Frederick Rauscher’s book Naturalism and Realism in Kant’s Ethics is a densely argued contribution to the debates on Kant’s metaethics (Rauscher Reference Rauscher2015; references will appear simply as page numbers). It locates Kant’s ethics within his overall system, and deals with virtually every aspect of Kant’s moral philosophy, the concepts of pure and empirical practical reason, his fact-of-reason doctrine, as well as God, free choice and value, which Rauscher interprets as ‘ideas of reason [which] merely ground the regulative principles of reason as heuristic devices’ (p. 154). Rauscher’s book is full of interesting claims and arguments, but one striking feature of this study is that the author consistently supports them with textual evidence from Kant’s works. Rauscher’s capability to navigate Kant’s practical philosophy – from published works to letters, fragments and lecture notes – is admirable.
In my comments I will focus on Rauscher’s views on the topic of naturalism in Kant’s ethics, as his argument in part relates to my paper on ‘Naturalism and Kant’s Resolution of the Third Antinomy’ (Pollok Reference Pollok2013). Our point of controversy concerns the version of metaphysical naturalism that Rauscher defends in his book, i.e. the claim that Kant’s ethics does not require metaphysical commitments that go beyond nature.
Following my distinction between methodological and metaphysical naturalism, Rauscher argues that Kant should be seen as a methodological anti-naturalist and a metaphysical naturalist. So we agree on Kant’s stance on methodological naturalism and differ on Kant’s stance on metaphysical naturalism.
By methodological naturalism we understand (first) the view that the goals and methods of philosophy are continuous with those of the sciences, and (second) that supernatural entities lack any explanatory force. Metaphysical naturalism is meant to designate the view that whatever exists must be some object that, at least in principle, can be accounted for by the mathematico-empirical sciences.
I will now quote from Rauscher’s critique of my view, in order to take the opportunity to argue why Kant should not be seen as a metaphysical naturalist:
[Pollok] understands metaphysical naturalism in relation only to the mathematical sciences, thus excluding Kant’s empirical psychology, and then invokes causal closure to argue that Kant restricts causal explanation to physical explanation. Pollok focuses on the Third Antinomy in claiming that Kant requires a different causal explanation other than physical causality to account for the actions of the mind. The problem with this approach is that in excluding the objects of empirical psychology from nature, Pollok applies a restriction that Kant himself did not use because Kant did consider inner sense to be a part of causally determined nature. (p. 31)
The gist of my response to this is that the problem is not so much about whether or not the objects of empirical psychology should be excluded from nature. On my (and Rauscher’s) view it should be granted that the objects of empirical psychology – whatever this exactly means for the critical Kant – indeed are parts of nature. In the first Critique, Kant writes, ‘If more than the cogito were the ground of our pure rational cognition of thinking beings in general; if we also made use of observations about the play of our thoughts and the natural laws of the thinking self created from them: then an empirical psychology would arise, which would be a species of the physiology of inner sense, which would perhaps explain the appearances of inner sense’ (A347/B405).Footnote
1
And in the amendments to the Anthropology, the so-called Rostock Manuskript H, Kant writes, ‘the I of inner sense, that is, of the perception and observation of oneself, is not the subject of judgment, but an object’ (7: 399). That is, an object of nature. This point should be uncontroversial, and any discrepancy might be cleared up as a misunderstanding.
However, what is important here is that, for Rauscher, ethics is not about different objects than those in nature but only about a different relation among those objects in nature, and hence, Kant’s ethics is seen as metaphysically naturalistic. By contrast, I argue that as an object of nature our soul (or whatever the object of inner sense may be called) is part of the sensible world, while our practical reason as the subject and the object of morality is not part of the sensible world.
In contrast to Rauscher, who sees the ‘practical standpoint as an action-oriented facet of experience’ (p. 28), I would argue that Kant’s idea of normativity – including practical normativity – fundamentally relates to judgements. Kant’s novel theory of the normativity of judgements, where the validity of our ordinary and scientific judgements is determined by certain synthetic principles a priori, is in part a response to early modern rationalism, where the perfection, i.e. the utmost clarity and distinctness, of God’s ideas served as the standard for the more or less ‘dark and confused’ ideas of finite reasoners like us. I argue for this interpretation in (Pollok Reference Pollok2017: chapter 1). What is relevant in the present context is that Kant’s novel theory of the normativity of judgements includes the practical. Here Kant’s normativity fundamentally relates to practical judgements, i.e. maxims as the subjective principles of our actions. Kant argues that for finite reasoners like us, it is the potential discrepancy between a maxim (subjective principle of action) and the moral law (objective principle of action) that characterizes the a priori syntheticity of the categorical imperative. For finite reasoners the consciousness of the moral law includes (1) their practical reason’s grasp of an ‘ought’, or more precisely, the stipulated universal form of their maxims; and (2) a specific attitude towards this ‘ought’. or more precisely, the acknowledgement of one’s own commitment to that ‘ought’. Kant’s concern about lawfulness in the practical context is not immediately about this or that justification of our willing, or even about justified actions, but rather about the justifiability of our maxims.
In a broader context, the judging subjects must be seen to exist, in Sellarsian terms, in a ‘space of giving and asking for reasons’. It is this space of reasons, or in Kant’s pre-analytic, and metaphysically more ambitious wording, this space of reason, for which he seeks to discover the a priori and constitutive structures. Kant’s space of reason is neither exhaustible by explanation and prediction, nor indicative of any transcendent existence. Here I agree with Rauscher who argues that ‘the practical point of view … abstracts from ontological claims’ (p. 163); I will come back to this point. Kant’s space of reason is rather the home of a community of intellectus discursivi (as opposed to a supernatural intellectus archetypus), and it is governed by laws that function as criteria for the validity of the judgemental claims of those finite reasoners.
Most relevant, when Kant considers the possibility of human action, is that only a rational cause, that is, a reason, can justify – rather than effect – the action. The authorship of an action is a non-natural attribute or status. Kant’s important point here is not that the moral realm consists of some non-natural moral facts. The moral realm, or the ‘intelligible world’ (or some part of it), as Kant calls it, rather stands for the evaluative reality of reason that can be recognized, or acknowledged, rather than cognized. The only dimension of the ‘intelligible world’ is validity, and its ‘existence’ is ‘perceivable’ or ‘cognizable’ only in terms of assessment.
On Rauscher’s reading, there is one set of objects, i.e. the objects of nature, and there are two different sets of causal order, one deterministic, the other transcendentally free. I think it is true that Kant does not argue for a second set of objects that would be somehow similar to natural objects – just not natural. However, this does not entail that the only existing kind of objects is natural. On my reading there is one set of objects, according to Kant, that may or may not have a sensible as well as a supersensible nature. The deterministic order refers to the sensible nature of things, including human beings (but excluding God), the moral order refers to the supersensible nature of things, including human beings (but excluding non-rational animals). Regarding human beings, Kant separates these two referents as the homo phaenomenon and the homo noumenon in us (see e.g. MS, 6: 335, 418). Whether we like it or not, when it comes to the moral order, the homo phaenomenon does not matter to Kant. Strictly speaking, Kant is not primarily concerned with the rescue of the drowning child. In his moral philosophy he is essentially concerned with the causality of a cause, pure practical reason, that either is or or is not the determining ground for the rescue.
On p. 120 Rauscher claims: ‘A transcendentally free reason can exist in nature just as other transcendental conditions for the possibility of experience such as causal relations can exist in nature’. On my reading, it is true that a human being at a certain time in a certain location is phenomenal. But this does not mean that a ‘transcendentally free reason can exist in nature just as other transcendental conditions for the possibility of experience exist in nature’. Causal relations exist in nature only insofar as they can be attributed to appearances, according to Kant. But a ‘transcendentally free reason’ cannot be attributed to appearances, it is not a condition of the possibility of experience. Its causality – the ‘empirically unconditioned causality (of freedom)’ (A561/B589) – must be distinguished from phenomenal causality. Again, justifying an action is something essentially different from explaining it. And so, even if the being that justifies an action in some sense exists in nature, this does not entail that the relevant intelligible causality of this being, pure practical reason, exists in nature as well. To the contrary, this causality is what stands in opposition to natural causality, or mechanism, just insofar as it is non-natural, i.e. unconditioned by other causes. In this sense, I cannot see how Kant could have claimed that a ‘transcendentally free reason can exist in nature just as other transcendental conditions for the possibility of experience … can exist in nature’.
Related to this is the problem of timelessness in nature. It is unclear to me how a ‘transcendentally free reason’, which must be seen as ‘timeless (p. 123, cf. p. 128) as Rauscher acknowledges, can exist in nature, given Kant’s concept of nature ‘as the sum total of all appearances’ (B163; cf. A418/B446), or ‘the sum total of all objects of experience’ (Prol, 4: 295), or ‘the sum total of all things, in so far as they can be objects of our senses, and thus also of experience. Nature, in this meaning, is therefore understood as the whole of all appearances, that is, the sensible world, excluding all non-sensible objects’ (MAN, 4: 467).
Rauscher concludes this subsection on ‘Reason as a transcendentally free cause’ by stating that ‘Pure reason can be a cause of actions in the sensible world by dictating that they ought to occur’ (p. 123). I think this is correct, since human beings, you and I, can acknowledge the validity of the moral law, i.e. follow the categorical imperative and act in the sensible world such that, from a noumenal standpoint, or from the standpoint of what Kant calls the intelligible world, this phenomenal occurrence can be given a justification (which, again, is different from an explanation). But Rauscher comments on this sentence as follows: ‘Reason as a faculty can be understood as providing a structure for experience, specifically, a structure for systematic and consistent actions’ (p. 123). This ‘pure structure of reason’ is ‘the moral law itself’ (p. 123). Here I disagree. The structure of experience is provided on the one hand by constitutive concepts of the understanding such as substance and causality, and on the other hand by regulative ideas of reason such as ‘the homogeneity, specification and continuity of forms’ (A658/B686) which together mark the ‘focus imaginarius’ (A644/B672) of our empirical research, as well as by the ‘four cosmological ideas’ that provide a unity for our understanding of the world as a whole (see A415/B443). These concepts and ideas together provide the structures for our explanations of nature. This includes actions, since actions like any other event in nature must follow some laws in order to be explained. But I cannot see why Kant would have included the moral law in order to explain actions. Psychologically my representation of the moral law may be invoked as such an explanation (however obscure and precarious psychology’s access to my representations may turn out to be). But the moral law itself and the standpoint required for our acknowledgement of the moral law must be distinguished from psychological (or any other) explanation. Building on Lanier Anderson’s distinction between normative rules, such as logical or moral laws, and natural laws, e.g. in psychology, we can say that:
Normative rules are such that their validity cannot entail the absence of exceptions; that is, the normative rule itself envisions the possibility of exceptions, even though we might find out later, on the basis of circumstances unforeseen in the rule, that the exceptions turn out not to be possible. So even if ‘ought’ implies ‘can’ (in some sense of ‘can’), it cannot imply ‘does.’ Natural laws, by contrast, are supposed to have just this implication. We can therefore propose a test: the direction of accountability between the rule and the facts it covers is different in the two cases. For descriptive empirical laws, if there is a mismatch between a proposed law and the facts, we find fault with the law, and revise it to fit the data; with normative rules, however, we hold the facts accountable to the rule, and lay the blame for mismatches on them (they are false judgments, or bad actions). Normative rules, then, have standing, independent of the facts they cover. They retain their validity and remain binding, even when violated in fact. (Anderson Reference Anderson2005: 293)
Now Rauscher may accept this description and try to relate it to Kant’s methodological anti-naturalism, leaving metaphysical naturalism intact. But from my point of view, he seems to underestimate the importance Kant places on the sensible/intelligible distinction. He writes, ‘When the transcendental structure of reason is embodied in a particular individual, it is identical with the particular empirical reason of that individual. All particular judgments are subject to causal determination by prior causes in nature. … Particular judgments can have causal antecedents in nature and still follow the proper structure of reason. There is no more a problem with human beings embodying valid judgments than there is with computers calculating’ (pp. 127–8). By contrast, I would argue that the utterance of a judgement has ‘causal antecedents in nature’ but not the judgement. The computer is entirely subject to natural laws – and in this sense its ‘calculation’ is normatively idle, i.e. neither correct nor incorrect. It is only through reason that calculations can be seen to be correct or incorrect. Hence, human beings are subject (a) to natural laws, and (b) as rational beings subject to logical and moral laws. If my calculator spits out 5+7=10 I do not blame the calculator but the manufacturer; if a truck hits a child, we do not blame the truck but the driver. Strictly speaking, praise and blame do not occur ‘in nature’, they are second-order objects of empirical psychology as it were. This means that psychology may analyse how human beings deal with the attribution of praise and blame, but empirical psychology makes no first-order judgement about praise and blame.
Similarly, it seems to me that neither concepts nor judgements nor inferences should be seen as ‘particular parts of the chain of natural causation’, as Rauscher maintains. Only words and sentences uttered or written by natural (human) beings can be such parts. Likewise, neither spoken nor written words or sentences, i.e. natural facts, can be ‘bearers of validity’ (p. 128). It is only logical entities such as concepts, judgements, inferences and theories that can be valid or invalid. Neither the words nor the full sentence ‘five plus seven equals twelve’ have any validity. Only our understanding’s grasp of these words as a judgement allows for assessment and hence validity.
Again, even though in Metaphysical Foundations of Natural Science Kant gives some reasons for denying empirical psychology the status of a science proper (see 4: 471), empirical psychology does deal with natural facts, that is facts of inner sense. But moral categories like praise or blame do not occur in empirical psychology in a normative sense. The theoretical, explanatory or descriptive approach of empirical psychology does not merely mark a different methodology from ethics, both of which would deal with natural objects, as Rauscher suggests. Rather, ethics deals with human beings insofar as they can be seen as persons, i.e. as beings with a supersensible nature. Hence, I disagree with the following argument that Rauscher presents in chapter 2:
Our awareness of the moral law and moral principles is in empirical consciousness. Empirical consciousness is the subject matter of empirical psychology, which is part of the theoretical study of nature. Kant believes that causal determinism is true of empirical psychology, even if we are unable to know any of the causal laws. Moral principles should, then, be available for use in causal explanations of human beings in nature. (p. 67)
Inferring from the empirical consciousness of the moral law to the empirical reality of the moral law seems to rest on what Kant calls a ‘dialectical confusion’, since our empirical consciousness is part of the sensible world but the moral law is not. The difference in German between Geltung and Gültigkeit (that Habermas for example emphasizes) might be of some help here. What has empirical reality is the actual Geltung of moral principles, i.e. whether or not, where and when, and how many people abide by them. This means that moral laws are at a given point in space and time of more or less general application. But this must be distinguished from the validity (Gültigkeit) of moral principles with which Kant’s ethics is concerned. So even if the actual consciousness of moral principles is empirically real and may figure in empirical, psychological explanations, moral principles are not empirical and cannot be invoked for empirical, psychological explanations. They cannot explain facts because they can only be invoked for assessing facts. Therefore, the empirical access to moral principles is irrelevant from an ethical standpoint. Psychologically, the empirical consciousness of the moral law is as much subject to causal determinism as any other fact of outer and inner senses. What is of ethical concern, however, is not the empirical reality of given laws but the lawfulness of laws. This difference is nicely captured by Kant’s distinction between quid sit iuris and iustum (see MS, 6: 205, 229).
Hence, ethically, Rauscher’s claim that ‘the practical point of view … abstracts from ontological claims’ (p. 163) should be emphasized even more strongly: existence is the wrong kind of predicate in relation to a law if the focus is on the moral point of view (i.e. the iustum). The ‘existence’ of a law is its validity, which is not an ontological predicate. Metaphorically, we may speak of the space of reason where laws ‘exist’ as rational structures, and which is populated by moral entities like persons, but also by pure, mathematical entities like the square root of 2 which come to ‘exist’ as postulates in the axioms of a theory. But this metaphorical talk of ‘existence’ must not be reified, neither naturally nor supernaturally. Kant speaks of the ‘sublimity of our own supersensible existence’ (KpV, 5: 88) that the moral law makes us feel in our respect for it. He writes, ‘the morally good as an object is something supersensible, so that nothing corresponding to it can be found in any sensible intuition’ (5: 68). This means that moral properties are not part of the sensible world, and hence cannot be treated by empirical psychology. Assuming that the moral law exists in the natural world suggests that it could be cognized – according to the principles of the understanding – and figure in (scientific) explanations. But, just like the square root of 2 that does not exist in the natural world as the length of the hypotenuse of a right triangle with legs of length 1 (because here the sum of the internal angles of a triangle is only approximately 180°), so does the moral law not exist in the natural world. Locating these laws and entities in the natural world violates Kant’s verdict that a priori concepts and judgements cannot be empirically derived.
On a larger scale, according to Kant, there can be only two kinds of material philosophy, a metaphysics of nature and a metaphysics of morals: ‘the legislation of human reason … has two objects, nature and freedom, and thus contains the natural law as well as the moral law’ (A840/B868). In the second Critique’s section titled ‘On the primacy of pure practical reason in its connection with speculative reason’, he writes, ‘if pure reason of itself can be and really is practical, as the consciousness of the moral law proves it to be, it is still only one and the same reason which, whether from a theoretical or a practical perspective, judges according to a priori principles’ (KpV, 5: 121). Obviously, ‘reason’ does not refer here to the particular capacity for inferences, but to the mind of finite discursive reasoners in general, just as this term was used in the title of the first Critique, and as it subsequently figures in the opening definition of the third Critique: ‘The faculty of cognition from a priori principles can be called pure reason” (5: 167). So it is one and the same reason, only differing in its referents, that ‘judges according to a priori principles’ (KpV, 5: 121), the principles of the understanding and the moral law, respectively.
In the Critique of Practical Reason Kant argues in greater detail for the conceptual connection between these two kinds of legislation warranted by the power of judgement – presumably not the third Critique’s reflective power of judgement but the first Critique’s determining power of judgement. At the end of the second chapter, right before he proceeds from practical concepts to the Triebfeder as the practical analogon to intuitions, Kant inserts a section called ‘Of the typic of pure practical judgment’, where he deals with the role of the power of judgment ‘by which what is said in the rule universally (in abstracto) is applied to an action in concreto’ (KpV, 5: 67). He introduces the idea of a ‘form of lawfulness in general’ (5: 70), provided by the power of judgement, which is common to the sensible and the practically intelligible: ‘no intuition can be put under the law of freedom (as that of a causality not sensibly conditioned) – and hence under the concept of the unconditioned good as well – and hence no schema on behalf of its application in concreto’ (5: 69). The reason for this claim is that in practical matters we deal with the ‘schema of a law itself’, rather than ‘with a schema of a case in accordance with laws’, because the will – not the ‘action with reference to its result’ (5: 69) – ought to be determined ‘through the law alone’ (5: 68–9). In other words, in practical matters we have moved up a level from the phenomena, such that we do not deal with natural events (human behaviour, actions) that were accessible to intuition themselves, but with the action-determining will, which cannot be given in any intuition. From this Kant concludes that:
the moral law has no cognitive faculty other than the understanding (not the imagination) by means of which it can be applied to objects of nature, and what the understanding can put under an idea of reason is not a schema of sensibility but a law, such a law, however, as can be presented in concreto in objects of the senses and hence a law of nature, though only as to its form; this law is what the understanding can put under an idea of reason on behalf of the power of judgment, and we can, accordingly, call it the type of the moral law. (KpV, 5: 69)
We are thus ‘permitted’, Kant emphasizes, ‘to use the nature of the sensible world as the type of an intelligible nature, provided that I … refer to it only the form of lawfulness in general … For to this extent laws as such are the same, no matter from what they derive their determining grounds’ (5: 70). The first usage of the term ‘nature’ in this quotation (‘the nature of the sensible world’) must be seen as referring to its formal concept, the essence of something, by which Kant means its determination by laws. The second usage of the term ‘nature’ (‘an intelligible nature’) refers to its material concept, that is, the sum total of objects standing under certain laws (cf. A418–19/B446; Prol, 4: 318, MAN, 4: 467).
This practical schematism argument may be even called Kant’s naturalism: it designates the unity of nature, which Kant distinguishes as ‘sensible nature’ and ‘supersensible nature’ (KpV, 5: 68, cf. 43–7) under the idea of lawfulness. This common idea of lawfulness rules out any sort of super-naturalism referring to entities inexplicable in causal terms. But it makes room for objects that have both a sensible or phenomenal nature and a supersensible or noumenal nature, which differs from Rauscher’s metaphysical naturalism that is confined to empirical objects/subjects. Metaphysical naturalism strips objects, including finite rational beings like humans, of their supersensible nature. In short, human beings are not empirical, or sensible, beings, but they have a sensible nature: ‘The sensible nature of rational beings in general is their existence under empirically conditioned laws and is thus, for reason, heteronomy. The supersensible nature of the same beings, on the other hand, is their existence in accordance with laws that are independent of any empirical condition and thus belong to the autonomy of pure reason’ (5: 43).
So, from the second Critique’s perspective, Kant’s philosophical system consists of a metaphysics of nature and a metaphysics of morals. What bridges the gap between the two, however, is not the metaphysical identity of natural objects with which both parts of metaphysics would be dealing, as suggested by Rauscher’s metaphysical naturalism. Kant’s ethics does not primarily deal with human beings as natural objects, or as Rauscher specifies, with human behaviour. On my view, Kant’s ethics deals with normative statuses of human beings as persons. The objects to which these statuses can be ascribed are not simply natural objects. Rather, these objects are as supernatural as they are natural, or more precisely, they are as much part of a ‘supersensible nature’ as they are part of a ‘sensible nature’. Our sensible nature makes us subject to laws of nature, our supersensible nature makes us subject to the moral law. But this means that, contrary to metaphysical naturalism, the ‘us’ here is not simply our empirical or phenomenal self. Rather, from the empirical standpoint it must be seen as a natural object, while from the moral standpoint it must be seen as a supernatural object (or rather ‘subject’). Finite rational beings like us are both, homines phaenomena and homines noumena. In contrast to Rauscher’s metaphysical naturalism that claims a natural object identity across the different parts of Kant’s ‘pure’ and ‘applied’ metaphysics (cf. MAN, 4: 469–70), what unites these parts according to my reading is that Kant’s metaphysics of nature and his metaphysics of morals are grounded on the idea of lawfulness. In a kind of architectonic chiasmus we may then characterize the unity of reason as follows: it is the practical use of reason that, on the one hand, has primacy over its speculative use, ‘since all interest is ultimately practical’ (KpV, 5: 121). But in order for the practical use to satisfy reason’s interest, that is, the ‘determination of the will with respect to the final and complete end’ (5: 120), reason must use ‘the form of lawfulness in general’, that is, ‘the nature of the sensible world as the type of an intelligible nature’ (5: 70). At the same time, by satisfying its practical interest, which ‘consists in the determination of the will with respect to the final and complete end’ (5: 120), it also satisfies the interest of its ‘speculative use’, which ‘consists in the cognition of the object up to the highest a priori principles’ (ibid.). What unifies the sensible and the supersensible realities into one nature is that their existence and their comprehensibility rest on their lawfulness. At least this is what Kant’s complete system of reason looks like from the standpoint of the second Critique.
In my comments I have mainly focused on the differences and disagreements between Rauscher’s and my own reading of Kant’s ethics. It would probably take more space to explain and discuss all the agreements between the two of us. I think there is a big overlap. Rauscher’s interpretation is a refreshing antidote to any spooky story about metaphysical entities in the realm of ends. In this respect, his emphasis on a non-ontological reading of Kant’s practical philosophy is very much in line with my own interpretation. The way in which Rauscher takes the first Critique’s account of the ideas of reason as relevant for Kant’s subsequent works, such that the regulative status of the idea of God prevents us from reifying it, definitely deserves closer attention. I am convinced that other readers will substantively disagree over this issue. But perhaps this should be seen as a great virtue of Rauscher’s work: it is in the best sense thought-provoking, as it will certainly stir up a very heterogeneous audience in the years to come.
You have
Access
Frederick Rauscher’s book Naturalism and Realism in Kant’s Ethics is a densely argued contribution to the debates on Kant’s metaethics (Rauscher Reference Rauscher2015; references will appear simply as page numbers). It locates Kant’s ethics within his overall system, and deals with virtually every aspect of Kant’s moral philosophy, the concepts of pure and empirical practical reason, his fact-of-reason doctrine, as well as God, free choice and value, which Rauscher interprets as ‘ideas of reason [which] merely ground the regulative principles of reason as heuristic devices’ (p. 154). Rauscher’s book is full of interesting claims and arguments, but one striking feature of this study is that the author consistently supports them with textual evidence from Kant’s works. Rauscher’s capability to navigate Kant’s practical philosophy – from published works to letters, fragments and lecture notes – is admirable.
In my comments I will focus on Rauscher’s views on the topic of naturalism in Kant’s ethics, as his argument in part relates to my paper on ‘Naturalism and Kant’s Resolution of the Third Antinomy’ (Pollok Reference Pollok2013). Our point of controversy concerns the version of metaphysical naturalism that Rauscher defends in his book, i.e. the claim that Kant’s ethics does not require metaphysical commitments that go beyond nature.
Following my distinction between methodological and metaphysical naturalism, Rauscher argues that Kant should be seen as a methodological anti-naturalist and a metaphysical naturalist. So we agree on Kant’s stance on methodological naturalism and differ on Kant’s stance on metaphysical naturalism.
By methodological naturalism we understand (first) the view that the goals and methods of philosophy are continuous with those of the sciences, and (second) that supernatural entities lack any explanatory force. Metaphysical naturalism is meant to designate the view that whatever exists must be some object that, at least in principle, can be accounted for by the mathematico-empirical sciences.
I will now quote from Rauscher’s critique of my view, in order to take the opportunity to argue why Kant should not be seen as a metaphysical naturalist:
[Pollok] understands metaphysical naturalism in relation only to the mathematical sciences, thus excluding Kant’s empirical psychology, and then invokes causal closure to argue that Kant restricts causal explanation to physical explanation. Pollok focuses on the Third Antinomy in claiming that Kant requires a different causal explanation other than physical causality to account for the actions of the mind. The problem with this approach is that in excluding the objects of empirical psychology from nature, Pollok applies a restriction that Kant himself did not use because Kant did consider inner sense to be a part of causally determined nature. (p. 31)
The gist of my response to this is that the problem is not so much about whether or not the objects of empirical psychology should be excluded from nature. On my (and Rauscher’s) view it should be granted that the objects of empirical psychology – whatever this exactly means for the critical Kant – indeed are parts of nature. In the first Critique, Kant writes, ‘If more than the cogito were the ground of our pure rational cognition of thinking beings in general; if we also made use of observations about the play of our thoughts and the natural laws of the thinking self created from them: then an empirical psychology would arise, which would be a species of the physiology of inner sense, which would perhaps explain the appearances of inner sense’ (A347/B405).Footnote 1 And in the amendments to the Anthropology, the so-called Rostock Manuskript H, Kant writes, ‘the I of inner sense, that is, of the perception and observation of oneself, is not the subject of judgment, but an object’ (7: 399). That is, an object of nature. This point should be uncontroversial, and any discrepancy might be cleared up as a misunderstanding.
However, what is important here is that, for Rauscher, ethics is not about different objects than those in nature but only about a different relation among those objects in nature, and hence, Kant’s ethics is seen as metaphysically naturalistic. By contrast, I argue that as an object of nature our soul (or whatever the object of inner sense may be called) is part of the sensible world, while our practical reason as the subject and the object of morality is not part of the sensible world.
In contrast to Rauscher, who sees the ‘practical standpoint as an action-oriented facet of experience’ (p. 28), I would argue that Kant’s idea of normativity – including practical normativity – fundamentally relates to judgements. Kant’s novel theory of the normativity of judgements, where the validity of our ordinary and scientific judgements is determined by certain synthetic principles a priori, is in part a response to early modern rationalism, where the perfection, i.e. the utmost clarity and distinctness, of God’s ideas served as the standard for the more or less ‘dark and confused’ ideas of finite reasoners like us. I argue for this interpretation in (Pollok Reference Pollok2017: chapter 1). What is relevant in the present context is that Kant’s novel theory of the normativity of judgements includes the practical. Here Kant’s normativity fundamentally relates to practical judgements, i.e. maxims as the subjective principles of our actions. Kant argues that for finite reasoners like us, it is the potential discrepancy between a maxim (subjective principle of action) and the moral law (objective principle of action) that characterizes the a priori syntheticity of the categorical imperative. For finite reasoners the consciousness of the moral law includes (1) their practical reason’s grasp of an ‘ought’, or more precisely, the stipulated universal form of their maxims; and (2) a specific attitude towards this ‘ought’. or more precisely, the acknowledgement of one’s own commitment to that ‘ought’. Kant’s concern about lawfulness in the practical context is not immediately about this or that justification of our willing, or even about justified actions, but rather about the justifiability of our maxims.
In a broader context, the judging subjects must be seen to exist, in Sellarsian terms, in a ‘space of giving and asking for reasons’. It is this space of reasons, or in Kant’s pre-analytic, and metaphysically more ambitious wording, this space of reason, for which he seeks to discover the a priori and constitutive structures. Kant’s space of reason is neither exhaustible by explanation and prediction, nor indicative of any transcendent existence. Here I agree with Rauscher who argues that ‘the practical point of view … abstracts from ontological claims’ (p. 163); I will come back to this point. Kant’s space of reason is rather the home of a community of intellectus discursivi (as opposed to a supernatural intellectus archetypus), and it is governed by laws that function as criteria for the validity of the judgemental claims of those finite reasoners.
Most relevant, when Kant considers the possibility of human action, is that only a rational cause, that is, a reason, can justify – rather than effect – the action. The authorship of an action is a non-natural attribute or status. Kant’s important point here is not that the moral realm consists of some non-natural moral facts. The moral realm, or the ‘intelligible world’ (or some part of it), as Kant calls it, rather stands for the evaluative reality of reason that can be recognized, or acknowledged, rather than cognized. The only dimension of the ‘intelligible world’ is validity, and its ‘existence’ is ‘perceivable’ or ‘cognizable’ only in terms of assessment.
On Rauscher’s reading, there is one set of objects, i.e. the objects of nature, and there are two different sets of causal order, one deterministic, the other transcendentally free. I think it is true that Kant does not argue for a second set of objects that would be somehow similar to natural objects – just not natural. However, this does not entail that the only existing kind of objects is natural. On my reading there is one set of objects, according to Kant, that may or may not have a sensible as well as a supersensible nature. The deterministic order refers to the sensible nature of things, including human beings (but excluding God), the moral order refers to the supersensible nature of things, including human beings (but excluding non-rational animals). Regarding human beings, Kant separates these two referents as the homo phaenomenon and the homo noumenon in us (see e.g. MS, 6: 335, 418). Whether we like it or not, when it comes to the moral order, the homo phaenomenon does not matter to Kant. Strictly speaking, Kant is not primarily concerned with the rescue of the drowning child. In his moral philosophy he is essentially concerned with the causality of a cause, pure practical reason, that either is or or is not the determining ground for the rescue.
On p. 120 Rauscher claims: ‘A transcendentally free reason can exist in nature just as other transcendental conditions for the possibility of experience such as causal relations can exist in nature’. On my reading, it is true that a human being at a certain time in a certain location is phenomenal. But this does not mean that a ‘transcendentally free reason can exist in nature just as other transcendental conditions for the possibility of experience exist in nature’. Causal relations exist in nature only insofar as they can be attributed to appearances, according to Kant. But a ‘transcendentally free reason’ cannot be attributed to appearances, it is not a condition of the possibility of experience. Its causality – the ‘empirically unconditioned causality (of freedom)’ (A561/B589) – must be distinguished from phenomenal causality. Again, justifying an action is something essentially different from explaining it. And so, even if the being that justifies an action in some sense exists in nature, this does not entail that the relevant intelligible causality of this being, pure practical reason, exists in nature as well. To the contrary, this causality is what stands in opposition to natural causality, or mechanism, just insofar as it is non-natural, i.e. unconditioned by other causes. In this sense, I cannot see how Kant could have claimed that a ‘transcendentally free reason can exist in nature just as other transcendental conditions for the possibility of experience … can exist in nature’.
Related to this is the problem of timelessness in nature. It is unclear to me how a ‘transcendentally free reason’, which must be seen as ‘timeless (p. 123, cf. p. 128) as Rauscher acknowledges, can exist in nature, given Kant’s concept of nature ‘as the sum total of all appearances’ (B163; cf. A418/B446), or ‘the sum total of all objects of experience’ (Prol, 4: 295), or ‘the sum total of all things, in so far as they can be objects of our senses, and thus also of experience. Nature, in this meaning, is therefore understood as the whole of all appearances, that is, the sensible world, excluding all non-sensible objects’ (MAN, 4: 467).
Rauscher concludes this subsection on ‘Reason as a transcendentally free cause’ by stating that ‘Pure reason can be a cause of actions in the sensible world by dictating that they ought to occur’ (p. 123). I think this is correct, since human beings, you and I, can acknowledge the validity of the moral law, i.e. follow the categorical imperative and act in the sensible world such that, from a noumenal standpoint, or from the standpoint of what Kant calls the intelligible world, this phenomenal occurrence can be given a justification (which, again, is different from an explanation). But Rauscher comments on this sentence as follows: ‘Reason as a faculty can be understood as providing a structure for experience, specifically, a structure for systematic and consistent actions’ (p. 123). This ‘pure structure of reason’ is ‘the moral law itself’ (p. 123). Here I disagree. The structure of experience is provided on the one hand by constitutive concepts of the understanding such as substance and causality, and on the other hand by regulative ideas of reason such as ‘the homogeneity, specification and continuity of forms’ (A658/B686) which together mark the ‘focus imaginarius’ (A644/B672) of our empirical research, as well as by the ‘four cosmological ideas’ that provide a unity for our understanding of the world as a whole (see A415/B443). These concepts and ideas together provide the structures for our explanations of nature. This includes actions, since actions like any other event in nature must follow some laws in order to be explained. But I cannot see why Kant would have included the moral law in order to explain actions. Psychologically my representation of the moral law may be invoked as such an explanation (however obscure and precarious psychology’s access to my representations may turn out to be). But the moral law itself and the standpoint required for our acknowledgement of the moral law must be distinguished from psychological (or any other) explanation. Building on Lanier Anderson’s distinction between normative rules, such as logical or moral laws, and natural laws, e.g. in psychology, we can say that:
Normative rules are such that their validity cannot entail the absence of exceptions; that is, the normative rule itself envisions the possibility of exceptions, even though we might find out later, on the basis of circumstances unforeseen in the rule, that the exceptions turn out not to be possible. So even if ‘ought’ implies ‘can’ (in some sense of ‘can’), it cannot imply ‘does.’ Natural laws, by contrast, are supposed to have just this implication. We can therefore propose a test: the direction of accountability between the rule and the facts it covers is different in the two cases. For descriptive empirical laws, if there is a mismatch between a proposed law and the facts, we find fault with the law, and revise it to fit the data; with normative rules, however, we hold the facts accountable to the rule, and lay the blame for mismatches on them (they are false judgments, or bad actions). Normative rules, then, have standing, independent of the facts they cover. They retain their validity and remain binding, even when violated in fact. (Anderson Reference Anderson2005: 293)
Now Rauscher may accept this description and try to relate it to Kant’s methodological anti-naturalism, leaving metaphysical naturalism intact. But from my point of view, he seems to underestimate the importance Kant places on the sensible/intelligible distinction. He writes, ‘When the transcendental structure of reason is embodied in a particular individual, it is identical with the particular empirical reason of that individual. All particular judgments are subject to causal determination by prior causes in nature. … Particular judgments can have causal antecedents in nature and still follow the proper structure of reason. There is no more a problem with human beings embodying valid judgments than there is with computers calculating’ (pp. 127–8). By contrast, I would argue that the utterance of a judgement has ‘causal antecedents in nature’ but not the judgement. The computer is entirely subject to natural laws – and in this sense its ‘calculation’ is normatively idle, i.e. neither correct nor incorrect. It is only through reason that calculations can be seen to be correct or incorrect. Hence, human beings are subject (a) to natural laws, and (b) as rational beings subject to logical and moral laws. If my calculator spits out 5+7=10 I do not blame the calculator but the manufacturer; if a truck hits a child, we do not blame the truck but the driver. Strictly speaking, praise and blame do not occur ‘in nature’, they are second-order objects of empirical psychology as it were. This means that psychology may analyse how human beings deal with the attribution of praise and blame, but empirical psychology makes no first-order judgement about praise and blame.
Similarly, it seems to me that neither concepts nor judgements nor inferences should be seen as ‘particular parts of the chain of natural causation’, as Rauscher maintains. Only words and sentences uttered or written by natural (human) beings can be such parts. Likewise, neither spoken nor written words or sentences, i.e. natural facts, can be ‘bearers of validity’ (p. 128). It is only logical entities such as concepts, judgements, inferences and theories that can be valid or invalid. Neither the words nor the full sentence ‘five plus seven equals twelve’ have any validity. Only our understanding’s grasp of these words as a judgement allows for assessment and hence validity.
Again, even though in Metaphysical Foundations of Natural Science Kant gives some reasons for denying empirical psychology the status of a science proper (see 4: 471), empirical psychology does deal with natural facts, that is facts of inner sense. But moral categories like praise or blame do not occur in empirical psychology in a normative sense. The theoretical, explanatory or descriptive approach of empirical psychology does not merely mark a different methodology from ethics, both of which would deal with natural objects, as Rauscher suggests. Rather, ethics deals with human beings insofar as they can be seen as persons, i.e. as beings with a supersensible nature. Hence, I disagree with the following argument that Rauscher presents in chapter 2:
Our awareness of the moral law and moral principles is in empirical consciousness. Empirical consciousness is the subject matter of empirical psychology, which is part of the theoretical study of nature. Kant believes that causal determinism is true of empirical psychology, even if we are unable to know any of the causal laws. Moral principles should, then, be available for use in causal explanations of human beings in nature. (p. 67)
Inferring from the empirical consciousness of the moral law to the empirical reality of the moral law seems to rest on what Kant calls a ‘dialectical confusion’, since our empirical consciousness is part of the sensible world but the moral law is not. The difference in German between Geltung and Gültigkeit (that Habermas for example emphasizes) might be of some help here. What has empirical reality is the actual Geltung of moral principles, i.e. whether or not, where and when, and how many people abide by them. This means that moral laws are at a given point in space and time of more or less general application. But this must be distinguished from the validity (Gültigkeit) of moral principles with which Kant’s ethics is concerned. So even if the actual consciousness of moral principles is empirically real and may figure in empirical, psychological explanations, moral principles are not empirical and cannot be invoked for empirical, psychological explanations. They cannot explain facts because they can only be invoked for assessing facts. Therefore, the empirical access to moral principles is irrelevant from an ethical standpoint. Psychologically, the empirical consciousness of the moral law is as much subject to causal determinism as any other fact of outer and inner senses. What is of ethical concern, however, is not the empirical reality of given laws but the lawfulness of laws. This difference is nicely captured by Kant’s distinction between quid sit iuris and iustum (see MS, 6: 205, 229).
Hence, ethically, Rauscher’s claim that ‘the practical point of view … abstracts from ontological claims’ (p. 163) should be emphasized even more strongly: existence is the wrong kind of predicate in relation to a law if the focus is on the moral point of view (i.e. the iustum). The ‘existence’ of a law is its validity, which is not an ontological predicate. Metaphorically, we may speak of the space of reason where laws ‘exist’ as rational structures, and which is populated by moral entities like persons, but also by pure, mathematical entities like the square root of 2 which come to ‘exist’ as postulates in the axioms of a theory. But this metaphorical talk of ‘existence’ must not be reified, neither naturally nor supernaturally. Kant speaks of the ‘sublimity of our own supersensible existence’ (KpV, 5: 88) that the moral law makes us feel in our respect for it. He writes, ‘the morally good as an object is something supersensible, so that nothing corresponding to it can be found in any sensible intuition’ (5: 68). This means that moral properties are not part of the sensible world, and hence cannot be treated by empirical psychology. Assuming that the moral law exists in the natural world suggests that it could be cognized – according to the principles of the understanding – and figure in (scientific) explanations. But, just like the square root of 2 that does not exist in the natural world as the length of the hypotenuse of a right triangle with legs of length 1 (because here the sum of the internal angles of a triangle is only approximately 180°), so does the moral law not exist in the natural world. Locating these laws and entities in the natural world violates Kant’s verdict that a priori concepts and judgements cannot be empirically derived.
On a larger scale, according to Kant, there can be only two kinds of material philosophy, a metaphysics of nature and a metaphysics of morals: ‘the legislation of human reason … has two objects, nature and freedom, and thus contains the natural law as well as the moral law’ (A840/B868). In the second Critique’s section titled ‘On the primacy of pure practical reason in its connection with speculative reason’, he writes, ‘if pure reason of itself can be and really is practical, as the consciousness of the moral law proves it to be, it is still only one and the same reason which, whether from a theoretical or a practical perspective, judges according to a priori principles’ (KpV, 5: 121). Obviously, ‘reason’ does not refer here to the particular capacity for inferences, but to the mind of finite discursive reasoners in general, just as this term was used in the title of the first Critique, and as it subsequently figures in the opening definition of the third Critique: ‘The faculty of cognition from a priori principles can be called pure reason” (5: 167). So it is one and the same reason, only differing in its referents, that ‘judges according to a priori principles’ (KpV, 5: 121), the principles of the understanding and the moral law, respectively.
In the Critique of Practical Reason Kant argues in greater detail for the conceptual connection between these two kinds of legislation warranted by the power of judgement – presumably not the third Critique’s reflective power of judgement but the first Critique’s determining power of judgement. At the end of the second chapter, right before he proceeds from practical concepts to the Triebfeder as the practical analogon to intuitions, Kant inserts a section called ‘Of the typic of pure practical judgment’, where he deals with the role of the power of judgment ‘by which what is said in the rule universally (in abstracto) is applied to an action in concreto’ (KpV, 5: 67). He introduces the idea of a ‘form of lawfulness in general’ (5: 70), provided by the power of judgement, which is common to the sensible and the practically intelligible: ‘no intuition can be put under the law of freedom (as that of a causality not sensibly conditioned) – and hence under the concept of the unconditioned good as well – and hence no schema on behalf of its application in concreto’ (5: 69). The reason for this claim is that in practical matters we deal with the ‘schema of a law itself’, rather than ‘with a schema of a case in accordance with laws’, because the will – not the ‘action with reference to its result’ (5: 69) – ought to be determined ‘through the law alone’ (5: 68–9). In other words, in practical matters we have moved up a level from the phenomena, such that we do not deal with natural events (human behaviour, actions) that were accessible to intuition themselves, but with the action-determining will, which cannot be given in any intuition. From this Kant concludes that:
the moral law has no cognitive faculty other than the understanding (not the imagination) by means of which it can be applied to objects of nature, and what the understanding can put under an idea of reason is not a schema of sensibility but a law, such a law, however, as can be presented in concreto in objects of the senses and hence a law of nature, though only as to its form; this law is what the understanding can put under an idea of reason on behalf of the power of judgment, and we can, accordingly, call it the type of the moral law. (KpV, 5: 69)
We are thus ‘permitted’, Kant emphasizes, ‘to use the nature of the sensible world as the type of an intelligible nature, provided that I … refer to it only the form of lawfulness in general … For to this extent laws as such are the same, no matter from what they derive their determining grounds’ (5: 70). The first usage of the term ‘nature’ in this quotation (‘the nature of the sensible world’) must be seen as referring to its formal concept, the essence of something, by which Kant means its determination by laws. The second usage of the term ‘nature’ (‘an intelligible nature’) refers to its material concept, that is, the sum total of objects standing under certain laws (cf. A418–19/B446; Prol, 4: 318, MAN, 4: 467).
This practical schematism argument may be even called Kant’s naturalism: it designates the unity of nature, which Kant distinguishes as ‘sensible nature’ and ‘supersensible nature’ (KpV, 5: 68, cf. 43–7) under the idea of lawfulness. This common idea of lawfulness rules out any sort of super-naturalism referring to entities inexplicable in causal terms. But it makes room for objects that have both a sensible or phenomenal nature and a supersensible or noumenal nature, which differs from Rauscher’s metaphysical naturalism that is confined to empirical objects/subjects. Metaphysical naturalism strips objects, including finite rational beings like humans, of their supersensible nature. In short, human beings are not empirical, or sensible, beings, but they have a sensible nature: ‘The sensible nature of rational beings in general is their existence under empirically conditioned laws and is thus, for reason, heteronomy. The supersensible nature of the same beings, on the other hand, is their existence in accordance with laws that are independent of any empirical condition and thus belong to the autonomy of pure reason’ (5: 43).
So, from the second Critique’s perspective, Kant’s philosophical system consists of a metaphysics of nature and a metaphysics of morals. What bridges the gap between the two, however, is not the metaphysical identity of natural objects with which both parts of metaphysics would be dealing, as suggested by Rauscher’s metaphysical naturalism. Kant’s ethics does not primarily deal with human beings as natural objects, or as Rauscher specifies, with human behaviour. On my view, Kant’s ethics deals with normative statuses of human beings as persons. The objects to which these statuses can be ascribed are not simply natural objects. Rather, these objects are as supernatural as they are natural, or more precisely, they are as much part of a ‘supersensible nature’ as they are part of a ‘sensible nature’. Our sensible nature makes us subject to laws of nature, our supersensible nature makes us subject to the moral law. But this means that, contrary to metaphysical naturalism, the ‘us’ here is not simply our empirical or phenomenal self. Rather, from the empirical standpoint it must be seen as a natural object, while from the moral standpoint it must be seen as a supernatural object (or rather ‘subject’). Finite rational beings like us are both, homines phaenomena and homines noumena. In contrast to Rauscher’s metaphysical naturalism that claims a natural object identity across the different parts of Kant’s ‘pure’ and ‘applied’ metaphysics (cf. MAN, 4: 469–70), what unites these parts according to my reading is that Kant’s metaphysics of nature and his metaphysics of morals are grounded on the idea of lawfulness. In a kind of architectonic chiasmus we may then characterize the unity of reason as follows: it is the practical use of reason that, on the one hand, has primacy over its speculative use, ‘since all interest is ultimately practical’ (KpV, 5: 121). But in order for the practical use to satisfy reason’s interest, that is, the ‘determination of the will with respect to the final and complete end’ (5: 120), reason must use ‘the form of lawfulness in general’, that is, ‘the nature of the sensible world as the type of an intelligible nature’ (5: 70). At the same time, by satisfying its practical interest, which ‘consists in the determination of the will with respect to the final and complete end’ (5: 120), it also satisfies the interest of its ‘speculative use’, which ‘consists in the cognition of the object up to the highest a priori principles’ (ibid.). What unifies the sensible and the supersensible realities into one nature is that their existence and their comprehensibility rest on their lawfulness. At least this is what Kant’s complete system of reason looks like from the standpoint of the second Critique.
In my comments I have mainly focused on the differences and disagreements between Rauscher’s and my own reading of Kant’s ethics. It would probably take more space to explain and discuss all the agreements between the two of us. I think there is a big overlap. Rauscher’s interpretation is a refreshing antidote to any spooky story about metaphysical entities in the realm of ends. In this respect, his emphasis on a non-ontological reading of Kant’s practical philosophy is very much in line with my own interpretation. The way in which Rauscher takes the first Critique’s account of the ideas of reason as relevant for Kant’s subsequent works, such that the regulative status of the idea of God prevents us from reifying it, definitely deserves closer attention. I am convinced that other readers will substantively disagree over this issue. But perhaps this should be seen as a great virtue of Rauscher’s work: it is in the best sense thought-provoking, as it will certainly stir up a very heterogeneous audience in the years to come.