1. Introduction
There is little consensus among scholars over how to interpret Kant’s deductions of morality and freedom in his mature ethics – with one striking exception. Nearly every commentator believes that Kant rejected his argument from Groundwork III (1785) soon after its publication.Footnote 1 Aside from their disagreements, scholars have come to unite on this basic point of interpretation.Footnote 2
The topic of this paper is whether or not this interpretation is true. Despite the fact that Kant never acknowledged a change in the structure of his arguments, the standard view in the literature is that his doctrine of the ‘Fact of Reason’ from the second Critique (1788) marks a turn in his thinking. For rather than establish our freedom independently – either on the grounds that we must engage in deliberation or on the grounds that theoretical reason is self-active within us – Kant now treats our consciousness of the moral law’s authority as an inexplicable ‘fact,’ one that admits of no deduction but that secures the reality of freedom in beings ‘who cognize this law as binding upon them’ (KpV, AA 5:47).Footnote 3 In an unexpected move, Kant appears to reverse his strategy of justification, working from freedom to morality in 1785 and then from morality to freedom in 1788.Footnote 4 In what follows, I will argue that this widely accepted reading of Kant’s deductions, what we may call the Reversal Reading, is mistaken. In its place, I will defend a Continuity Reading.Footnote 5
Admittedly, there is only so much I can hope to accomplish in the following pages. What I plan to offer is a sketch of an alternative interpretation, the details of which would take us well beyond the scope of a single paper. I say this now in order to forestall worries readers might have about the completeness of my proposal. As much as I will try to cover all the key points of dispute over the relationship between Groundwork III and the second Critique, the fact remains that we lack sufficient space to satisfy every point of interest. My aim is not so much to settle the dispute once and for all, but to make the Continuity Reading a plausible (though perhaps unorthodox) candidate for understanding Kant’s project of moral justification. I shall be content, then, merely to shift the burden of proof: by the end of this paper, I hope it is clear that proponents of the Reversal Reading must defend, rather than assume, the truth of their position. Whatever the outcome, I believe the current literature stands in need of such re-examination.
I begin (in Section 2) by reviewing the early reception of Kant’s ethics in the late eighteenth and early nineteenth centuries. A curious historical point is that Kant’s first interpreters saw no fundamental difference between the proof-structure of Groundwork III and the second Critique. In fact, the Reversal Reading did not acquire uniform shape in the literature until 1960, the year Dieter Henrich and Lewis White Beck published their studies on Kant’s project of justification in ethics. With this background in place, I turn attention to the details of Groundwork III itself (in Section 3), showing that contemporary readers have found the Reversal Reading attractive because it purports to explain Kant’s worry about a ‘hidden circle.’ I then proceed to outline an alternative reading of this circle, one that is consistent with Kant’s argument from morality to freedom in the second Critique. My full defense of the Continuity Reading comes next (in Section 4), and I devote a further section to answering potential objections to it (in Section 5). Then, after explaining why Kant’s two works still differ in their presentation (in Section 6), I close by briefly considering what we can learn from Kant’s project of justification today (in Section 7).
2. Historical background
Those familiar with the current literature will no doubt be surprised to learn that Kant’s first interpreters saw no change in his mature ethical works. In the decades to follow the publication of the second Critique, commentators embraced the doctrine of the ‘Fact of Reason’ (Faktum der Vernunft) as a paradigm for philosophical inquiry, not only in the domain of morality, but also – more controversially – in the domain of knowledge. As early as Reference Reinhold1786/87, Karl Reinhold, a devoted advocate of Kant’s philosophy, introduced the concept of ‘facts of consciousness’Footnote 6 (Tatsachen des Bewußtseins) – a phrase that was quickly picked up by Johannes Gottlieb Fichte in his formulation of a ‘science of knowledge’ (Reference Fichte1794/95). Neither of these authors spoke of Kant’s ‘Faktum’ as a novel development within his system of ethics. Nor was this unusual for the time. In his book-length commentary on the second Critique (Reference Michaelis1796–97), for example, Christian Friedrich Michaelis discussed why our consciousness of the moral law is a ‘fact’ (in the sense of a Tatsache) that admits of no further proof, a point he contrasted more than once to Kant’s transcendental deduction of the categories in the first Critique.Footnote 7 In this two-volume work, amounting to over 600 pages in length, Michaelis only mentioned Kant’s Groundwork five times, and each time he did so without reference to the Fact of Reason.Footnote 8
This is not to say Kant’s arguments from Groundwork III and the second Critique won the immediate approval of his contemporaries. Many interpreters of these books were of a critical bent.Footnote 9 Yet what is interesting is that none of these early writers even suggested that Kant’s later doctrine of the Fact of Reason marked a change, let alone a reversal, in his project of justification. Looking at them now, one gets the impression that commentators were either uninterested in the question of Kant’s intellectual development from 1785 to 1788, or they were willing to accept his word that the second Critique ‘presupposes’ the Groundwork (KpV, AA 5:8). Indeed, this is how Georg Albert Mellin organized the relationship between these texts in his monumental Encyclopedic Dictionary of Critical Philosophy at the turn of the nineteenth century. After summarizing the three sections of the Groundwork, Mellin paraphrased Kant’s remark that the second Critique presupposes the Groundwork insofar as it ‘constitutes preliminary acquaintance with the principle of duty and provides and justifies a determinate formula of it’ (KpV, AA 5:8).Footnote 10 Like Reinhold, Fichte, and Michaelis before him, Mellin also spoke of the Fact of Reason as a kind of Tatsache without ever mentioning Kant’s earlier deductions. Not a single sentence in the Encyclopedic Dictionary’s six volumes hinted at anything like a rupture between Groundwork III and the second Critique.
All of this stands in sharp contrast to the Reversal Reading popular today. So we must ask: When did commentators come to believe that Kant’s project of justification underwent a turn? The answer is elusive in part because attitudes toward Kantian philosophy in the early nineteenth century were inconsistent. Not everyone was as enthusiastic as Reinhold and Fichte, for instance. Still, what we can say for certain is that by the early 1800s, 20 years after the publication of the second Critique, the positive reception of Kant’s ethics was on the wane. There was a growing cynicism among scholars of the time, not only toward the system of Kantian philosophy, but also toward the specific doctrine Kant’s early advocates cherished: the doctrine of the Fact of Reason. In his book On the Spirit of Philosophy (1803), for example, Jakob Salat complained that Kant’s appeal to a Faktum reduced the foundations of his practical philosophy to ‘mysticism’ – anticipating Hegel’s famous remark that ‘cold duty is the final undigested lump left within the stomach, the revelation given to reason.’Footnote 11 If there was a shared sentiment among scholars, it was that the doctrine of the Fact of Reason committed a gross error, although there was no consensus at the time over the exact nature of this error.
Surprisingly, the Reversal Reading did not acquire uniform shape in the literature until 1960, the year Dieter Henrich published his essay, ‘Der Begriff der sittlichen Einsicht und Kants Lehre vom Faktum der Vernunft,’Footnote 12 and Lewis White Beck published his book, A Commentary on Kant’s Critique of Practical Reason. Each author reached the same verdict on Kant’s deductions. As Beck put it, the argument of the second Critique ‘takes a truly astonishing turn’ (Reference Beck1960, 173). By having said ‘that the principle [of morality] needs no deduction, he apparently stands the argument of the Groundwork on its head. He uses the moral law, the Fact of Reason, as the prius to deduce something else, namely, freedom, which is its ratio essendi’ (Reference Henrich, Henrich, Schulz and Volkmann-Schluck1960, 172).Footnote 13 Henrich went further to explain the cause of this apparent reversal, suggesting that the argument of the Groundwork was still driven by Kant’s pre-critical project of deriving morality from theoretical reason. ‘The difference in the structure of the Groundwork and the Critique of Practical Reason can be understood,’ he wrote, ‘only if we realize that Kant did not yet see clearly in the Groundwork that a deductive justification of ethics must necessarily turn out to be unsatisfactory and contradictory.’Footnote 14 What is strange, given the reception of Kant’s ethics I have just reviewed, is that the opinion Beck and Henrich expressed has gone largely unchallenged by scholars over the last 50 years.Footnote 15
3. The argument of Groundwork III
I believe many contemporary readers have found the Reversal Reading attractive because it purports to explain Kant’s worry about a ‘hidden circle’ (geheimer Zirkel) in Groundwork III. In what follows, I will outline an alternative reading of this circle, one that is consistent with Kant’s argument from morality to freedom in the second Critique (Section 3.1).Footnote 16 I will then argue that Kant’s concept of an ‘intelligible world’ (intelligiblen Welt) provides the key to his deduction of freedom, but only if we interpret this concept as a moral ideal (Section 3.2). As we shall see, one virtue of this reading is that it makes sense of Kant’s opening remark in Groundwork III that he plans to offer a ‘deduction of the concept of freedom from pure practical reason, and with it the possibility of a categorical imperative’ (GMS, AA 4:447; my emphasis).Footnote 17
3.1. A ‘hidden circle’
Kant’s worry about a circle in Groundwork III is a bit mystifying, since he says little to clarify what the problem amounts to. At first, all he says is that ‘one must freely confess that a kind of circle appears here from which, as it seems, there is no way out’ (GMS, AA 4:450). Now if we assume that the threat of a circle emerges from using a moral premise to establish our freedom, then we can agree that our only means to escape the circle would be to find non-moral evidence for thinking we are free. On this assumption, it would be natural to hear Kant’s remarks in the second Critique as departing from his earlier approach, as when he says that the moral law, while admitting of no deduction, ‘conversely itself serves as the principle of the deduction of an inscrutable faculty,’ that is, the ‘faculty of freedom’ (KpV, AA 5:47). Yet the naturalness of this reading rests on how we characterize Kant’s worry in Groundwork III. If the threat of a circle does not come from using a moral premise to establish our freedom, then it is far from obvious that our only means of escape is to find a non-moral route.
Let us consider what Kant says more closely:
One must freely confess that a kind of circle appears here from which, as it seems, there is no way out. We take ourselves as free in the order of efficient causes in order to think of ourselves under moral laws in the order of ends, and we afterwards think of ourselves as subject to these laws because we have attributed to ourselves freedom of will. (GMS, AA 4:450)
Upon reading this, it seems Kant is saying that a circle arises if we proceed from a moral starting point – for example, if we consider ourselves under moral laws in the order of ends. Consequently, it is tempting to think that our only hope to escape the circle lies in finding a non-moral premise, one that would establish the validity of moral laws without begging the question. While the content of this premise is a topic of much controversy, the majority of scholars agree that Kant’s argument has this proof-structure. They believe we should find non-moral evidence for thinking we are free, either on the grounds that we must engage in deliberation or on the grounds that theoretical reason is self-active within us (see Section 5.2 below).
While tempting, I find this familiar characterization of Kant’s worry problematic. When we look at how he describes the circle, it is by no means clear that our only hope of escape lies in finding a non-moral route. As we read further, Kant suggests that our only hope of escape lies in finding a non-analytic or ‘synthetic’ route. Consider what he says, for instance, in the remainder of the passage quoted above:
[F]reedom and the will’s own lawgiving are both autonomy, hence reciprocal concepts, yet for this very reason one cannot be used to explain the other or to specify its ground, but can at most be used only for the logical aim of bringing different representations of the same object to a single concept (as different fractions of the same value are brought to their lowest expression). (GMS, AA 4:450; my emphasis)
The problem of a circle emerges, then, because the concepts of freedom, autonomy, and morality are all co-entailing. Analyzing freedom leads us to the concept of autonomy, and analyzing autonomy leads us to the concept of morality (GMS, AA 4:447). Yet, the movement from one to another is not a movement of justification: it is a movement of clarification. At no point have we established a ‘ground’ (Grund) for anything. All we have shown, rather, is that apparently different representations reduce to a common source. And that is true regardless of our starting point. Whether we start from freedom and end up with morality, or whether we start from morality and end up with freedom, we cannot conceptualize our way to the validity of a concept.
This begins to explain why Kant voices his worry in the first place. What commentators often overlook is that Kant is trying to motivate the final ‘transition’ (Übergang) of the Groundwork, the transition from a ‘metaphysics of morals’ (the task of Groundwork II) to a ‘critique of pure practical reason’ or what he also calls a ‘critique of the subject’ (the task of Groundwork III) (GMS, AA 4:440).Footnote 18 In Kant’s view, a metaphysics of morals is effective for investigating our commonly received notions of duty. When we consider the structure of imperatives and their relation to rational beings in general, we can specify the formula of the moral law in terms of autonomy. We would never have been able to see this connection had we – like the popular philosophers of Kant’s day – adopted a mixed method of inquiry. And yet, as we learn by the end of Groundwork II, the method of a metaphysics of morals is also limited. Since it does nothing more than analyze and clarify concepts, it is unfit to justify our everyday notions of duty (GMS, AA 4:445). What we need, Kant thinks, is a new method, what he calls a ‘critique.’ Only a critique – that is, an investigation into the bounds of our faculty of reason – will decide whether we are justified in thinking we are free.
My suggestion is that Kant is advancing a methodological point at the beginning of Groundwork III. When we employ the method of a metaphysics of morals, we are entitled to make analytic claims about freedom, autonomy, and morality since we are only attempting to explicate their conceptual structure. The moment we actually claim to be free, however, we are no longer within the sphere of analysis: we are asserting a synthetic claim of real possibility, and such a claim now demands justification. Without this justification, our claim to be free would lack proper warrant – in short, we would be begging the question. Yet, my point is that we would be begging the question either way: so long as we only employ the method of a metaphysics of morals, we have no basis to assert the reality of freedom or the validity of the categorical imperative. So Kant’s worry about a circle is his way of telling us that we need to go beyond the sphere of analysis and examine our faculty of reason instead, but it is left open at this point whether he thinks the ultimate proof of our freedom will be moral or non-moral in character.
3.2. The intelligible world: a moral ideal
After introducing the problem of a circle, Kant tells us there is only ‘one way out, namely to inquire whether we do not take a different standpoint when by means of freedom we think ourselves as causes efficient a priori than when we represent ourselves in terms of our actions as effects that we see before our eyes’ (GMS, AA 4:450). This different standpoint leads him to the concept of an intelligible world, which we arrive at by abstracting from every condition of the world of sense, including our needs, impulses, and inclinations. What is left over from this abstraction is an idea, the idea of a world governed by rational laws. But Kant is clear that the force of the idea is entirely practical: it gives us nothing to intuit or cognize.Footnote 19 Thus, in a revealing passage near the end of Groundwork III, he writes that the idea of an intelligible world
signifies only a ‘something’ that is left over when I have excluded from the determining grounds of my will everything belonging to the world of sense, merely in order to limit the principle of motives from the field of sensibility by circumscribing this field and showing that it does not include everything within itself but that there is still more beyond it; but of this something more I have no further cognizance. As for pure reason, which thinks this ideal, after its isolation from all matter, that is, cognition of objects, nothing is left for me but the form of it – namely the practical law of the universal validity of maxims – and to think of reason, conformably with this, with reference to a pure world of understanding as a possible efficient cause, that is, a cause determining the will. (GMS, AA 4:462)Footnote 20
The general direction of Kant’s argument comes out clearly here. He is saying we can think of the law of an intelligible world – as the ‘practical law of the universal validity of maxims’ – and we can think of our reason in conformity with this law – as a ‘cause determining the will.’ In doing so, we recognize our capacity to act independently of all motives coming from the world of sense. We recognize that our faculty of reason has a pure use beyond the satisfaction of our needs, impulses, and inclinations. As the cited passage also makes clear, Kant is saying that when we think of ourselves in conformity with the law of an intelligible world, we have grounds to ascribe freedom to ourselves. From the standpoint of this ideal, we can go beyond mere analysis and claim to be free, but – and this is the crucial point – the basis of our claim is of a moral character, for it rests on our consciousness of a pure practical law.Footnote 21
Support for this reading also comes from what Kant says about our common presupposition of freedom earlier in the text. The ‘legitimate claim even of common human reason to freedom of the will,’ he writes, ‘is founded on the consciousness and the granted presupposition of the independence of reason from merely subjectively determining causes’ (GMS, AA 4:457). In the very next paragraph, Kant links this consciousness of independence to the ‘law’ we give ourselves in the intelligible world:
So it is that the human being claims for himself a will which lets nothing be put to his account that belongs merely to his desires and inclinations, and on the contrary thinks as possible by means of it, indeed as necessary, actions that can be done only by disregarding all desires and sensible incitements. The causality of such actions lies in him as intelligence and in the laws of effects and actions in accordance with principles of an intelligible world, of which he knows nothing more than that in it reason alone, and indeed pure reason independent of sensibility, gives the law. (GMS, AA 4:457)Footnote 22
Our warrant for thinking we are free therefore comes not merely from our membership in an intelligible world, but from our consciousness of the law we give to ourselves in this world – the moral law. No longer limited to the world of sense and the practical perspective it affords us, then, we can at last see the real possibility of determining our will to action on the basis of pure reason alone. And since we are no longer analyzing a relationship between concepts, the threat Kant used to motivate his discussion, the threat of a ‘hidden circle,’ is now out of place.
Of course, one might still wonder: Even if we are entitled to think of ourselves as free for the reasons just spelled out, what right do we have to speak of those moral laws governing the intelligible world? Do they not stand in need of justification as well? Kant’s answer is clear, but it only becomes explicit at the end of Groundwork III. Given the nature of human reason, he explains, we can only comprehend what happens or what ought to happen on the basis of prior conditions (GMS, AA 4:463). Yet the laws of morality convey absolute necessity. So when it comes to those laws themselves, the most Kant says we can do is appreciate the fact that our philosophical investigation has come to an end. That means we cannot secure a deduction of morality proper, but we can at least see why this is no fault of our own. ‘It is,’ he writes, ‘no criticism of our deduction of the supreme principle of morality, but an accusation that would have to be brought against human reason in general, that it cannot make comprehensible, as regards its absolute necessity, an unconditional practical law (such as the categorical imperative must be)’ (GMS, AA 4:463).Footnote 23
Bringing these pieces together, we can now reconstruct Kant’s argument into the following steps:
-
Step 1When we think of ourselves as ‘causes efficient a priori,’ we presuppose our membership in a world beyond the influences of our sensible nature.
-
Step 2This ‘merely intelligible’ world gives us nothing to intuit or cognize. But it gives us an ideal for thinking of our will in conformity with pure practical laws.
-
Step 3When we think of our will in this way, we have a standpoint for cognizing our freedom: our capacity to act by reason alone. This capacity is equivalent to autonomy.
Having reached this point in his discussion, Kant is able to connect his deduction back to the formula of morality he had established earlier in Groundwork II:
-
Step 4Autonomy is equivalent to the formula of morality: ‘so act as if your maxims were to serve at the same time as a universal law (for all rational beings).’
With this outline in hand, we are ready to turn to the second Critique. But first, let us recall that Kant had also promised us a second argument, writing that he would offer a ‘deduction of the concept of freedom from pure practical reason, and with it the possibility of a categorical imperative’ (GMS, AA 4:447; my emphasis). After avoiding the threat of a circle, Kant goes on to say that the moral law takes the form of an imperative for us because it has its source within us: what we ‘would’ do were we only members of an intelligible world is what we ‘ought’ to do as beings who belong to the world of sense ‘at the same time’ (GMS, AA 4:454).Footnote 24 While unpacking this argument would take us beyond the scope of this paper, I mention it here only to make the following point salient: in Groundwork III, Kant never equates a deduction of the categorical imperative with a deduction of the moral law itself. As he makes clear, the latter is beyond the comprehension of human reason and so beyond the reach of any philosophical proof.Footnote 25
4. The argument of the second Critique
4.1. A ‘vainly sought deduction’
Those sympathetic to the Reversal Reading may now wonder how I can explain Kant’s apparent confession in the second Critique that his earlier strategy of justification was a failure. In a much-discussed passage, Kant writes that ‘the moral law cannot be proved by any deduction,’ adding:
But something different and quite paradoxical takes the place of this vainly sought deduction of the moral principle, namely that the moral principle conversely itself serves as the principle of the deduction of an inscrutable faculty which no experience could prove but which speculative reason had to assume as at least possible [ … ] namely the faculty of freedom, of which the moral law, which itself has no need of justifying grounds, proves not only the possibility but the reality in beings who cognize this law as binding upon them. (KpV, AA 5:47)
As far as confessions go, this passage seems to present us with strong evidence that Kant changed his mind by the time he composed the second Critique. For what could this ‘vainly sought deduction of the moral principle’ refer to, if not the very deduction Kant had ventured in 1785?Footnote 26
This passage acquires a different meaning, I believe, when we read it alongside the ‘concluding remark’ of Groundwork III just mentioned. In this final paragraph, Kant explains why our faculty of reason faces a dilemma (GMS, AA 4:463). On the one hand, our reason is driven to seek the condition of what happens or what ought to happen with the aim of finding something absolutely necessary, something that would end the regress of conditions. On the other hand, due to the very limitations of reason itself, we can only comprehend what happens or what ought to happen when we discover its underlying condition. So the very principle driving the activity of reason itself, namely, to find something unconditionally necessary, can never be satisfied. As Kant puts it, reason ‘restlessly seeks the unconditionally necessary and sees itself constrained to assume it without any means of making it comprehensible to itself, fortunate enough if it can discover only the concept that is compatible with this presupposition,’ that is, the concept of freedom (GMS, AA 4:463).
This is why Kant thinks we should not hope to explain the moral law’s necessity, such as a deduction would require of us, because its necessity is by definition unconditional. In other words, a deduction of the moral law would be vainly sought since we cannot ‘comprehend the practical unconditional necessity of the moral imperative’ (GMS, AA 4:463). When Kant makes this point explicit in the second Critique, he may have Groundwork III in mind, but not because he is repudiating his earlier approach. On my reading, Kant may be wanting to draw the same lesson in each text, that human insight reaches an end as soon as we are dealing with fundamental laws or fundamental faculties (KpV, AA 5:47). That is why he thinks seeking to comprehend the moral law would be futile. All we can do is comprehend its ‘incomprehensibility’ (Unbegreiflichkeit), and that is, Kant says, ‘all that can be justly required of a philosophy that strives in its principles to the boundary of human reason’ (GMS, AA 4:463). On this point at least Groundwork III and the second Critique stand in agreement: a deduction of the moral law is beyond our explanatory reach.Footnote 27
Granted, when Kant says that the moral law ‘cannot be proved by any deduction,’ it is hard to rid ourselves of the impression that he is referring to his earlier strategy of argument. Yet further study reveals that this impression is deceiving since the argument of Groundwork III refers to a deduction of the categorical imperative, not to a deduction of the moral law (GMS, AA 4:454). The insight Kant is seeking in the final section of the Groundwork is insight into how the moral law takes the form of an imperative (or ‘ought’) for finite rational beings like us. Strictly speaking, then, he is seeking a deduction of the law’s ‘necessitation’ (Nötigung), not a deduction of its ‘necessity’ (Notwendigkeit). As we have seen, Kant denies that we could ever attain the latter, given the subjective limitations of our reason (the fact that we only comprehend things on the basis of prior conditions). As a result, Kant’s remark about a ‘vainly sought deduction’ does not support the Reversal Reading since the deduction he affirms in Groundwork III and the deduction he denies in the second Critique are not one and the same.
Interestingly, the only place where Kant makes a direct reference to the Groundwork is in the Preface to the second Critique, but what he says there is seemingly positive:
I must leave it to connoisseurs of a work of this kind to estimate whether such a system of pure practical reason as is here developed from the critique of it has cost much or little trouble, especially so as not to miss the right point of view from which the whole can be correctly traced out. It presupposes, indeed, the Groundwork for the Metaphysics of Morals, but only insofar as this constitutes preliminary acquaintance with the principle of duty and provides and justifies [rechtfertigt] a determinate formula of it; otherwise, it stands on its own. (KpV, AA 5:8)
For decades after the publication of the second Critique, readers would refer to this passage to explain the relationship between Kant’s two works, as we saw with Mellin’s Encyclopedic Dictionary. Looking at this passage now, there is no trace of a rupture in Kant’s conception of the path he took from 1785 to 1788. Instead of confessing a change of mind, he underlines the continuity of his project, not only with his formulation of the moral principle (the task of Groundwork I and II), but also with his justification of it (the task of Groundwork III). What is more, the language of KpV, AA 5:8 is strongly reminiscent of what Kant says in the Preface to the Groundwork: that the text is ‘nothing more than the search for and establishment of’ the principle of morality (GMS, AA 4:392). When Kant later says that the second Critique ‘presupposes’ the Groundwork, he may not be restricting his statement to the task of clarification, for in each place, he speaks of justification (‘Rechtfertigung’ or ‘Festsetzung’).Footnote 28 In any case, there is no direct evidence here to support the Reversal Reading.Footnote 29
4.2. The continuity reading
All the pieces for my alternative account are now in place. To begin with, it is clear that Kant still accepts the first step of his 1785 deduction:
-
Step 1When we think of ourselves as ‘causes efficient a priori,’ we presuppose our membership in a world beyond the influences of our sensible nature.
One difference is that in the second Critique, Kant does not move directly to:
-
Step 2This ‘merely intelligible’ world gives us nothing to intuit or cognize. But it gives us an ideal for thinking of our will in conformity with pure practical laws.
Instead, Kant appeals to the Fact of Reason: ‘It is therefore the moral law,’ he writes, ‘that first offers itself to us and, inasmuch as reason presents it as a determining ground not to be outweighed by any sensible conditions and indeed quite independent of them, leads directly to the concept of freedom’ (KpV, AA 5:29–30; cf., 5:4n). Nevertheless, it is clear that Step 2 of Groundwork III plays a similar role as the Fact of Reason, for in both cases, Kant is appealing to our consciousness of pure practical laws to establish our freedom. So both the arguments of Groundwork III and the second Critique use a moral premise to infer:
-
Step 3When we think of our will in this way, we have a standpoint for cognizing our freedom: our capacity to act by reason alone. This capacity is equivalent to autonomy.
Nor do the parallels between Kant’s two works end at this point. Later, in the second Critique, he explains that the law of autonomy ‘is the moral law, which is therefore the fundamental law of a supersensible nature and of a pure world of the understanding’ (KpV, AA 5:42). Here, it is worth recalling that Kant made a similar move in Groundwork III when he connected the deduction of freedom back to his formula of the moral principle,Footnote 30 leading us to:
-
Step 4Autonomy is equivalent to the formula of morality: ‘so act as if your maxims were to serve at the same time as a universal law (for all rational beings).’
Finally, while the question of how a moral ‘ought’ is possible is not central to the second Critique, Kant does make a point of addressing it. In the case of human beings, he explains, ‘the law has the form of an imperative, because in them, as rational beings, one can presuppose a pure will but, insofar as they are beings affected by needs and sensible motives, not a holy will, that is, such a will as would not be capable of any maxim conflicting with the moral law’ (KpV, AA 5:32).Footnote 31 Just as we saw in Groundwork III, Kant is saying that moral imperatives are possible for us because they have their source within us: what we ‘would’ do were we only members of an intelligible world is what we ‘ought’ to do as beings who belong to the world of sense ‘at the same time’ (GMS, AA 4:454). On the whole, then, there is remarkable consistency in the layout of each argument. Kant is not venturing a deduction of the moral law itself, but in both places, he is offering a deduction of freedom, followed by an account of how the moral law takes the form of an imperative.Footnote 32
5. Objections and replies
Stepping back, there are at least three objections one could raise against the preceding sections of this paper. One could argue that I have overlooked Kant’s solution to the circle problem in Groundwork III (Section 5.1), and one could argue that I have overlooked evidence that Kant was seeking a non-moral route to freedom, both in the Groundwork (Section 5.2) and in his earlier writings (Section 5.3).
5.1. First objection: the circle problem (again)
Kant in fact gives us two characterizations of the circle problem in Groundwork III, and commentators often draw upon the second to support the view that he was seeking a non-moral argument. On this occasion, Kant writes:
The suspicion that we raised above is now removed, the suspicion that a hidden circle was contained in our inference from freedom to autonomy and from the latter to the moral law – namely that we perhaps took as a ground the idea of freedom only for the sake of the moral law, so that we could afterwards infer the latter in turn from freedom, and that we were thus unable to furnish any ground at all for the moral law but could put it forward only as a petitio principii [or ‘question-begging’ claim] disposed souls would gladly grant us, but never as a demonstrable proposition. (GMS, AA 4:453)
Here, one might worry that if my reading is correct, and Kant was appealing to a moral premise to establish our freedom of will, then he should be saying something else in this passage – that the suspicion of a circle has been confirmed. On my reading, it may seem that Kant was starting with the moral law (something ‘disposed souls would gladly grant us’) in a way that would be question-begging if directed to a skeptic (a ‘petitio principii [ … ] but never a demonstrable proposition’). When we return to the cited passage, however, Kant is clear in saying that he has established the moral law as a demonstrable proposition, and this suggests he was arguing to, rather than from, a moral claim, just as the standard reading tells us.Footnote 33
As compelling as this evidence appears, let us first consider the meaning behind Kant’s terminology more closely. To be sure, talk of a ‘demonstrable proposition’ seems to be precisely what Kant denies in the second Critique: that we can grasp the moral law’s necessity as a principle for all rational beings. And if this is what he has in mind in Groundwork III, there would be good reason to think his project of justification changes significantly in later works. Yet on reflection, it turns out that Kant’s reference at GMS, AA 4:453 is open to interpretation. For example, the proposition he says requires demonstration could refer to his formula of the moral law, rather than to the law’s supreme normative authority (which Kant later says we cannot comprehend).Footnote 34 This would be consistent with my reading because the ‘moral premise’ I have claimed Kant employs to establish our freedom of will (the pure rational law we cognize from the standpoint of an intelligible world in Step 2) is distinct from the ‘moral conclusion’ he works toward (the formula of this law in terms of a principle of autonomy in Step 4). In view of this distinction, there is no conflict between the account I favor and Kant’s second characterization of the circle problem. There are other ways to understand his talk of a ‘demonstrable proposition,’ and for the purposes of this paper, that all I need to make the Continuity Reading a plausible alternative.
5.2. Second objection: a non-moral route in the Groundwork
A second objection one could raise is that my account of Groundwork III overlooks two passages where Kant appears to be offering non-moral evidence for our freedom of will. The first concerns his remark that rational beings must act under the idea of freedom (Section 5.2.1); the second concerns his remark that theoretical reason is self-active within us (Section 5.2.2).
5.2.1. First passage: the idea of freedom
Starting with the first passage, Kant writes:
I say now: every being that cannot act otherwise than under the idea of freedom is just because of that really free in a practical respect, that is, all laws that are inseparably bound up with freedom hold for him just as if his will had been validly pronounced free also in itself and in theoretical philosophy. (GMS, AA 4:448)Footnote 35
Given its placement in the text, it may look like Kant is advancing the following argument: (i) that freedom and morality are co-entailing concepts, (ii) that we must think of ourselves under the idea of freedom, and (iii) that morality is therefore valid for us. The problem with this interpretation is that Kant does not intend his remark about acting ‘under the idea of freedom’ as a premise in his deduction. The remark occurs in subsection two where Kant is offering nothing more than ‘preparatory elucidation’ (GMS, AA 4:447). Afterwards, he explains that all he has done in subsection two is trace the ‘determinate concept of morality back to the idea of freedom,’ indicating that his discussion belongs to the analytic side of Groundwork III, that is, before the transition to a critique. Kant also remarks that ‘we could not, however, prove freedom as something actual even in ourselves or in human nature; we saw only that we must presuppose it if we want to think of a being as rational and endowed with consciousness of its causality with regard to its actions, i.e. with a will’ (GMS, AA 4:447). Kant’s claim is only that we must attribute freedom to the concept of a rational being with a will, and the question of whether we must attribute freedom to ourselves remains open.
Still, one might think we can reconstruct an argument for freedom on the basis of subsection two, regardless of how well it captures Kant’s intentions in Groundwork III. Indeed, a number of contemporary philosophers have thought there is a deeper truth to Kant’s assertion that a rational being must act ‘under the idea of freedom’ since this treats freedom as a ‘necessary condition of playing the game of deliberation’ (Hill Reference Hill and Guyer1998, 265) or as a ‘fundamental feature of the standpoint from which decisions are made’ (Korsgaard Reference Korsgaard1996b, 163). Along these lines, it seems Kant could have said:
-
(1) We must play the game of deliberation.
-
(2) In playing this game, we must think of ourselves under the idea of freedom.
-
(3) A free will and a will under moral laws are co-entailing concepts.
-
(4) Therefore, given (1)–(3), morality is valid for us.
Even as a reconstruction, it is not clear if this argument works since the sense of ‘freedom’ in premise (2) is too weak to support the inference to premise (3). To see why, suppose that I am caught between two desires, say, between working for the rest of the afternoon and going to the movies. We can agree that there is a sense of ‘freedom’ relevant to my decision-making: when I give reasons for my action, whether to work or go to the movies, I am guiding myself by concepts, and by doing so, I am presupposing my distance from the two desires I feel at the time. Accepting this, we are still not committed to saying that I must regard myself as independent from all external influences when I play the game of deliberation, for I could very well be calculating the comparative degree of pleasure I will receive from one activity verses the other. Nor are we committed to saying that I must regard myself as autonomous in this game, i.e. as legislating myself by laws of reason, for the principles I appeal to in making up my mind may be prudential and so dependent on my natural sensibility after all. In short, the status of ‘freedom’ as a practical assumption – even if a true characterization of agency in general – falls short of both the negative and positive senses of freedom in premise (3), and so falls short of supporting the conclusion in (4).
5.2.2. Second passage: the spontaneity of theoretical reason
Aside from whether the above reconstruction works, those interested in capturing Kant’s intentions in Groundwork III will no doubt press me on another passage where he appears to offer non-moral evidence for our freedom.Footnote 36 Consider what he says, for example, about the spontaneity of theoretical reason:
Now, a human being actually does find in himself a capacity by which he is distinguished from all other things, even from himself, in so far as he is affected by objects, and that is reason … [This faculty] under the name of the ideas shows a spontaneity so pure that thereby he goes far beyond anything that sensibility can ever afford him. (GMS, AA 4:452; cf., KrV A546/B574; Prol AA, 4:345)
A few sentences earlier, Kant had claimed that ‘what reaches our consciousness not by affection of the senses, but immediately’ exhibits a ‘pure activity’ (reine Thätigkeit) within us, and that such activity indicates that we do not belong wholly to the world of sense. After identifying the source of this activity in our capacity to think ideas of reason, Kant then concludes: ‘On account of this a rational being must view itself, as an intelligence (thus not from the side of its lower powers), as belonging not to the world of sense, but to that of understanding’ (GMS, AA 4:452). Here, the argument seems to be something like this:
-
(1) We are conscious of reason’s pure self-activity within ourselves.
-
(2) This activity secures our membership in the intelligible world.
-
(3) As members of this world, we are negatively free (i.e. free from all external influences).
-
(4) The principle of a negatively free will is the principle of self-legislation.
-
(5) The principle of self-legislation is equivalent to morality.
-
(6) Therefore, given (1)–(5), morality is valid for us.
Now it is unclear to me if this conclusion goes through since the spontaneity we display in thinking ideas is not yet a causality, that is, it is not yet freedom to act on pure practical reasons.Footnote 37 However, for our purposes here, I need not deny that Kant appeals to the spontaneity of theoretical reason in Groundwork III, and so I need not deny the textual evidence supporting premises (1)–(3). The question we must ask is whether these premises are meant to warrant our claim to freedom, as many commentators think, or whether they are merely intended to prepare the way for such a warrant. My own view is that the second option is more charitable and also more consistent with the range of texts we have examined above,Footnote 38 and while I lack space to explore this issue at length, it is certainly one that deserves further discussion.
5.3. Third objection: a non-moral route pre-1785
A final objection I would like to consider is that my reading lies in tension with the trajectory of Kant’s pre-1785 writings, all of which suggest he was seeking a non-moral route to freedom. In the larger context of these writings, it seems that Kant hoped to derive freedom from theoretical reason, either by appealing to the spontaneity of ‘I think’ (Section 5.3.1) or by appealing to the spontaneity of the understanding (Section 5.3.2).
5.3.1. First passage: the spontaneity of ‘I think’
The first piece of evidence is from Kant’s lectures on metaphysics during the mid-1770s:
When I say, ‘I think, I act, etc.’ either the word ‘I’ is employed improperly, or I am free [ … ] But now I am conscious to myself that I can say: ‘I do’; therefore I am conscious of no determination in me, and thus I act absolutely freely [ … ] All practical objective propositions would have no sense if human beings were not free. (V-Met/Heinze, AA 28:269)
The present argument seems to be (i) that by expressing ‘I think, I do’ we display a pure form of spontaneity, (ii) that this spontaneity proves our freedom from all external influences, and (iii) that we therefore have a secure foundation for ‘practical objective propositions,’ i.e. for moral laws. As Dieter Henrich interprets this passage in his 1960 essay, Kant is proceeding from a strictly theoretical starting point. ‘That moral obligation is something real,’ Henrich explains, ‘is certain because the consciousness of the self in thought, which alone could doubt that obligation, can be conceived only as freedom. The principle of theoretical reason gives foundation to the possibility of moral existence.’Footnote 39
Other passages from this lecture show that Henrich’s reading is less decisive than it first appears. At one point, Kant says freedom is beyond our comprehension, but that we still have grounds to attribute freedom to ourselves because we are subject to practical imperatives (V-Met/Heinze, AA 28:270). After claiming that ‘all practical objective propositions would have no sense if human beings were not free,’ he adds: ‘But now there are such imperatives according to which I should do something; therefore all practical imperatives, problematic as well as pragmatic and moral, must presuppose a freedom in me’ (V-Met/Heinze, AA 28:269). This point is all the more important when we return to the passage Henrich cites, for we can see that Kant is using a normative premise – the consciousness of our will under imperatives – to warrant our claim to ‘absolute spontaneity’ (V-Met/Heinze, AA 28:268). What makes Kant’s argument in the 1770s different from his mature position, then, is not that he appeals to the spontaneity of thinking to infer our freedom in the practical sphere, as Henrich and others suppose. What makes his argument different is that he thinks we have access to our freedom via practical imperatives in general, not via moral imperatives in particular.
5.3.2. Second passage: from thinking to willing
Another place scholars have thought Kant was seeking a non-moral route to freedom is in his ‘Review of Schulz’ from 1783 – roughly the time he was planning to write the Groundwork. Near the end of this review, Kant says that even a strict fatalist
has assumed in the depths of his soul that understanding is able to determine his judgment in accordance with objective grounds that are always valid and is not subject to the mechanism of merely subjectively determining causes [ … ] hence he always admits freedom to think, without which there is no reason. In the same way [Eben so] he must also assume freedom of the will in acting, without which there would be no morals. (RezSchulz, AA 8:14)
Scholars have read this passage as saying that the spontaneity we exhibit in the cognitive sphere (where we govern our understanding) warrants the assumption that we are free in the practical sphere (where we govern our will). However, a few sentences above the cited passage, Kant writes that the ‘most confirmed fatalist, who is a fatalist as long as he gives himself up to mere speculation, must still, as soon as he has to do with wisdom and duty, always act as if he were free, and this idea also actually produces the deed that accords with it and can alone produce it’ (RezSchulz, AA 8:13; my emphasis). Thus, Kant’s point is that a fatalist must assume he is free as soon as he becomes conscious of the demands of morality (i.e. ‘wisdom and duty’), for such demands presuppose his independence from sensible inclinations. Compared to his lectures on metaphysics, it is clear that by 1783, Kant had attained a version of his mature position – that only moral laws provide grounds for cognizing our freedom – since he must have realized that technical and prudential imperatives do not presuppose our absolute spontaneity.
If anything, the passages I have discussed above show that Kant’s commitment to finding moral evidence for freedom is much earlier than commentators have assumed. In addition to the ‘Review of Schulz’ from 1783,Footnote 40 we find Kant arguing along these lines in the 1781 edition of the first Critique:
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 (not merely hypothetically under the presupposition of other empirical ends), 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. (KrV A807/B835; my emphasis)Footnote 41
By 1781, then, we have evidence that Kant believed moral laws offer us grounds for cognizing our freedom, and what is more, we have evidence that he considered the ‘judgment of every human being’ sufficient to assume the existence of such laws.Footnote 42 Of course, it is difficult to say precisely when Kant adopted this view since the lectures on metaphysics from the mid-1770s suggest he had yet to see the special status of moral imperatives compared to technical and prudential imperatives. From the documents we have of this time, signs foreshadowing Kant’s mature view begin to appear in the late 1770s, and on inspection, it is striking how close they come to stating a thesis normally associated with the doctrine of the Fact of Reason. In Reflexion 6849, for example, Kant writes that the ‘primary ought (original = absolute or the universal idea of duty) cannot be comprehended,’ adding in Reflexion 6850: ‘The primary ought is a condition under which alone freedom becomes a capacity in accordance with constant rules that determine a priori.’ But even if we treat such literary fragments as inconclusive, the fact remains that Kant’s moral orientation shines through in the early 1780s. And we should not overlook the fact that in the Groundwork itself, Kant states clearly that reason ‘first becomes aware that it can of itself also be practical’ through ‘the pure thought of duty and in general of the moral law, mixed with no foreign addition of empirical inducements’ (GMS, AA 4:410; my emphasis). So whether we focus on the official or unofficial documents of this time, there is ample evidence to believe Kant was committed to a moral argument for freedom much earlier than 1788.
6. The real difference between Kant’s two works
If the interpretation I have defended above is correct, then we have no reason to believe Kant changed his approach to moral justification between 1785 and 1788. Instead, we have good grounds to take Kant at his word, as readers did for many decades, when he claimed that the second Critique ‘presupposes’ the Groundwork in its formulation and justification of the moral principle. As we have seen, the real difference between Kant’s two works is a matter of presentation: the second Critique says little about the formula of the moral law, and even less about how it appears to us as an ‘ought.’ In my view, we can explain this difference in terms of the two genres Kant self-consciously worked in: classical moral philosophy, on the one hand, and critical transcendental philosophy, on the other hand.
In the Preface to the second Critique, Kant tells us his goal is to show ‘that there is pure practical reason’ (KpV, AA 5:3). This differs from his stated goal in the Groundwork, namely, to explain and establish ‘the supreme principle of morality’ (GMS, AA 4:392). Now if the second Critique presupposes the Groundwork, we should not be surprised to discover abbreviated versions of the 1785 work turning up in the 1788 work. Bearing in mind Kant’s different aims, we can see why he would devote an entire section to explaining the principle of morality as a law of autonomy in Groundwork II, only to compress this discussion in the second Critique. In the latter work, Kant bypasses his complex analysis of the formulas, and proceeds directly to what he calls ‘the fundamental law’ of pure practical reason: ‘So act that the maxim of your will could always hold at the same time as a principle in a giving of universal law’ (KpV, AA 5:30). Similarly, we can also see why Kant would devote an entire subsection to establishing the principle of morality as a categorical imperative in Groundwork III, only to compress this discussion in the second Critique. In the latter work, Kant bypasses the details of his second deduction, and proceeds directly to say that morality takes the form of an imperative for us because we are, unlike holy wills, imperfectly rational (KpV, AA 5:32).
My hypothesis, then, is that the Groundwork and the second Critique diverge in their presentations because they belong to different genres of exposition. On the one hand, the Groundwork was written to prepare for a ‘future metaphysics of morals,’ a project Kant had conceived for many years but only got around to publishing in 1797. What this shows is that Kant’s Groundwork belongs to the tradition of classical moral philosophy: it aims to support a system of normative principles that could inform real-life decision-making.Footnote 43 What Kant must have realized was necessary before the publication of such a system was a clearing of foundations – in particular, a clearing away of popular moral philosophy – in order to articulate the true method of moral inquiry (cf., KrV Axx). On the other hand, the second Critique was written to complete a project Kant had already begun in 1781 with his first Critique. What this shows, by way of comparison, is that the second Critique belongs to the genre of Kant’s own critical philosophy: it aims to show that pure reason, though wholly dialectical in the speculative sphere, has a legitimate use in the practical sphere. From the perspective of genres, Kant’s brevity in the second Critique makes perfect sense. He did not feel compelled to offer a long analysis of the categorical imperative, nor a deduction of its possibility, because his aim was no longer to prepare for a future metaphysics of morals. His aim was to render the transition from pure speculative to pure practical reason more perspicuous, and to that end, his first objective was simply to show that pure practical reason exists.Footnote 44
7. Closing remarks
What I hope to have achieved in this paper is a plausible alternative to the standard story of Kant’s deductions of morality and freedom. After showing that it is relatively new in the reception of Kant’s ethics (Section 2), I suggested that contemporary readers have found this story compelling because it purports to explain Kant’s worry about a ‘hidden circle’ in Groundwork III (Section 3.1). On my reading, the idea of an ‘intelligible world’ avoids the circle, but only because our consciousness of its laws, i.e. the laws of morality, legitimates our claim to freedom (Section 3.2). Turning to the second Critique, I showed that Kant’s references to the Groundwork are positive on every score, and that his comment about a ‘vainly sought deduction’ of the moral law is consistent with the conclusion of Groundwork III (Section 4.1). With these details in place, I offered my full defense of the Continuity Reading, showing that the arguments of Groundwork III and the second Critique share the same general structure (Section 4.2). Then, after defending my reading from three potential objections (Section 5), I offered a hypothesis for why Kant’s two works still differ in their presentation (Section 6).
Questions of interpretation aside, it is difficult to say how we should evaluate Kant’s project of justification today, if not in terms of his failure to derive the moral law from non-moral premises. This impression is only reinforced by the fact that a number of contemporary Kantians have thought we can reconstruct a successful defense of the moral law from an account of action and agency in general.Footnote 45 In this light, Kant’s admission in the second Critique that no deduction of the moral law is possible, even if consistent with his earlier work, appears to fall short of what we ultimately want: a strict philosophical proof. At the same time, we must not lose sight of the fact that Kant drew attention to the limits of his argument in Groundwork III, saying that it is ‘a reproach that must be brought against human reason in general, that it cannot make comprehensible as regards its absolute necessity an unconditional practical law (such as the categorical imperative must be)’ (GMS, AA 4:463). Whether or not the reproach is just, I do not think we should assume in advance that our ideal strategy is one of finding a non-moral route to freedom. Kant believed our consciousness of the moral law is the only viable orientation a project of justification can take because it is the only orientation that works within the bounds of human reason.
At this final juncture, it may seem that we are left with a standoff of intuitions about what the task of justification in ethics should accomplish. Those who think it should refute all forms of moral skepticism will likely find my version of the Continuity Reading disappointing. If I am right, Kant was not seeking to justify the necessity of the moral law; indeed, that is precisely what he thought goes beyond our comprehension and so beyond the scope of any deduction. Those who insist that we should be seeking such a deduction will wonder what kind of ‘justification’ of morality Kant ends up with in the end, if any. To call our awareness of the moral law a ‘fact’ hardly seems satisfying, even if we follow Kant’s repeated claim that this fact is part of our common moral consciousness. However, it is worth emphasizing that Kant did not regard refuting the skeptic to be the measure of success for his moral philosophy. He was very much speaking to those of us who are committed to morality, but who lack the self-understanding necessary to make this commitment intelligible. That in itself does not settle the standoff of intuitions we have, but it does bring an important point to light. It shows that how we frame the task of justification in ethics comes down to where we locate the center of our discussion – whether in philosophical speculation or in ordinary life. In arguing from morality, Kant was, I believe, attempting to begin with the latter.
Abbreviations
In the case of the Critique of Pure Reason, I follow the standard practice of referring to the 1781 (A) and 1787 (B) editions. For all other texts, citations appear in the order of abbreviation, volume number, and page number from the Akademie Ausgabe (AA), Kants Gesammelte Schriften, edited by Königlich Preussische Akademie der Wissenschaften (29 vols. Berlin: de Gruyter, 1900–). Translation decisions are my own, though I have consulted (and sometimes followed) The Cambridge Edition of the Works of Immanuel Kant, edited by Paul Guyer and Allen Wood (Cambridge: Cambridge University Press, 1992–).
BrCorrespondence
FMWhat Progress has Metaphysics Made in Germany? (1791)
GMSGroundwork for the Metaphysics of Morals (1785)
KpVCritique of Practical Reason (1788)
KrVCritique of Pure Reason (1781/1787)
KUCritique of Judgment (1790)
MSMetaphysics of Morals (1797)
ProlProlegomena to Any Future Metaphysics (1784)
RGVReligion Within the Boundaries of Mere Reason (1793)
ReflFragments and Reflections
RezSchulz‘Review of Schulz’ (1783)
VAMS‘Preparatory Work for the Metaphysics of Morals’
V-Met/HeinzeLectures on Metaphysics: Heinze
V-MS/VigilLectures on Ethics: Vigilantius
VT‘On a Newly Raised Superior Tone in Philosophy’ (1796)
Funding
This work was supported by the Alexander von Humboldt-Stiftung.
Acknowledgements
For helpful feedback on previous incarnations of this paper, I wish to express my gratitude to Gabriele Gava, Andree Hahmann, Dai Heide, Bernd Ludwig, Sven Nyholm, Joe Saunders, Martin Sticker, Sergio Tenenbaum, Evan Tiffany, Leah Ware, Marcus Willaschek, Benjamin Yost, Ariel Zylberman, two anonymous CJP reviewers, as well as audience members at the Goethe-Universität Frankfurt and the Georg-August-Universität Göttingen. Thanks also to the Alexander von Humboldt-Stiftung for offering me a fellowship during the final stages of writing this paper.
