The Meaning of 'Refutation' in Kant's Refutation of Idealism

1. Introduction

In the ‘Refutation of Idealism’ in the B-edition of the Critique of Pure Reason, Kant promises what he calls a ‘refutation [Widerlegung]’ of Cartesian skepticism, which he also calls ‘psychological idealism’ (Bxxxix), ‘empirical idealism’ (A371–72) and ‘problematic idealism’ (B274).^1,2 Kant takes himself to turn the tables on the Cartesian skeptic by showing how ‘inner experience in general’ undoubted by the skeptic turns out to be ‘possible only through outer experience in general’ (B278–79), that is, the existence of extended beings in space.

There are many difficult questions surrounding the Refutation, including the relation between the B-edition of the Refutation and its initial formulation in the A-edition Paralogisms; why in the B-edition Kant inserts it after the Second Postulate of Empirical Thinking in General; the soundness of the argument; and the extent to which the Refutation is a non-question-begging argument against Cartesian skepticism.³

In this article, I focus on a question that, to my knowledge, has escaped sustained examination: what does Kant mean, exactly, by a ‘refutation’?

It is not obvious that this question merits an article-length treatment. By ‘refutation’ Kant presumably has in mind its ordinary meaning, to demonstrate the falsity of the claim in question.⁴ However, insofar as the Refutation is read as a refutation in this ordinary sense, the consensus seems to be that it has some serious lacunas and thus must be supplemented with substantive and controversial premises concerning transcendental idealism or self-knowledge.⁵

Examining the legal, logical, and what I call the critical uses of the notion of ‘refutation’ in Kant’s work will reveal a more modest sense of refutation, namely a diagnostic one. Kantian refutations first identify the ‘origins’ or grounds of the claim to be refuted in a way that explains why someone might have been initially attracted to the claim in question. They then show how such grounds contain a difficult-to-see fallacy that, once made explicit, should lead to a suspension of judgment. This will give us a better sense of the Refutation’s argumentative aims and strategy.

Here is a brief roadmap. I explain Kant’s legal, logical, and critical uses of refutations (§2). I then show how these uses clarify three aspects of the Refutation’s argumentative strategy (§3–5). I conclude (§6) by explaining how Kantian refutations share an affinity with contemporary ‘Moorean’ arguments against external world skepticism. The real promise of such arguments, I show, is to help us resist the initial appeal of skeptical arguments, as opposed to seeking to convince the skeptic of the external world’s existence.

2. Legal, Logical, and Critical Uses of ‘Refutation’ in Kant’s Work

Legal contexts. Dieter Henrich famously emphasized the significance of deduction-writings (Deduktionsschriften) for appreciating why Kant calls his transcendental deduction of the categories as such (Henrich 1989). Such deduction-writings justified rights by adducing the relevant legal fact, for example, my right to inherit my parent’s property would be ‘deduced’ by presenting an authentic will (the relevant legal ‘fact’ or Faktum).

Kant evinces familiarity with a similar legal use of ‘refutation’ which is synonymous with ‘dismissing’ a challenge to a particular right (see A430/B458, A501/B529), most notably in which he ‘refutes’ merely empirical derivations or justifications of the categories ‘by the fact [das Faktum].’ Kant’s point is that, since categories are by definition pure concepts that are independent of experience, any attempts to justify their use by appealing to experience can be ‘dismissed’ or rejected by fiat, based on the ‘actuality’ of our possession of such concepts as demonstrated in mathematics and general natural science.

Here, I shall focus on Kant’s familiarity as demonstrated in his ‘On the Wrongfulness of Unauthorized Publication of Books [Von der Unrechtmäßigkeit des Büchernachdrucks]’ (8:79–87), henceforth the ‘Wrongfulness essay’ (see also 6:289–91), which is divided into three parts, of which for present purposes I shall focus on the first two.⁶ The first part is titled ‘Deduction of the Right of a Publisher against an Unauthorized Publisher’ (my italics), which justifies an authorized publisher’s right to reproduce a given text for profit by adducing as the relevant legal fact the original consent of the author to the publisher, coupled with proof as to how such consent can only be legitimately given to one publisher.

The second part relevant to our purposes is titled ‘A Refutation of the Pretended [vorgeschützten—‘alleged’ is an alternative translation] Right of an Unauthorized Publisher Against the Publisher’ (my italics). In this refutation, Kant addresses a natural line of thinking that could justify an unauthorized publisher’s illegal activity, namely by appealing to ownership: her legitimate purchase of the book entails, the thought goes, that she ought to be able to do whatever one wishes to do with it—which includes lending it, using it as firewood, as well as copying its contents and selling them for profit.

In this refutation, Kant shows that this argument contains a fallacy of equivocation (sophisma figure dictionis) or a conflation of two meanings of ‘book’ (again, see 6:290–91). Purchasing a book grants someone the rights of ownership of the book as a ‘corporeal artifact,’ for example, lending it to someone else or burning it as firewood. However, such ownership does not imply ownership of the rights to the book as a ‘discourse of the publisher to the public,’ that is, its intellectual contents, which are owned solely by the publisher in virtue of the original consent of the author. By distinguishing these two senses the refutation makes manifest a difficult-to-see error that is contained in the grounds of the unauthorized publisher. The right of an unauthorized publisher is what Kant calls a ‘pretended’ or alleged right, a mere appearance of a right to publish.

In sum, if a deduction-writing justifies a certain right by adducing the relevant legal fact, a legal refutation reconstructs the grounds for why someone might challenge the relevant right, and then demonstrates how those grounds contain a fallacy that, once made clear, shows why the challenge can be ‘refuted’ or dismissed.

Logical contexts. The second related context in which Kant demonstrates his understanding of ‘refutations’ appears in his lectures on logic. Logical refutations reconstruct the claim to be refuted as the conclusion of a possible syllogism, making perspicuous the fallacious reasoning at stake. Like legal refutations, logical refutations aim to expose a covert error in the grounds of the claim to be refuted. In doing so, they help move the reasoner away from error and set them toward the path of truth.

Within Kant’s logic, good inferences are truth-preserving while bad or fraudulent syllogisms merely appear or seem to be truth-preserving (see A303/B360, A790/B818). Kant calls these fraudulent syllogisms fallacies or ‘an inference of reason that is wrong as to form, although it has for itself the illusion of a correct inference (fallacia)’ (9:134, my underline; see also 23:38, 24:96, 24:287, 24:777). A fallacy thus generates the ‘illusion [Schein]’ of truth-preservation. Kant thus distinguishes between truth and falsity as a property of judgments and truth and illusion (error) as a property of the relation between the judger and the judgment. An error arises when a subject ‘holds-to-be-true [correct]’ a false [incorrect] judgment due to a fallacy, and thus leads a thinker or reasoner astray by deceiving a thinker into inferring something false from true premises.

In turn, Kant distinguishes between what we might call a ‘direct’ refutation, which demonstrates a certain judgment as false, and a ‘proper’ refutation, which initially concedes that the judgment to be refuted could be true as a ‘provisionally good judgment’ and then seeks to show how such a judgment is the conclusion of a syllogism that contains an error:

. . . to avoid errors, then, one must seek to disclose and to explain their source, illusion. Very few philosophers have done that, however. They have only sought to refute the errors themselves, without indicating the illusion from which they arise. This disclosure and breaking up of illusion is a far greater service to truth, however, than the direct refutation of errors, whereby one does not block their source and cannot guard against the same illusion misleading one into errors again in other cases because one is not acquainted with it . . . (9:56)⁷

A logical refutation is didactic: it helps direct the reasoner away from error to help them reason towards the truth, without yet determining how that will be, by appealing to general principles (in this case, principles of general logic).

Like legal refutations, logical refutations reconstruct the grounds for the claim initially acknowledged as true, present the would-be grounds for the judgment to be refuted, and then show how such grounds contain a fallacy that the judger was unable to see, leading to the judger to suspend judgment.

Critical refutations. While my previous discussions of legal and logical refutations appealed to direct textual evidence, the following notion of a ‘critical refutation’ is something I am attributing to Kant on his behalf, although I am certainly not the first to recognize this tendency in Kant’s work. My textual evidence for this notion will thus be comparatively indirect.

Throughout the Critique, Kant tends to assess philosophical positions in teleological and normative terms, that is, distinguishing their true and false parts in terms of how close he believes they were to recognizing the necessity of a critique of pure reason. A critical refutation thus seeks to refute or dismiss a given philosophical position in a charitable manner, namely showing why a thinker might have been led to the position in the first place. I show this is the best way to read Kant’s relation to Descartes, and by extension the Refutation: assessing how Cartesian skepticism helps us get clearer about the nature and scope of our cognitive power. I shall develop this notion of a critical refutation by examining Kant’s critical refutation of Humean skepticism, especially those found in the Discipline of Pure Reason, and then show how Kant extended such an attitude towards Cartesian skepticism too.

There is a growing appreciation that contrary to previous readings that sought to frame Hume as a direct opponent or even foil for Kant (principally in the Transcendental Deduction or in the Second Analogy), Kant’s attitude towards Hume is ‘critical’ in the sense I am describing. On the one hand, Kant credits Hume as rightly recognizing the excesses of dogmatic rationalist metaphysics, indeed going so far as to suggest that Hume’s skepticism concerning our right to apply the concept of causality does not go far enough and must be generalized for all metaphysical concepts (see 4:312).⁸ On the other hand, Kant also suggests that when Humean skepticism is not constrained without prior critique, it can lead us to illegitimately reject the possibility of pure knowledge, a possibility that Kant claims that even the ‘common understanding’ can recognize as actual (see B3–5). For Kant, Humean skepticism must be harnessed correctly to enable a critique of pure reason to get off the ground.

Call a generalized Humean skepticism an argument-strategy that raises doubts about reason’s right to apply pure concepts in general (their ‘objective validity’) given that they are independent of experience: the categories (‘metaphysical skepticism’), mathematical concepts (‘scientific skepticism’), and the moral law (‘practical skepticism’). The Humean challenge is that the purported necessity of these concepts can be explained as useful and even psychologically necessary fictions or useful illusions that originate from our empirical faculty of reproductive imagination—what Kant calls mere subjective necessity. That is, we can at best show that we are so constituted to think of objects as causally related, or to think of ourselves and others as acting freely, as opposed to being able to cognize real causal relations or to cognize acts as genuinely originating from free subjects.

Kant critically refutes this generalized Humean skepticism as follows. In very broad strokes, Kant credits Hume’s metaphysical skepticism for making clear the need for a transcendental deduction as a special argument strategy that is otherwise required to justify our right to use the categories. Such an argument consists of two steps: first, demonstrating the pure (as opposed to empirical) origins of the categories by deriving them from the logical forms of judgment; and second, showing that such categories, despite such pure origins, are nevertheless restricted in their legitimate application for appearances given to us through sensibility, and thereby denying reason their possible use for cognition of supersensible objects. In remarks in the Discipline chapter, Kant credits this metaphysical skepticism for making possible the ‘censorship of reason’ or ‘subjecting the facta of reason to examination’ which leads to necessary ‘doubts about all transcendent use of principles’ (A760–61/B788–89). Such skepticism is ‘preparatory’ for reason in ‘arousing its caution and showing it fundamental means for securing it in its rightful possession.’ Humean skepticism, understood in its guise of metaphysical skepticism, has a genuinely important role to play as a catalyst for initiating a project of critique.

On the other hand, Kant ‘refutes’ Humean skepticism in its scientific and practical modes, because each leads to doubts concerning our right to use pure concepts that we already can affirm by virtue of theoretical and practical facts of reason. As I mentioned earlier, doubts about our possession of synthetic a priori cognitions in mathematics (and natural science) can be ‘refuted’ by the ‘fact’ of the activities of these two sciences, though to be sure Kant still thinks one can have legitimate doubts as to how such concepts, given their being pure concepts, must nevertheless apply to objects given to us through sensibility (that a deduction can help explain).

As for practical skepticism, in the second section of the Groundwork, Kant suggests that it originates from a difficult-to-see error: they falsely infer from epistemological conditions, or that we might never know or be able to determine which of our acts originate from the moral law, to metaphysical ones, or the impossibility of an action that originates from the pure will. Denying this inference at least, by Kant’s lights, leaves open the possibility that our actions are governed by an otherwise pure, non-sensible moral law. As Kant puts this: ‘it would be an even greater absurdity for us not to allow any things in themselves at all, or for us to want to pass off our experience for the only possible mode of cognition of things . . . and so to want to take principles of the possibility of experience for universal conditions on things in themselves’ (4:350–51). When Humean skepticism leads to doubts concerning this fact of practical reason, Kant then thinks this skepticism itself manifests into a kind of unacceptable ‘dogmatism’ that ‘boldly denies whatever lies beyond the sphere of its intuitive cognitions . . . it itself makes the same mistake of immodesty, which is all the more blameable here, because it causes an irreparable disadvantage to the practical interests of reason’ (A417/B499). This disadvantage manifests in what Kant calls errors of materialism, that is, no cognition of soul, atheism (no God) and fatalism (no freedom) (see Bxxxiv). Practical skepticism can thus be dismissed or refuted by appealing to a counterpart ‘fact’ of pure practical reason, or our consciousness of the moral law that Kant thinks must be available even to the ‘most hardened scoundrel’ who wishes to be disposed to act in accordance with it.

Kant thus ‘critically refutes’ Humean skepticism by drawing a clear boundary between its legitimate and illegitimate uses: when properly used, it leads a reasoner to appreciate the necessity of a transcendental deduction, which is the key move into Kant’s system. However, if improperly generalized or extended to all pure concepts, it would deprive us, by Kant’s lights, of the scientific status of mathematics and natural science. To be sure, one might have doubts about the merits of Kant’s appeal to such a fact, and how it remains an open possibility that our moral practice can be explained by an empirical practical reason, and I shall revisit this move in the conclusion.

***

In what follows, I shall now apply these legal, logical, and critical uses of refutations to clarify three aspects of Kant’s intended argumentative strategy in the Refutation.

3. The Legal Aim of the Refutation

Recall a deduction-writing justifies a certain right by adducing the relevant ‘fact’, while a legal refutation acknowledges a challenge to this right and then shows how this challenge originates from a fallacy. By doing so, a refutation shows how a certain putative right lacks justification and is thus merely a ‘pretended right’ or the mere appearance of a right.

On this legal model, I propose we read the Refutation as addressing Cartesian skepticism as a challenge to the transcendental deductions of space and time in the Transcendental Aesthetic, which justifies our right to apply space to objects of outer sense and not things-in-themselves by adducing the fact of space as an original a priori intuition. In Kant’s terms, Cartesian skepticism raises doubts about the objective validity of space. The legal aim of the Refutation is to meet this challenge and thereby ‘refute’ or dismiss it.

The transcendental deductions of space and time are easy to overlook.⁹ Based on Kant’s remarks in which he first introduces the notion of a transcendental deduction in §13, however, such a deduction almost certainly has taken place in the Transcendental Aesthetic, although not called one as such, presumably a conscious decision by Kant not to confuse the reader too early (see A88/B121). In what follows, then, I shall explain the metaphysical and transcendental expositions of space in the context of its transcendental deduction (the same applies to time): those expositions together demonstrate space as an a priori intuition as the relevant ‘fact’ or origin of space that grounds our right to use it as a form of sensibility, from which Kant ‘deduces’ the objective validity of space in his general remarks concerning the empirical reality and transcendental ideality of space, in which the validity of space is restricted to appearances and not things-in-themselves.

While the details of the metaphysical and transcendental expositions are controversial, it is generally agreed that the First and Second Metaphysical Expositions show that space is a pure representation, that is, prior to experience rather than an empirical one, while the Third and Fourth Expositions show that space is a singular, intuitive unity that is a whole prior to its parts which makes it an a priori intuition (as opposed to a concept). In my view, I am sympathetic to readings that treat Kant’s argument in the metaphysical expositions as undertaking a special kind of ‘analysis’ or examination of our capacity to grasp spatial relations in general.¹⁰ For example, in the Third Exposition, by reflecting upon an exercise of our capacity to represent different spaces Kant argues that this exercise necessarily presupposes a further capacity to conceive of these parts as part of a single, overarching spatial framework, which in turn must always be contained in a larger space; and so forth. Instances of successful exercises of our capacity to grasp spatial relations are presupposed, for which then Kant seeks the conditions of such successful exercises. By doing so, the metaphysical expositions show that the ‘original representation of space is an a priori intuition’ (A25/B40, my underline), whereby I emphasize ‘original’ to show that these expositions determine the ‘origin’ or ground of this a priori intuition of space from our cognitive faculties rather than being derived from experience.

The transcendental exposition of space then shows that ‘the representation of space . . . must originally be intuition’ that ‘must be encountered in us a priori, i.e., prior to all perception of an object, thus it must be pure, not empirical intuition’ (B41). Now, the transcendental exposition corroborates the analysis of our cognitive capacity found in the metaphysical exposition and concludes that space must originate a priori from sensibility to explain the ‘possibility of geometry as a synthetic a priori cognition’ (B41). Again, Kant appeals to the ‘reality’ of geometry and regresses towards the conditions of its possibility, namely that space must originally be an a priori intuition and not derived from experience. Taken together, the metaphysical and transcendental expositions confirm on independent grounds that space is ‘originally acquired’ (8:221): since it cannot be derived from experience, it must be derived instead from our cognitive faculties.

The necessity of a transcendental deduction is occasioned by a challenge concerning rightful or authorized use. Recall this challenge as it arises for the categories. The question Kant seeks to answer is ‘how subjective conditions of thinking should have objective validity’ (A89–90/B122). As a priori concepts, they cannot be derived from experience but must instead be derived from the logical forms of judgments as merely its ‘subjective’ forms, where by ‘subjective’ Kant means ‘originating from the subject’ and not derived from the object. Such subjective origins raise the possibility that these categories could be merely what Kant calls Hirngespinste or figments of the imagination that, while certainly valid for the subject, might not be valid for objects given to me through sensibility; they could be ‘empty, nugatory, and without significance’ or meaning (see A90–91/B123).

A similar worry arises for space insofar as it, too, like the categories is an a priori representation. Having demonstrated that space is an a priori intuition, the transcendental deduction of space can be found in the following remarks, in which Kant infers that, given the expositions rest upon a faculty of sensibility as a capacity for mathematical cognitions, the legitimate scope of space as a form of sensibility pertains to (outer) appearances and not things-in-themselves (see A87–88/B120):

Our expositions accordingly teach the reality (i.e., objective validity) of space in regard to everything that can come before us externally as an object, but at the same time the ideality of space in regard to things when they are considered in themselves through reason, i.e., without taking account of the constitution of our sensibility. (A272–78/B43–44)

That is, ‘the principles of the transcendental aesthetic . . . [show that] space and time are the conditions of the possibility of all things as appearances, as well as the restriction of these principles, namely that they cannot be related to things in themselves’ (A149/B189). Space thus only applies to both pure and empirical intuitions, while it cannot be applied to objects beyond experience, for example, the soul or God, as someone like Crusius might have thought.

On this legal model, Kant frames the challenge of Cartesian skepticism as raising a worry that is internal to the Aesthetic: conceding that both space and time are subjective forms of our sensibility, but then going on to raise doubts about the validity of space as applying to any object in general. In the A-Refutation, Kant frames ‘empirical idealism’ as a ‘false scruple concerning the objective reality of our outer perceptions’ (my italics) and that space, ‘though in itself it is only a mere form of representation, nevertheless has objective reality to all our outer appearances (which are also nothing but mere representations)’ (A376–77). Perhaps, the thought goes, given the method of doubt in which all I can determine with certainty is that I can only become conscious of my representations of things and not of the objects represented, it follows that, while the Cartesian can grant that time is a form of our inner sense (our capacity for representations is not in doubt), it is certainly possible that there does not exist anything to which the concept of space applies, that is, it remains open that space is a Hirngespinste or figment of the imagination that has no validity.

The legal model of a refutation suggests two points. First, the deductions of space and time are taken for granted. There is thus no real threat for Kant that we cannot cognize external objects. Such doubts concerning the validity of space run afoul of the fact of theoretical reason as it pertains to the ‘reality’ of geometry as a legitimate source of synthetic a priori cognitions: ‘geometry nevertheless follows its secure course through strictly a priori cognitions without having to beg philosophy for any certification of the pure and lawful pedigree of its fundamental concept of space’ (A87/B120). Its status as a source of synthetic a priori cognitions cannot be in doubt. Since doubts concerning the validity of space lead to doubts concerning this fact, Kant thinks we can dismiss it.

Second, however, Kant thinks that a ‘refutation’ of Cartesian skepticism is warranted because it follows from what he takes to be a line of reasoning that would be natural and understandable for someone not yet fully initiated into transcendental idealism: one who, while correctly recognizing that we can only be certain of our representations of objects and thereby implicitly endorsing time as a form of inner sense, then falsely infers that such a doctrine entails that we cannot know outer objects as they are. The aim of the Refutation is thus to shed light on a difficult-to-see error that, once made explicit by reiterating the significance of transcendental idealism and in particular the dependence of time upon space, should make clear why this natural inference is ultimately mistaken. Making this error explicit will then be the aim of the next section, which addresses the logical aim of the Refutation.

4. The Logical Aim of the Refutation

By Kant’s lights, the Cartesian skeptic begins from a premise that Kant would accept—that we can only become conscious of all our possible representations (B131–32)—and then concludes with doubts concerning our right to apply space to objects of outer experience but not time to objects of inner experience. The logical aim of the Refutation is thus to show how this reasoning contains an illusion or fallacy that the Cartesian skeptic is unable to see—but one that a transcendental idealist like Kant can show. I shall seek to explain the proof of the Refutation as revealing this error.

Befitting a logical refutation, here is a syllogism that Kant seems to attribute to the Cartesian skeptic:

• (1) Only inner perceptions can be immediately cognized with certainty;

• (2) If only inner perceptions can be immediately cognized with certainty, then outer perceptions cannot (i.e., must be indirectly cognized or inferred from inner perceptions);

• (3) If outer perceptions cannot be immediately cognized with certainty (i.e., must be indirectly cognized or inferred), then outer perceptions can be doubted;

Therefore,

• (5) Outer perceptions can be doubted.

Kant thinks that the Cartesian skeptic arrivesat (1) through a justified, legitimate method of doubt. Kant also seems to concede (3). The problematic premise, then, is (2), and the kind of implicit error it contains.

In the A-edition Refutation Kant explains that (2) conflates the empirical and transcendental meanings of ‘outer’ (A373). (2) is true only if by ‘outer perception’ one takes it in a transcendental sense, namely a capacity to cognize things-in-themselves as objects outside of reason. However, (2) is false if by ‘outer’ perception the premise was taken in an empirical sense, for example, ‘outside’ our bodies. The Cartesian skeptic doubts the existence of outer objects because they conflate outer objects with things-in-themselves; Kant concedes that such a skeptic would indeed be correct to note how it is ‘absolutely impossible to comprehend how we are to acquire cognition of their reality outside us’ (A378). The A-edition Refutation is thus a proper refutation, removing the grounds that otherwise would lead someone to doubt the existence of external objects in the first place.

In the B-Refutation, however, on my view Kant seeks to demonstrate the fallacy of equivocation at stake more directly by presenting what he calls a ‘direct’ or ‘ostensive’ proof which ‘combines with the conviction of truth and simultaneously with insight into its source’ (A789/B817) and is proper in matters of pure reason. The proof of the B-Refutation now directly demonstrates the negation of (2), namely that inner perceptions can be immediately cognized with certainty only if outer perceptions (as objects outside of us in space) must also be immediately cognized with certainty. And to show this, Kant thinks there are two notions of priority at stake: the ‘genetic priority’ of time over space as form of our sensibility and the ‘explanatory priority’ of time over space, or how the latter explains the possibility of the former as a form of sensibility. The Cartesian conflates these two types of priorities.

To illustrate this distinction between genetic and explanatory priority, consider Kant’s famous claim arguing that the understanding ‘can make no other use of these concepts than that of judging by means of them’ and that the understanding is a faculty or capacity for judging (A69/B94). As Thomas Land helpfully explains, what Kant must be claiming here given his other philosophical commitments is that a capacity for concept use presupposes, or is to be explained by, capacity for concept use as predicates of possible judgment, rather than a stronger claim that every actual exercise of our capacity for concept use involves an exercise of judgment (Land 2015). This thesis explains, for example, how geometrical concepts can be used for acts of constructing objects in pure imagination, or even explaining our standalone concept use when I use the concept of red to think of red trucks, flowers, what angers bulls, and so forth—what Kant’s empiricist contemporaries would have called an apprehension of ideas. Kant is here reducing in explanatory terms distinct acts of our intellect—concept use, judging, and syllogisms—in terms of a single capacity—namely our capacity for judging.

On the one hand, Kant concedes that concepts are ‘genetically prior’ to judgments insofar as judgments are composed of concepts; logical forms take as their ‘matter’ different concepts. However, Kant thinks that a crude empiricism that explains the logical forms of judgments as arising from composition or through patterns of repeated association is an obvious failure since it could not explain how such forms stand in objective a priori relations to other forms within general logic. By Kant’s lights, empiricist accounts of judging fail to explain the logical form of generality insofar as they conflate genetic priority with explanatory priority. Rather, Kant thinks that concepts acquire their characteristic logical form (generality) in virtue of being usable as predicates of possible judgments.¹¹ Instead, the logical form of judgments is explanatory prior to the forms of concepts (see B141).

I take Kant to be making an analogous philosophical move in the Refutation. On the one hand, time as a form of inner sense is genetically prior to space, in that representations are ordered successively via the form of time through inner sense:

Wherever our representations may arise, whether through the influence of external things or as the effect of inner causes, whether they have originated a priori or empirically as appearances—as modifications of the mind they nevertheless belong to inner sense, and as such all of our cognitions are in the end subjected to the formal condition of inner sense, namely time, as that in which they must all be ordered, connected, and brought into relations. (A98–99)

However, given Kant’s independent commitments in his philosophy of mathematics, he also endorses the explanatory priority of space over time, or how a capacity for temporal determination (or to distinguish different states within me in relations of succession) presupposes a capacity for spatial determination (objects in different spaces). This explanatory possibility is how I read Kant’s claims that suggest, for example, ‘we cannot even represent time without, in drawing a straight line which is to be the external figurative representation of time’ attending merely to the action of the synthesis of the manifold through which we successively determine the inner sense, and thereby attending to the succession of this determination in inner sense’ (B154; see also B292). For Kant, a capacity to represent and distinguish moments in time involves a capacity to represent and attend to each successive moment by which I construct a pure intuition of a line in productive imagination which, in turn, is a kinematic act—attending to the manner in which a point is moving in space. Again, this need not imply that every exercise of a temporal determination requires a representation of it in space, for example, in counting or even representing my internal representations as succeeding one another, I need not explicitly appeal to spatial relations.

Granting this thesis of the explanatory priority of space over time, however, we do not yet have Kant’s intended conclusion of the Refutation which suggests he wants to show something stronger than our merely having a capacity for cognizing outer objects: he says that inner experience is ‘possible only through a thing outside me and not through the mere representation of a thing outside me’ (B275–76) or how ‘the consciousness of my own existence is at the same time an immediate consciousness of the existence of other things outside me’ (B276). Demonstrating that someone has a certain capacity to do something need not imply that they will ever be in the right conditions to exercise such a capacity.

Just as we saw in the metaphysical and transcendental expositions, having a capacity for spatial determination is sufficient to show that the objects of such a capacity, namely external things, exist given the ‘reality’ of the facts of mathematics and natural science. For Kant, a determinate exercise of this capacity for spatial determination involves being able to construct objects in pure intuition, for example, to determine two spaces as different locations is to be able to draw a straight line between the two points in question. In turn, Kant insists that this capacity for geometric objects cannot be a source of mere figments of the brain or Hirngespinste unlike, say, the concepts of fortune or fate that lack such objective validity. In his discussion of the Postulates of Empirical Thought (directly preceding the Refutation), Kant tells us that geometric concepts are objectively valid precisely because they must be applicable to possible experience:

It may look, to be sure, as if the possibility of a triangle could be cognized from its concept in itself (it is certainly independent of experience); for in fact we can give it an object entirely a priori, i.e., construct it. But since this is only the form of an object, it would still always remain only a product of the imagination, the possibility of whose object would still remain doubtful, as requiring something more, namely that such a figure be thought solely under those conditions on which all objects of experience rest. (A223–24/B271, my underline)

Kant’s general point here is that geometric objects—produced by what he calls a figurative synthesis (B151)—are merely ‘forms of an object in general’ and not representations of objects as such. However, insofar as such mathematical objects (circles, conic sections, parabolas) are seemingly always realized within our best empirical-scientific theories, such objects cannot merely products of the imagination but must otherwise be really applicable to existing objects of experience if we are to explain the exact application of mathematics to external objects, that is, ‘. . . all mathematical concepts are not by themselves cognitions, except insofar as one presupposes that there are things that can be presented to us only in accordance with the form of that pure sensible intuition. Things in space and time, however, are only given insofar as they are perceptions (representations accompanied with sensation), hence through empirical representation’ (B147; see also A165–66/B206).¹²

Broadly speaking, then, Kant does not maintain a distinction between what we might now call pure and applied mathematics; all mathematical concepts (and their objects) must apply to objects of empirical intuitions to explain their synthetic character, though they must also be products of a pure intuition to explain their a priori character. The worry being raised—someone maintaining we have a capacity for spatial determination but doubts the existence of objects upon which such a capacity would be exercised—cannot be reconciled, as it were, with what we might now call an ‘indispensability argument’ within the philosophy of mathematics, or the fact that geometric concepts such as lines and circles are exactly applicable to objects within empirical science (distance, orbits) implies a commitment to the existence of such objects. What Kant is saying is that if mathematical objects are not to be mere Hirngespinste we are committed, so to speak, to their necessary application for empirical objects—a project that Kant outlines in greater detail in his Metaphysical Foundations of Natural Science and his quasi-mathematical treatment of the empirical concept of matter.

To sum, Kant begins by supposing on behalf of the Cartesian skeptic that they grant that time is a form of inner sense. Such a capacity for temporal determinations presupposes a further capacity for spatial determination. In turn, Kant thinks that objects of such a capacity must exist to explain the synthetic character of mathematical claims—the latter which, in turn, is presupposed by Kant as a kind of theoretical ‘fact’ of reason. Since for Kant doubts concerning the existence of external objects leads to doubts concerning the possibility of applied mathematics, which in turn cannot be coherently doubted, Cartesian skepticism can be cast aside.

5. The Critical Aim of the Refutation

In both editions of the Refutation, Kant concedes that Cartesian skepticism is a serious problem that must be addressed by a critique of pure reason. Kant is crediting Cartesian skepticism and its method of doubt that proceeds upon the basis that ‘no decisive judgment until a sufficient proof has been found,’ or not to accept anything unless one can establish the grounds of a particular claim (B274–75), and he finds this to be a ‘rational and appropriate for a thorough philosophical manner of thought.’ In doing so the skeptical idealist exhibits the right kind of humility or caution required in matters of pure reason and guarding against any ‘surreptitious’ acquisition of concepts, for example, confusing the concept of the soul with the unity of apperception (A377–78).

In ‘critical’ terms, then, Kant thinks that the Cartesian method of doubt is a legitimate exercise of reason as it originates from what he takes to be a proper philosophical attitude, namely to accept only those claims that can be demonstrated with certainty. This method of doubt, when used correctly, can help a reasoner distinguish those judgments that originate solely from pure reason from those judgments that do not, or what Kant calls prejudices which are judgments accepted out of custom, imitation, or inclination. By Kant’s lights, this method of doubt leads to an acceptable kind of representationalism that implies we cannot cognize things-in-themselves. Kant commends this true part of Cartesian skepticism as a ‘benefactor for human reason’ (A377) as it contains the beginnings of a genuinely critical idealism. In the A-edition Kant credits this kind of thoughtful doubt as regarding all perceptions merely as representations and not things-in-themselves:

The utility created by these idealistic projects is now clearly before our eyes. They drive us forcefully . . . to regard all perceptions, whether they are called inner or outer, merely as a consciousness of something that depends on our sensibility, and to regard their external objects not as things in themselves but only as representations, of which we can become immediately conscious like any other representation, but which are called external because they depend on that sense which we call outer sense; its intuition is space, but it is itself nothing other than an inner mode of representation, in which certain perceptions are connected with one another. (A378)

When the method of doubt is used correctly, as beginning from a premise concerning how we can only be conscious of our possible representations (see B131–32), Cartesian skepticism can lead a reasoner to Kant’s crucial distinction between appearances and things-in-themselves and thus to a version of a view that is genuinely Kantian—that we cannot cognize things-in-themselves—and, by extension, to an alternative route to the beginnings of a genuinely critical or transcendental idealism.¹³

For Kant, then, the kernel of truth in Cartesian skepticism is that it seemingly raises legitimate doubts about our right to apply certain pure representations: namely, a non-intellectual, that is, sensible a priori representation of space as the form of our sensibility. When the method of doubt that characterizes Cartesian skepticism is used correctly or legitimately, it contains the beginnings of a genuinely critical idealism by Kant’s lights, that is, such a method is a ‘thorough philosophical manner of thought’ (B275) insofar as it shows that we can only become immediately conscious of our representations in general, which thus should lead to legitimate doubts concerning our capacity to become conscious of things-in-themselves or objects independent of our forms of representing them. The legitimate use of Cartesian skepticism is ‘preparatory’, to arouse the reader towards the necessity of a critique.

The critical aim of the Refutation is to show that, when this method of doubt is indiscriminately applied to lead to doubts about both things-in-themselves (as objects ‘outside’ of reason) and outer appearances (as objects ‘outside’ of us), it manifests as an illegitimate empirical idealism as opposed to what Kant takes to be the right and indeed only possible form of idealism—transcendental idealism as the ‘only refuge remaining to us’ (A378). Properly construed, Cartesian skepticism contains the beginnings of an idealism that could be deemed ‘critical’ by Kant’s lights—this is why Kant calls it a ‘problematic’ idealism, or an exercise of reason that raises the possibility of the kind of transcendental idealism that Kant finds acceptable.

However, if such Cartesian skepticism were to then be illegitimately used in a way that would leave us in what Kant calls the ‘scandalous’ position of being unable to justify, and thus take on mere ‘faith’ what we must already claim to know, namely our direct perception of objects existing outside of us (see Bxxxix)—something that presumably Kant believes would be rejected by even the ‘common understanding’—then it can be refuted or dismissed. The utility of Cartesian skepticism can be constrained: while Cartesian skepticism is nevertheless certainly ‘preparatory’ for arousing the caution of reason regarding our capacity to cognize things-in-themselves, when it leads to ‘distrust’ of reasoning cognizing external objects, it can be dismissed or refuted.

Compare, for example, Kant’s favorable attitude towards Cartesian skepticism with his dismissive attitude towards Berkeleyan idealism, ‘who declares space, together with all the things to which it is attached as an inseparable condition, to be something that is impossible in itself, and who therefore also declares things in space to be merely imaginary.’ Unlike Cartesian skepticism, which raises doubts about the objective validity of space, Berkeleyan idealism raises doubts about the possibility of space itself. A ‘refutation’ of Berkeleyan idealism—though Kant does not call it as such—is thus comparatively terse: such an idealism presupposes ‘space as a property that is to pertain to the things in themselves; for then it, along with everything for which it serves as a condition, is a non-entity.’ In fact, Berkeleyan idealism falls short insofar as it fails to even distinguish between appearances and things-in-themselves: thus ‘the ground for this [Berkeleyan] idealism, however, has been undercut by us in the Transcendental Aesthetic’ (B274–75) and its distinction between appearances and things-in-themselves. Cartesian skepticism, and its method of doubt, is a way to yield this distinction and to this extent is, by Kant’s lights, comparatively more ‘critical’ than Berkeleyan idealism. It thus warrants a sustained response, or a ‘refutation’ in the legal and logical senses already explained.¹⁴

So, Cartesian skepticism (unlike Berkeleyan phenomenalism) merits a ‘refutation’ precisely because such skepticism is obliquely raising the problem of the necessity of a transcendental deduction of space in general, or a determination of the legitimate boundaries to which space can apply. Insofar as Cartesian skepticism implies the necessity of a transcendental deduction of a certain a priori concept (though certainly neither Descartes nor any other external world skeptic would frame it as such), for Kant it contains the beginnings of a truly critical idealism.

The Refutation thus takes Cartesian skepticism seriously insofar as it is generated by reason alone, rather than a skepticism concerning our justification of the existence of external objects.¹⁵ Drawing upon the work of Stephen Engstrom, the argumentative form of the Refutation can be helpfully compared to a theodicy.¹⁶ A theodicy is addressed to a believer undergoing a crisis of faith because of an initial inability to reconcile an internal conflict between her belief in God and her recognition of the existence of evil. A theodicy seeks to help the would-be believer reconcile this conflict to help them rediscover or recommit, so to speak, to their faith in God, and that the conflict between these two beliefs is in fact only apparent and has an underlying common ground. It presupposes, so to speak, that belief in God is the default position, and seeks to reconcile the apparent conflict in its favor; such an argument form would not convince a committed atheist to change their mind about God.

Just as a theodicy is not meant to convince, say, an atheist that God exists, the Refutation is not addressed to persuade a committed Cartesian skeptic. The existence of external objects is presupposed. On this model, the task of the Refutation is instead dialectical, in the sense of seeking to reconcile two seemingly legitimate but conflicting exercises of reason: one that, through the expositions and transcendental deduction of space in the Transcendental Aesthetic, shows that space as our subjective form of sensibility must only apply to outer appearances and not things-in-themselves; and another that, through the method of doubt, shows that only objects of inner and not outer appearances can be cognized with immediate certainty, and thus raises doubts about the objective validity of space. The Refutation aims to reconcile this conflict by showing that this apparent conflict can be reconciled through an appeal to a common ground: appreciating the significance of the former exercise shows that the latter conflates what I have called the genetic priority of time with the explanatory priority of space. By doing so, the reasoner can appreciate, in a new light, the significance of transcendental idealism outlined in the Aesthetic, and reconcile this conflict of reason.

6. Conclusion: Kantian Refutations in General

In the Refutation, Kant is using the threat of Cartesian skepticism as a foil to reiterate the significance of his doctrine concerning the transcendental ideality of space and time first articulated in the Transcendental Aesthetic. The Refutation makes clear that by the ‘subjectivity’ of space he is not suggesting that our cognitive access to the external world is somehow mediated by, and thus possibly even distorted or constrained, by these forms, or that such forms ‘impose’ some structure onto external content. Rather, such forms—in conjunction with the categories as forms of thought—supply us with just those necessary rules which constitute our direct cognition of such objects. To deploy a Hegelian turn of phrase, the ‘subjective validity’ of space is a subjectivity that is also, immediately at once, an objectivity, or that space and time are just those modes (and indeed, the only modes) through which we can directly cognize external sensible objects.

I shall now conclude by explaining what we might stand to gain from my examination of Kantian refutations. Past readers have read Kant’s arguments, and especially the Refutation, as a model for a class of ‘transcendental arguments’.¹⁷ Standardly, such arguments typically begin with a claim that is conceded by the skeptic, and then ‘regress’ by demonstrating certain necessary conditions for the claim that the skeptic otherwise denies.

In what follows, however, I shall conclude by showing that Kantian refutations actually share a closer affinity with what we might call broadly ‘Moorean’ strategies against external world skeptics. From my examination, Kantian refutations have seemingly two characteristic features: first, they seek to diagnose an internal instability within the skeptic’s position that should at least give them pause by identifying a difficult-to-see error. Second, in diagnosing this instability Kant appeals to a fundamental class of facts—what I have identified as theoretical and practical facts of reason—as fixed, initial points of inquiry.

In a series of papers, Thomas Kelly outlines two similar features of Moorean arguments as follows (Kelly 2005). First, for Kelly the dialectic between the skeptic and the Moorean is not a first-order disagreement between what would, in fact, justify our ordinary experiences, but a second-order disagreement concerning who ought to be attributed the burden of proof. Second, Kelly thinks we should read Moore’s well-known appeals to, for example, the existence of his hands as appealing to a special class of ‘Moorean’ facts in the context of this second-order disagreement: those facts that have a kind of ‘epistemic standing which renders it peculiarly resistant to being rationally undermined’ (Kelly 2005: 180). For Moorean arguments, then, insisting upon the special epistemic significance of these facts is not aimed at the skeptic, but helping us ordinary reasoners resist the claims of the skeptic. The key move for Moorean arguments, then, is to argue that the burden is squarely on the skeptic: ‘the onus is on the skeptic to provide a compelling argument for his conclusion, and Moore is providing reasons for thinking that such a project will inevitably end in failure’ (Kelly 2005: 182).

The Kantian ‘facts of reason’ that I have discussed in this paper play analogous roles with regards to Moorean facts in helping the reasoner resist the force of skeptical arguments, and showing that the burden of proof is on the skeptic, not the reasoner. And the real point of contention, then, between Kant and a skeptic would be a second-order disagreement concerning the nature of reason more generally, especially concerning the relation between these facts of reason and what that means for the possibility of a faculty of pure reason.

As ‘facts of reason’ Kant seems to frame them as capturing core commitments entailed by ordinary reasoners. In the Introduction to the Critique, Kant seems to take for granted that we do in fact have a faculty for pure cognitions. Kant distinguishes between cognitions beginning with the senses to their being totally constituted by experience, and begins with the possibility that experience is a ‘composite . . . of that which we receive through impressions and that which our own cognitive faculty (merely prompted by sensible impressions) provides out of itself’ (B1–2). Kant seems to emphasize the significance of this fact based on a standard Leibnizian view whereby, since mathematical judgments are necessary truths, and such necessary truths can never be derived from experience (experience can only tell us what is, not what must be), it follows that such necessity points to the presence or existence of an ‘a priori’ faculty. Kant also seems to think that this is something that ‘even’ the common understanding can recognize, and again he points to examples of mathematics as well as examples from natural science (alteration). As he puts it:

strict universality belongs to a judgment essentially; this points to a special source of cognition for it, namely a faculty of a priori cognition. . . .

Now it is easy to show that in human cognition there actually are such necessary and in the strictest sense universal, thus pure a priori judgments. If one wants an example from the sciences, one need only look at all the propositions of mathematics (B3–4).

Kant thinks these examples suffice for proof of the ‘reality’ of pure a priori principles and a faculty of a priori cognition more generally (B5–6). A similar consideration seems to apply for the fact of practical reason: the broader project of the Groundwork rests upon Kant’s belief that the intelligibility of our ordinary moral practices, which involve attributing free acts and responsibility to ourselves and other individuals, must thus be underwritten by at least the possibility of a pure morality, or that at some level our actions are intelligible as being moved independent of our inclinations. Again, Kant appeals to ordinary intuitions:

I assume that there are really pure moral laws, which determine completely a priori (without regard to empirical motives, i.e., happiness) the action and omission, i.e., the use of the freedom of a rational being in general, and that these laws command absolutely . . . and are thus necessary in every respect. I can legitimately presuppose this proposition by appealing not only to the proofs of the most enlightened moralists but also to the moral judgment of every human being, if he will distinctly think such a law. (A807/B835)

The problem with this move, however, and with Moorean strategies more generally, is how to argue for the distinct epistemic status of such facts of reason, and how appealing to such facts does not constitute a return to the kind of unacceptable dogmatism that Kant sought to redress.

By Kant’s own lights, for example, one might wonder whether the inability to deny something, it would seem, says nothing about either its epistemic status (whether we should in fact believe it) or its truth (whether it, in fact, is true). It may well be maintained that perhaps all Kant has been able to establish is a conditional statement to the effect of something like, if mathematical judgments are thought to be necessary, then skepticism can be met; but the skeptic could argue that the ‘necessity’ in question can be explained without an appeal to a mysterious faculty of pure reason. Moreover, one of Kant’s accomplishments was to distinguish between the conditions of a concept (analysis) and the conditions of their representing objects (which requires ‘synthesis’ or intuitions in general). The kind of conceptual analysis undertaken in Groundwork I reveals, at best, a capacity for acting purely out of a good will is possible. It can be conceded that an analysis of our concept of duty reveals the form by which a capacity would act solely in accordance with the moral law. But why think that such a capacity is actual? Why not think we are motivated merely by empirical practical reason, rather than the pure practical reason Kant thinks we are?

Again, an empiricist like Hume need not be committed to the view that our possession of mathematical and moral cognitions imply a pure faculty. Hume would have to simply insist that the purported ‘necessity’ of mathematical judgments need not indicate a pure faculty of reason but instead point towards our subjective necessity, that we are so constituted that we cannot think otherwise. Someone like Quine, for example, explicitly puts this view as follows:

Mathematics and logic, central as they are to the conceptual scheme, tend to be accorded such immunity, in view of our conservative preference for revisions which disturb the system least; and herein, perhaps, lies the ‘necessity’ which the laws of mathematics and logic are felt to enjoy . . . the laws of mathematics and logic may, despite all ‘necessity’, be abrogated. (Quine 1982: 2–3)

Even the purported necessity attributed to claims of mathematics and logic, then, would be subject to confirmation by experience; but this confirmation is not done piecemeal, but rather wholesale; logic and mathematics are, in principle, revisable but their indispensability for scientific theorizing means that their claims must be last, so to speak, in their revision—but their necessity is thus one of degree rather than in kind.

So we are left here, then, with a second-order disagreement concerning the possibility of a pure faculty. In his What Progress essay Kant comes closest to entertaining the consequences of an empiricist view of reason seriously—which he takes to lead to the conclusion that ‘transcendental philosophy is an absurdity’—and then argues against it as follows:

But since, however, of those propositions which prescribe a priori the rule to possible experience, such as, e.g., All change has its cause, it cannot be denied that they are strictly universal and necessary, and yet are nevertheless synthetic, it follows that empiricism, which declares all this synthetic unity of our representations in cognition to be a mere matter of custom, is totally untenable, and there is a transcendental philosophy firmly grounded in our reason . . . (20:275)

Here Kant’s response is simply to insist, yet again, upon the universal and necessary status of judgments in pure mathematics and natural science. Against the Quinean view, Kant simply seems to maintain that

the very concept of a cause so obviously contains the concept of a necessity of connection with an effect and a strict universality of rule that it would be entirely lost if one sought, as Hume did, to derive it from a frequent association of that which happens with that which precedes and a habit (thus a merely subjective necessity) of connecting representations arising from that association. (B4–5)

I think a more promising line of response, however, lies in picking up a thread that I briefly discussed in section 4. For Kant thinks anyone genuinely committed to an empiricism concerning reason, that is, denying a pure faculty, will always be left in a state of lingering ‘dissatisfaction’ (A805/B833) concerning questions of immortality, God and freedom, which in turn for Kant points towards a general ‘natural predisposition’ for metaphysics that is common to all of us as rational beings:

For human reason, without being moved by the mere vanity of knowing it all, inexorably pushes on, driven by its own need to such questions that cannot be answered by any experiential use of reason and of principles borrowed from such a use; and thus a certain sort of metaphysics has actually been present in all human beings as soon as reason has extended itself to speculation in them, and it will also always remain there. (B21)

For Kant this natural predisposition seems to be an anthropological fact about human reason, partly empirical and part a priori. It is ‘empirical’ insofar as it says something about human tendencies in general, but a priori to the extent that it captures a necessary feature concerning any rational thinker as such. And for Kant, any reasoner will, at some point or another, ask questions and demand answers concerning the ultimate grounds of reality, especially that of the possibility of souls, God, and acting freely. The fact of reason, especially the practical fact concerning consciousness of the moral law, explains both the universal persistence of such questions as well as the seeming sense of unease or dissatisfaction that accompanies Humean empiricist explanations that concepts like freedom are merely useful fictions. And the only possible answer for such questions, for Kant, is to be found in transcendental idealism.

Competing Interests

The author has no competing interests to declare.