Allison distinguishes two ways the two-aspect view can be understood: (1) metaphysically, as a form of property dualism, where objects have phenomenal and noumenal properties, or (2) “methodologically, as a contrast between two ways in which such objects can be considered in a philosophical reflection on the conditions of their cognition.”
Allison thinks the best way to finding the correct interpretation of Kant starts with the thesis that opposes transcendental idealism—namely, transcendental realism. The correct interpretation of transcendental realism will lead us to the correct interpretation of transcendental idealism. Allison aims to show that transcendental realism is not a metaphysical thesis; and from that it will follow that transcendental idealism is not a metaphysical thesis.
In part 1, Allison proposes that Kant does not commit to a specific ontology when he defines transcendental realism. According to Kant, transcendental realism is the view that “space and time are given in themselves independently of the conditions of sensibility”. Transcendental realists “interpret outer appearances … as things in themselves, which would exist independently of us and our sensibility.” By contrast, transcendental idealism is “the doctrine that all appearances are to be regarded as mere representations and not things in themselves, and accordingly that time and space are only sensible forms of our intuition, but not determinations given for themselves or conditions of objects as things in themselves” (A369). Transcendental realism also needs to be contrasted with empirical realism which limits spatiotemporal predicates to objects of possible experience. (Allison says his interpretation of transcendental idealism has affinities to Hilary Putnam’s internal realism.)
Allison points out that in the Antimonies Kant provides an indirect argument for transcendental idealism based on the negation of transcendental realism. For this argument to work, it has to be that the two are contradictories rather than contraries. If this is right, transcendental realism can’t be equated with the scientific realism of the Descartes and Newton, since the indirect argument in the Antimonies would equally be an argument for Berkeleyan idealism, which Kant opposed. In order for his indirect argument to work, it has to rule out Berkeleyan idealism as well.
Two distinct definitions of transcendental can found in Kant’s use of “transcendental cognition”: one has to do with the necessary conditions of possible experience; the second has to do with things in general, or general metaphysics. The first use of “transcendental” can be clarified by distinguishing between transcendental and a priori cognition. A priori cognitions are those of mathematics. Transcendental cognitions pertain to the possibility of these a priori cognitions. The second use of “transcendental” has to do with the use of spatiotemporal predicates about thing in general, as apart from only using spatiotemporal predicates about objects of experience. In other words, the first use of transcendental is a matter of level, and the second use is a matter of scope.
Allison wants to take this second use of “transcendental” and apply it to the distinction between transcendental realism and empirical realism. In effect, the difference between the two is a matter of scope rather than level: transcendental realists apply spatiotemporal predicates to things in general, while empirical realists apply it only to objects of possible experience.
Allison argues that transcendental realism is not a metaphysical thesis since it doesn’t commit itself to a particular ontology of space and time. For further support, Allison cites Kant’s question: What are space and time? (A23,B37-38) Here four views are distinguished: space and time are substances, Newton’s view, Leibniz’s view, and Kant’s view. Allison takes the first three views to exhaust the recognized ontological alternatives. Therefore, on his view, Kant is not as offering up a distinct ontology, but, rather, regardless of the ontology, space and time are the conditions of our cognition of things. In other words, it’s an epistemological position rather than a metaphysical one.
In part 2, critics to Allison’s two-aspect view point out that Kant accuses transcendental realists with “conflating appearances with things in themselves rather than with inflating claims about objects of possible experience into claims about things in general.” In other words, Allison’s critics are saying that Kant takes the transcendental realists to be mistaken about things in themselves rather than things in general. But Allison says if predicates were applicable to things in general, then ipso facto they would be applicable to things in themselves. (He provides some supporting evidence which I’ll skip.)
In part 3, Allison addresses two distinct objections to his view. The first objection is that on his two-aspect view, the non-spatiotemporality of things in themselves reduces Kant’s thesis to a trivial definition. In other words, the objector says: Of course, we can’t know things in themselves, because, by definition, things in themselves represent the knowing of things apart from the conditions of our knowledge. Allison says his view is not trivial, because there is a substantive claim in that space and time are the forms of our sensibility for which we represent the world.
The second (distinct) objection is that just because our sensibility determines that things appear us spatiotemporally, it doesn’t mean the things in themselves are not spatiotemporal. Allison agrees that just because one can consider x without considering its y-ness does not entail that x doesn’t have y-ness. But he thinks this misses the point.
The point is rather that in considering things in this manner one is, ex hypothesi, considering them apart from the condition under which alone such predicates are applicable to them, namely, in their relation to human sensibility. Accordingly, it is not simply the case that spatiotemporal predicates are ignored or set aside when things are considered in this manner, but that they are denied of the object qua considered apart from that relation.
In other words, in the way we can’t consider the weight of Newtonian bodies apart from relations of attraction between bodies, we also can’t consider things as having spatiotemporal properties apart from the relation of the form of our sensibility.
In part 4, Allison considers transcendental realism within the context of the Antimonies. In the Antimony of Pure Reason Kant claims “If we would give in to the deception of transcendental realism, then neither nature nor freedom would be left” (A543/B571). The assumption of transcendental realism in conjunction with certain considerations lead to the loss of nature as explained in the first and second antinomies—the mathematical antimonies. The first antimony deals with the finitude or infinitude or space and time, and the second antimony deals with whether the world consists of divisible or indivisible parts. In both antinomies we arrive at two equally warranted but contradictory conclusions. Kant’s solution is to say that both transcendental realist solutions are false and that the question is malformed under transcendental idealism, thus not in need of a solution. This provides an indirect argument for transcendental idealism.
The transcendental realist problem in the first antimony stems from a seemingly intuitive principle: “If the conditioned is given, then the whole series of all conditions for it is also given“(A497/B525). (If we apply this principle to the first antinomy, we get: if parts of the sensible world are given, then the world as a whole is given.) The problem, according to transcendental idealists, is that the world is a pseudo-empirical object in that it claims to be a spatiotemporal object, yet it is not given. Categories only apply to what is given. And, to bring this back to a point in part 2 of the paper, this “hypostatization” is the transcendental realist mistake of inflating claims about of objects of possible experience into claims about things in general. This is a result from transcendental realist ignoreing the contributions of the sensibility.
The next question is: What must transcendental idealism be like, if it is to come from the rejection of transcendental realism? Allison notes that since the rejection of transcendental realism was independent of the metaphysical views, transcendental idealism must also be independent of metaphysical views.
This further suggests that the force of transcendental idealism is deflationary and its function, particularly in the Dialectic, largely therapeutic. As I have argued above, it provides the requisite means for resolving the contradiction of reason with itself, which, if unresolved, would spell ‘the euthanasia of pure reason ‘ (A40 7/B434).
Allison explains how this therapeutic function also applies to the case of transcendental freedom. On the two-object view, the phenomenal self is causally determined, while the noumenal self is transcendentally free. On the two-aspect view, there is a single self that is causally determined qua phenomenon and free qua noumenon. A problem for the two-object view is justifying why we should punish the phenomenal self if it is causally determined. A similar problem holds for the two-aspect view: if the phenomenal self is causally determined, then second-aspect view of a noumenal self is of no assistance to the freedom of the phenomenal self.
Allison thinks he can solve this dilemma in a similar way the first antimony was solved.
My claim, then, is that not only is transcendental idealism not committed to the assumption that there must be some noumenal fact of the matter regarding freedom, but it functions therapeutically to disabuse us of any such assumption. … Basically, what Kant finds necessary is a warrant to assert our freedom from ‘the practical point of view ‘, which is quite distinct from a justification of the metaphysical thesis that we are ‘noumenally free ‘. Accordingly, the problem is to explain how we can be warranted to assert something from the practical point of view that we are explicitly prohibited from asserting theoretically. And, assuming that this requires preserving both something like the Kantian conception of freedom and the normativity of the principles of theoretical reason, the only conceivable way in which this could be accomplished is by limiting the scope of the latter. In other words, it is done by deflating a transcendental to a merely empirical realism, which, once again, is just what transcendental idealism does. With this deflation in place, it becomes possible to view both the theoretical and the practical points of view as having their own set of norms, while avoiding the assumption that there must be some context-independent truth or fact of the matter. Otherwise expressed, Kantian dualism is normative rather than ontological.
I think this solution is a little better than the previous two-aspect solution, but it still maintains the worry that the transcendental idealist is ignoring the theoretical side. To be fair, I don’t think it works on the two-world view either.
Allison, Henry. Transcendental Realism, Empirical Realism, and Transcendental Idealism.