We argue that honesty in assertion requires non-empirical knowledge that what one asserts is what one believes. Our argument proceeds from the thought that to assert honestly, one must follow and not merely conform to the norm ‘Assert that p only if you believe that p’. Furthermore, careful consideration of cases shows that the sort of doxastic self-knowledge required for following this norm cannot be acquired on the basis of observation, inference, or any other form of detection of one's own doxastic states. It is, as we put it, transparent rather than empirical self-knowledge.

It has become standard to treat Kant'scharacterization of pure apperception as involving the claim that questions about what I think are transparent to questions about the world. By contrast, empirical apperception is thought to be non-transparent, since it involves a kind of inner observation of my mental states. I propose a reading that reverses this: pure apperception is non-transparent, because conscious only of itself, whereas empirical apperception is transparent to the world. The reading I offer, unlike the standard one, can accommodate Kant'sclaim that the I of pure apperception is the same as the I of empirical apperception.

According to the transparency approach, achievement of self-knowledge is a two-stage process: first, the subject arrives at the judgment ‘p’; second, the subject proceeds to the judgment ‘I believe that p.’ The puzzle of transparency is to understand why the transition from the first to the second judgment is rationally permissible. After revisiting the debate between Byrne and Boyle on this matter, I present a novel solution according to which the transition is rationally permissible in virtue of a justifying argument that begins from a premise referring to the mental utterance that is emitted in the course of judging ‘p.’

According to some influential readings of the Groundwork of the Metaphysics of Morals, the view presented there of the kind of spontaneity we are conscious of through theoretical reason and the significance of such self-consciousness is irremediably at odds with the Critical theory, and thus roundly and rightly rejected in the second edition of the Critique of Pure Reason and the Critique of Practical Reason. This paper argues, on the contrary, that the Groundwork can be read as articulating for the first time the account of self-consciousness and spontaneity that Kant goes on to develop in the B-Critique, especially the B-Transcendental Deduction.

The paper argues that Kant'sdistinction between pure and empirical apperception cannot be interpreted as distinguishing two self-standing types of self-knowledge. For Kant, empirical and pure apperception need to co-operate to yield substantive self-knowledge. What makes Kant'saccount interesting is his acknowledgment that there is a deep tension between the way I become conscious of myself as subject through pure apperception and the way I am given to myself as an object of inner sense. This tension remains problematic in the realm of theoretical cognition but can be put to work and made productive in terms of practical self-knowledge.

This paper relates Kant'saccount of pure apperception to the agential approach to self-knowledge. It argues that his famous claim ‘The I think must be able to accompany all of my representations’ (B131) does not concern the possibility of self-ascribing beliefs. Kant does advance this claim in the service of identifying an a priori warrant we have as psychological persons, that is, subjects of acts of thinking that are imputable to us. But this warrant is not one to self-knowledge that we have as critical reasoners. It is, rather, an a priori warrant we have, as thinkers, to prescribe to given representations their conformity to principles of thinking inherent in our capacity of understanding itself.

Much recent work on self-knowledge has been inspired by the idea that the ‘transparency’ of questions about our own mental states to questions about the non-mental world holds the key to understanding how privileged self-knowledge is possible. I critically discuss some prominent recent accounts of such transparency, and argue for a Sartrean interpretation of the phenomenon, on which this knowledge is explained by our capacity to transform an implicit or ‘non-positional’ self-awareness into reflective, ‘positional’ self-knowledge.