Practical Cognition and Knowledge of Things-in-Themselves

Practical Cognition and Knowledge of Things-in-Themselves Karl Schafer 8/15/16 Famously, in the second Critique, Kant claims that our consciousness of the moral law provides us with sufficient grounds for the attribution of freedom to ourselves as noumena or things-in-themselves. In this way, while Kant insists that we have no rational basis to make substantive assertions about things-in-themselves from a theoretical point of view, it is rational for us to assert that we are noumenally free from a practical one. This much is uncontroversial. What is controversial is the cognitive relation to things-in-themselves that is possible from a practical point of view. Interpreters have tended to regard such “practical cognition” of things-in-themselves as a poor step-cousin of its theoretical counterpart — as a sort of mere “rational faith” unworthy of comparison with genuine theoretical knowledge or cognition. Recent work on these issues has begun to shift this picture.1 But this reassessment is in its early stages and remains highly controversial. My aim here is to take this tendency and push it farther than most have generally been willing to go - at least with respect to our practical self-understanding as noumenally free agents.2 Indeed, I believe that, far from representing an impoverished cousin of theoretical cognition and knowledge, our practical awareness of ourselves as free possesses the central marks of cognition and knowledge in Kant’s sense of these terms, albeit on practical grounds. Thus, it is no surprise that Kant is willing to speak of our consciousness of the moral law as enabling genuine cognition of noumenal freedom. And that he stresses that such cognition is importantly different from the rational faith he endorses with respect to the existence of 1 Karl Ameriks's work on these issues is particularly significant here. For Ameriks's views, see Chapter 11 of Ameriks(2012). The excellent Kain(2010) independently argues for broadly similar conclusions about these matters. As will become plain, I am in general agreement with much of what Kain says, but there are important differences between our interpretations, which I note below. Moreover, even someone wholly sympathetic to Kain’s understanding of these issues should (I hope) find much of use in the discussion that follows. Abaci(forthcoming) also thematizes some of these issues, although he seems to me to neglect some of the obstacles to practical cognition of freedom which I discuss below. 2 Plainly Kant takes the cognitive relation we can bear to God and our immorality, even from a practical point of view, to be weaker than what is possible with respect to our own freedom. For the most part I will focus on the case of freedom here, which is where practical cognition and knowledge comes closest to its theoretical counterpart. Elsewhere, I discuss these other cases in more detail. !1 God and the afterlife.3 In this way, the negative conclusions of the first Critique are far from Kant’s final word on our ability to achieve cognitive contact with things as they are in themselves.4 Rather, to evaluate Kant’s understanding of the limits of cognition, we need to restore his comments about practical cognition to a central place in our understanding of his views.5 In examining these questions, I’ll proceed as follows. First, I’ll discuss a variety of ways in which Kant distinguishes theoretical and practical cognition. This will help us to characterize the sense in which our cognition of noumenal freedom is practical and the sense in which it is not. Next, I’ll discuss the nature of theoretical cognition and knowledge. In doing so, I’ll focus on extracting from Kant’s account several basic constraints that explain why it is impossible for us to achieve either theoretical cognition or substantive theoretical knowledge of things-in-themselves. Then, I’ll explain why it is possible for us to overcome each of these constraints on theoretical cognition from a practical point of view, at least with respect to our nature as noumenally free agents. Finally, I’ll discuss some ways in which this sort of “practical cognition” of ourselves as things-in-themselves might nonetheless be taken to be less of a cognitive achievement than genuine theoretical cognition.6 1. Theoretical and Practical Cognition 3 Nonetheless, Kant is much more reticent about speaking of our consciousness of the moral law as providing us with Wissen of noumenal freedom. But this seems to mainly be a product of a terminological tendency on Kant’s part to restrict Wissen to the theoretical case. (This also, I think, explains why Kant is comfortable with describing nominal freedom as something that we postulate as opposed to know.) Thus, it may be wrong to speak of us as having Wissen of nominal freedom in a strict sense. But, as I discuss below, the moral law does provide us with grounds for assenting to our noumenal freedom that are objective in the senses central to Kant’s understanding of Wissen. Thus, even if we restrict the use of “Wissen” to the theoretical case, our practically grounded assent to nominal freedom is in many ways more like theoretical Wissen than it is like theoretical Glauben. For a helpful discussion of Kant's conception of belief in this context, see Pasternack(2011). I'm sympathetic to the general tenor of Pasternack's discussion, although, I differ from him in my understanding of the Fact of Reason and its relevance in this context. 4 My discussion here will focus on Kant’s views at the time of the second Critique. I leave to another time the development of Kant's views about these issues. 5 My discussion will be independent of one’s understanding of the distinction between appearances and things-in-themselves, provided one does not take this distinction to be merely epistemic in character. Personally, I am inclined to what might be called a “metaphysical one object” interpretation, on which it is correct to say (as Kant does) that the very same object can be both appearance and thingin-itself, but where this distinction nevertheless involves a genuine metaphysical difference between properties or relations or metaphysically real “aspects” of that one thing. But almost everything below is compatible with any reading of Kant that allows that there is metaphysical substance to this distinction. 6 The possibility of such cognition helps to explain Kant’s otherwise puzzling inclusion of supersensible “Ideas” under the heading of “cognitions” in the Stufenleiter passage of the first Critique. For given the argument of this paper, while such Ideas do not provide us with any new theoretical cognition, they can (at least in principle) provide us with such cognition from a practical point of view. !2 One of the difficulties in discussing these issues is the complexity of Kant’s distinction between “theoretical” and “practical”. In fact, there are at least five interrelated ways Kant distinguishes “theoretical” from “practical” cognition, all of which are relevant in some way to what follows: 1. Kant distinguishes theoretical and practical cognitions in terms of their relationship to their object. This distinction is roughly the following. Theoretical cognitions depend upon an object that is in some sense given to cognition. Now, this does not mean that this object’s form is independent from our cognitive capacities. But at least in the canonical case, it does mean that the existence of our representation of that object depends either on its existence or the existence of something appropriately related to it. In practical cognition, on the other hand, this relation of dependence is reversed — so that, in the canonical case, the existence of its object depends on our representation of it. 7 2. Closely related to this is a distinction between cognitions that represent their object as something that is (theoretical cognitions) and cognitions that represent their objects as something that ought to be (practical cognitions). 8 Kant sometimes writes as though this distinction is identical with the first, but this connection is more complicated than it might seem. 3. Kant describes these two distinctions as formal. But Kant also sometimes distinguishes theoretical and practical cognitions on the level of content. 9 In this sense, theoretical cognitions concern the world of appearances — i.e. their principles are laws of nature.10 And practical cognitions in this sense are cognitions that concern the 7 “Insofar as there is to be reason in these sciences, something in them must be cognized a priori, and this cognition can relate to its object in either of two ways, either merely determining the object and its concept (which must be given from elsewhere), or else also making the object actual. The former is theoretical, the latter practical cognition of reason.” (Bix-x) C.f. 5:10, 5:20, 5:46, 5:57, 5:171, 5:177-8, 20:197, 20:199, 20:230. 8 “Here I content myself with defining theoretical cognition as that through which I cognize what exists, and practical cognition as that through which I represent what ought to exist. According to this, theoretical use of reason is through which I cognize a priori (as necessary) that something is; but the practical use is that which it is cognized a priori what ought to happen.” (A633/B661) C.f. A802/ B830, A840/B868, 4:387, 5:5, 5:109, 5:176, 20:246-7, 9 C.f. “.. hence all of this is represented as an immediate consequence from the theory of the object in relation to the theory of our own nature (ourselves as cause): hence the practical precept here differs from a theoretical one in its form, but not in its content.” (20:196-7) 10 !3 Or, in the case of mathematical cognition, more abstract principles that underlie these laws. world of things insofar as they are free — i.e. cognitions whose principles are given by the laws of freedom. 11 4. In addition to these differences in form and content, Kant also distinguishes theoretical and practical uses of reason. While Kant conceives of reason as a unified faculty — namely, the faculty of principles or inferences — he nonetheless contrasts two uses of this faculty. These uses are distinguished by the “interest” or aim that guides the activity of reason. In this sense, the interest of theoretical reason lies in cognition of what is (from basic principles), while the interest of practical reason lies in the determination of the will (again from principles of pure practical reason). 12 5. Finally, Kant distinguishes theoretical and practical grounds for assenting to a judgment. 13 This is closely related to the fourth distinction. For we can say that a ground for assent is theoretical when it is generated by reason’s interest in cognition of what is. And a ground is practical when it is generated by reason’s interest in determining the will in accordance with principles. 14 These distinctions can be combined in a variety of ways. Thus, while Kant sometimes moves quite freely between them — and while they have various connections with each other — they need to be kept separate in discussing these issues. For example, for Kant, rational belief in God is practical in the third and fifth senses, but not in the first or second. For this belief concerns the realm of free noumena. And (for us) it is based on distinctively practical grounds. But nonetheless, the existence of God does not depend on our representation of his existence. And when we claim that God exists, we are not merely asserting something about what ought to be, we are asserting something about 11 “In a word: all practical propositions that derive that which nature can contain from the faculty of choice as a cause collectively belong to theoretical philosophy, as cognition of nature; only those propositions which give the law to freedom are specifically distinguished from the former in virtue of their content. One can say of the former that they constitute the practical part of a philosophy of nature, but the latter alone ground a special practical philosophy.” (20:197) Cf. A800/B828, 4:387, 4:420, 5:57, 5:171, 20:199, 20:246-7. 12 “The interest of its speculative use consists in the cognition of the object up to the highest a priori principles; that of its practical use consists in the determination of the will with respect to the final and complete end.” (5:120) 13 “Here there is a ground of assent that is, in comparison with speculative reason, merely subjective but that is yet objectively valid for a reason equally pure but practical; by means of the concept of freedom objective reality is given to the ideas of God and immortality and a warrant/indeed a subjective necessity (a need of pure reason) is provided to assume them, although reason is not thereby extended in theoretical cognition and, instead, all that is given is that their possibility, which was hitherto only a problem, here becomes an assertion and so the practical use of reason is connected with the elements of the theoretical.” (5:4-5) C.f. A823-4/B851-2, 5:134-5 14 “Now yet another experiment remains open to us: namely, whether pure reason is also to be found in practical use, whether in that use it leads us to the ideas that attain the highest ends of pure reason which we have just adduced, and thus whether from the point of view of its practical interest reason may not be able to guarantee that which in regard to its speculative interest it entirely refuses to us.” (A804/B832) !4 what is. Thus, in cases like this, the practical use of reason can lead to an extension of our cognition of what is, even though its interest does not lie in this extension. These relationships are even more complex in the case of our practical cognition of noumenal freedom. This cognition is practical in our third and fifth senses. For it concerns the world of free things-in-themselves. And it is based on distinctively practical grounds — namely, our consciousness of the moral law — or, more precisely, of the categorical imperative as it applies to us. But what of the first and second distinctions? On the one hand, our representation of ourselves as noumenally free is based in our consciousness of ourselves as following under the categorical imperative. And, according to Kant, the fact that we fall under this imperative is metaphysically grounded in the fact that we are noumenally free. Thus, there is a sense in which the existence of this representation is dependent upon the thing or property it represents. In this way, this representation counts as theoretical in our first sense of the term. But while this is correct, it also ignores the fact that our consciousness of ourselves as free is part of that in virtue of which we are free. For this freedom is that of a self-consciously free being. Someone who is free in this sense is someone who self-consciously determines themself in accordance with their representation of the principles of autonomy or the moral law. Given this, we should not place too much space between our consciousness of our capacity for freedom and our consciousness of ourselves as subject to the moral law. After all, as Kant repeatedly stresses, these are reciprocal representations such that one immediately grounds the other. Thus, rather than thinking of this case in terms of a representation of what is (our representation of ourselves as free) that rests on a distinct representation of what ought to be (our representation of ourselves as subject to the categorical imperative), we might be tempted to think of our consciousness of ourselves as subject to the categorical imperative and our consciousness of ourselves as free as two aspects of a single representation — a representation that simultaneously represents what is (our capacity for noumenal freedom) and what ought to be (the categorical imperative). But this would also be too quick. For we need to do justice to Kant’s claim that our consciousness of ourselves as subject to the categorical imperative is the ratio cognoscendi of our cognition of ourselves as free. And this seems to require that these are two distinct representations, one of which is the ground of the other. I’ll return to this issue in Section 3, where I offer an account of these two representations, on which they are formally distinct, but nonetheless both make us consciousness of a single underlying aspect of reality. This will have important implications for the application of the first distinction to this case. Here again, the crucial thing to remember is that our capacity for freedom is an essentially self-conscious one. I believe that this means that possession of this capacity requires some sort of consciousness of the principles governing it. 15 If so, then if our consciousness of the moral law and our consciousness of freedom are linked in the manner just sketched, it would only be in virtue of representing ourselves as subject to the moral law (and so implicitly as free) that we are in fact free. So our representation of ourselves as free would be a partial ground of our freedom — in which case such 15Less demandingly, this might be understood to imply that possession of this capacity requires the capacity to become so conscious. This would weaken the present point. But it would remain true that it is only in virtue of having the capacity to represent ourselves as free that we are in fact free. !5 representations might also count as practical in the first sense. If so, the cognition of ourselves as free that is at issue here would be akin to maker’s knowledge in at least one sense of this term. Indeed, in this case, it might be that we possess something like the sort of intellectual intuition that is otherwise limited to higher cognitive powers. 16 2. The Limitations of the Theoretical Perspective I return to these complexities below. But before doing so, we need to get clear on the meaning that Kant attaches to two terms that will be crucial in what follows: “cognition” (Erkenntnis) and “knowledge” (Wissen).17 The interpretation of either of these terms is complicated by the fact that Kant seems to use them in a variety of ways. But I believe that we can identify a core notion of each which sits at the center of Kant's discussion. Cognition in this central sense may be glossed as “objective representation with consciousness”. 18 But what does this mean? Objective representation with consciousness requires representing something to one's consciousness in a manner that makes one conscious of it as something objective — that is, as something more than a mere episode in one’s subjective experience. That is, it is the sort of representation that makes one conscious of objects — as opposed to merely providing one with a consciousness of one's own subjective state. Whenever we have such a representation, we will be in a position to recognize that it possesses — not merely a formal or internal standard of truth or correctness — but also a “material” standard of truth that holds in virtue of the manner in which it represents object(s). For Kant, this in turn requires that this representation possess two further features. First, it can only count as cognition to the degree that it represents its object to our consciousness in a determinate fashion.19 For only insofar as it does will it provide our consciousness with a non-trivial material standard of correctness. And, second, it can only count as cognition if we are conscious of its object as something that is, not merely logically, 16 Nick Stang has stressed to me the centrality of this idea to Kant’s account of a priori cognition in general. See, for example: Gott erkennet Alles a priori. Wir können nur wenige Dinge, ohne vorhergegangene sinnliche Anscheuungen, erkennen; ja es ist solches bei allen solchen Dingen für uns ganz unmöglich, wovon wir nicht selbst die Urheber sind. (28:1051) Und folglich ist die göttliche Erkenntniß aller Dinge keine andere, als die jenige Erkenntniß, die Gott von sich selbst hat als eine wirksame Kraft. (28:1054) 17 With respect to the translation of these terms, I do want to stress that I think that both represent a sort of “knowledge” in the expansive ordinary sense of the word. 18 The following discussion, and in particular the unification of the two constraints noted below in terms of their connection to “material truth”, draws heavily on Schafer(forthcoming). For related discussion, and some criticism of the details of my own view, see Watkins and Willaschek(forthcoming) and Tolley(forthcoming). 19 !6 See Ameriks(2011) and Kain(2010). but also really possible.20 For only in this case, will we be conscious of this object as generating a material, as opposed to a merely logical, standard of correctness. Thus, for Kant, a representation can only count as a cognition insofar as it makes us conscious of its object as a reasonably determinate, really possible thing.21 We can call these two constraints, respectively, the Determinate Content Constraint and the Real Possibility Constraint. We can make the first of these requirements more precise by noting two important respects in which our conscious representation of something can be more or less determinate: (i) with respect to its identity and difference with other things and (ii) with respect to the qualities it possesses. Once again, the relevance of these forms of determinacy follows from the idea that the object of a cognition must provide that representation with a standard of “material truth” and make us conscious of this standard. For an object will be capable of playing this role only if two things are true of it. First, it must be determinate which object (or class of possible objects) is relevant to the truth or falsity of judgments involving the cognition in question.22 And second, for any quality, it must be determinate whether or not the object in question possesses this quality. Thus, a cognition will only provide us with consciousness of an object in the sense relevant here insofar as it provides us with consciousness of a determinate object in both these senses. Of course, such determinacy is a matter of degree. But a representation will count as a genuine cognition only insofar as it possesses it.23 Now, it is the interaction of the Real Possibility Constraint with the Determinate Content Constraint that explains why theoretical cognition of things-in-themselves is impossible for us. For it is possible — even from a theoretical point of view — for us to form representations of things-in-themselves that satisfy one of these constraints in isolation from the other, at least to some degree. For instance, we know that appearances are really possible in virtue of their presence in intuition. And Kant believes that we know that any appearance must be grounded in some thing or things-in-themselves.24 Thus, by forming 20 See Chignell(2010) for this constraint in the context of knowledge as opposed to cognition. Note that this constraint might be strengthened to require, not just real possibility, but also actuality. Since real possibility will be established below via an appeal to actuality, this won't affect what follows. 21 The qualification “reasonably determinate” is needed because, strictly speaking, this constraint on cognition is a matter of degree. 22 “Material truth must consist in this agreement of a cognition with just that determinate object to which it is related.” (9:51) “If truth consists in the agreement of a cognition with its object, then this object must thereby be distinguished from others; for a cognition is false if it does not agree with the object to which it is related even if it contains something that could well be valid of other objects.” (A58/B83) 23 Confirmation that cognition is a matter of degree in this sense can be found in the following: “What it means that in order to cognize a thing completely one has to cognize everything possible and determine the thing through it, whether affirmatively or negatively. Thoroughgoing determination is consequently a concept that we can never exhibit in concreto in its totality..” (A573/ B601) As this passage indicates, complete cognition of something is an ideal that finite cognizers like us can only hope to approximate. 24 Often this seems to be an analytic truth about the very concept of an appearance. See A251, 4:314. !7 the definite description “the thing- or things-in-themselves that ground my present appearances”, I can form an indeterminate representation of things-in-themselves that assures me of their real possibility, even from a theoretical point of view.25 Thus, what prevents this representation from being a cognition is not the Real Possibility Constraint. Rather, it is its lack of Determinate Content. On the other hand, although here we are limited by the thinness of the unschematized categories, I could certainly form a richer conception than this of the things-in-themselves that ground appearances. For example, simply through use of the unschematized categories, I could think of all of my current appearances as grounded in a single thing-in-itself in the manner that would be true if these appearances were the phenomena of a single Leibnizian monad. But while this would give my representation of this thing-in-itself greater determinacy of content, I would thereby sacrifice any justification for asserting that such things are really possible — since I have no theoretical reason to think that such things is really, as opposed to merely logically, possible. Thus, it is the combination of these two constraints that explains Kant’s insistence that cognition of noumena is impossible from a theoretical point of view.26 In this context, intuition is especially significant because it provides us with a way of giving determinate content to a concept, while also demonstrating its real possibility. Indeed, the ability to perform these two tasks simultaneously is distinctive of a representation that is both singular and immediate in its relationship to its object. Thus, it is not just that intuition is capable of playing both these roles — it is the theoretical representation that is capable of doing so for creatures like us. A central issue in the practical domain will be whether our consciousness of the moral law is capable of playing both these roles in a similar way. But at least in the theoretical domain, a reliance on intuition appears to be the only way for us to simultaneously satisfy these two requirements. Thus, we arrive at the following claims about why theoretical cognition of things-inthemselves in general — and ourselves as noumenally free in particular — is impossible for us: Lack of Real Possibility: We lack sufficient theoretical grounds to assert that noumenal freedom is really possible. Lack of Determinate Content: What we can rationally assert of our noumenal self from a theoretical point of view is insufficient to form a determinate conception of that self as noumenally free.27 25 Of course, one might challenge the claim that we can have theoretical knowledge of this connection between appearances and things-in-themselves. But this makes it difficult to understand how Kant is in position to so much as articulate his views from a theoretical point of view. 26 27 For a fuller presentation of this argument, see again Schafer(forthcoming) This limitation is, again, the product both of the paucity of our (unschematized) concepts for thinking of things-in-themselves and the extremely restrictive limits on what we have rational grounds for asserting about things-in-themselves using these limited resources. !8 As above, we can make the second of these more precise by focusing on one important respect in which our representation of something can be determinate: Lack of Individuating Content: What we can rationally assert of our noumenal self from a theoretical point of view is insufficient to determine when our noumenal self is identical with other noumenal things and when it is not. 28 To my mind, these limitations on cognition (Erkenntnis) form the heart of Kant’s views about the scope of theoretical reason. But there is a second limitation to the theoretical point of view that is a natural corollary to this first one: namely, that we cannot achieve much theoretical knowledge (Wissen) about things-in-themselves. Unlike cognition, which primarily applies to representations, knowledge (Wissen) primarily applies to the “taking to be true” of some representation. What is distinctive of knowledge in this sense is that it is a taking to be true for which one has “objectively sufficient grounds”. Such grounds go beyond subjectively sufficient grounds insofar as they hold equally for any possible rational subject, no matter their particular subjective constitution. But Kant also claims that grounds can be objectively sufficient only insofar as they are themselves grounded in the nature of the object known. 29 For example, consider this passage, which immediately precedes the definition of knowledge: Truth, however, rests upon agreement with the object, with regard to which, consequently, the judgments of every understanding must agree (consentientia uni tertio, consentiunt inter se). The touchstone of whether taking something to be true is conviction or mere persuasion is therefore, externally, the possibility of communicating it and finding it to be valid for the reason of every human being to take it to be true; for in that case there is at least a presumption that the ground of the agreement of all judgments, regardless of the difference among the subjects, rests on the common ground, namely the object, with which they therefore agree and through which the truth of the judgment is proved. (A821/B849) Now, we must be careful when reading such passages. For while they indicate that we can have objectively sufficient grounds for assent just in case these grounds can be located in the object or objects this assent concerns, the notion of object that is at issue here is the one 28 This would be true even if (as Kant sometimes seems to say) we can tell by introspection that we are a substance and not an accident. For this is still insufficient to determine our relations to other noumenal substances. In any case, I don’t think this represents Kant’s mature view of these issues. 29 See Chignell(2007). Like Chignell, I take Kant’s discussion towards the end of the first Critique to represent his considered views— at least in broad outlines. That being said, I think that Chignell is wrong to draw the distinction between objective and subjective grounds in terms of a freestanding distinction between epistemic and non-epistemic merit. Rather, insofar as something like this is at work here, it seems to me to be captured by the fact that objectively sufficient grounds must be grounded in the object known. Thus, as we will see, it is possible to acquire objectively sufficient grounds for assent on a distinctively practical basis. In addition, pace Chignell, I am doubtful that there are merely subjective grounds rational grounds that hold for all rational subjects. For example, as I discuss elsewhere it is not strictly speaking true that the grounds Kant cites for practical faith in God and immortality hold for all possible rational subjects (as Chignell claims). !9 discussed above — namely, something that places objective material conditions on the correctness of cognition. Thus, we should not imagine that this object is wholly independent of our cognition of it. Indeed, it is essential to Kant’s “Copernican turn” that one of the most important sources of objective constraints on cognition are the necessary conditions on the possibility of cognition itself. Thus, in many cases, the ultimate source of objectively sufficient grounds will be the nature of human cognition as such, even if these grounds can also be described as rooted in the nature of the appearances being cognized. In this way, we should be careful not to read too much into Kant’s insistence that objectively sufficient grounds must be grounded in the nature of the object of cognition. Rather, this is best understood as establishing a formal connection between the nature of objects of cognition and grounds of this sort — as opposed to as providing us with a robust theory of the nature of such grounds. Nonetheless, this establishes a close relationship between cognition (Erkenntnis) and knowledge (Wissen). For given Kant’s account, taking a judgment to be true can be based on objectively sufficient grounds only if we are conscious of these grounds as themselves rooted in the nature of the objects this judgment concerns. Thus, we can have objectively sufficient grounds for assenting to a judgment only if it provide us with a consciousness of the features of its objects that ground its truth. In some cases, such as very abstract judgments about things-in-themselves, this will be possible even though these concepts do not rise to the level of genuine cognitions. But when the judgment in question goes beyond a few very abstract claims, this requires that the concepts involved in it are cognitions in the sense defined above. Thus, for the most part, judgments can only be candidates for knowledge insofar as they provide us with cognition of their objects. Given this, it follows from Kant’s insistence that theoretical cognition of things-inthemselves is impossible that we cannot have objectively sufficient theoretical grounds for all but a few very abstract claims about things-in-themselves. Thus, we arrive at our last constraint: Lack of Grounds: We lack objectively sufficient theoretical grounds to assert of ourselves that we are noumenally free. Moreover, now that we distinguished between objectively and subjectively sufficient grounds, we can return to the Real Possibility Constraint and ask whether it requires that we have objectively sufficient grounds for real possibility. Although I don't want to defend this claim in detail here, it seems plausible to me that it does, in which case we can re-write Lack of Real Possibility as follows: Lack of Real Possibility: We lack objectively sufficient theoretical grounds to assert that noumenally freedom is really possible. Of course, this interpretation makes the task to follow harder, rather than easier. So operating with this more challenging conception will not result in any loss of generality. 3. The Fact of Reason and Objectively Sufficient Grounds !10 I believe that our awareness of the moral law allows us to overcome all these barriers to theoretical cognition and knowledge, at least to a substantial degree. In particular, given this awareness, we can form a conception of ourselves as noumenally free that has sufficient determinacy of content to count as cognition. And we can establish that this is a conception of something that is really possible. Finally, this awareness provides us with objectively sufficient, if practical, grounds for the claim that we are free. Thus, from the practical point of view, we are capable of achieving the analogues of both theoretical cognition and theoretical knowledge of ourselves as noumenally free. If this is right, then in the practical philosophy, Kant gives us back much of what he has taken away in his critique of theoretical cognition — at least with respect to practical cognition of noumenal freedom.30 On the other hand, for Kant, our ability to form beliefs about God and the afterlife is much more limited, even from a practical perspective. For instance, Kant is clear that our grounds for assenting to God’s existence are at most subjectively sufficient. Similarly, our ability to form a determinate conception of God’s nature is limited in ways that can only partially be overcome by an appeal to analogical forms of “symbolic” cognition. I’ll say a bit about these others cases below. But for the most part freedom is my focus here. And the constraints on rational faith in God and immortality should not detract from the radical increase in our ability to cognize the nature of things-in-themselves that practical cognition provides. But how is this increase possible? To understand this, we need to consider three questions: 1. How does the moral law provide us with a determinate conception of ourselves as noumenally free beings? 2. How does the moral law establish the real possibility of the objects of this conception? 3. How does the moral law provide us with objectively sufficient grounds to assent to the existence of these objects? In some ways, the outlines of the answers to the second and third of these questions is relatively clear. According to Kant, the capacity for consciousness of the principles of pure practical reason is essential to the possession of practical reason — just as the possibility of becoming conscious of the principles of theoretical reason is essential to the possession of theoretical reason: We can become aware of practical laws just as we are aware of pure theoretical principles, by attending to the necessity with which reason prescribes them to us and to the setting aside of all empirical conditions to which reason directs us. (5:30, my emphasis) Thus, to gain this consciousness, we must simply reflect on the operations of our own rational faculty — and, in particular, on the manner in which it determines us independently of empirical constraints. Thus, much as anyone who is in position to judge that P is also in a position to attach the ‘I think’ to that judgment, anyone who possesses the faculty of reason is in a position to become conscious of the principles that govern its activity. 30 Of course, to say this is not to say that Kant has given everything that we might want here. For, as I discuss below, there are limits to our “insight” into the nature of noumenal freedom, even from a practical point of view. !11 It is this consciousness of the principles at work in the activity of reason that provides Kant with the famous “Fact of Reason”, which is the foundation of his discussion of these issues: Consciousness of this fundamental law may be called a fact of reason because one cannot reason it out from antecedent data of reason, for example, from consciousness of freedom (since this is not antecedently given to us) and because it instead forces itself upon us as a synthetic a priori proposition that is not based on any intuition, either pure or empirical... (5:31, my emphasis) According to Kant, “freedom and unconditional practical law reciprocally imply each other.” (5:29) In short, the Fact of Reason provides us with immediate, but nonetheless purely rational grounds for assenting to the proposition that we are noumenally free.31 For example, in the famous “gallows” thought experiment, he writes: He would perhaps not venture to assert whether he would do it or not, but he must admit without hesitation that it would be possible for him. He judges, therefore, that he can do something because he is aware that he ought to do it and cognizes freedom within which, without the moral law, would have remained unknown to him. (5:30) This expresses Kant’s conviction that the moral law is the ratio cognoscendi of freedom. But it does not characterize the transition from the one to the other as involving a multi-premise syllogistic inference.32 Indeed, as Kant suggests just prior to this example, our consciousness of the moral law, as a principle of pure practical reason, seems closer to simply being the selfconsciousness of a noumenally free being: Now I do not ask here whether they are in fact different or whether it is not much rather the case that an unconditional law is merely the self-consciousness of a pure practical reason, this being identical with the positive concept of freedom; (5:29) This would fit well with the earlier suggestion that we should regard our consciousness of ourselves as subject to the categorical imperative and our consciousness of our own freedom, not as two wholly distinct representations, but rather as two aspects of a single form of self-consciousness. But, as before, we need to be careful here. For Kant’s conception of the moral law as the ratio cognoscendi of freedom requires that there are, in some sense, two representations here, one of which grounds the other. Thus, we need to understand the immediate connection between these representations, without simply identifying them. 31 The literature on the Fact of Reason is extensive. But the primary focus of this discussion has tended to be what the Fact of Reason tells us about the foundations of Kant's mature moral theory, in general, and the relationship between the second Critique and the Groundwork, in particular. My focus will not be on these issues, but instead on the sorts of grounds the Fact of Reason provides for the existence of freedom. For some of the relevant discussion of the Fact of Reason, see Henrich(1960), Americks(1981), Allison(1990), Proops(2003), Sussman(2008), Schönecker(2013), Tenenbaum(2012), and Ware(2014). 32 To be clear, this does not mean that it would be impossible to make this transition by means of such a syllogism. !12 I believe we can best do so by thinking of our initial consciousness of ourselves as subject to the categorical imperative (of what ought to be) as involving an implicit representation of our noumenal freedom (of what is). If so, the transition from consciousness of the categorical imperative to consciousness of freedom would simply make explicit something that was implicit in our consciousness of ourselves as subject to the categorical imperative all along. In this case, each of these cognitions would contain the other implicitly — although not merely as a further concept within its content, since the difference between these representations is at least in part a matter of their form. Indeed, it might be more accurate to speak of these two cognitions as differing from one another primarily in their form. That is, we might think that there is a sense in which they both make us conscious of a single aspect of reality — one in a practical (what ought to be) guise and the other in a theoretical (what is) guise. This would fit well with Kant’s explicit comments about what it is for an agent to be subject to an ought claim. For, as indicated by the idea that freedom is the ratio essendi of the moral law, Kant claims that what it is for someone to be subject to categorical imperative is just for them to be a creature who possesses the capacity for rational self-determination, but is also subject to sensible inclination, and so not determined to act on rational grounds: A practical rule is always a product of reason because it prescribes action as a means to an effect, which is its purpose. But for a being in whom reason quite alone is not the determining ground of the will, this rule is an imperative, that is, a rule indicated by an ‘ought’ [Sollen], which … signifies [bedeutet, dass] that if reason completely determined the will the action would without fail take place in accordance with this rule.” (5:20)33 … this ‘ought’ is strictly speaking a ‘will’ that holds for every rational being under the condition that reason in him is practical without hindrance; but for beings like us … that necessity of action is called only an ‘ought,’ and the subjective necessity is distinguished from the objective (4:449) In this way, the essence of the fact that we ought to comply with the moral law is that we are both sensible and rational creatures — that we are capable of rational self-determination, but also capable of deviating from the demands of rationality.34 But it is just this that we attribute to ourselves when we regard ourselves as noumenally free, but also susceptible to sensible inclination. So our consciousness of ourselves as falling under the categorical 33 Compare A294-95/B350-51 on errors of theoretical cognition. 34 Compare Schönecker(2006) and Marshall(forthcoming). Marshall claims that this involves a reduction of the normative to the descriptive. But I think this is too quick. For one, the notion of a rational capacity that is at issue here is still a teleological one which carries with it an evaluative aspect. And, in any case, it would be better to say that our distinction between the normative and the descriptive does not apply to the domain of perfectly rational beings. Thus, it would be misleading to describe this as a reduction of one to the other. !13 imperative and our consciousness of ourselves as phenomenal beings who are nonetheless noumenally free are two ways of being conscious of a single state of affairs.35 That independence, however, is freedom in the negative sense, whereas this lawgiving of its own on the part of pure, and as such, practical reason is freedom in the positive sense. Thus, the moral law expresses nothing other than the autonomy of pure practical reason, that is, freedom … (5:33) If so, the difference between these representations would primarily be a matter of our second formal difference between theoretical and practical representations — namely, that between claims about what ought to be and claims about what is. Of course, this would involve a difference in both form and content. For what we represent in one case as something that is — namely, our imperfect capacity to act autonomously — is different from what we represent in the other as something that ought to be — namely, that we ought to act in an autonomous fashion. But at the same time, there would be a deeper sense in which both of these representations are make us conscious of a single underlying state of affairs, in virtue of which they both are valid. In short, for sensible and rational creatures like us, there would be two ways of becoming consciousness of our imperfect capacity for freedom — one theoretical in form, and the other practical — with the latter serving as the ratio ratio cognoscendi of the other. On this reading, while these representations would be distinct, they would be reciprocal in a strong sense. Thus, this deepens the sense in which the immediacy of the “Fact of Reason” is akin to that of sensible intuition. Just as sensible intuition provides us with immediate cognition of appearances, the Fact of Reason provides us with cognition of ourselves as free that does not require any further mediating element, but instead arises out of the selfconsciousness of practical reason as a power to determine the will in accordance with rational principles. Moreover, it’s plausible that part of what makes intuitions “immediate” representations of appearances is that they (qua appearances) are grounded in our intuitive capacities.36 In this way, the “immediacy” of intuition is not merely epistemological, but metaphysical as well. That is, Kant seems to explain the epistemological immediacy of intuition by appealing to the idea that there is a metaphysical sense in which intuitions are immediate.37 Thus, to return to the themes above, our cognition of the formal features of appearances seems to 35 It is important to stress that this formal difference would not exist (at least in the same way) for a purely rational being. For such beings are not subject to ought claims for Kant and would not understand the moral law in such terms. Thus, for such a being there would be no formal difference between the representation of itself as free and the representation of itself as subject to the moral law. In this way, the formal difference at issue here is in part the product of our own limitations as rational beings. For related discussion, see Stern(2012). 36 Of course, an appearance only depends on intuition with respect to its form. 37 Thus, it is best to understand Kant’s arguments in the Aesthetic as presupposing only the epistemological conception — and arguing from this to a certain metaphysical sense of immediacy. !14 involve a sort of "maker's knowledge” — insofar as these features are grounded in our representational capacities.38 Similarly, suppose, as above, that the capacity for freedom at issue here is an essentially selfconscious capacity. If so, our consciousness of the moral law must be a partial ground of the fact that we free. Thus, in creatures like us, who cognize the moral law (when applied to ourselves) as an ought claim, this cognition will ground our freedom.39 But now suppose that I am right that a consciousness of our own freedom is implicit in our consciousness of ourselves as subject to the categorical imperative. Then, an implicit consciousness of ourselves as free will be a partial ground of that very freedom.40 Much as in the case of intuition, this means that there is a sense in which our cognition of our own freedom may be regarded as something like a species of maker's knowledge, — one which is akin in some ways to an intuitive intellect’s cognition of things in themselves.41 And similarly, here to the epistemological immediacy of this consciousness can be explained in part by the fact that this consciousness has an “immediate” metaphysical connection with what it represents. Thus, there is a sense in which the Fact of Reason is immediately “given” to us — but this “givenness” should not be interpreted in terms of the rational intuition of an independent moral domain. Rather, because of the self-conscious character of practical reason, in possessing the capacity for practical reason we are conscious of our capacity to rationally determine our will in accordance with principles. Thus, the Fact of Reason involves nothing more than our consciousness that we possess a genuine faculty of practical reason — a consciousness that exhibits our capacity for free self-determination to us. But even if we do not accept all this, the “Fact of Reason” clearly provides us with immediate, but purely rational grounds for assenting to the proposition that we are noumenally free. And since actuality implies real possibility, it thereby provides us with such 38 But note that it the capacity for intuition that ground these features, not any intuition in particular. That stronger claim might be true of figures constructed in intuition, which would be maker's knowledge in a stronger sense. 39 Once again, a perfectly rational being would not cognize the moral law (as applied to themself) as an ought claim. But it would remain the case that their status as free is partially grounded by their cognition of the moral law (in whatever form that takes). 40 It is important to see that this does not conflict with Kant’s claim that freedom is the ratio essendi of the moral law. For what is being claimed here is that our noumenal freedom is partially constituted by our consciousness of the moral law, and this is wholly compatible with the idea that this law holds of us in virtue of the fact that we are free. In other words, on the picture here, part of what it is to be noumenally free is to be conscious of this law, but the fact that this law applies to one is grounded in the fact that one is so free. So the applicability of this law to us depends in our part on our representation of it to ourselves as something we ought to comply with. 41 Just how strong this parallel is will depend on the question, discussed above, of the sense in which our capacity for freedom is essentially self-conscious. I'm inclined to understand this as requiring some sort of representation of the moral law - as opposed to merely the capacity for such representations - in which case, the form of maker's knowledge that we possess in this special case would indeed be very close to that possessed by an intuitive intellect more generally. !15 grounds for the real possibility of noumenal freedom. 42 In saying this, it is important to stress that the Fact of Reason only secures our capacity to act in an autonomous fashion. For it is only this that is required for the categorical imperative to apply to us. We only act in accordance with this law insofar as we act in a fully autonomous fashion. Thus, the Fact of Reason only establishes the actuality of this capacity. It does not provide us with similar grounds for thinking that we are fully actualizing this capacity in the manner the moral law requires. It only tells us that we ought to act in this fashion, and that we have the capacity to do so. Thus, this does not conflict with Kant’s views about the pervasiveness of moral selfdeception.43 Moreover, while these grounds are of course rooted in practical reason, this need not mean that they are not “objective” in the sense defined above. Indeed, according to Kant, although merely practical, these grounds “perfectly justify” us in taking ourselves to be noumenally free: The objective reality of a pure will or, what is the same thing, of a pure practical reason is given a priori in the moral law, as it were by a fact — for so we may call a determination of the will that is unavoidable even though it does not rest upon empirical principles. In the concept of a will, however, the concept of causality is already contained, and thus in the concept of a pure will there is contained the concept of a causality with freedom, that is, a causality that is not determinable in accordance with laws of nature and hence not capable of any empirical intuition as proof of its reality, but that nevertheless perfectly justifies its objective reality a priori in the pure practical law, though not (as is easily seen) with a view to the theoretical use of reason but only to its practical use. (5:55-6, my emphasis) In the case of freedom, this is not surprising. After all, if we think of our grounds for asserting that we are free as our consciousness of the moral law, this ground will support the same first-personal proposition for every rational subject, insofar as they are considering these matters from a practical point of view.44 And, as I discuss below, the same will be true of third-personal ascriptions of freedom in general as well. Thus, such grounds possess what Kant takes to be the primary “external mark" of objectively sufficient grounds. Moreover, since our consciousness of the moral law is the self-consciousness of our capacity to be freely self-determining, these grounds are also grounded in the nature of the 42 Does this amount to a “deduction” of the objective reality of freedom on the basis of our consciousness of the moral law? Yes and no. On the one hand, we are conscious of freedom as the necessary condition of the moral law — but at the same time the connection between these two is so close that one might wonder whether it is appropriate here to really speak of a deduction of one on the basis of the other. 43 Nonetheless, in establishing the actuality (and real possibility) of this capacity, it does establish the real possibility of our actualizing this capacity, at least in general. 44 Of course, if we describe the ground here less abstractly as “one’s consciousness of oneself as subject to the categorical imperative”, it will only apply to imperfect rational beings. !16 object that is thereby known.45 Thus, the grounds we have for asserting that we are noumenally free possess both of the features Kant that takes to be distinctive of grounds that are objectively — as opposed to merely subjectively — sufficient.46 Of course, the same is not true of the grounds we have for assenting to the existence of God or our own immortality. Unfortunately, there is not space here to discuss these cases, so I will leave the differences between them to the side here.47 4. The Determinate Content of Practical Cognition of Freedom But establishing objectively sufficient grounds for assenting to the actuality of noumenal freedom is only half our story. For to achieve genuine cognition of ourselves as noumenally free, we need to locate grounds that can simultaneously perform two functions: (i) establish the real possibility of the thing being cognized and (ii) do so in a manner that gives determinate content to our conception of that thing. Again, it is the ability of intuition to perform both these functions that makes it so central to theoretical cognition. Thus, if the Fact of Reason is to play a similar role in our practical cognition of freedom, it must not only provide us with objectively sufficient grounds to regard freedom as really possible, it must also give content to this concept so as to make genuine cognition of freedom possible. In order to do so, the Fact of Reason must help us answer at least four basic questions about ourselves as noumenally free: 1. How are things-in-themselves individuated from one another? 2. Which elements of the phenomenal world are grounded in which things-inthemselves and how? 3. How are the properties or states of things-in-themselves determined by other things? 4. What are the internal principles that govern the activity of things-in-themselves? 45 This is true even though our primary consciousness of this object is not mere “object consciousness”, but rather consciousness of our own free activity of self-determination. An important question that this raises, which is the focus of some of Pasternack(2010)'s discussion, is whether these claims commit us to the claim that the Fact of Reason is itself based on objectively sufficient grounds. For instance, it might be thought that our assent to X can provide us with objectively sufficient grounds for assenting to Y only if our assent to X is itself based on objectively sufficient grounds. But I think this inference can and should be resisted in this case. For the reason why we should not say that Fact of Reason is based on objectively sufficient grounds is that it is not based on any grounds at all. But this need not prevent it from providing us with objectively sufficient grounds to assent to other propositions — provided that it satisfies the requirements above. Thus, it seems to me that the mere fact that the Fact of Reason is "groundless" should not prevent it from being a source of objectively sufficient grounds. (Compare Ware(2014).) 46 See, in particular, Kant’s discussion of these issues at the close of the third Critique (e.g. 5:468-474) where he characterizes our morally grounded recognition of our own freedom as “a matter of fact” or “scibilia” as opposed to an instance of mere “faith”. Compare Ameriks(2011). 47 The original version of this paper contained further discussion of this issue, which I explore elsewhere. !17 I believe that the Fact of Reason provides us with a basis for answering each of these questions to a greater degree than we might expect. Thus, remembering that cognition is a matter of degree, it provides us with something like genuine cognition of our nature as noumenally free things-in-themselves. Not surprisingly, the answer to the first of these questions will be largely determined by the answers to the others. So I want to postpone discussion of it — beginning instead with the second. In considering it, we should remember that, while the most fundamental role of the moral law is to guide our conduct, it also guides the attribution of moral responsibility for phenomenal actions to ourselves and others.48 Indeed, as we will see in a moment, these two roles are closely connected for Kant. Now, whenever we attribute moral responsibility for some phenomenal action to an agent, we thereby, for Kant, regard this action as (at least partially) grounded in the free choice of that agent as a noumenal being. For example, if the action is morally good it will be grounded (in part) in the agent’s acceptance of the moral law. And if it is morally bad, it will be grounded (in part) in the agent’s refusal to be guided by that law: Now even if one believes the action to be determined by these causes, one nonetheless blames the agent, and not on account of his unhappy natural temper, not on account of the circumstances influencing him, not even on account of the life he has led previously; for one presupposes that it can be entirely set aside how that life was constituted, and that the series of conditions that transpired might not have been, but rather that this deed could be regarded as entirely unconditioned in regard to the previous state, as though with that act the agent had started a series of consequences entirely from himself. This blame is grounded on the law of reason, which regards reason as a cause that, regardless of all the empirical conditions just named, could have and ought to have determined the conduct of the person to be other than it is. And indeed one regards the causality of reason not as a mere concurrence with other causes, but as complete in itself, even if sensuous incentives were not for it but were indeed entirely against it; the action is ascribed to the agent's intelligible character: now, in the moment when he lies, it is entirely his fault; hence reason, regardless of all empirical conditions of the deed, is fully free, and this deed is to be attributed entirely to its failure to act! (A555/B583, my emphasis) And: In default of this intuition, the moral law assures us of this difference between the relation of our actions as appearances to the sensible being of our subject and relation by which this sensible being is itself referred to the intelligible substratum in us. From this perspective, which is natural to our reason though inexplicable, appraisals can be justified which, though made in all conscientiousness, yet seem at first glance quite contrary to all equity. (5:99) 49 48 5:98-99. 49 Compare 5:100. !18 Thus, when we make judgments about who is morally responsible for what, we are committed to views about which aspects of the world of appearances are grounded in the activity of which noumenally free agents. For in apportioning moral responsibility with respect to phenomenal actions, I am assigning each of these actions to a noumenal agent who is its (partial) ground. But these judgments about moral responsibility can only be rooted in our consciousness of the moral law, plus the relevant empirical facts.50 Thus, our consciousness of the moral law, when combined with our empirical knowledge, must provide us with sufficient grounds for making them. For if our consciousness of the moral law did not provide us with such grounds, our ordinary practice of holding others morally responsible for their accounts would not be preserved by Kant’s account. 51 As I will discuss in a moment, this is not an accidental feature of the moral law for Kant. Rather, the ability of the moral law to (partially) ground such determinations is essential to its practical significance. Of course, this process will not assign a noumenal ground to every element of the world of appearances — nor given Kant’s complex views about our responsibility for our own character, will it necessarily give us a complete account of the manner in which appearances are grounded in the activity of any noumenal agent. 52 But it will provide us with a great deal of information about which agents ground which elements in the world of appearances. And by doing so, it provide us with a rich basis for individuating these noumenal agents from one another. But there is another, in some ways deeper, reason why the moral law must provide us with grounds for attributing phenomenal actions to noumenal agents. After all, many of the moral law’s most familiar demands upon us tacitly appeal to attributions of this sort. For example, consider the requirement that we keep our promises. In order to act upon this demand, we must be able to determine whether some phenomenal action is the action of the same noumenal agent as some promise was. For one’s actions are only constrained by a promise if that promise was one’s own. More generally, any attempt to use the Formula of Humanity to guide action requires the attribution of noumenal freedom by us to the various phenomenal “agents” we come into 50 Compare Kohl. 51 Of course, Kant famously insists that experience alone is insufficient to determine whether some action is the product of a good or evil will. (4:407) But this is mainly because any such determination must ultimately rest on our consciousness of the moral law, not because these determinations are impossible for us from a practical point of view. In addition, remember that the Fact of Reason immediately secures our capacity to act autonomously, not that we actually do so in any particular instance. Also compare Kant’s discussion of “imputation” in the Metaphysics of Morals, where he describes this as “the judgment by which someone is regarded as the author (causa libra) of an action, which is then called a deed (factum) and stands under laws.” (6:227) Note that Kant distinguishes here between imputations that have “rightful force” and those merely “appraising the deed” — but not in order to treat the latter category as illegitimate. Rather, Kant’s point here is the only the former involve the right to apportion reward and punishment in accordance with the judgment. See also “On a Supposed Right to Lie from Philanthropy” (8:426-7). 52 See, e.g., 5:97-8. See Watkins for helpful discussion of this issue. !19 contact with. Thus, from a moral point of view, we must be able to attribute distinct phenomenal actions over time to a single noumenal agent: It is not enough that we ascribe freedom to our will on whatever ground, if we do not have sufficient ground for attributing it also to all rational beings. For, since morality serves as a law for us only as rational beings, it must also hold for all rational beings; and since it must be derived solely from the property of freedom, freedom must also be proved as a property of all rational beings; and it is not enough to demonstrate it from certain supposed experiences of human nature (though this is also absolutely impossible and it can be demonstrated only a priori), but it must be proved as belonging to the activity of all beings whatever that are rational and endowed with a will. (4:447-8) In short, it is impossible to act upon moral law without some way of determining which noumenal agents ground which phenomenal actions. So once again, the moral law must provide us with some basis (clearly lacking from a purely theoretical point of view) for accomplishing this task.53 Moreover, the nature of these rational grounds must be fundamentally akin in the first- and third- or second-personal cases. For contrary to the way Kant’s moral theory is sometimes presented, it is not as if we begin with a certain sort of purely first-personal practical self-consciousness and then extend this self-consciousness so that it also includes other rational beings within its scope. Rather, as both the Formula of Humanity and the Formula of the Kingdom of Ends make clear, our consciousness of the moral law is essentially and thoroughly interpersonal in character from the start.54 In this way, in my conscious of myself as subject to the moral law, I must both conceive of myself as free and conceive of the behavior of phenomenal creatures like me as grounded in their own free noumenal activity in the same way.55 Thus, the classical problem of other minds does not really arise for Kant, at least from the practical perspective — for it is essential to the sort of practical self-consciousness we possess that we regard other phenomenal rational beings as possessing a free noumenal self in the same manner we do. Thus, it is no accident that the moral law provides us with purely rational grounds for attributing noumenal freedom to other agents. And while it may not seem as though Kant tells us very much 53 Against this, one might suggest that we can simply appeal here to our ordinary (non-moral) conception of when two phenomenal actions are attributable to the same agent here. But given Kant’s account of the limits of theoretical point of view, it is plain that we have no adequate basis from a purely theoretical point of view for making such determinations. Thus, insofar as we do draw upon our ordinary understanding of these issues in making these determinations, in doing so we must be understood to be drawing upon an element of our ordinary understanding that is partially moral in character. 54 Something made clear by the association of the category of reciprocity (“of one person to the condition of others”) with distinctively moral practical thought in the Table of the Categories of Freedom with Respect to the Concepts of the Good and Evil. (5:66) 55 But which sorts of creatures are “like me” in the relevant way? As Evan Tiffany points out, this raises a number of fundamental questions about Kant’s ethics. So all I want to stress again is that for the moral law to do the work it must for Kant, it must provide us with a basis for making these sorts of determinations (against the background of our empirical knowledge of the world). !20 about how this occurs, as a matter of fact, nearly all of his discussion of the actual content of the moral law for how to treat others can be seen as fleshing this out.56 But what is the nature of our grounds for so individuating agents? In particular, are they objectively sufficient? In the case of the phenomenal actions we attribute to ourselves these grounds do seem to have the main marks of objective sufficiency — since here, as before, our grounds for attributing some phenomenal action to our noumenal self will themselves be grounded in our own self-conscious free activity. But this question is more complicated in the case of other agents. For instance, suppose I attribute some phenomenal action P to an agent A other than myself. In doing so, I treat his noumenal self as grounding this phenomenon. And, as just discussed, I do so on grounds that hold for any rational being — namely, the moral law.57 Thus, these grounds possess the universality which is the external mark of objective sufficiency for Kant. But nonetheless one might still wonder about the other dimension of Kant’s conception of objective sufficiency. For are these grounds grounded in the nature of the object cognized in the manner is characteristic of objectively sufficient grounds? After all, my grounds for attributing P to A are grounded, not in A’s free activity and its connection with P, but rather in my own free activity. Thus, one might insist that these grounds lack the requisite connection with the object. But this would circumscribe the class of objectively sufficient grounds too narrowly. For while my grounds for attributing P to A are not grounded in the particular connection between A and P, they are grounded in the activity of something that grounds phenomenal actions in the same manner that A grounds P. So while these grounds are not based in A (and P) as particular individuals, they are based in something whose nature is the same as A and P. Thus, these grounds are grounded in the nature of A and P insofar as these natures are relevant to the connection between them. This, I submit, is sufficient to make these grounds objectively sufficient. After all, it would be too demanding to claim that our grounds for assenting to a claim about a particular X are only objective when they are based in X itself. For example, natural science can give me objectively sufficient grounds to assert of an unobserved particle that it possesses some property, even though these grounds are not based in the particular particle in question. As before, this is possible because, while these grounds are not based in that particular particle, they are based in its nature - in virtue of being based in other things with the same nature. 56 This, of course, is not to say that it is always easy to make the distinctions relevant here — only that they themselves must be made on fundamentally moral grounds. In particular, I certainly don’t mean to suggest that Kant is right (for example) to treat animals in the manner he does on this score. (See 27:459) 57 Again, Kant is not as clear as one would like about how this occurs, but his commitment to the idea that the moral law grounds such judgments of responsibility is clear enough. !21 Thus, it seems to me that there is no reason to regard our grounds for attributing phenomenal actions to other agents as less than objective.58 If this is right, then our consciousness of the moral law provides us with objectively grounded information about the relationships between noumenal agents and the world of phenomenal appearances — one that can go a considerable way towards answering the demand that we be able to individuate these agents. But the Fact of Reason provides us with something even more significant than this — namely, a partial, but non-trivial, understanding of the nature of ourselves as things-inthemselves and our relations with other things-in-themselves. On the second of these points, the significance of the moral law is clear. If morality demands of us that we regard our noumenal will as self-determining and unconditioned by anything external, then we must regard the activity of each noumenal self as independent of purely external determination by other (finite) things-in-themselves. In this way, morality does demand of us that we regard ourselves as something like Leibnizian monads, whose activity is wholly grounded in their own internal principles.59 But the Fact of Reason also provides us with a positive supplement to this negative cognition of noumenal agents. For it provides us with consciousness of the internal principle that governs our autonomous activity as noumenal agents. Thus, while we can only achieve a negative conception of the nature of noumenal freedom from a theoretical point of view, from a practical point view we achieve a positive conception of our essence as free beings: The preceding definition of freedom is negative and therefore unfruitful for insight into its essence; but there flows from it a positive concept of freedom, which is so much the richer and more fruitful. (4:446) This is an understanding — not merely of the relational properties of our noumenal selves — but also of their intrinsic nature. Thus, this is, not just cognition of our noumenal self, but also cognition of this self as it is in itself.60 Indeed, what we are able to cognize about ourselves in this case seems to represent the essence of noumenal agency (at some level of abstraction). Thus, not only does the Fact of Reason allow us to gain a great deal of individuating knowledge about things-in-themselves, it also allows us to gain a certain level of cognition of their very nature — at least insofar as they are free. In order to understand this point, it is important to distinguish two questions that Kant often discusses in quick succession. Consider, for example, the following quotation: Hence this disclosure does not help us in the least for speculative purposes, although with respect to the practical use of pure reason it does help us to extend this cognition of ours. The above three ideas of speculative reason are in themselves still not cognitions; nevertheless they are (transcendent) thoughts in which there is nothing impossible. (5:135) 58 Kant seems to assert just this when he writes that, “it is not enough for the judge merely to believe that the man accused of a crime actually has committed it. He must (juridically) know it, or he acts unconscionably.” (9:70) 59 Compare Jauernig(2008). 60 Compare Langton(2001) !22 It is easy to be puzzled by such quotes. After all, its first sentence seems to assert that our consciousness of the moral law does extend our cognition of things-in-themselves, if only from a practical point of view. But the second sentence seems to contradict this claim by denying that the ideas of reason are cognitions at all. The key to resolving this tension is to note that the second claim is restricted to these ideas considered in themselves. Thus, the point Kant is making in the second sentence is that even from a practical point of view, these ideas considered on their own remain mere thoughts as opposed to genuine cognitions. But this is compatible with the claim that they gain the status of cognitions through their relationship with the moral law. Thus, while Kant stresses that (say) our idea of freedom does not provide us with theoretical insight or cognition into our nature as a thing-in-itself, these passages are best read as involving a (usually explicit, but sometimes implicit) restriction to this idea considered either on its own or only in relation to theoretical reason.61 5. Conclusion That being said, there are important limits to our ability to grasp the nature of things-inthemselves from a practical point of view — even in the case of freedom. For while the moral law does provide us with a positive understanding of our nature as free beings, and while it also proves the real possibility of freedom in this sense, it does not provide us with an understanding of how such freedom is possible: But among all the ideas of speculative reason freedom is also the only one the possibility of which we know a priori, though without having insight into it, because it is the condition of the moral law, which we do know. (5:4) But once again, the reason for this is quite general. For as Kant remarks elsewhere, But all human insight is at an end as soon as we have arrived at basic powers or basic faculties/for there is nothing through which their possibility can be conceived, and yet it may not be invented and assumed at one’s discretion. (5:46-7) Thus, although the Fact of Reason provides us with a grasp of what it is to be a creature with a genuine faculty of practical reason, it does not provide us with an understanding of how such a faculty is really possible. But this limitation does not mark a deep difference between this case and appearances. For our understanding of how appearances are possible is also grounded (for Kant) in our grasp of our basic cognition faculties — faculties whose possibility we similarly cannot explain. Thus, while there remains something opaque about 61 A similar reading seems to me to be called for with respect to the following passage, which also seems to both assert and deny that we can cognize freedom from a practical point of view: But is our cognition really extended in this way by pure practical reason, and is what was transcendent for speculative reason immanent in practical reason? Certainly, but only for practical purposes. For we thereby cognize neither the nature of our souls, nor the intelligible world, nor the supreme being as to what they are in themselves, but have merely unified the concepts of them in the practical concept of the highest good as the object of our will, and have done so altogether a priori through pure reason but only by means of the moral law and, moreover, only in reference to it, with respect to the object it commands. But how freedom is even possible and how this kind of causality has to be represented theoretically and positively is not thereby seen; that there is such a causality is only postulated by the moral law and for the sake of it. (5:133) !23 the nature of freedom, the same opaqueness attaches to our understanding of our spontaneous theoretical capacities — and, by extension, to appearances as well. So, while there are real limits to our ability to cognize the nature of noumenal freedom, at a certain point we encounter very similar limitations in trying to understand the nature of anything at all. Once again, it is important to stress the disanalogy between the case of freedom and that of God and immortality in this respect. For while the moral law does contribute to our ability to form a determinate conception of (say) God, its ability to do so is far more limited than in the case of freedom.62 But, in the central case of our own freedom, the Fact of Reason appears to provide us with all of the main marks of both knowledge and cognition noted above. In particular, we can see now how it allows us to form a quite determinate conception of our nature as things-in-themselves, while also establishing that things with this nature are really possible. And in doing so, it provides us with objectively sufficient (if practical) grounds for affirming the actuality of freedom in this sense. Thus, unlike in the theoretical case, we have both of the following: Real Possibility: We have sufficient practical grounds to assert that noumenally freedom is really possible. Determinate Content: What we can rationally assert of our noumenal self from a practical point of view is sufficient to form a reasonably determinate conception of our noumenal self. And in particular: Individuating Content: What we can rationally assert of our noumenal self from a practical point of view is often sufficient to determine whether certain appearances are grounded in the same noumenal self or different selves. Now, this aspect of cognition is a matter of degree. So the cognition of ourselves as noumenally free, which the moral law makes possible, will not rise to the level that is possible in some cases of empirical cognition. But it will nonetheless be superior in this way to our cognition of many empirical objects. And insofar as the moral law provides us with knowledge of the fundamental principles at work in us on the noumenal level, it gives us something that we cannot fully achieve for any empirical object. For the nature of appearances are,, of course, always grounded in the nature of things that we cannot cognize via empirical means. Finally, we also have: Grounds: We possess objectively sufficient practical grounds to assert of ourselves that we are noumenally free. 62 “The ideas of God and immortality, however, are not conditions of the moral law but only conditions of the necessary object of a will determined by this law, that is, of the mere practical use of our pure reason; hence with respect to those ideas we cannot affirm that we cognize and have insight into - I do not merely say the reality but even the possibility of them.” (5:4) !24 Of course, all this is based on a practical use of the faculty of reason. For it is based in a use whose interest lies in the determination of the will. Thus, the extension of cognition that occurs here comes into existence for purposes other than itself. And this does mean that this cognition is rooted in a non-epistemic use of our faculty of reason. But this should not be taken to imply that this extension is somehow second-rate. Rather, in rationally determining our will, we also come to determine our cognition — not just of what ought to be — but also, in certain ways, of what is. Thus, a non-epistemic use of reason can nonetheless result in a genuine extension of our cognition of what is. So, for Kant, in order to get something of epistemic value out of the exercise of our rational capacities, we need not exercise these capacities with distinctively epistemic aims in mind. To some ears, the very possibility of generating epistemic goods through a non-epistemic use of reason may seem paradoxical. But given Kant’s insistence on the fundamental unity of our rational capacities, it should not be surprising to encounter this idea in his work. In any case, whatever we think of the philosophical plausibility of this idea, it is central to Kant’s discussion of these issues. And we will not do justice to it if we attempt to water-down his defense of this idea so as to make it more philosophically palatable. A further attractive feature of this view is that it enables us to see a way in which Kant is both more like Leibniz and more like Fichte than we might have thought. In doing so, it helps to bridge a basic divide between interpreters who tend to think of Kant as the culmination of trends in early modern thought and those who focus on the manner in which he anticipates the main ideas of German Idealism. But further exploration of this thought, as well as its potential metaethical implications will have to wait for another time.63 63 For a discussion of some related ideas in a contemporary context, see Tiffany(2013). !25 References All references to Kant are to the Academy addition pagination, within the exception of references to the first Critique. Abaci, Uygar. (forthcoming) “Kant’s Enigmatic Transition: Practical Cognition of the Supersensible”. Allison, Henry E. (1990). Kant’s Theory of Freedom. Cambridge: Cambridge University Press. Ameriks, Karl. (1981). “Kant’s Deduction of Freedom and Morality.” Journal of the History of Philosophy 19: 53-79. ------. (2011). Kant’s Elliptical Path. Oxford: Oxford University Press. Chignell, Andrew. (2010). “Real Repugnance and Belief about Things in Themselves.” in Kant's Moral Metaphysics, ed. Krueger/Lipscomb, De Gruyter. -------. (2007). Kant’s Concept of Justification.” Nous. Engstrom, Stephen. (2002). “Kant's Distinction Between Theoretical and Practical Knowledge,” The Harvard Review of Philosophy 10 (1):49-63. Evans, Gareth. (1982). The Varieties of Reference. Oxford: Oxford University Press. Hawthorne, John and Manley, David. (2012). The Reference Book. Oxford: Oxford University Press. Henrich, Dieter. (1960). “Der Begriff der sittlichen Einsicht und Kants Lehre vom Faktum der Vernunft.” in Die Gegenwart der Griechen im neueren Denken. Festschrift für Hans-Georg Gadamer zum 60. Geburtstag, eds. Dieter Henrich, W. Schulz, and K.-H. Volkmann-Schluck. Tübingen: Mohr Siebeck. Jauernig, Anja. (2008). “Kant’s Critique of the Leibnizian Philosophy: Contra the Leibnizians, but pro Leibniz,” in Dan Garber and Béatrice Longuenesse (eds). Kant and the Early Moderns. Princeton: Princeton University Press: 41—63. Kain, Patrick. (2010). “Practical Cognition, Intuition, and the Fact of Reason” in Kant’s Moral Metaphysics: God, Freedom and Immortality, ed. Benjamin Lipscomb and James Krueger, Berlin: DeGruyter: 211-230. Kohl, Markus. (2016) “Kant on Idealism, Freedom, and Standpoints.” Archiv für Geschichte der Philosophie 98 (1):21-54. Langton, Rae. (2001). Kantian Humility: Our Ignorance of Things-in-Themselves. Oxford: Oxford University Press. !26 Marshall, Colin (forthcoming) "Kant's derivation of the moral 'ought' from a metaphysical ‘is'." in N. Stang and K. Schafer (eds.). Sensible and Intelligible Worlds, Oxford; Oxford University Press. Pasternack, Lawrence. (2011). "The Development and Scope of Kantian Belief: The Highest Good, The Practical Postulates and The Fact of Reason," Kant-Studien 102 (3):290-315. Proops, Ian (2003) “Kant's Legal Metaphor and the Nature of a Deduction.” Journal of the History of Philosophy 41 (2), 209-229. Schafer, Karl. (forthcoming). “Kant’s Conception of Cognition and our Knowledge of Things-in-Themselves”. Schönecker, Dieter. (2006). “How is a Categorical Imperative Possible?” in Groundwork for the Metaphysics of Morals, eds. Christoph Horn and Dieter Schönecker. Berlin: De Gruyter. 40 ——. (2013). “Kant’s Moral Intuitionism: The Fact of Reason and Moral Predispositions.” Kant Studies Online 1: 1-38. Stern, Robert (2012). Understanding Moral Obligation: Kant, Hegel, and Kierkegaard. Cambridge: Cambridge University Press. Sussman, David. (2008). “From Deduction to Deed: Kant’s Grounding of the Moral Law.” Kantian Review 13: 52—81. Timmermann, Jens. (2010). “Reversal or Retreat? Kant’s Deductions of Freedom and Morality.” in Kant’s Critique of Practical Reason: A Critical Guide, eds. Jens Timmermann and Andrews Reath, Cambridge: Cambridge University Press. Tenenbaum, Sergio. (2012). “The Idea of Freedom and Moral Cognition in Groundwork III.” Philosophy and Phenomenological Research 84: 555-589. Tiffany, Evan (2013). “Choosing Freedom: Basic Desert and the Standpoint of Blame.” Philosophical Explorations. Tolley, Clinton. (forthcoming). "Kant on the Place of Cognition in the Progression of our Representations.” Watkins, Eric (2005). Kant and the Metaphysics of Causality. Cambridge: Cambridge University Press. Watkins, Eric and Willaschek, Marcus. (forthcoming) . "Kant’s Account of Cognition." Ware, Owen. (2014). "Rethinking Kant's Fact of Reason," Philosophers' Imprint 14 (32). !27