In 323 elegantly written pages, Manifest Reality manages both to present two major topics of the Critique of Pure Reason in a manner accessible to non-specialists and to make a major contribution to the scholarly literature on these topics. It succeeds at this delicate balancing act because it both presents a convincing interpretation of the text and attributes to Kant an intrinsically plausible and interesting philosophical view. Part of the appeal of the book, and one of its themes, is that very balance: Allais gives all of the aspects of Kant’s complex doctrines their due, not slighting one in favour of others (as is often done). Thus its subtitle: Kant’s Idealism and his Realism. Allais shows us how to read Kant as precisely what he claims to be: an idealist (who thinks that concrete spatio-temporal reality depends in some fashion upon minds like ours) who is also a realist (who thinks that there is more to reality than could ever be discovered by minds like ours).
This brings me to the focus of my remarks, Allais’s interpretation of Kant’s doctrine of transcendental idealism. Allais’s ambition in her interpretation of transcendental idealism and the ontological status of appearances is to navigate between two historically influential extremes: the Scylla of ‘phenomenalist’ readings and the Charybdis of ‘epistemological’ readings (the latter made famous in English-language scholarship by Graham Bird and Henry Allison). Allais’s interpretation, which I will flesh out below, is that the objects we experience in space and time have two kinds of properties. Essentially manifest properties are properties the possession of which is grounded in how a thing is experienced, or would be experienced, by a human subject. But the natures of these things are not exhausted by their essentially manifest properties: they also have an inner nature that, together with the inner nature of our minds, grounds (in some way we will never fully grasp) why they have these essentially manifest properties. The essentially manifest properties are all and only the properties we can know objects to have (aside from certain very general or structural properties, like the property of having some inner nature or other). Talk of things in themselves is not talk of some additional realm of objects ‘above and beyond’ the objects whose properties are manifest to us; it is talk of the inner nature of these objects, which can never manifest itself to us in experience, and which consequently we can never know.
While I am sympathetic to much of her positive account of Kant’s idealism, I think that Allais’s negative project is not as successful. In particular, I will argue (§1) that her attempts to distinguish her reading from the phenomenalist reading are not as successful as they initially appear, and (§§2 and 3) that she has failed to give a satisfying non-phenomenalist account of the actuality of appearances.
1. Identity and Non-Identity
One issue that has often been thought central to the interpretation of transcendental idealism (henceforth, TI) is whether ‘appearances’ and ‘things in themselves’ refer to one and the same set of objects under different descriptions, or to distinct sets of objects. As has been pointed out by other scholars, there are reasons to think that Kant cannot be introducing these technical terms so that every object that is an appearance is also a thing in itself (has a non-manifest inner nature). For this would mean that there are at least as many things in themselves as there are appearances, which would entail that we have some very substantive knowledge of things in themselves: a lower bound on their number. If there are seven spatio-temporal objects in the room, then there are at least seven things in themselves. If this does not violate Kant’s doctrine that we lack knowledge of the determinate properties of things in themselves, I do not know what does.
Allais is sensitive to these issues, writing at one point:
However, I suggest that it should rather be taken to show that the question of whether things in themselves and appearances are numerically distinct or numerically identical objects is not the right question: the question is well posed only where we have objects individuated in two different ways and we want to know whether we can identify them. This is what we do not have: we have no way of individuating things in themselves. We cannot pose a question about the numerical identity or distinctness of things in themselves and appearances without having some way of identifying the objects that are things in themselves. We do not have any way of doing this. (p. 72)Footnote 1
I take this to mean that she rejects the coherence of the question of whether appearances are numerically identical to things in themselves. However, this poses a problem for the ways in which she typically formulates Kant’s idealism:
[F]rom there [Kant] argues that we cannot know these things’ [things with intrinsic and relational properties] intrinsic natures and can know only their relational appearances. (pp. 72–3, emphasis added)
He asserts that there are entirely mind-independent things, that these things are given to us as objects of the senses, existing outside us, but that we know nothing about them as they are in themselves. (p. 128, emphasis added)
Throughout the Critique Kant talks about his distinction as being one between things as objects of experience and these same things as they are in themselves. (p. 130, emphasis added)
In these passages, and throughout the book, Allais writes as though the very same things that have manifest qualities (e.g. spatio-temporal properties) also have non-manifest inner natures. This entails that every appearance is numerically identical to a thing in itself (a thing that has non-manifest inner properties)! The issue of numerical identity is not, as Allais presents it, about establishing some relation of correspondence between two sets of objects we know how to individuate separately. It is instead the simple issue of whether every object that is an appearance is also a thing in itself. And given the way she has answered the identity question (in order to avoid the ‘counting’ problem) she is not entitled to say that the very same things that have manifest properties have inner non-manifest properties.
Why is this important? Here is one of the ways Allais distinguishes her view from phenomenalism: ‘there is no sense in which merely mental ideas, modifications of subjects’ states, can be said to be the aspects of the very same things that exist outside us. On the other hand, this is entirely straightforward for the essentially manifest view’ (p. 130). But as we have seen, Allais’s own considered view on the matter rejects the coherence of the claim that the bearers of essentially manifest properties are the very same things as (hence numerically identical to) the bearers of non-manifest inner natures. Why is it not open to the phenomenalist, likewise, to argue that it is incoherent to identify a construction of mental states (or whatever exactly the phenomenalist thinks appearances are) with an object that exists independently of human mentality (a thing in itself)? On this point, then, the difference between Allais’s view and phenomenalism is not as stark as it first appeared.
2. The Existence of Appearances and the Postulate of Actuality
The core idea of Allais’s interpretation of TI, to which the title of the book refers, is that the only substantive positive properties we know objects to have are ‘essentially manifest qualities’: qualities that are essentially such as to be manifestable to human subjects in experience. To possess such a property is to stand in a certain relation to possible or actual human subjects. Therefore, in knowing such properties of objects we do not thereby know how they are independently of us, that is, how they are in themselves.
But this is an account of the kinds of properties that can appear to us and the kinds of properties we can know objects to have. It is not an account of the properties objects actually have, for sometimes objects appear to us to have manifest properties they actually lack. What grounds the fact that appearances have the properties they actually possess? What fills in the blank in the following principle?
(1) If appearance a actually has (essentially manifest) property F, this is so in virtue of the fact that ________.
Allais might respond by saying that what fills in the blank in (1) is some fact about the an sich nature of objects and the an sich nature of our minds. The grounds that explain why, for instance, the world manifests itself to us as having the actual distribution of matter in space and time it does are in principle inaccessible to us. All we can know is the causal relations among the manifest properties of these objects.
But this is in tension with Allais’s characterization of them as essentially manifest properties. For the fact that objects possess the properties they actually do is not metaphysically grounded, according to this proposal, in anything about how they actually or potentially manifest to us. If there is no accessible ground of this fact, what on Allais’s view excludes the following possibility: even at the idealized limit of scientific inquiry we could be wrong about the essentially manifest properties of objects? Call this the sceptical scenario. If this remains a possibility on her view, I have lost track of the sense in which these properties are essentially manifest or even manifestable. Allais develops her account of essentially manifest qualities in part through an analogy with colour. But consider this possibility: normal human observers in appropriate viewing conditions are systematically in error about the colours of objects and no further science of vision we can achieve will ever discover this fact. That does not seem like a coherent possibility about colour, if colour is an essentially manifest quality. Consequently, I do not think the corresponding sceptical scenario about the actual properties of appearances is a coherent one. Allais needs to be able to fill in the blank in (1) with something that (i) is in principle accessible or at least informatively statable by us and (ii) explains why the sceptical scenario is impossible. In the rest of this section I will canvass some ways she might do this.
The natural place to turn to in the Critique is the second Postulate of Empirical Thinking in General, corresponding to the modal category actuality: ‘that which is connected with the material conditions of experience (sensation) is actual’ (A218/B266). One thing to note at the outset is that the Postulates are not merely epistemic criteria for possibility and necessity, but give the meaning of these concepts, at least when applied to phenomena. ‘The principles of modality’, Kant writes, ‘are nothing further than definitions of the concepts of possibility, actuality, and necessity in their empirical use’ (A219/B267, emphasis added). So we should not read the postulate of actuality as merely giving us epistemic conditions on how we acquire epistemic warrant for making judgements about which appearances there actually are. Instead, in line with the other postulates, we should read it as giving us the conditions under which it is the case that some appearance actually exists.
The postulate of actuality is mainly concerned with the existence of appearances (e.g. whether magnetic matter actually exists). Since existence is not a real predicate, we cannot read it directly as a way of filling in the blank in (1), but instead the blank in (2):
(2) If there actually is an appearance that is F (an empirically manifest property, e.g. being magnetic matter), this is so in virtue of the fact that _______.Footnote 2
We understand the ontological status of appearances only if we know in virtue of what (if anything) they exist and in virtue of what (if anything) they have the properties they actually have. In other words, understanding the ontological status of appearances requires knowing how to fill in the blanks in (1) and (2).
How then on Allais’s view should we fill in the blank in (2)? I am somewhat unsure what her answer is, so I am going to canvass a couple of answers she considers. In her discussion of this issue, she at one point (p. 143) appears tempted to give this answer:
(2a) If there actually is an appearance that is F, this is so in virtue of the fact that I intuit an appearance that is F.
This relies on Kant’s doctrine that we intuit all the objects there are in space and time, consciously or otherwise.Footnote 3 But that is not the answer he actually gives in the Postulates, where he does not appeal to his doctrine that we intuit every object in space and time.
Kant’s answer in the Postulates has two parts. First, he points out that what exists in space and time is not limited to what we perceive or what we can perceive, given our particular sense organs. Instead, what actually exists is limited by the nature of discursive spatio-temporal intellect in general. Any object that exists in time is such as to be possibly intuited by some spatio-temporal discursive intellect. But that is a necessary rather than sufficient condition on the actual existence of objects in space; it is after all possible for a discursive spatio-temporal intellect to intuit a unicorn, but there are actually no unicorns.
Instead, Kant claims that what is actual is what is ‘connected’ with perception according to laws. If we read this as an epistemic condition on our warrant for making judgements about what exists in space and time, it is a relatively anodyne claim: we are epistemically warranted in positing the existence of objects we cannot directly perceive (e.g. magnetic matter) because those objects, given the laws of nature, explain the existence of the objects we do perceive (e.g. the motion of iron filings in the presence of a magnet). But the postulate of actuality does not state a merely epistemic condition on judgements of existence; rather, it gives the definition of actuality when this category is restricted to objects of possible experience. So it must give necessary and sufficient conditions on what actually exists.
Some other things Allais says suggest she might want to fill in the right-hand side of (2) as follows:
(2b) There actually is an appearance that is F if and only if there is an F in space and time that is either (i) directly perceived or (ii) connected to what is directly perceived according to natural laws.
Now this raises the obvious question: which laws? Well, presumably not the laws that merely might obtain or possibly obtain, but those that actually do obtain. But what are the actual laws of nature? What grounds the fact that it is these laws that are actual, and not some others? Could there be objects in space and time connected to what we perceive, but connected by laws that we could never uncover, even at the ‘idealized limit’ of scientific inquiry (see above)? Claim (2b) is incomplete without an explanation of why this sceptical scenario is impossible. But there is an even deeper way in which this claim is incomplete: the right-hand side is of the same logical form as the left, ‘there is an appearance such that ...’ So nothing of the form of (2b) gives us an informative account of that in virtue of which there are appearances of a given sort.
Allais’s discussion of the postulate of actuality suggests she may be reading it in a third way. She may be reading Kant as claiming that what exists is what we could perceive if our senses were finer or different (while still being spatio-temporal). But is this what grounds facts about what exists or is it merely a necessary condition on what exists? Every object in space and time could be directly perceived by us, if our senses were finer; if this is what grounds such facts, then we need some account of this notion of what ‘could be’ perceived by us. Clearly it is not equivalent to mere compatibility with the form of discursive spatio-temporal intellect (formal possibility, for short), for reasons I gave above (unicorns are formally possible, but we cannot perceive them because they do not actually exist). Allais’s answer, I take it, is that the notion of ‘possibility’ here is what she calls ‘the actually possible’, or what we might gloss as: what is possible for us to perceive, given what is actually the case (so in this sense if it is possible that we perceive that p then it is actually the case that p). But this is to presuppose the very notion we are trying to define: which objects actually exist in space and time. What is more, it conflicts with the ‘definition’ of actuality Kant gives in the Second Postulate (quoted above). Allais, I think, has conflated a consequence of the postulate with the postulate itself. What exists in space and time is what is connected with our perceptions according to the actual laws of nature (remaining neutral on what makes them actual). Anything in space and time is such that a discursive spatio-temporal intellect with appropriate sense organs could directly perceive it. So for any object in space and time, even if we cannot directly perceive it, we could directly perceive it if our senses were finer or different (while still spatio-temporal). The possibility of the direct perception of an object in space and time is not constitutive of its actuality, but is only a consequence of its actuality. This is why Kant writes:
one can also cognize the existence of the thing prior to the perception of it . . . if only it is connected with some perception in accordance with the principles of their empirical connection (the analogies) . . . for in that case the existence of the thing is still connected with our perceptions in a possible experience, and with the guidance of the analogies we can get from our actual experience to the thing in a series of possible perceptions. (A225-6/B273, emphasis added)
It is because the thing is actual (causally connected to our actual perceptions) that we could perceive it if our senses were finer, not vice versa.
3. The Resolution of the Cosmological Dialectic
Aside from the Postulate of Actuality, probably the most important single section of the Critique for these issues (also aside from the subsequently deleted Fourth Paralogism in the A edition) is ‘Transcendental idealism as the key to solving the cosmological dialectic’ (A491–7/B519–25). Allais devotes a substantial amount of attention to it, arguing that it does not support a phenomenalist reading of TI. There are indeed passages in this section that can sound bad for the phenomenalist, for example:
That there could be inhabitants of the moon, even though no human being has ever perceived them, must of course be admitted; but this means only that in the possible progress of experience we could (könnten) encounter them . . . (A493/B521)
The reason this might be thought bad for phenomenalism is, as Allais puts it: ‘Kant’s analysis does not consist of a construction out of what is actually in minds, and in this respect its connection with what is in consciousness is less demanding [than the phenomenalist view]’ (p. 143). Kant appears to give the following account of the grounds of the truth of an existential judgment for appearances:
(2c) If there actually is an appearance that is F then this is so in virtue of the fact that (i) we directly perceive an F, or (ii) in the further course of experience we could experience an F.
Now the question is: what is the relevant notion of possibility here? For reasons we discussed earlier, it cannot be formal possibility (compatibility with the forms of experience). Allais suggests an answer: what is relevant here is actual possibility. But this is circular for the very reason discussed earlier: if it is actually the case that there is an F then this is so in virtue of the fact that, given that there is actually an F, it is possible for us to experience it. Considered abstractly the problem is this: formal possibility (see above) is too inclusive a notion (many formal possibilities are not actual), so how do we constrain formal possibility and give a non-circular account of what is actual in space and time? This is a question for Kant just as much as for Allais: what is the meaning of ‘could’ in the passage quoted earlier?
Elsewhere in the same section Kant suggests an analysis of the actual not in terms of what we could experience but in terms of the ‘transcendental object’:
Meanwhile we can call the merely intelligible cause of appearances in general the transcendental object, merely so that we have something corresponding to sensibility as a receptivity. To this transcendental object we can ascribe the whole extent and connection of our possible perceptions, and say that it is given in itself prior to our experiences. (A494/B522–3)
This suggests that what determines, within the space of formal possibility, which possibility is actualized is something about the ‘in itself’ nature of the things that are affecting us.Footnote 4 But what facts are these? One might be tempted to answer that we cannot say, but that will not suffice. For if we cannot say at all what the relevant noumenal facts are, we will not be able to explain why the ‘sceptical situation’ (see above) is impossible.
As far as I can tell, this exhausts the answers to these questions that Allais explicitly or implicitly considers in her book: (i) actuality is grounded in non-manifest noumenal properties, (ii) it is grounded in what we could experience or (iii) it is grounded in what we ‘actually could’ experience. I have argued that none of these is adequate. What is more, I have argued that she must answer these questions, for without answers, she cannot explain why the sceptical scenario is not possible. Without such answers she potentially leaves the door open to a phenomenalist reading of Kant’s idealism, which can answer (1) and (2) in terms of a construction of subjects’ actual and counterfactual experiences. Allais does argue at length against a phenomenalist reading of Kant’s idealism, and I have not discussed those objections here. But if a sophisticated phenomenalist could both counter those objections and provide a non-circular grounding of the actuality of appearances that explains why the sceptical scenario is impossible, it might carry the day against Allais’s moderate metaphysical reading.Footnote 5
