I just finished reading Ameriks's essay "A Common-Sense Kant?" (Chapter 5 of Kant and the Historical Turn
, OUP: 2006). This is a nice paper in which he compares the philosophies of Reid and Kant, arguing that there are many methodological, epistemological and metaphysical positions shared by these two thinkers that have not been properly appreciated.
In his discussion of the ideality of the self, Ameriks makes an interesting claim. "Once Kant's idealism, even of the self, is expressed as relying essentially on the claim that our determinate self-knowledge is inevitably parasitic on spatial
knowledge, then the issue between Kant and his opponents needs to be focused on making sense of the basic claim that even spatiality (and therefore the self simply in so far as it and its temporality is dependent on spatiality) can be 'mere appearance'." (Ameriks: 2006, pp. 126-127)
However, I do not agree that the ideality of the phenomenal self is in this way connected to the ideality of space. Kant thinks that we know ourselves only as we appear and not as we are in ourselves because of the ideality of time. Time is the form of inner intuition. Since time is transcendentally ideal, it follows that all knowledge gained by inner sense is restricted to knowledge of the phenomenal realm, in particular to knowledge of the phenomenal self.
Kant's emphasis on inner sense being parasitic on outer sense is part of the Refutation of Idealism and the General Note on the System of the Principles, both of which are additions to the B-Edition. The restriction to knowledge of the self as appearance, however, is already part of the Transcendental Aesthetic of the A-Edition, where it seems to be purely based on considerations regarding time. Moreover, even if self-knowledge is parasitic on spatial knowledge and if spatial knowledge is only of mere appearances, it does not seem to follow that we only know ourselves as we appear. To reach this conclusion we need to appeal to specific considerations regarding the transcendental ideality of time. According to the Refutation of Idealism, self-knowledge is parasitic on spatial knowledge insofar as consciousness of our existence as determined in time requires there to be an outer permanent in perception. This is because all time-determination presupposes a permanent and because permanence can only be outer. But merely because we appeal to something that is ideal in determining our existence in time does not yet imply that we can only know ourselves as appearances. For this we must appeal to the ideality of time. But given the ideality of time, no further reference need be made to how we determine things in time and whether or not this time-determination requires appealing to spatial knowledge of mere appearances. That is, the transcendental ideality of time is both necessary and sufficient for establishing Kant's conclusions and hence no appeal to spatial knowledge needs to be made.
Thus Ameriks seems to be mistaken to say that the temporality of the self is dependent on spatiality. We appeal to spatial knowledge in determining the temporality of the self, but this is not equivalent to its temporality depending on spatiality and therewith does not imply that ideality of spatiality transfers to ideality of temporality. The ideality of time is established independently of the ideality of space in the Transcendental Aesthetic.