her predecessor concepts that he or she might not have possessed. Let me give a stark example of someone taking that risk. Suppose one were to offer the following account of the distinction between things in themselves and appearances; the toy account that I will offer is an instance of the so-called ‘one world’ family of views, according to which things in themselves are numerically identical with appearances. On this toy account, the distinction is a distinction between things considered with properties that are modally independent of us and things considered with properties that are modally dependent on us; a property is modally dependent on us just in case there is no possible world in which something