Marcus Willaschek’s new book is a penetrating analysis of the Transcendental Dialectic of Kant’s Critique of Pure Reason that should set the standard for further work on the subject for years to come. I found much of the book completely convincing and will not pretend to offer some devastating criticism. I do want to propose that looking at the second Critique more fully than Willaschek does in a brief epilogue, and continuing on into the third Critique, might offer some interesting variations on and/or additions to his conclusions. In particular, carrying the research on into these sources will reveal more roles for the idea of God in Kant’s reconstruction of traditional metaphysics than is suggested by Willaschek’s treatment of it in the first Critique, as just one more idea of pure reason alongside the ideas of the soul and the world-whole. And only if we fully recognize all the roles in our thought that Kant assigns to the idea of God can we also recognize the importance of his ultimate insistence, in the Opus Postumum, that the idea of God is nothing but our own idea – although I will not have room to explore that thought here.
I did find the framework of the book slightly confusing in one regard. Willaschek’s thesis is that the argument of the Transcendental Dialectic is comprised of two strands that have not always been clearly distinguished: the criticism of traditional speculative metaphysics, which ordinarily sucks up all the oxygen in discussions of the material, but also Kant’s explanation of the central ideas of metaphysics as no mere delusions of misguided philosophers but as natural products of human reason that have their proper use if properly understood. I would add, on Kant’s underlying teleological presumption, that everything in nature has its proper use if properly understood, indeed must have its proper use if properly understood. As Willaschek puts it, ‘Kant’s aim in the Transcendental Dialectic is not just to criticize traditional forms of metaphysics but equally to show that they arise naturally out of indispensable and epistemically unproblematic employments of reason and are thus inscribed into the very structure of rational thinking itself’ (pp. 9–10). The first, purely critical or destructive strand of Kant’s argument needs no special name: it is the critique of speculative metaphysics as transcending the limits of sensibility, specifically by claiming knowledge of the unconditioned when sensibility, which must confirm all our knowledge-claims, is always conditioned, with which we are all familiar. The second strand, because it needs more emphasis than it usually receives, gets a special name, namely the ‘Rational Sources Account’ (p. 3 et passim). The characterization of the second strand of Kant’s argument as ‘epistemically unproblematic’ works only if the two strands of Kant’s argument are rigorously separated, but I at least had some trouble keeping them apart: Willaschek’s overall argument is actually that our construction of the ideas of reason, which is rooted in epistemically unproblematic tendencies of human reason, leads to metaphysical illusion by the natural but unwarranted assumption of transcendental realism, that is, the assumption that the structure of our thought must reveal the structure of reality because there is no difference between appearance and reality. Here is a point where I will subsequently play one variation on Willaschek’s argument, because I will suggest that part of Kant’s argument for the existence of God in the third Critique turns on an application of one element of transcendental idealism, namely the assumption that all necessity in the world must be a product of thought, and if not of our own then of an intellect other than and greater than our own. Willaschek actually suggests this point, on the basis of one of Kant’s notes from the 1770s (p. 143) rather than on the basis of the third Critique, but does not make much of it. But it is important, I would suggest, because while the entire third Critique is written in the key of ‘reflecting judgement’, and thus every claim in it is carefully marked off from any form of traditional speculative metaphysics, this nicety was not observed by all subsequent idealists, from Hegel to Josiah Royce, who held that human thought is part of some larger thought. Kant’s explanation of the generation of this natural but misleading thought should not be omitted from his critique of speculative metaphysics, at least if we take that as prospective as well as retrospective.
But before I explain this claim, let me say why I found it difficult to keep Willaschek’s two strands apart. The purely critical strand of Kant’s argument is supposed to be represented by three steps, which Willaschek finds in what he calls the ‘transition passage’ (A307–8/B364) and sums up thus:
(1) We start from a logical principle or concept (belonging to the logical use of reason) and make a rationally necessary transition to a corresponding transcendental principle or concept (that belongs to the real use of reason). (2) That transition will appear to justify a metaphysically committed constitutive use of the transcendental principle or concept (and not just its metaphysically harmless regulative use) (3) because of the tacit variation of some variant of TR [Transcendental Realism]. (p. 151)
Properly speaking, for Kant a ‘logical’ principle concerns cognitions of some kind – representations, concepts, judgements – while a ‘transcendental’ principle concerns the objects of such concepts, or more precisely, necessary claims about such objects. But Kant puts the distinction slightly differently in the Introduction to the Transcendental Dialectic in the Critique of Pure Reason, where he characterizes the logical principle as ‘to find the unconditioned for given cognitions of the understanding’, and the transcendental principle as the assumption ‘that when the conditioned is given, then so is the whole series of conditions … also given (i.e., contained in the object or its connection)’ (A307–8/B364). This seems to make the distinction that between seeking the unconditioned and assuming it actually exists and can be known, i.e. is given. That is, the logical principle or concept to which Kant refers is a principle such as, for example, that, for anything conditioned, seek its condition, for anything caused, seek its cause; the corresponding transcendental principle is that, for anything for which we seek its cause the cause actually exists; and the claim is that we make this inference to a constitutive rather than merely regulative conclusion because we unwittingly but naturally assume that our appearances are reality itself and thus that the structure of our thought about our appearances necessarily reflects the structure of reality itself. This unwarranted assumption is of course the fatal error of traditional metaphysics.
Yet Willaschek also often presents the Rational Sources Account as comprising three steps, in a way that may make its difference from the traditional critical interpretation of the Dialectic difficult to discern:
(1) Rational reflection on empirical questions necessarily raises metaphysical questions about ‘the unconditioned’.
(2) Rational reflection (by ‘pure reason’) on these metaphysical questions necessarily leads to metaphysical answers that appear to be rationally warranted.
(3) The rational principles that lead from empirical to metaphysical questions and from there to metaphysical answers are principles of ‘universal human reason’; that is, they belong to rational thinking as such. (p. 5, cf. p. 157)
This looks awfully similar to the critical strand.
However, in Part II of the book, Willaschek fills out the skeleton of the Rational Sources Account in detail by showing how the Paralogisms, Antinomies and Ideal of Pure Reason turn on the ideas of the soul, the world and God as unconditioned. It seems to me that it is this level of detail that comprises the real difference between the merely critical strand and the Rational Sources Account in the Dialectic, thus the latter but not the former brings out that pure reason’s ideas of the unconditioned are natural products of reason’s discursive, iterative and systematic character (chapters 1–2). The difference between the critical strand and the Rational
Sources Account thus does not lie at a ‘fourth level’, as Willaschek suggests, at which
Kant argues that even though the transcendental principles and ideas derived at the previous level have a legitimate “regulative” use in guiding empirical scientific research and our search for unity in the diversity of natural phenomena, they are easily mistaken for being constitutive, that is, for being true representations of objects. (pp. 7–8)
For this is much the same as the second level of the merely critical account. The difference is rather that the Rational Sources Account inserts its additional level of detail before it reaches this point. Thus Willaschek’s clearest overview of the Rational Sources Account runs thus:
On the first, most general level, there is the transition from the ‘Logical Maxim,’ which requires us to find a condition for each conditioned cognition, to the ‘supreme principle of pure reason’ … according to which if something conditioned is given, then so is the complete series of conditions, where this series is itself unconditioned. On Kant’s view, it is this latter principle that drives human reason to metaphysical speculation. …
On the second level, Kant derives the system of transcendental ideas … concepts of objects that, if they do exist, are unconditioned (such as the soul, the world, and God). … This system consists of the concept of the unconditioned …, the three classes of transcendental ideas (psychological, cosmological, and theological …), and nine modes (ways in which objects can be thought to be unconditioned): substantiality, simplicity, personality, and spirituality in the case of the soul; the absolute completeness of composition, division, origin and mutual dependence in the case of the world; and finally the idea of an ens realissimum in the case of God. On this second level, Kant only derives the systematic order of transcendental ideas, not these ideas themselves.
On the third level, there are the specific ‘dialectical’ (that is, fallacious) inferences that purport to provide us with a priori knowledge about the soul, the world, and God …
Finally, on the fourth level, Kant argues that even though the transcendental principles and ideas have a legitimate ‘regulative’ use in guiding empirical scientific research and our search for unity in the diversity of natural phenomena, they are easily mistaken for being constitutive, that is, for being true representations of objects. (pp. 7–8)
The first and second strands of Kant’s argument differ chiefly in that, while both insist that the principles of reason should be understood regulatively rather than constitutively, only the second fills in the details about the actual derivation of the ideas of unconditioned soul, world and God, but only the first explicitly invokes transcendental realism to explain why we mistake regulative for constitutive principles. In a nutshell, then, the Rational Sources Account is Kant’s derivation of the regulative principles of reason from its most basic operations, while the critical core of the Dialectic is its explanation that we transform the legitimate regulative principles of reason into illegitimate constitutive principles because of our natural but unwarranted attraction to transcendental realism. I think that the overall argument of Willaschek’s book might have been clearer if he had put it thus rather than speaking of two separate strands.
That organizational point is independent of the substantive suggestions to which I now turn. First, I suggest that the idea of God has a broader role in both the theoretical and practical uses of reason than Willaschek proposes. On the theoretical side, Willaschek focuses on Kant’s analyses of arguments for the existence of God in the fourth Antinomy of Pure Reason and in chapter three, section two, of the Ideal of Pure Reason. The first could be considered a variation of the cosmological argument, which Kant will consider in its more ordinary form in his critique of the three traditional arguments for the existence of God in chapter three, section four, of the Ideal of Pure Reason: whereas the traditional ontological argument infers the existence of a necessary being from the existence of any contingent being, the thesis of the fourth Antinomy purports to be able to infer the existence of a necessary being outside of the completed series of all contingent beings as the unconditioned condition for the totality of conditioned beings. In the Antinomies, this argument for the existence of God stands alongside of but is independent of the theses of the other antinomies, which argue for the existence of a world that is complete and finite in both its spatial and temporal extension and division and for the existence of a spontaneous will that is the unconditioned condition of conditioned events (and by implication, not just one such will, but many, thus not just for God but for us); all of these arguments are of course opposed by their antitheses, and as is well known Kant resolves these two disputes in two different ways, arguing that the dispute whether the world is finite or infinite in spatial and temporal extension and division is rendered moot by transcendental idealism, the recognition that everything in space and time is only indefinitely extendable appearance, neither finite nor genuinely infinite, while in the case of free will and God, those ideas may be thought although not known to have real objects in the realm of things in themselves but not in appearance, where they are precluded by the conditions of sensibility. Thus transcendental realism is blocked with regard to the first two antinomies, but allowed as an indemonstrable possibility for the theses of the last two. The argument that Kant criticizes in chapter three, section two, may be considered a variant or refinement of the ‘only possible basis for a demonstration of the existence of God’ that Kant had introduced in his 1763 work of that name: whereas in the earlier work Kant had simply argued that all possibility requires a ground in a necessary being, in the first Critique he presents a more elaborate account of possibility, arguing that the complete determination of the concept of any particular requires a ‘sum total of all possibility’ (A573/B601) from which the predicates of the particular being can be chosen, and that this sum total of all possibility must in turn be grounded in the necessary existence of a being possessing all ‘positive’ predicates (whatever exactly they might be). This applies the idea of the unconditioned twice, in the form of the idea of the sum total of all possibility and in the idea of the ground of the latter in the existence of a being possessing all positive predicates; and transcendental realism leads to the metaphysical assertion of the existence of God by means of the assumption that these ideal conditions for the complete determination of concepts must actually exist.
Willaschek presents these matters in much more detail than I have here, and I have no argument with his presentation. What I do want to question is only the suggestion, at least in the Antinomies, that the argument for the existence of God just stands alongside the other arguments for metaphysical conclusions, and further the suggestion that the metaphysical tendency to assert the actual existence of God is always impelled by transcendental realism. As to the first point, I think more emphasis might be needed on Kant’s claim in the third section of the first book of the Transcendental Dialectic, on the ‘system of the transcendental ideas’, that there are ‘three classes’ of transcendental ideas, ‘of which the first contains the absolute (unconditioned) unity of the thinking subject, the second the absolute unity of the series of conditions of appearance, the third the absolute unity of the condition of all objects of thought in general’ (A334/B391). In other words, however the ideas of the unconditioned subject and world-whole might be reached, the idea of God is inevitably reached as the unconditioned condition of those, which in that regard turn out not to be entirely unconditioned after all. The idea of the absolute subject is unconditioned with regard to its own predicates, and that of the world-whole is unconditioned in relation to its constituent parts, but both of these are themselves also conditioned in the sense of depending upon God. The idea of God is thus not just the idea of one more unconditioned alongside others, but in Kant’s mind all roads lead to the idea of God. Then transcendental realism can kick in and lead from the unconditioned idea of God to the assertion of the existence of God, although of course we should also step back from this natural tendency and recognize that the idea of God can serve only as a regulative idea.
But the other point that I want to make in this regard is that if we carry the story forward to the third Critique, we see that Kant thinks that belief in the existence of God, which has to be recast as a regulative idea, also arises from a natural misinterpretation of transcendental idealism. What I have in mind is Kant’s argument in section four of the published Introduction to this work. Here Kant argues that particular laws of nature, that is, laws beneath the level of and more determinate than the ‘universal transcendental laws’ of nature established in the System of Principles of the first Critique, most of all the principles of the Analogies of Experience, ‘as empirical, may indeed be contingent in accordance with the insight of our understanding’, but still, ‘if they are to be called laws (as is also required by the concept of a nature), must be regarded as necessary on a principle of the unity of the manifold’ (5: 180), that is, some principle of the unity of a manifold; and, since all necessity must be considered to be the product of some understanding, that is, of the synthesis of a manifold in accordance with the necessary principles of some understanding, then
This principle can be nothing other than this: that since universal laws of nature have their ground in our understanding, which prescribes them to nature (although only in accordance with the universal concept of it as nature), the particular empirical laws, in regard to that which is left undetermined in them by the former, must be considered in terms of the sort of unity they would have if an understanding (even if not ours) had likewise given them for the sake of our faculty of cognition, in order to make possible a system of experience in accordance with particular laws of nature. (5: 180)
Here Kant is arguing that we naturally arrive at a conception of God by an extension of the fundamental insight of transcendental idealism with regard to the most general laws of nature, that they are the product of our own understanding, to all laws of nature, which, as laws, must be necessary and, as necessary, must be the product of an understanding, although one that is not ours and is obviously greater in capacity than ours. Perhaps transcendental realism must then also kick in to take the unsuspecting subject to the actual assertion of the existence of such a divine law-giver, but if so it can kick in only after this natural extension of transcendental idealism has taken place. And then, of course, the subject must be warned to make only regulative use of the idea of this understanding.
One way to bring out the difference between this argument and Kant’s argument in the thesis of the fourth Antinomy is that, while the latter infers the existence of God as the unconditioned condition of the series of all contingent beings, the present argument infers the existence of God from the imagined complete system of all necessary laws of nature, which is to say all laws of nature. Because of this difference, I think it is worth noting this addition to the analysis in the first Critique, to which Willaschek confines himself.
I now turn to Willaschek’s brief Epilogue on Kant’s use of the idea of the unconditioned in practical philosophy. Willaschek rightly says that ‘an adequate discussion of this aspect of Kant’s philosophy would require a book of its own’ and proposes to offer only ‘a glimpse of Kant’s “practical metaphysics”’ (p. 270). So I do not want to tax him with the incompleteness of his discussion. But I do want to make two points. First, I want to dispute his characterization of the epistemic status of the postulates of pure practical reason. Willaschek says that ‘speculative reason, considered on its own, requires us to reject (i.e. not to accept) these postulates because they are theoretically undecidable. But we cannot both accept and not accept the postulates, so either speculative or practical reason must give’, and then argues that because of the categorical character of the demands of practical reason, speculative reason must give way to practical reason and its postulates (p. 273). Surely Kant’s position is rather that undecidability and rejection are not the same thing, and that it is only because speculative reason must leave the postulates undecided, for example, it can neither prove nor disprove the existence of God, that practical reason is free to step into the vacuum left by theoretical reason. Theoretical reason does not have to give way to practical reason; in Kant’s view, his doctrine of the primacy of practical reason is entirely consistent with his view of the unity of reason in general.
My second remark concerns the highest good. Willaschek’s lapidary statements that the ‘highest good is “the unconditioned totality of the object of pure practical reason,” that is, the ultimate and inclusive end of rational agency’ and that as such ‘it must be considered to be realizable through our own actions, since otherwise we could not be obligated to bring it about’ (p. 271) seem entirely correct. And that is to say that the demand of morality is not unconditional only in the sense of being absolutely binding no matter what anyone’s desires might be, but also that it is unconditional in the further sense of demanding completion of its application: in words from the first Critique, morality demands that we complete the transformation of the natural world into a moral world, in which ‘a system of happiness proportionately combined with morality can also be thought as necessary, since freedom, partly moved and partly restricted by moral laws, would itself be the cause of the general happiness’ (A809/B837). But Willaschek then confines himself to the further argument suggested in the first Critique that, since virtuous individuals cannot in fact count on the moral cooperation of all others in the transformation of the natural world into a moral world, those of us who do our moral best must defer our hopes of happiness proportionate to our own virtue to ‘a world that is future for us’ in which God does reward the virtuous with happiness (A811/B839). I believe that Kant was already moving away from this lame, consolatory approach to the highest good in the treatment of the postulate of the existence of God in the second Critique and then even more clearly in the treatment of the highest good in the third Critique, for in those places he suggests that we must postulate the existence of God as the author of a nature in which human efforts to be moral can after all be successful and can produce happiness in nature (5: 125, 450), as the natural consequence of establishing a realm of ends in which not only is each person treated as an end in herself but also their particular ends are realized to the greatest compossible extent (4: 433). That is, the consoling thought is not that virtuous individuals, but only those, will be rewarded with happiness in an afterlife by the fiat of God, but rather that the human species as a whole is capable of morality, which when complete will naturally be accompanied with happiness even though happiness – that of any individual agent or that of all human beings – is not its fundamental motivation (see also 8: 279). This is further to say that the solution to the paradox of the highest good is not to say that the necessarily inadequate efforts of human beings towards the highest good will be supplemented in an afterlife, but that humans must be considered capable of bringing about the highest good by their own efforts after all because they are part of a nature authored by God with that possibility in mind. Of course, this thought too involves a probable deferral beyond the natural life of many individuals, but not a deferral of the happiness of the virtuous to an afterlife, rather a deferral of the full realization of both complete virtue and complete happiness to some indeterminate later date in the natural history of the human species. And whether Kant puts this point in the language of postulation as in the second Critique or of reflecting judgement and regulative principles as in the third, it offers us only a possibility that can suffice to make our efforts toward the realization of the highest good rational rather than a proof that could satisfy the most dogmatic metaphysician that this happy state must come about. Nevertheless, it seems a more attractive conception of the highest good than the one to which Willaschek confines himself on the basis of his study of the Critique of Pure Reason.
This is a further thought about Kant’s use of the idea of the unconditioned to which Willaschek’s work pushes me. It is not meant to detract from my almost unconditional admiration for what he has accomplished in his outstanding book.
Marcus Willaschek’s new book is a penetrating analysis of the Transcendental Dialectic of Kant’s Critique of Pure Reason that should set the standard for further work on the subject for years to come. I found much of the book completely convincing and will not pretend to offer some devastating criticism. I do want to propose that looking at the second Critique more fully than Willaschek does in a brief epilogue, and continuing on into the third Critique, might offer some interesting variations on and/or additions to his conclusions. In particular, carrying the research on into these sources will reveal more roles for the idea of God in Kant’s reconstruction of traditional metaphysics than is suggested by Willaschek’s treatment of it in the first Critique, as just one more idea of pure reason alongside the ideas of the soul and the world-whole. And only if we fully recognize all the roles in our thought that Kant assigns to the idea of God can we also recognize the importance of his ultimate insistence, in the Opus Postumum, that the idea of God is nothing but our own idea – although I will not have room to explore that thought here.
I did find the framework of the book slightly confusing in one regard. Willaschek’s thesis is that the argument of the Transcendental Dialectic is comprised of two strands that have not always been clearly distinguished: the criticism of traditional speculative metaphysics, which ordinarily sucks up all the oxygen in discussions of the material, but also Kant’s explanation of the central ideas of metaphysics as no mere delusions of misguided philosophers but as natural products of human reason that have their proper use if properly understood. I would add, on Kant’s underlying teleological presumption, that everything in nature has its proper use if properly understood, indeed must have its proper use if properly understood. As Willaschek puts it, ‘Kant’s aim in the Transcendental Dialectic is not just to criticize traditional forms of metaphysics but equally to show that they arise naturally out of indispensable and epistemically unproblematic employments of reason and are thus inscribed into the very structure of rational thinking itself’ (pp. 9–10). The first, purely critical or destructive strand of Kant’s argument needs no special name: it is the critique of speculative metaphysics as transcending the limits of sensibility, specifically by claiming knowledge of the unconditioned when sensibility, which must confirm all our knowledge-claims, is always conditioned, with which we are all familiar. The second strand, because it needs more emphasis than it usually receives, gets a special name, namely the ‘Rational Sources Account’ (p. 3 et passim). The characterization of the second strand of Kant’s argument as ‘epistemically unproblematic’ works only if the two strands of Kant’s argument are rigorously separated, but I at least had some trouble keeping them apart: Willaschek’s overall argument is actually that our construction of the ideas of reason, which is rooted in epistemically unproblematic tendencies of human reason, leads to metaphysical illusion by the natural but unwarranted assumption of transcendental realism, that is, the assumption that the structure of our thought must reveal the structure of reality because there is no difference between appearance and reality. Here is a point where I will subsequently play one variation on Willaschek’s argument, because I will suggest that part of Kant’s argument for the existence of God in the third Critique turns on an application of one element of transcendental idealism, namely the assumption that all necessity in the world must be a product of thought, and if not of our own then of an intellect other than and greater than our own. Willaschek actually suggests this point, on the basis of one of Kant’s notes from the 1770s (p. 143) rather than on the basis of the third Critique, but does not make much of it. But it is important, I would suggest, because while the entire third Critique is written in the key of ‘reflecting judgement’, and thus every claim in it is carefully marked off from any form of traditional speculative metaphysics, this nicety was not observed by all subsequent idealists, from Hegel to Josiah Royce, who held that human thought is part of some larger thought. Kant’s explanation of the generation of this natural but misleading thought should not be omitted from his critique of speculative metaphysics, at least if we take that as prospective as well as retrospective.
But before I explain this claim, let me say why I found it difficult to keep Willaschek’s two strands apart. The purely critical strand of Kant’s argument is supposed to be represented by three steps, which Willaschek finds in what he calls the ‘transition passage’ (A307–8/B364) and sums up thus:
(1) We start from a logical principle or concept (belonging to the logical use of reason) and make a rationally necessary transition to a corresponding transcendental principle or concept (that belongs to the real use of reason). (2) That transition will appear to justify a metaphysically committed constitutive use of the transcendental principle or concept (and not just its metaphysically harmless regulative use) (3) because of the tacit variation of some variant of TR [Transcendental Realism]. (p. 151)
Properly speaking, for Kant a ‘logical’ principle concerns cognitions of some kind – representations, concepts, judgements – while a ‘transcendental’ principle concerns the objects of such concepts, or more precisely, necessary claims about such objects. But Kant puts the distinction slightly differently in the Introduction to the Transcendental Dialectic in the Critique of Pure Reason, where he characterizes the logical principle as ‘to find the unconditioned for given cognitions of the understanding’, and the transcendental principle as the assumption ‘that when the conditioned is given, then so is the whole series of conditions … also given (i.e., contained in the object or its connection)’ (A307–8/B364). This seems to make the distinction that between seeking the unconditioned and assuming it actually exists and can be known, i.e. is given. That is, the logical principle or concept to which Kant refers is a principle such as, for example, that, for anything conditioned, seek its condition, for anything caused, seek its cause; the corresponding transcendental principle is that, for anything for which we seek its cause the cause actually exists; and the claim is that we make this inference to a constitutive rather than merely regulative conclusion because we unwittingly but naturally assume that our appearances are reality itself and thus that the structure of our thought about our appearances necessarily reflects the structure of reality itself. This unwarranted assumption is of course the fatal error of traditional metaphysics.
Yet Willaschek also often presents the Rational Sources Account as comprising three steps, in a way that may make its difference from the traditional critical interpretation of the Dialectic difficult to discern:
(1) Rational reflection on empirical questions necessarily raises metaphysical questions about ‘the unconditioned’.
(2) Rational reflection (by ‘pure reason’) on these metaphysical questions necessarily leads to metaphysical answers that appear to be rationally warranted.
(3) The rational principles that lead from empirical to metaphysical questions and from there to metaphysical answers are principles of ‘universal human reason’; that is, they belong to rational thinking as such. (p. 5, cf. p. 157)
This looks awfully similar to the critical strand.
However, in Part II of the book, Willaschek fills out the skeleton of the Rational Sources Account in detail by showing how the Paralogisms, Antinomies and Ideal of Pure Reason turn on the ideas of the soul, the world and God as unconditioned. It seems to me that it is this level of detail that comprises the real difference between the merely critical strand and the Rational Sources Account in the Dialectic, thus the latter but not the former brings out that pure reason’s ideas of the unconditioned are natural products of reason’s discursive, iterative and systematic character (chapters 1–2). The difference between the critical strand and the Rational Sources Account thus does not lie at a ‘fourth level’, as Willaschek suggests, at which
Kant argues that even though the transcendental principles and ideas derived at the previous level have a legitimate “regulative” use in guiding empirical scientific research and our search for unity in the diversity of natural phenomena, they are easily mistaken for being constitutive, that is, for being true representations of objects. (pp. 7–8)
For this is much the same as the second level of the merely critical account. The difference is rather that the Rational Sources Account inserts its additional level of detail before it reaches this point. Thus Willaschek’s clearest overview of the Rational Sources Account runs thus:
On the first, most general level, there is the transition from the ‘Logical Maxim,’ which requires us to find a condition for each conditioned cognition, to the ‘supreme principle of pure reason’ … according to which if something conditioned is given, then so is the complete series of conditions, where this series is itself unconditioned. On Kant’s view, it is this latter principle that drives human reason to metaphysical speculation. …
On the second level, Kant derives the system of transcendental ideas … concepts of objects that, if they do exist, are unconditioned (such as the soul, the world, and God). … This system consists of the concept of the unconditioned …, the three classes of transcendental ideas (psychological, cosmological, and theological …), and nine modes (ways in which objects can be thought to be unconditioned): substantiality, simplicity, personality, and spirituality in the case of the soul; the absolute completeness of composition, division, origin and mutual dependence in the case of the world; and finally the idea of an ens realissimum in the case of God. On this second level, Kant only derives the systematic order of transcendental ideas, not these ideas themselves.
On the third level, there are the specific ‘dialectical’ (that is, fallacious) inferences that purport to provide us with a priori knowledge about the soul, the world, and God …
Finally, on the fourth level, Kant argues that even though the transcendental principles and ideas have a legitimate ‘regulative’ use in guiding empirical scientific research and our search for unity in the diversity of natural phenomena, they are easily mistaken for being constitutive, that is, for being true representations of objects. (pp. 7–8)
The first and second strands of Kant’s argument differ chiefly in that, while both insist that the principles of reason should be understood regulatively rather than constitutively, only the second fills in the details about the actual derivation of the ideas of unconditioned soul, world and God, but only the first explicitly invokes transcendental realism to explain why we mistake regulative for constitutive principles. In a nutshell, then, the Rational Sources Account is Kant’s derivation of the regulative principles of reason from its most basic operations, while the critical core of the Dialectic is its explanation that we transform the legitimate regulative principles of reason into illegitimate constitutive principles because of our natural but unwarranted attraction to transcendental realism. I think that the overall argument of Willaschek’s book might have been clearer if he had put it thus rather than speaking of two separate strands.
That organizational point is independent of the substantive suggestions to which I now turn. First, I suggest that the idea of God has a broader role in both the theoretical and practical uses of reason than Willaschek proposes. On the theoretical side, Willaschek focuses on Kant’s analyses of arguments for the existence of God in the fourth Antinomy of Pure Reason and in chapter three, section two, of the Ideal of Pure Reason. The first could be considered a variation of the cosmological argument, which Kant will consider in its more ordinary form in his critique of the three traditional arguments for the existence of God in chapter three, section four, of the Ideal of Pure Reason: whereas the traditional ontological argument infers the existence of a necessary being from the existence of any contingent being, the thesis of the fourth Antinomy purports to be able to infer the existence of a necessary being outside of the completed series of all contingent beings as the unconditioned condition for the totality of conditioned beings. In the Antinomies, this argument for the existence of God stands alongside of but is independent of the theses of the other antinomies, which argue for the existence of a world that is complete and finite in both its spatial and temporal extension and division and for the existence of a spontaneous will that is the unconditioned condition of conditioned events (and by implication, not just one such will, but many, thus not just for God but for us); all of these arguments are of course opposed by their antitheses, and as is well known Kant resolves these two disputes in two different ways, arguing that the dispute whether the world is finite or infinite in spatial and temporal extension and division is rendered moot by transcendental idealism, the recognition that everything in space and time is only indefinitely extendable appearance, neither finite nor genuinely infinite, while in the case of free will and God, those ideas may be thought although not known to have real objects in the realm of things in themselves but not in appearance, where they are precluded by the conditions of sensibility. Thus transcendental realism is blocked with regard to the first two antinomies, but allowed as an indemonstrable possibility for the theses of the last two. The argument that Kant criticizes in chapter three, section two, may be considered a variant or refinement of the ‘only possible basis for a demonstration of the existence of God’ that Kant had introduced in his 1763 work of that name: whereas in the earlier work Kant had simply argued that all possibility requires a ground in a necessary being, in the first Critique he presents a more elaborate account of possibility, arguing that the complete determination of the concept of any particular requires a ‘sum total of all possibility’ (A573/B601) from which the predicates of the particular being can be chosen, and that this sum total of all possibility must in turn be grounded in the necessary existence of a being possessing all ‘positive’ predicates (whatever exactly they might be). This applies the idea of the unconditioned twice, in the form of the idea of the sum total of all possibility and in the idea of the ground of the latter in the existence of a being possessing all positive predicates; and transcendental realism leads to the metaphysical assertion of the existence of God by means of the assumption that these ideal conditions for the complete determination of concepts must actually exist.
Willaschek presents these matters in much more detail than I have here, and I have no argument with his presentation. What I do want to question is only the suggestion, at least in the Antinomies, that the argument for the existence of God just stands alongside the other arguments for metaphysical conclusions, and further the suggestion that the metaphysical tendency to assert the actual existence of God is always impelled by transcendental realism. As to the first point, I think more emphasis might be needed on Kant’s claim in the third section of the first book of the Transcendental Dialectic, on the ‘system of the transcendental ideas’, that there are ‘three classes’ of transcendental ideas, ‘of which the first contains the absolute (unconditioned) unity of the thinking subject, the second the absolute unity of the series of conditions of appearance, the third the absolute unity of the condition of all objects of thought in general’ (A334/B391). In other words, however the ideas of the unconditioned subject and world-whole might be reached, the idea of God is inevitably reached as the unconditioned condition of those, which in that regard turn out not to be entirely unconditioned after all. The idea of the absolute subject is unconditioned with regard to its own predicates, and that of the world-whole is unconditioned in relation to its constituent parts, but both of these are themselves also conditioned in the sense of depending upon God. The idea of God is thus not just the idea of one more unconditioned alongside others, but in Kant’s mind all roads lead to the idea of God. Then transcendental realism can kick in and lead from the unconditioned idea of God to the assertion of the existence of God, although of course we should also step back from this natural tendency and recognize that the idea of God can serve only as a regulative idea.
But the other point that I want to make in this regard is that if we carry the story forward to the third Critique, we see that Kant thinks that belief in the existence of God, which has to be recast as a regulative idea, also arises from a natural misinterpretation of transcendental idealism. What I have in mind is Kant’s argument in section four of the published Introduction to this work. Here Kant argues that particular laws of nature, that is, laws beneath the level of and more determinate than the ‘universal transcendental laws’ of nature established in the System of Principles of the first Critique, most of all the principles of the Analogies of Experience, ‘as empirical, may indeed be contingent in accordance with the insight of our understanding’, but still, ‘if they are to be called laws (as is also required by the concept of a nature), must be regarded as necessary on a principle of the unity of the manifold’ (5: 180), that is, some principle of the unity of a manifold; and, since all necessity must be considered to be the product of some understanding, that is, of the synthesis of a manifold in accordance with the necessary principles of some understanding, then
This principle can be nothing other than this: that since universal laws of nature have their ground in our understanding, which prescribes them to nature (although only in accordance with the universal concept of it as nature), the particular empirical laws, in regard to that which is left undetermined in them by the former, must be considered in terms of the sort of unity they would have if an understanding (even if not ours) had likewise given them for the sake of our faculty of cognition, in order to make possible a system of experience in accordance with particular laws of nature. (5: 180)
Here Kant is arguing that we naturally arrive at a conception of God by an extension of the fundamental insight of transcendental idealism with regard to the most general laws of nature, that they are the product of our own understanding, to all laws of nature, which, as laws, must be necessary and, as necessary, must be the product of an understanding, although one that is not ours and is obviously greater in capacity than ours. Perhaps transcendental realism must then also kick in to take the unsuspecting subject to the actual assertion of the existence of such a divine law-giver, but if so it can kick in only after this natural extension of transcendental idealism has taken place. And then, of course, the subject must be warned to make only regulative use of the idea of this understanding.
One way to bring out the difference between this argument and Kant’s argument in the thesis of the fourth Antinomy is that, while the latter infers the existence of God as the unconditioned condition of the series of all contingent beings, the present argument infers the existence of God from the imagined complete system of all necessary laws of nature, which is to say all laws of nature. Because of this difference, I think it is worth noting this addition to the analysis in the first Critique, to which Willaschek confines himself.
I now turn to Willaschek’s brief Epilogue on Kant’s use of the idea of the unconditioned in practical philosophy. Willaschek rightly says that ‘an adequate discussion of this aspect of Kant’s philosophy would require a book of its own’ and proposes to offer only ‘a glimpse of Kant’s “practical metaphysics”’ (p. 270). So I do not want to tax him with the incompleteness of his discussion. But I do want to make two points. First, I want to dispute his characterization of the epistemic status of the postulates of pure practical reason. Willaschek says that ‘speculative reason, considered on its own, requires us to reject (i.e. not to accept) these postulates because they are theoretically undecidable. But we cannot both accept and not accept the postulates, so either speculative or practical reason must give’, and then argues that because of the categorical character of the demands of practical reason, speculative reason must give way to practical reason and its postulates (p. 273). Surely Kant’s position is rather that undecidability and rejection are not the same thing, and that it is only because speculative reason must leave the postulates undecided, for example, it can neither prove nor disprove the existence of God, that practical reason is free to step into the vacuum left by theoretical reason. Theoretical reason does not have to give way to practical reason; in Kant’s view, his doctrine of the primacy of practical reason is entirely consistent with his view of the unity of reason in general.
My second remark concerns the highest good. Willaschek’s lapidary statements that the ‘highest good is “the unconditioned totality of the object of pure practical reason,” that is, the ultimate and inclusive end of rational agency’ and that as such ‘it must be considered to be realizable through our own actions, since otherwise we could not be obligated to bring it about’ (p. 271) seem entirely correct. And that is to say that the demand of morality is not unconditional only in the sense of being absolutely binding no matter what anyone’s desires might be, but also that it is unconditional in the further sense of demanding completion of its application: in words from the first Critique, morality demands that we complete the transformation of the natural world into a moral world, in which ‘a system of happiness proportionately combined with morality can also be thought as necessary, since freedom, partly moved and partly restricted by moral laws, would itself be the cause of the general happiness’ (A809/B837). But Willaschek then confines himself to the further argument suggested in the first Critique that, since virtuous individuals cannot in fact count on the moral cooperation of all others in the transformation of the natural world into a moral world, those of us who do our moral best must defer our hopes of happiness proportionate to our own virtue to ‘a world that is future for us’ in which God does reward the virtuous with happiness (A811/B839). I believe that Kant was already moving away from this lame, consolatory approach to the highest good in the treatment of the postulate of the existence of God in the second Critique and then even more clearly in the treatment of the highest good in the third Critique, for in those places he suggests that we must postulate the existence of God as the author of a nature in which human efforts to be moral can after all be successful and can produce happiness in nature (5: 125, 450), as the natural consequence of establishing a realm of ends in which not only is each person treated as an end in herself but also their particular ends are realized to the greatest compossible extent (4: 433). That is, the consoling thought is not that virtuous individuals, but only those, will be rewarded with happiness in an afterlife by the fiat of God, but rather that the human species as a whole is capable of morality, which when complete will naturally be accompanied with happiness even though happiness – that of any individual agent or that of all human beings – is not its fundamental motivation (see also 8: 279). This is further to say that the solution to the paradox of the highest good is not to say that the necessarily inadequate efforts of human beings towards the highest good will be supplemented in an afterlife, but that humans must be considered capable of bringing about the highest good by their own efforts after all because they are part of a nature authored by God with that possibility in mind. Of course, this thought too involves a probable deferral beyond the natural life of many individuals, but not a deferral of the happiness of the virtuous to an afterlife, rather a deferral of the full realization of both complete virtue and complete happiness to some indeterminate later date in the natural history of the human species. And whether Kant puts this point in the language of postulation as in the second Critique or of reflecting judgement and regulative principles as in the third, it offers us only a possibility that can suffice to make our efforts toward the realization of the highest good rational rather than a proof that could satisfy the most dogmatic metaphysician that this happy state must come about. Nevertheless, it seems a more attractive conception of the highest good than the one to which Willaschek confines himself on the basis of his study of the Critique of Pure Reason.
This is a further thought about Kant’s use of the idea of the unconditioned to which Willaschek’s work pushes me. It is not meant to detract from my almost unconditional admiration for what he has accomplished in his outstanding book.