Things in Themselves and Metaphysical Grounding:
On Allais’ Manifest Reality[1]

James Kreines - jkreines@cmc.edu

Allais’ Manifest Reality is wonderfully ambitious, and in my view there isn’t any comparably comprehensive interpretation of Kant’s transcendental idealism that is more successful. Debate about Kant will be much the better for it.

I think of the book as aiming to resolve two distinct but linked problems. A first problem focuses on Kant’s notion of appearances, as distinct from things in themselves. Kant, of course, claims that appearances are transcendentally ideal and yet empirically real. The problem is to make sense at once of texts that seem to push toward a mind-dependence of appearances, and toward a mind-independence of appearances. Second, there is a problem more focused on things in themselves: Kant argues that we are unavoidably ignorant of things in themselves; it can be hard to see why this should not block Kant’s own seeming assertion that there must be things in themselves. Some would defend extreme solutions, giving up on the face value of the texts on one side or the other. For example, some would deny that appearances are mind-dependent in any sense, or that Kant is an idealist in this sense.[2] In all cases, Allais seeks what she calls a “moderate” interpretation, showing that she can navigate between the extremes, doing justice to the face value of texts that initially seem to conflict. Pulling this off with respect to either problem is difficult; coming up with a joint solution of both seems to me exponentially more so. And that is the great ambition of Manifest Reality.

With respect to the first problem, the short story of Allais’ reading is this: where Kant says that we can cognize only appearances, he is referring to essentially manifest features of objects.[3] This limits us to what are, in a sense, “mind-dependent” properties; but it “allows that the direct objects of perception are objects in space, not constructions out of merely mental states” (2015, 135). I find this approach extremely powerful.

With respect to the other problem, more focused on things in themselves, matters are different. Here Allais sees, as one extreme, deflationary readings that would exclude two metaphysical commitments:

1.      There are things in themselves.

2.     Things in themselves stand to appearances in a relation of metaphysical grounding.

I will note a sense in which the idea of Kant as concerned with metaphysical grounding might lead us astray, but I do not think that it need mislead; I agree with Allais that, other things equal, an ideal interpretation of transcendental idealism will be better if it can find a way to include some version of these claims.

The other extreme, in Allais’ view, would be a “noumenalist” reading, according to which Kant would be more specifically “committed to the existence of non-sensible, non-spatio-temporal entities which are distinct from the objects of our knowledge… intelligibilia” (2015, 10). I will defend a broader understanding of the extreme to be avoided, but I agree at least that noumenalism is among the views that the Critique does not endorse. 

How does Allais navigate between those extremes? With an interpretation that is metaphysical and also more “two-aspect” rather than noumanalism’s “two worlds” or “two objects”. Different versions of this approach will look to texts distinguishing the “outer” from the “inner” or “in itself” of things, and I will here reserve the terms “inner” and “outer” as neutral between such interpretations. One specific version would be Langton’s, on which this distinction is between “intrinsic” and “extrinsic” properties, where “intrinsic” properties are those compatible with loneliness. But Allais gives a different account, and one that is remarkable because (in part) of the way it fits her powerful and balanced approach to the first problem, above. The idea is this: Kant’s limitation of our cognition to appearances refers, again, to the essentially manifest features of things. These are relational properties: those whose essence involves relations with others.[4] Allais includes dispositions and powers as helpful cases of relational properties in this sense; and in this vein she calls the contrasting inner properties “categorical” as well as “non-relational”.[5] Further, Allais’ Kant thinks that “there must be something non-relational which grounds relational appearances” (2015, 18). This construal of the assertion that there are things in themselves is supposed to be consistent with our ignorance: we know that there are further properties, but not what they are. There is an “unknown ground”; “Kant has a metaphysical commitment to an unknown reality which grounds the world we experience”.[6] Grounding, here, is an explanatory relation: the categorical properties of something are responsible for its having the powers and dispositions that it has (2015, 254).

But while there may be much right in this general orientation, I think there is a problem, and that the details cannot really get Kant right. Noumenalist readings are too extreme, specifically because “Kant clearly argues that … we do not have knowledge such objects exist” (Allais 2015, 61). I agree, and find this central to Kant’s very idea of a critique of reason. The problem is that what the Critique says by way of a critique of reason—the topic of the “Transcendental Dialectic”—does not specifically target “noumenalist” metaphysics. It targets all metaphysics involving claims supported by a supposed need for “the unconditioned”. And Allais’ metaphysics of categorical grounds fits precisely the general kind of view targeted by Kant’s critique. Perhaps the many other advantages of Allais’ approach could be retained in combination with a new solution. But the metaphysics described by Allais cannot, I think, be precisely the view of the Critique. Or so I argue.

1. The Critique of Reason and Allais’ Things in Themselves

I begin by noting some broadly familiar features of Kant’s account of the faculty of reason, in the Transcendental Dialectic, borrowing from Allais’ own interpretation. To begin with, the faculty of reason is characterized by an interest, or an aim: “For everything caused, contingent, or not entirely self-explanatory, reason seeks for an explanation, ground, or cause”. Ultimately, reason seeks a ground that is complete, or “something unconditioned, something with respect to which no further explanation or ground could be asked for”. Kant strikes a balance: On the one hand, reason’s interest “drives science”, and “without this we would not have a coherent use of reason”. On the other hand, this makes us subject to an illusion:

…we mistake the (unavoidable) principle which tells us to seek for the conditions for … the claim that there is a condition...

...natural and unavoidable dissatisfaction with explanatory incompleteness misleads us into thinking that we have insight into the existence of the unconditioned…

The illusion in the sense of a temptation toward certain conclusions stems directly from reason’s interest and is supposed to be natural and ineliminable; the Critique seeks to teach us—even while still guided by reason’s demand for the unconditioned—to “avoid making the erroneous conclusions” (Allais 2015, 90).

Now consider the forms that these erroneous conclusions can take. There are supposed to be three kinds of “ideas of reason”—psychological, cosmological, and the theological idea of God. I focus especially on the Antinomy, concerning the cosmological case, because of its unusual form of argument. Here reason supports conflicting conclusions, by means of arguments with one and the same form:

The entire antinomy of pure reason rests on this dialectical argument: If the conditioned is given, then the whole series of all conditions for it is also given; now objects of the senses are given as conditioned; consequently, etc.

The major premise is a version of the deceived principle of reason, demanding conditions and completeness; thus the “major premise seems so natural and evident” (A497/B525).

To take one example, the Second Antinomy concerns the regress from wholes to parts. We can think of this in terms of the image of the earth resting on a stack of turtles, with turtles all the way down. The Antithesis argues that the regress to parts must be turtles all the way down, or specifically composition all the way down: “[n]o composite thing in the world consists of simple parts” (A435/B463). The Thesis argues that this regress must have a last turtle: “[e]very composite substance in the world consists of simple parts” (A434/B462). In the case of the first two, or “mathematical” antinomies, Kant will try to argue that such arguments bring us to contradictions. Resolving the contradictions is supposed to require finding a sense in which both sides are false, requiring in turn transcendental idealism (A507/B535).

The last two, or “dynamical”, antinomies allow consideration of an additional scenario, on which there is a ground “different in kind … outside the series” (A530/B558). For example, in the fourth antinomy considers the possibility of a necessary being grounding from outside the regress of contingency (whether it is infinite or not). This ground would be different in kind, in a respect that does not allow the same question about conditions to arise again: a necessary being does not raise the question of what it is contingent on. Instead of last turtle, this is more like a turtle with a jetpack. Such conceptions are of course a familiar in rationalist metaphysics. And this should be no surprise: metaphysical rationalism is a paradigmatic case of the forms of metaphysics which Kant takes to make a critique of reason necessary. I have argued that the deceived way of using the principle of reason, described by Kant, is simply arguing from the unrestricted principle of sufficient reason, so central to rationalist metaphysics; the idea is that, “for anything that is not a sufficient reason for itself, or for anything conditioned, there must be a complete series of conditions that provides for it a sufficient reason” (Kreines 2008, 49). Kant allows theoretical reasoning from something like this principle only in the case restricted to time (e.g. A201/B246).[7]

In the dynamical antinomies, Kant is not ruling out either side (A532/B560). Still, the point—applied to the necessary being—is that, if we could have legitimate theoretical justification for its existence, then the “the same ground of proof” would “prove its nonexistence, and indeed with equal rigor” (A459/B487). So for a theoretical philosophy to assert one side of such an antinomy would be unacceptable, or one form of “dogmatic stubbornness, setting its mind rigidly to certain assertions without giving a fair hearing to the grounds for the opposite” (A407/B434).[8]

With this account of Kant’s critique of reason in hand, consider again Allais’ account of things in themselves. First, I think it is clear that the issues Allais highlights, about relational properties like dispositions, concern a case of what Kant calls a regress of conditions. For example, salt has the disposition to dissolve in water. Is this itself brute, or a last turtle? No. It depends on the properties and arrangement of the underlying electrons and protons. And so we have our characteristic question about which pattern the regress fits. Can there be a last turtle, or a brute disposition, such as a brute dormitive virtue? If we make the deceived use of reason’s demand, or “the claim that there is a condition” (Allais 2015, 90), then we will conclude that there cannot be brute dispositions. Could the regress be only dispositional or relational all the way down? Allais’ crucial claim answers directly: “there cannot be relations all the way down; there must ultimately be intrinsic natures which ground powers” (2015, 255). The “ultimately” asserts that “all the way down” scenarios lack completeness of grounding. I think that similar formulations can be found in general where philosophers see reason for demanding categorical grounds.[9] Compare the Thesis of the second Antinomy: if there were only composition, all the way down, we could consider all of that composition together, and then there would be nothing left to enter into relations of composition; so there must ultimately be non-composed grounds of composition.

Granted, the position of the Critique is that the nature of reason itself makes this kind of reasoning attractive and commonsensical, and even nearly irresistible. But this is bad news, at least when it comes to Kant interpretation. For what explains the appeal of the conclusion Allais sees in Kant would be the deceived use of reason’s demand for completeness of conditions—the use against which Kant is warning us. Or it would be another instance of the argument schema driving the antinomies: there must be completeness in a regress for something conditioned; relational properties are conditioned; and so on. Further, this seems to me like this would be clear to Kant, given that he is so taken with the idea of a critique of reason that he uses this idea to frame the entire book. So it seems to me that this is a powerful reason against reading the Critique as advancing the account of things in themselves that Allais sees there.

Another way to make the same point is this: while we can indeed give an intuitively compelling thesis-style argument for the existence of categorical grounds, the same principle could offer support for an antithesis-style denial of existence. Compare the antithesis of the fourth antinomy: On the one hand, a necessary being as ground cannot be either a member of a regress of contingent alterations, or the sum of all members of this; in either case, it would be merely contingent.[10] On the other hand, a necessary being cannot ground a series of contingent alterations from outside of the series and time; for then it would be outside of time, and the final ground of the alteration would leave unexplained why it happens at a particular time[11]; deceived use of reason’s demand for completeness of grounds rules this out. So with respect to the idea of a necessary being grounding contingent alterations, this would either not be necessary, or it would not be the ground.

Parallel reasoning applies to any proposed categorical ground (CG) for a disposition (D): On the one hand, it cannot be in virtue of the very nature or essence of CG that an object having the property CG must also have the property D; for then CG would be relational, not categorical. On the other hand, if the nature of CG is such that it does not require D, in that way, then CG would not completely explain why the object possesses D. Anyone willing to reason from a demand for completeness would then have to deny that CG could really be the ground of D. So either the categorical ground is not categorical, or it is not the ground.[12]

Allais reads Kant as holding that there must be categorical grounds, and then adding that our epistemic limits leave us unable to “understand how they ground appearances, or even what kind of relation this grounding might be” (2015, 231). I think that this kind of position is common for rationalists, like Leibniz: there must be sufficient reasons “although these reasons usually cannot be known by us” (Monadology #32). I certainly grant that one can formulate an internally consistent position here; my worries concern possible reasons for it. First, as a philosophical matter, Kant’s worries about Antinomy style reasoning seem to me powerful when applied to categorical grounds: theoretical philosophy can give a case in favor of existence, but it will rest, whether explicitly or not, on a demand for completeness; the same demand can equally well support a case for “nonexistence, and indeed with equal rigor” (A459/B487). So it seems that it would indeed be an unacceptable form of dogmatic stubbornness for theoretical philosophy to side with the thesis in asserting existence and then saying that the antithesis fails to prove non-existence but only limits our understanding. Another way to put the point: precisely if our understanding of the relevant grounding relation is so tenuous, we cannot know that an infinite descent of powers or relational grounds is insufficient to offer completeness of such grounding. Second, as an interpretive manner, whatever we think of them, these worries about the kind of reasoning Allais uses are certainly Kant’s worries.

2. Objections and Replies

Objection: The point of the Critique is not entirely negative, nor entirely about epistemic limits. It includes a metaphysical assertion that there are things in themselves. In fact, it includes the metaphysical assertion that there must be “inner” grounds of the “outer” properties of objects known to us. And even the Antinomy itself is supposed to justify this metaphysics. Further, Kant’s complex epistemic claims about the limits of different epistemic states—Erkenntnis, Wissen, etc.—can be read in accordance with those assertions.[13]

Reply: I deny none of this. My point is that there is a powerful reason to deny that “things in themselves”, in such assertions, could refer to unconditioned grounds,[14] and so specifically to non-relational or categorical grounds.[15] And I did not reason from any claims about epistemic limits; the point of the Dialectic is to show that our direct pursuit of metaphysics conflicts with itself;[16] whatever the complex details about different epistemic states, they are meant in part to serve the aim of resolving the Dialectic problem about the unconditioned.  

Objection: The alternative to categorical grounds would be what Allais calls an “entirely relational ontology”[17], which Allais shows is not an “intuitive” view (34). The argument above succeeds at justifying neither that rival ontology, nor the attribution of it to Kant.

Reply: This objection to my case would misconstrue the options. On my account, the Dialectic shows that theoretical philosophy should withhold assertion of either existence or non-existence—of unconditioned grounds in general, and so categorical grounds in particular.

Objection: Allais argues that Kant “simply assume”[s] non-relational natures as an “intuitive starting point”, so her Kant cannot be making the mistake of arguing on grounds of anything like a deceived use of reason, because he is not arguing at all.[18]

Reply: First, it is crucial to Kant’s idea of a critique of reason, from the first words of the A-preface, that it will target views that are intuitive, for example in the sense of being “common sense” (Aviii). Second, it will target presuppositions. For example, there is a “deception” that is “not artificial, but an entirely natural mistake of common reason… we presuppose … the conditions and their series as it were sight unseen” (A500/B528). Here is another way to think about this: Imagine you assumed that there are non-relational natures of things, but then realized that we would be unavoidably ignorant of them. You might still hold on to that existence claim, or you might give it up. But either way, one can ask why?[19] Allais’s Kant holds on to existence, and has a reason why: “we cannot have a complete ontology with relations only” (226). On my view, Kant aims to convince theoretical philosophy to give up views resting on that intuitive assumption about completeness.

Objection: Kant himself does not apply his critique of reason specifically to the topic of inner, non-relational properties; thus it is no reason to doubt Allais’ reading.

Reply: I do not accept the premise. But since interpretation of texts bearing on the matter, like the Amphiboly discussions of inner properties noted below, would be controversial, it is important to see that the conclusion would not follow from the premise anyway. The Critique is a book that argues, and arguments bring their own level of generality. Kant argues that it is the faculty of reason that can lead us astray, and into metaphysical error—and that it can do so because of principles Kant himself formulates very generally in terms of the notions of conditions and the unconditioned. I agree with Allais interpretation of these notions, cited above, in terms of connections to the notion of explanation and complete explanation. What follows from Kant’s argument, then (if successful at all) is that theoretical philosophy should in general resist assertions whose appeal rests on such a deceived principle demanding unconditioned grounds. Given the argument as written, and highlighted by the title of the book, the relevant question is whether Allais’ assertion (that there are categorical grounds) would be one whose appeal rests on that kind of principle of reason demanding completeness in a regress of conditions. I have argued that it is.

Objection: Kant’s Dialectic criticizes only theories positing distinct, unknowable objects; the posit of unknowable properties is exempt because it is “not making new, unjustified existential commitments” (Allais 2010, 15).

Reply: Allais thinks that deflationary readings cannot adequately capture the face value of the texts where Kant asserts that there are things in themselves; she does better by construing these texts in terms of the claim that there are additional non-relational properties of things, over and above the essentially manifest properties; so I think she needs this to be precisely a “new … existential commitment”. I cede that it is not “unjustified”, but its justification takes precisely the general form targeted by Kant’s critique of reason. Again, as above, Kant’s own argument concerns general principles about the unconditioned; there is no limitation to cases of objects in any sense opposed to properties.

Objection: Allais argues that the “Amphiboly” of the Critique asserts the existence of non-relational properties.

Reply: There is an approach to the Amphiboly that does at least as well the text of that section—I would think better—and in any case much better with respect to coherence with the Dialectic. First, Allais holds that “intelligibilia” are objects “about whose existence Kant is agnostic” (2015, 234). In the Amphiboly, Kant reconstructs Leibnizian arguments for monads. Allais thinks that Kant accepts these arguments up to the point of positing that there must be non-relational properties of things, but then rejects the jump to there being “monads (things with only inner natures, things that are not spatial)” (235). But the text of Kant’s reconstructions generally begin by asking us to consider the notion of “an object of the pure understanding”; or thinking things “merely in relation to the understanding” or “[a]ccording to mere concepts”.[20] It is at least as natural, then, to take all of the steps following to follow only specifically or objects of the pure understanding, or objects given independently of all sensible intuition—or intelligibilia. The Critique’s own view, then, would be—now in accordance with precisely what is required by the Dialectic—that intelligibilia would need categorical grounds, but that theoretical philosophy should neither assert nor deny existence of such.[21]  

Objection: Some would argue that the resolution of the Antinomies asserts that the principle demanding the unconditioned is deceived only when applied to appearances, but does apply to things in themselves, thus guaranteeing unconditioned grounds there.[22] Kant says, for example:

If the conditioned as well as its conditions are things in themselves, then when the first is given … the latter is thereby already given along with it. (A498/B526; Allais 92)

Reply: On that proposed reading, the Critique would clearly have to endorse a thesis argument to an unconditioned necessary being: appearances require grounding in things in themselves; these things in themselves could not all be contingent beings, because then reason’s demand, applied to things in themselves as this reading requires, would allow us to bundle all contingent things in themselves and require a further ground for them, and so a necessary being. But, as above, there is a reading that does better with respect to the stated aims of the antinomy, withholding assertion from that thesis argument, while doing better with the text of that passage. The passage above considers what would follow if a thing in itself “is given”. The givenness in question would have to be independent of all sensible intuition, so these would be intelligibilia. Thus the passage says that the principle in question would require unconditioned grounds for intelligibilia—about which the Critique neither asserts nor denies existence. And that is precisely the view that would be consistent with the stated aims of the Antinomy: the aim of showing that “it lies beyond our reason to settle” (A481/B509) such questions about whether there exist unconditioned grounds, or that it would be “dogmatic stubbornness” for theoretical philosophy to side with thesis over antithesis here, and so on.

3. A Cost-Benefit Rejoinder and Kant’s Indirect Proof

One might worry that my criticism cuts off all hope of claiming the other considerable advantages of Allais’ approach, including those concerning appearances; thus (one might argue) it would be worth paying any costs stemming from the conflict I note with the Dialectic. To defend my criticism, then, I need to show that there is at least hope of taking the criticism to heart while retaining the other advantages of Allais’ approach. 

Note first that there can seem to be two different notions of things in themselves at work in Allais:

* TII-broad:    “there actually is a way things are in themselves … there exists an aspect of reality that is independent of us.” (2015, 27)

* TII-narrow: “things have a way that they are independent of their relations to other things, including us” (2015, 34)

Allais’ precise position is as follows. First, “Kant primarily uses” the broad notion. But, second, “Kant’s reasons for thinking that there must be a way things are as they are in themselves” latch onto the narrow sense: “things must have intrinsic, non-relational natures”; this allows that there might also things that fall in the broad space but not the narrow, or “relations between them as they are in themselves” (2015, 233).

Here is yet another difficulty for Allais here: she seeks to distance herself from readings which can seem out of keeping with the face value of many texts, insofar as ignorance of things in themselves would follow merely from the fact that our knowledge is receptive, bypassing with what Ameriks calls a “short argument” all the specifics concerning space and time being the pure forms of our sensibility. Allais affirms Ameriks’ point (2015, 291). I agree. But on Allais’ just-noted account, Kant’s “reasons” for asserting existence are reasons for asserting existence specifically of properties that involve no relations to anything; if a short argument from receptivity alone demonstrates ignorance, then it would demonstrate ignorance of things in themselves in precisely this sense.

The fix to all the above would be a reading putting more emphasis on Allais’ own broad and so less demanding sense. Here is the idea: We can know how some powers are grounded by comparatively more inner properties,[23] in the form of other powers. We today might think in terms of the idea that what is responsible for the power of water to dissolve salt is the power of the underlying electrons and protons to attract and repel. The Kantian position—here following Allais’ powerful account of appearances—would be that all these knowable powers are properties that essentially involve relations to epistemic subjects of our kind, dependent on sensible intuition, with the a priori forms of space and time. But the metaphysical assertion in the Critique, that there are things in themselves, would then mean this: the regress of powers extends to properties that ground the knowable properties, more inner properties of things, unknown to us. As Kant says, appearances “must have grounds that are not appearances” (A537/B565). What we would know about these is that they involve no essential relations to epistemic subjects of our kind—although they may well (for all we can know) be further powers, and so relational in other respects. For this reason, we cannot know what these further properties are. What the Critique would not assert, then, would be any commitment about how or whether that regress of powers might end, or any commitment between these options:

1.      Perhaps the regress of powers ends with an unknowable brute power.

2.     Perhaps there is an infinite regress of powers all the way down, or (better) all the way in.

3.     Perhaps the regress comes to categorical grounds, which in some manner beyond our comprehension more completely ground everything resting on them.

Thus the Critique would make the metaphysical assertion about grounds. But this would include no assertion or denial of the existence of unconditioned grounds, and so no conflict with the Dialectic. And, further, there would be no assertion of existence of anything for which our ignorance would follow from a short argument concerning merely receptivity, if at all.

But I don’t think that Allais could simply amend her view, in this way; what would be needed is a wholesale replacement for Allais’ account of Kant’s reasons for the assertion that there are things in themselves, which (as is) latch onto the narrow notion. And the replacement reasons will have to be radically different. To see why, consider this attempt at a slight alteration, arguing from this seemingly modest principle: where something stands in a relation to X, this must also have as a ground a property that is not a relation to X. Given appearances as essentially involving relations to us, it would follow immediately that there must be properties that are essentially independent of us (even if not independent of all relations)—just the modest claim we seek. But that principle would also push us back into antinomy arguments, by means of the bundling step central in a cosmological argument, or the thesis of the second antinomy: just take all the relational properties of an object, and make out of them a conjunctive super-property, P; since we have bundled all the relational features together, it would now still follow from the modest principle that objects must specifically have properties that do not essentially involve relations to anything: categorical grounds—just the dogmatic assertion we seek to avoid. As far as I can see, the problem will equally affect other attempts, whether two-aspect or two-world, to find reasons that extend from the known, by intuitive principles requiring grounds, to the existence of things in themselves.[24]

I think that Kant provides the needed and radically different form of reasoning in the Antinomy. Here the argument does not begin with the known and attempt to expand by intuitive principles about grounds. It is more the reverse, or more that the argument experimentally makes very extensive knowledge claims—insofar as it considers the possibility that appearances are things in themselves, it considers everything as continuous with the knowable. It then demonstrates that contradictions follow. For we must always at least think the unconditioned, insofar as we must, as rational, always be guided by regulative principles demanding that we seek the unconditioned; perhaps, as Kant often suggests, such guidance requires us to presuppose for the sake of inquiry that there are unconditioned grounds.[25] And yet the thought or presupposition would generate the contradictions of the first two antinomies. Resolving the contradictions is supposed to require transcendental idealism, which the Antinomy is thus “proving indirectly” (A506/B534). And transcendental idealism, on the face of it, means rolling back the extent of our knowledge in a very distinctive manner, precluding knowledge of things in themselves, but leaving knowledge that there are things in themselves grounding appearances. Kant sketches the relevant argument in glossing the overall case of the Critique in the B-Preface:

…if we find that on the assumption that our cognition from experience conforms to the objects as things in themselves, the unconditioned cannot be thought at all without contradiction … if we assume that our representation of things as they are given to us does not conform to these things as they are in themselves… then the contradiction disappears…  (Bxx-xxi)

On this reading, the point of the metaphysical assertion that there are things in themselves does have to do with ideas of unconditioned grounds; but that assertion does not assert the existence of unconditioned grounds; the point is to save room for thoughts or presuppositions about, or belief in, some forms of them.[26]

Needless to say, much more would need to be done to develop such an interpretive approach. My point here has not been to completely develop or defend this, but to deal with the objection that my criticism would leave no hope for claiming the otherwise considerable advantages of Allais’ interpretive approach.

Finally, I should concede that I have here been thinking about the Critique of Pure Reason in a manner that may be unfamiliar. Perhaps some will think that my evidence here, concerning the idea of a natural illusion, must carry much less weight than consideration of other aims. For instance, some would introduce the Critique in terms of a familiar anti-skeptical aim. Perhaps some today would introduce it as aiming to construct a metaphysics of grounding. I do not deny that something like these aims is present in the Critique. But I emphasize that there is also an organizing frame that Kant himself uses to integrate together the book’s many aims. That frame concerns the critique of reason announced by the title, and this should impact our interpretations.[27] So the idea that Kant aims to construct a metaphysics of grounding could lead us astray. It will mislead if we think, in terms that come so naturally to us, of an ambition to construct a complete metaphysics of grounding—and if we then set about supplying Kant with the resources we would need to that end. This is yet another juncture at which I find Allais’ counsel of moderation so important, even if I would follow it in a slightly different direction. For while Kant does argue that reason itself is always concerned with metaphysical grounds, and ultimately with complete, fundamental grounds, we should understand Kant’s ambitions concerning this topic first of all in terms of the idea of a critique—or the idea that the most natural ways of pursuing a complete metaphysics of grounding will always generate antinomy contradictions or the other problems discussed in Kant’s Dialectic. Kant’s antidote might well include some more modest claims about metaphysical grounding, carefully tailored to resolve the contradictions, blocking rather than satisfying our natural ambitions to completeness. This critical aim of using metaphysics against itself, or to limit itself, is no doubt today less familiar. But, first, this should make it all the more interesting: it provides us the occasion to consider Kant’s arguments in favor of pursuing philosophy differently. And, second, regardless of our orthodoxies, thinking first of all in in terms of an antinomy and an illusion itself is essential to Kant’s organizing frame for the Critique of Pure Reason.

I conclude, then, as follows: There is no comparatively comprehensive interpretation of transcendental idealism that is, in my view, more successful than Allais’ ambitious Manifest Reality. With respect to Kant on things in themselves, however, I think Allais’ account retains one disadvantage common to its competitors. Escape, as I see it, requires different understanding of the extremes that a “moderate” interpretation should avoid: On the one extreme, there are indeed deflationary readings, precluding a metaphysical assertion of the existence of things in themselves as grounds of appearances. But the other extreme is not precisely noumenalism, as Allais claims, but any metaphysics asserting the existence of unconditioned grounds, or metaphysical fundamentality in this sense; the argument of the Dialectic rules these all to be unacceptably dogmatic, and for what seem to me powerfully philosophical reasons. I would think that the big interpretive challenge is to navigate between these extremes concerning things in themselves—hopefully in a manner that can retain access to the many unmatched advantages of Allais’ interpretation of Kant on appearances. 

End