STEPHEN PALMQUIST | Comprehensive Commentary on Kant’s Religion within the Bounds of Bare Reason | Wiley-Blackwell 2015 


By Stephen R. Palmquist

Before I respond to the four essays that have each offered valuable feedback on my Comprehensive Commentary on Kant’s ‘Religion within the Bounds of Bare Reason‘ (hereafter CCKR),[1] a meta-critical question calls for an answer: Why was yet another commentary on Kant’s book, Religion within the Bounds of Bare Reason (hereafter RGV), needed in 2015,[2] given the unprecedented fact that each of the three previous years had seen the publication of a commentary on the same book? The short answer is that work on my commentary began several years before James DiCenso (2012), Lawrence Pasternack (2013), or Eddis Miller (2014) embarked on their own versions of such a project. After completing work on Kant’s Critical Religion in 2000 (Palmquist 2000), I initially planned to move directly into full-time work on Kant’s Critical Science, volume three in my planned Kant’s System of Perspectives series. However, I put that plan on hold a few years later, after I realised that Kant’s Critical Religion, despite its length, only scratched the surface of numerous issues relating to Kant’s theory of religion—especially various textual complexities in RGV. Moreover, some critics of Kant’s Critical Religion, even among those who agreed with my overall claim—namely, that Kant was attempting to affirm and even bolster Christianity (albeit in a significantly reformed interpretation)—tended to misconstrue and therefore reject key arguments I had sketched. Clearly, more work needed to be done if the scholarly community was to be convinced that anything like the position I defended there was accurate.

In 2006, I therefore set myself the task of writing the first comprehensive commentary in English on RGV. At the APA Central Division Meeting in 2007, having applied for a major grant from the Hong Kong government, I had lengthy conversations with the philosophy editors representing various publishers in the book room—including Cambridge, Routledge, and Bloomsbury—and pointed out to them that scholarship on RGV had been mushrooming in recent years, yet nobody had yet written an English-language commentary on Kant’s book. The publishers all seemed interested, yet none took the bait—or so I thought. A few months later, I paused my work on the project because the grant application was rejected; I worked on other projects for two years before I applied again, submitting an improved version of the previous application, and received the grant in 2010. In the seven years since that time, DiCenso, Pasternack, and Miller have each told me that the publisher of their respective commentaries invited them to write a commentary on RGV. Fortunately, Wiley did not take my suggestion elsewhere, but offered me a contract in 2012—and the rest is history.

Putting aside the foregoing explanation for my commentary appearing to be a late-comer on the scene, I shall now sketch briefly a few of its distinctive features. All four recent commentaries are ‘comprehensive’ in the sense that they claim to comment on Kant’s entire text. However, only my work actually includes Kant’s entire text within it. I present a revised English translation in a series of block quotes, a few lines at a time, followed in each case by a paragraph or two of commentary on the quoted passage. When I approached the Hackett table at that 2007 APA Meeting, the philosophy editor told me that, although they don’t do commentaries, they had contracted Werner Pluhar to prepare a translation of RGV, and were looking for someone to write the scholarly Introduction. My efforts on that occasion were not wasted, as I was eventually invited to write that Introduction. Moreover, Hackett then gave me permission to use Pluhar’s translation as the basis for the text quoted in CCKR. After comparing Pluhar’s translation not only with Kant’s original but also with di Giovanni’s Cambridge translation and with the older Greene and Hudson translation, I ended up revising 8–10% of Pluhar’s words.

Those who have not yet seen the text of CCKR may wish to take note of several other distinctive features it exhibits before considering my response to the critics. First, many of my revisions of Pluhar’s text relate to technical terms whose traditional translations tend to obscure Kant’s meaning, in some cases downplaying certain deeply religious features of Kant’s text. Thus, for example, I use ‘conviction’ for Gesinnung, in place of the traditional ‘disposition’ or Pluhar’s ‘attitude’,[3] while for Kant’s Lebenswandel I use ‘lifestyle’ instead of the range of expressions used by other translators, such as ‘life-conduct’ and ‘way of life’. I am also the only translator to acknowledge that Gottesdienst is the standard German term for ‘liturgy’; I translate it as such whenever it is combined as a single word, reserving the more literal ‘service of God’ for passages where Kant uses two separate German words. Kant’s key distinction is therefore no longer between moral religion and the confusing “religion of the service of God”, but is now more straightforwardly presented as a contrast between moral and liturgical religion. The list of such changes is extensive, so to assist the reader CCKR includes a Glossary that lists every revised word that occurs more than once, and offers a paragraph or more of justification for 40 to 50 of the key technical terms. Incidentally, I also follow Pluhar in using two new translations that he used at my recommendation: ‘bare’ for bloßen (as in the title, though I rendered his sporadic usage more consistent by using ‘bare’ each time the word appears as an adjective and ‘merely’ whenever it appears as an adverb), and ‘piece’ for Stück, the term Kant uses for the book’s four major divisions, probably to remind his readers that the four parts were originally composed as journal articles.

Aside from revisions to technical terms, most of my revisions were of three types. First, on many occasions I found that Pluhar and usually also the other translators were inconsistent, translating a single German word with two or more different English words (even though their meanings are not significantly distinct) or translating two or more German words with the same English word. Wherever I could do so without creating unbearably awkward phrasings, I aimed to standardise the usage in such cases, so that the reader knows which German word Kant is using by seeing which English word I have used. (Incidentally, in the block quotes from RGV I use dotted underlining to indicate every word I have revised from Pluhar’s translation.)

In the course of revising Pluhar’s translation of RGV, I also assessed the many amendments that Wobbermin, the original editor for volume 6 of the Akademie Ausgabe, introduced into Kant’s text over a century ago, in places where Kant’s original seemed to have a typo or Kant’s word choice made no apparent sense. In all such cases, the other English translators have simply followed the Akademie text blindly, almost never telling the reader that what they are translating is not actually what Kant wrote. In about half of the 100 or so occurrences of such variants, I have argued either that Kant’s original does make sense, once we understand and take seriously the fine details of the position he was defending in RGV or that Kant’s error was likely different from the one Wobbermin assumed. One of the most significant examples occurs in the first sentence of the General Comment to the Third Piece (RGV, AA 6:137.01–02), where the three previous translators all follow the Akademie Ausgabe in a change that has Kant saying that mysteries arise whenever we probe the inward make-up of historical faiths; but what Kant actually writes, if taken at face value (despite being awkward in its grammar) would have to mean that mysteries arise when we probe the inward make-up of rational religion. This, therefore, is what I take Kant to be saying—a change that makes the four parerga sections far more important than they would otherwise be: they are necessary appendages to rational religion, not necessary appendages to historical faith.

Another type of change I made to Pluhar’s translation concerns passages where Kant’s meaning was obscured by his use of rare terms whose meanings the translators have to guess. Probably the most significant example of this type of re-translation comes in the first paragraph of Subsection A of Section One of the Second Piece, where Kant quotes from several scripture passages in the process of introducing the archetype of human perfection. Alluding to the opening verses of John’s Gospel, Kant at one point observes that the archetype expresses itself as “the Word” (RGV, AA 6:60); in a parenthetical insertion he then seems to state the word he has in mind, but nobody seems to know what word Kant intends to specify. Greene and Hudson translate Kant’s “(das Werde!)” as “(the Fiat!)”—an expression that always struck me as being grossly out of place here; di Giovanni merely follows Greene and Hudson, despite the lack of any coherent meaning in this context. Pluhar does better, explaining in a footnote that Kant’s term is a special use of the verb meaning “to become,” but he laments that Kant here treats the verb idiosyncratically, as a noun. Assuming that the exclamation mark implies that Kant intended the parenthetical insertion to serve as an imperative, Pluhar translates it loosely, as “the Let it be so!” Taking Pluhar’s hint further than he himself was willing to do, I translate Kant’s German literally, as “the Become!” This harmonises well with the emphasis Kant goes on to adopt in the later sections of Section One of the Second Piece (and throughout the remainder of the book) on the importance of moral improvement: to follow the archetype of perfect humanity is to have the firm resolve to become the person one ought to be—this resolve being the core of true religion.[4] If my re-translation is correct, then Kant’s parenthetical insertion indicates that any person who exemplifies the archetype of perfect humanity will be so focused on the process of putting the good principle into practice that such a person could aptly be named the Become!

A second distinctive feature of CCKR (aside from presenting a thoroughly revised translation of the full text of RGV) is that I pay close attention to the differences between the first and second editions. Other than the First Critique, RGV is the only book that Kant published in a significantly revised second edition. What no other commentator has noted in any context, as far as I am aware, is that Kant’s reason for composing the second edition, less than a year after the first edition appeared, was that a highly influential Protestant theologian from Tübingen, Gottlob Christian Storr, wrote a short book in late 1793,[5] criticising Kant’s theory of religion in depth, from the perspective of a religiously-oriented follower of Kant’s Critical philosophy—Storr being the professor who taught Kant’s philosophy to Schelling, Hegel, and Hölderlin. CCKR provides evidence that on numerous occasions a new footnote or some other change that Kant made in the second edition comes as a direct response to a comment Storr raised in his book.

A third feature worth noting here is that I make a concerted effort to trace the origin of various sources that Kant probably used or alludes to without telling us the details. A prime example is his reference to the trilogy written by “one ancient” (probably the same author whom he referred to as “that Greek” in Perpetual Peace [ZeF, AA 8:365]), which has previously stumped commentators on Kant. Thanks to the wonders of Google, a search for several key terms enabled me to identify the ancient work that Kant alludes to at AA 6:34n as Agamemnon: the first work in the trilogy, Oresteia, written by Aeschylus and first performed in Athens in 458 BC, which dramatises the spiraling evils of the Trojan War (see CCKR, p. 89n). In some cases, such as Kant’s much-neglected quip about avoiding the company of one’s Muslim friends or family members in direct proportion to the number of times they have visited Mecca (AA 6:189n), I was unable to trace a source, but I did come across various later texts that cite the same “Persian proverb”, and the evidence from these sources suggests that Kant may well have invented it himself, as an attempt at humor (see CCKR, p. 474n).

Finally, one of the reasons I wanted to write CCKR was to clear up a point that seems to have been misunderstood by quite a few readers of my previous works. Some Kantians have labelled me as one of those Kantians who are trying to ‘Christianise Kant’. Such a goal could fairly be inferred from Kant’s Critical Religion, for I do there argue that Kant was attempting to revolutionise Christianity, and I do take that to be a serious effort on his part and an effort that Christians ought to take seriously. If that is what ‘Christianising Kant’ means, then I fully embrace it in CCKR as well; but most who raise this point seem implicitly to be assuming that the recent scholars who read Kant in this way are trying to make Kant into a Christian Evangelical or even a Fundamentalist. In response to this misunderstanding, I have at various points in CCKR attempted to make clear that viewing Kant as defending a revolutionary new form of Christianity does not (and was never meant to) imply that he sought to condone (or to refute) specific dogmas defended by any given Christian theologian (or by the countless religious believers who may have followed their lead) in the past.

1. Response to Ronald Green

It is virtually inevitable that, in an exchange of critical essays that focus on RGV, questions would be raised about Kant’s theory of evil. This is not only because Kant’s arguments relating to evil—insofar as they can be called arguments—are notoriously obscure and difficult to follow. It is also because the First Piece, where most of Kant’s comments on evil appear, has given rise to more secondary literature than any other part of the book—indeed, perhaps more than all the other secondary literature on RGV, combined. It was primarily for this reason that I intentionally decided not to allow CCKR to be bogged down with an extensive and independent discussion of all the interpretative options relating to Kant’s theory of evil. I do not ignore the issue, of course; I could not have done that while still calling the book comprehensive. Rather, the rule I followed was to apportion the length of my comments on each Piece to the length of Kant’s original text. Likewise, following Kant’s exposition, my discussion of the Second Piece is the shortest of the four main parts, while my discussion of the Third Piece is the longest.

That Ronald Green is unconvinced by my account of Kant’s ‘quasi-transcendental’ argument for the necessity and universality of the propensity to evil may be due in part to the fact that CCKR presents only a relatively brief and summary account of my full argument. This is because I had previously published an article that discusses my understanding of Kant’s argument in detail,[6] contrasting it with other interpretations—including the approach that Green himself prefers. Another rule I followed in composing CCKR was that if I had already discussed an argument extensively in a previously published article or book, then I merely summarised that argument and referred the reader to the relevant publication for further details. This was the only possible way I could manage to comment on every passage in Kant’s Religion and still keep the main text to a mere 500 pages. This strategy may also account for Green’s rather surprising claim that my interpretation of Kant’s theory of evil has negative implications for understanding his account of grace and of the role historical faith plays in true religion, for here too CCKR merely summarises arguments defended more fully elsewhere.[7]

Green’s objection to my summary exposition of Kant’s argument is based on one of the three main interpretative approaches found in the literature, where Kant’s position is portrayed as being either merely analytic but in need of a transcendental supplement (e.g. Allison), or transcendental in some qualified sense (e.g. Morgan), or an argument based on empirical and anthropological evidence (e.g. Wood and Anderson-Gold). I adopt an approach similar to Morgan’s, calling Kant’s argument not transcendental in the most technical sense of the term but nevertheless quasi-transcendental; I admit that the third approach, a version of which Green defends in his essay, is grounded in some evidence from Kant’s text. That evidence, however, is misunderstood when interpreters regard it as Kant’s proof of the evil propensity: Kant admittedly cites an impressive string of examples of empirical evil in the First Piece; but he never claims to build a formal proof out of them on their own; instead, they foreshadow an argument he develops in the Third Piece, which is not about the nature of evil at all, but about the reason we must look for a community-oriented remedy to the existential problem posed by evil. I shall come back to this point toward the end of this section.

On the basis of his assumption that Kant has empirical reasons for believing in the necessity and universality of the origin of the evil propensity, Green specifically challenges my claim (CCKR, p. 112) that Kant’s argument implies that

if just one of us is guilty of an evil action, then the evil propensity is imputed to the human nature shared by us all.

He says this formula should be reversed:

It is because ‘the evil propensity is reasonably imputed to the human nature shared by us all’ that ‘one of us is guilty of an evil action’.

My response is that Green’s reformulation of my statement is not wrong, because these two versions are reciprocal: they can both accurately represent Kant’s position; yet one of them must have priority over the other. The question is, which explanation has the priority when considered philosophically, the transcendental or the empirical?

That Kant regards the propensity to evil as necessary seems indisputable, given the textual evidence, which CCKR repeats almost ad nauseam, due to its frequent occurrence in RGV’s First Piece. Throughout the relevant sections of the First Piece Kant repeatedly claims that the human propensity ‘must’ be evil; he even appeals explicitly to the concept of necessity in numerous statements regarding the status of this propensity. As Part One of CCKR cannot be read without having this evidence thrown in the reader’s face, I shall not attempt to rehearse it here. The paradox, and the reason Kant’s position is so difficult to unpack, is that he also repeatedly insists that each individual’s evil, being self-chosen, is contingent and therefore avoidable. So how can the evil propensity be both necessary and contingent? The key to answering both this and the previous question, regarding the proper priority between the two versions of the individual-species relation, lies in understanding Kant’s Copernican view of the conditions for calling something ‘necessary’.

In the First Critique Kant famously describes “the altered method of our way of thinking” (Bxviii n.) as entailing “that we can cognize of things a priori only what we ourselves have put into them”. He then develops his refutation of Hume’s scepticism regarding the application of the principle of causality to empirical events by arguing that the only way we can make sense out of the claim that even a single case of causal relation—such as the cause of a roof being supported by the walls of a house (A190/B235), or the cause of a ship floating downstream (A192/B237)—is to presuppose that all events are subject to the a priori principle of necessary connection. That is, one would not be able to recognise necessity anywhere in the empirical world, if it were not for the transcendental condition whereby the human mind reads necessity into what we experience. Becoming aware of such transcendental conditions does not prevent us from recognising that the very same events are empirically contingent; rather, it means only that if we can detect and confirm at least one relation that exhibits necessary connection, then we can be sure that the principle of causality holds transcendentally. Essentially the same Copernican logic, I claim, applies to Kant’s argument for the necessary and universal applicability—note the carefully chosen word, applicability, not application—of the propensity to evil. In both cases, the same event can be viewed from two quite distinct perspectives: either transcendentally, as exhibiting something that is necessarily true about the genus of any rational being who experiences such an event, or empirically, as exhibiting something that is contingently true about the individual person or object having the experience.

The sentence from CCKR that Green focuses on most intently is part of my interpretation of the status of the propensity to evil in Kant’s system of religion; it is not, as Green wrongly assumes, about the necessary or contingent status of particular evil deeds. It is intended more as a rhetorical device than as a defining feature of my interpretation. I don’t mean that I wrote it carelessly; I stand by its accuracy as an implication of one side of Kant’s two-perspective position. The point might be more clearly expressed by translating it into the language of traditional Christian ethics, which Green himself (unlike Kant) tends to use: sin (or what Kant calls ‘radical evil’) is a necessary condition of being human, to the extent that someone who claims to have no trace of what theologians call ‘original sin’ would not be recognisably human; yet sinning at any given point of time (what Kant would call exercising one’s Willkür to choose to do something evil) is not necessary but thoroughly contingent, chosen by each individual on each and every occasion of doing evil—or good, for that matter. Kant puts it thus:

Every evil action must be regarded, when one seeks its rational origin, as if the human being had fallen into it immediately from the status of innocence. (RGV, AA 6:41)

Perhaps the best justification for the claim that my version of the maxim (i.e. the transcendental one) has primacy for Kant, not Green’s version (i.e. the empirical one) is to recognise the incoherence of claiming the opposite. For Kant no amount of empirical reflection on our own past evil could possibly determine the necessity of our propensity to evil. Necessity, as I have argued above, can only be determined transcendentally. And making such a transcendental determination does not commit Kant to the view that original sin makes a person sin, as in the proverbial excuse: “The devil made me do it!” As I interpret his argument, Kant’s quasi-transcendental alternative is that we know that the human race is beset by original sin because this must be true, given the fact that I, at least, have sinned. So, while it is not wrong to say that empirical reflection on one’s own (or any given human being’s) sinful acts plays a role in Kant’s argument, this role is as a subordinate step in an argument that has a transcendental structure overall: the presence of the propensity to evil as a feature of the human species is what makes it possible for you and me to act out of self-love instead of following the moral law in all its purity.

In defending the empirical origin of radical evil, Green fails to distinguish between two terms that many interpreters assume to be coextensive, if not synonymous: ‘propensity’ (Hang) and ‘heart’ (Herz). For Kant, the propensity is a transcendental and therefore necessary aspect of our nature; we, as individuals at least, do not choose to have it; instead, as members of the genus, we find ourselves to have it already, even when we trace our moral pilgrimage back to its earliest beginnings—at or even before birth.[8] Thus, when Kant quotes Horace at the beginning of Section III of the First Piece, just before listing some “of the multitude of screaming examples that experience places before our eyes in the deeds of human beings” (RGV, AA 6:32–3), he is not preparing the reader for an empirical proof of evil, as Green contends. Rather, Horace’s “No one is born without vices” (RGV, AA 6:32; my translation of Vitiis nemo sine nascitur) clearly implies that there is a kind of vice that is not empirical, because it is already there when we are born! Moreover, Kant insists that this transcendental vice (i.e. radical evil) remains with us throughout our life and cannot be ‘unchosen’, as it were. By contrast, the ‘heart’ is empirical, in the sense that it is always open to experiencing a revolution that converts us from being a self-absorbed, evil-hearted person to being an other-affirming, good-hearted person. In other words, the choice to be either good-hearted or evil-hearted is entirely contingent, and is determined by the nature of the freely chosen ‘conviction’ (Gesinnung) that informs the individual’s heart.

That the propensity to evil is not chosen by the human individual—if chosen at all, then it is somehow ‘chosen’ timelessly, by the whole human genus (see RGV, AA 6:31–2; cf. CCKR, p. 82n), whatever that means!—becomes more evident once we explore the relevance of Kant’s theory of the three “predispositions to good” (RGV, AA 6:26–8) to this whole issue. In the first numbered section of the First Piece Kant argues that human nature consists of animality, humanity, and personality: we share the former (i.e. the fact that we are alive) with other animals; we share the latter (i.e. the fact that we are responsible for what we do) with God; and our unique status as humans consists in our rational capacity, through which we stand in between these two extremes. As soon as Kant finishes providing brief accounts of each component of this three-fold distinction, he launches immediately into his lengthy discussion of the propensity to evil. My claim in CCKR is that one of the main reasons he is so elusive about what makes us radically evil is that he sees it as being built in to our nature. In some sense our genus ‘chooses’ this nature; just how this happens is not something that Kant claims we can know through philosophical reasoning, though narratives such as those found in the Bible express helpful symbols that enable us to picture the process to ourselves. What philosophy does enable us to know, though, is that we (and this ‘we’ is crucial, as the Third Piece argues) must have chosen it, and that this therefore faces each human individual with a crucial empirical choice: Will you give in to the self-love that is prioritised by your very nature, or will you appeal to the rational potential that lies within you in order to effect a conversion from the original state of being controlled by self-love (i.e. the state of being ‘evil hearted’) and to the new-born state of being self-controlled (i.e. the state of being ‘good hearted’)? This choice is empirical and therefore contingent; so Green is entirely correct to say that, according to Kant, every human being, with the possible exception of Jesus (see below), must make it. This, however, does not make the evil propensity empirical and contingent, otherwise a converted person would thereby be rejecting evil in its entirety and would thus become holy—something I am confident that Green would not want to claim.

When Kant refers to the “anthropological probing” (RGV, AA 6:25) that would be necessary in order to discover whether or not there are any individual exceptions to the general rule that human nature is radically evil, he is not saying that our justification for calling human nature evil is empirical. Many interpreters assume this is Kant’s position, because of the “long melancholy litany of charges against humanity” (RGV, AA 6:33) that he provides in Section III of the First Piece. But a careful analysis of Kant’s wording reveals that he does not say that those various examples prove that human nature is evil; rather, he says that they are so easy to come by that “we can spare ourselves the formal proof” (RGV, AA 6:32–3). That is, if one were not a philosopher, then the a priori proof that he elsewhere claims to provide (RGV, AA 6:35.31–7) would not be necessary. He never states that his own exposition will dispense with the need for such a proof; quite to the contrary, he explicitly states that a “formal proof” is needed (RGV, AA 6:32–3; see CCKR 2.4) and at the end of Section III affirms that the “actual [a priori] proof” in question (RGV, AA 6:39n) has already been given.

Instead of serving as a surrogate for an a priori proof, Kant’s paragraph of examples serves as a crucial (and indeed, empirical) step in Kant’s quasi-transcendental argument. It demonstrates that—in case any of his readers might perhaps think of themselves as being so enlightened that they are beyond reproach—at least someone has done evil, even if it wasn’t me. This is enough to prove that human nature is necessarily evil (or as Christian theology puts it, that original sin applies to the whole genus). When he finally comes to the issue of whether or not one person (e.g. Jesus) might have escaped that fate, Kant responds that, in order to be an efficacious example for us, a religious hero must be regarded as sharing our evil propensity even if he never gave into the temptation to act upon it. Now this argument is clearly anthropological and empirical, based specifically on the historical records found mainly in the Bible. But this part of Kant’s argument is about how historical religion can assist us in overcoming the debilitating effects of the human propensity to evil; it is not about how that propensity arises in the first place.

I have some qualms of my own with Green’s reconstructed version of Kant’s argument in the First Piece, but here I shall only briefly outline my concerns, as I have dealt with this alternative more fully elsewhere (see note 6). First, in order to maintain that Kant’s arguments about evil are “based not on an a priori argument” but solely “on an assessment of the empirical reality of human experience”, Green must simply ignore the fact that Kant repeatedly states that he provides a formal, a priori proof. My reading, by contrast, takes both the a priori and the empirical aspects of Kant’s argument into account. Second, Green replaces Kant’s consistent use of language that attributes necessity to the evil propensity with the assumption that Kant is only arguing that human beings are “possibly” evil. But Kant never even comes close to arguing in the modality that Green attributes to him: he never asks us to argue from the “consistent pattern of choices” we see in “the record of this [human] species” to the possibility that we have an evil propensity. Rather, he argues that we must have an evil propensity in order to explain any such departure from our good predisposition. Third, Kant’s ‘rigorism’ is not about whether or not a person can live a good life in an ‘exceptionless’ way, as Green claims. Properly understood, rigorism is the view that every specific moral choice is either good or evil, never both and never neither. Fourth, Kant’s understanding of what it means to live a virtuous life does not require absolute perfection; the latter would constitute holiness, which he attributes only to God. And holiness (not virtue) would be Kant’s name for a life lived in exceptionless pursuit of goodness. Kant fully and repeatedly recognises that, for nearly all (if not all) human beings, being ‘virtuous’ involves occasional departures from the path of goodness. What makes some people virtuous, even though they still sometimes do evil, while others remain vicious even though they sometimes do good, is that the person’s inward conviction is different in each case. That is, the virtuous person for Kant is the good-hearted person who humbly recognises his or her moral failings but still does his or her best to be good. The role of the specifically religious notion of ‘holiness’ in Kant’s argument is to awaken us to the impossibility of ever living up to God’s standards in our empirical choices, so that we can empower ourselves to be virtuous, thanks to the religious belief that God will judge our conviction rather than our individual deeds.

In the end, I do not think that Green and I are as far apart as his comments seem to suggest. For I agree with Green that Kant’s “argument doesn’t claim that I am corrupt because I am a member of a genus with some necessary fault”. I do not believe Green, or anyone else who takes what Kant says in the First Piece at face value, would dispute that Kant’s theory of radical evil entails that the human genus has evil as part of its nature. He apparently took my interpretation to imply that this makes every human being empirically guilty before they even exercise any moral choice. But I never drew this inference from Kant’s position. Rather, the crucial point I defended in CCKR is that I can know that the human genus has a propensity to evil because I know that I, at least, have made evil choices. That is, Kant is calling each human being to recognise: I am Adam; my deeds provide ample evidence for the radical evil that must be imputed to the entire genus (in the form of an evil propensity), in order to understand how my evil choices were even possible. This is the necessary correlate to Kant’s doctrine of the archetype, whereby we are offered a path of salvation from our absolute guilt for the radical evil of our genus: only by appealing to the archetype of perfection that lies within our reason can we pull ourselves out of this situation and learn to walk the path of virtue.

This, then, is the reason Kant mentions the possibility, near the beginning of his discussion of evil in the First Piece, that ‘anthropological probing’ might lead us to the conclusion that not every human person actually chooses to allow his or her moral choices to be governed by the evil principle. Indeed, the most frequently cited evidence for the ‘anthropological’ thesis (i.e. the claim many Kant interpreters make, that for Kant evil has an entirely empirical source) actually comes from the beginning of the Third Piece, where Kant points out that even a good-hearted, archetypally-informed individual will tend to fall back into evil “at once when he is among human beings” (RGV, AA 6:94). This argument is empirical, and is crucial to Kant’s understanding of the community-based nature of the proper antidote to evil. But this empirical argument, urging us to participate in religious communities, says nothing about the origin of evil, for there is absolutely nothing in Kant’s various examples that would justify the attribution ‘radical’ to the evil deeds that we commit; instead, Kant’s anthropological probing in Religion is about how to overcome radical evil—an issue I have discussed in detail elsewhere (see note 7).

Finally, Green’s closing comments about the recovery from evil call for a response. Green claims

that the possibility of a human being’s not sinning is essential to Kant’s Christology. At least one individual avoided all temptations.

But this again conflates the evil propensity with evil deeds or choices. If ‘sinning’ here refers merely to our empirical deeds and choices, then Green is correct but is not disagreeing with any feature of my interpretation. Yet Green appears also to be claiming that Kant’s Christology requires there to be a human being who stood above the entire propensity to evil. Far from affirming this latter position, however, Kant argues in Subsection B of Section One of the Second Piece that if one is to believe in a perfect example of human goodness, then what is essential is that we must regard such a person as being just like the rest of us—i.e. as having the evil propensity—otherwise Jesus (or any other religious hero) would not be an example that we could follow and would consequently contribute nothing to our moral empowerment. Thus Kant’s Jesus cannot avoid temptations, as Green claims, but must fully experience them in all of their weight, coming face to face with the propensity to evil. Here Green incorrectly assumes

that Kant explicitly denies that this archetype requires a sensible presence—schematisation—in time and space.

Kant does deny the possibility of schematisation, because this would mean we could obtain empirical knowledge of the moral status of a person’s character. But this does not mean that Kant denies all forms of sensible presence. Rather, in the footnote that Green refers to (RGV, AA 6:64–5n) Kant argues that in the religious realm schematisation is to be replaced by symbolism, and in the main text Kant adds that the reason he denies the legitimacy of viewing one person as the only possible example of the archetype of perfection is that this would discourage us ordinary individuals from becoming examples of the archetype ourselves—a position Kierkegaard did indeed later adopt and develop, making it the basis for a new (but perhaps not quite so radically new) approach to the philosophy of religion.[9]

2. Response to Christina Drogalis

Christina Drogalis offers a fairly accurate summary of my interpretation of the account Kant gives, in Section V of the First Piece[10] and in Section One of the Second Piece, of how we can be empowered to overcome the effects of radical evil. A slight correction or clarification I would make to her presentation of my position is that I do not think Kant argues that we can know “that humans are incapable of restoring this [moral] predisposition to good within us on our own”, as she claims. Rather, on my reading Kant argues as follows: we cannot know whether or not we are capable of completely overcoming the effects of our propensity to evil, yet we do know that we ought to restore our moral predisposition to its original status of primacy; and since ought implies can, we may believe in divine assistance, if this belief enables us to resolve the impasse that otherwise keeps us mired in radical evil, as long as we do not regard such belief as replacing our need to choose the new conviction that sets us on the path to goodness. In CCKR, I sometimes express Kant’s position as being that we must believe such divine assistance; but this is a conditional ‘must’. That is, we must believe in God’s assistance if we wish to move beyond this impasse, whereby we cannot understand how it is possible to effect a change of heart. As long as we regard this ‘must’ as conditional, we will understand it, as Drogalis rightly suggests, in terms of the ‘may’ of hope.

Kant acknowledges that some people also believe that such a change of conviction is not possible without God’s assistance. Again, Kant’s own view is that the limitations of human knowledge prevent us from being certain as to whether or not God really does play a role in this process, but that the belief that God plays such a role can be a rational and even wise belief, if (and only if!) it empowers us to improve our lifestyle (our Lebenswandel), for the latter is the only evidence we have that a genuine change of heart has occurred. So, Kant is not claiming to know that divine assistance is necessary, nor is he claiming that divine assistance absolutely must occur; rather, he is explaining how we must conceive of its operation, if it takes place. He also seems to admit that he can think of no other way in which radically evil human beings can have a rational hope to become good.

With this clarification in mind, the path to a response to the two issues Drogalis raises should be fairly straightforward. First, she points out that overcoming radical evil requires “only the commitment to improve our actions in the future and not actual perfection in our actions”, so that on my reading “virtue […] appears to be adequate for this overcoming”. Here I would appeal to Green’s comments regarding the uncompromising nature of the moral law as Kant sees it: while I disagree that this is what Kant means by the term ‘rigorism’ (see above), I fully agree with his claim that the demand is for exceptionless goodness—not merely improvement of our actions. This is because the moral law appeals to our intelligible character: in the above-quoted statement from Drogalis’s essay, the reference to a ‘commitment’ is, for Kant, an appeal to a choice that is intelligible, in the sense that it is made without reference to the time-conditions with which all human actions are intertwined. Drogalis’s first point is therefore a good one: holiness for God entails not having any inclinations that compete with the moral law; but we humans cannot reasonably hope to attain an inclination-free life, so the moral law does not command this. The fact that we have inclinations, whereas God does not, is not a legitimate excuse for wrongdoing: “Do the best you can, if the circumstances allow” is not a valid formulation of the categorical imperative! I agree with Drogalis that

the issue at hand in [Section V of the First Piece] […] is not one’s ability to become morally good in one’s actions but merely to restore the predisposition to good, or to change one’s intelligible character from bad to good.

The problem, which Drogalis does not take fully into consideration, is that this change of conviction, being intelligible, is beyond the grasp of our theoretical knowledge.

Kant’s logic goes like this: the moral law (or for the religious believer, God) requires holiness in our intelligible character (i.e. our conviction); we as individuals must make the change to a good conviction, otherwise we would not be responsible for it and it would not make us holy; but because we cannot know our intelligible character, we are constrained to infer it from our actions; if we have changed our intelligible character entirely on our own initiative, then we do not need divine grace; but the presence of radical evil in our nature militates against this possibility; this situation puts us in a very real risk of experiencing despair and simply giving up on the attempt to be good; belief in divine grace, however, is a wise option for those who are humbly aware of their own ignorance, for belief in at least the possibility of divine aid is alone what enables us even to conceive of our ability to experience a genuine change of heart and thereby to avoid despair; the hope that such divine aid is indeed available is therefore practically necessary for us human beings. This is Kant’s basic strategy for justifying faith-based commitment.

Drogalis concludes her discussion of this first issue by expressing it as a ‘concern’:

It is unclear to me how this robust sort of belief is possible to acquire, if the need for it arises due to moral commands. This seems contrary to the way in which sincere religious belief typically arises.

I find it difficult to respond to this concern, because Drogalis does not explain how she thinks “sincere religious belief typically arises”. If she is referring, as Green does near the end of his essay, to the fact that the religious believer is confronted with some historical truth-claims (such as Jesus’s resurrection, in the case of Christianity), then my response would be that Kant does not deny that such empirical factors are relevant to how conversion actually occurs in each given religious tradition. Rather, his view is that such historical factors are irrelevant to the concerns of the philosophical theologian and should instead be left in the able hands of biblical theologians to decide. Kant leaves such questions undecided because he is not interested in promoting any one historical faith (or any one account of what historical facts authentic believers should regard as true) over the others, but rather with the moral implications of whatever historical faith one might end up having, once a change of heart has occurred. Indeed, Kant explicitly states in the first Preface (RGV, AA 6:9) that the philosophical theologian must leave all decisions concerning the ‘public teachings’ of religion to the biblical theologians. He then makes essentially the same point in the second Preface by explaining that religion within the bounds of bare reason will not adopt “the theoretical standpoint (under which must be included also the technically practical standpoint of instructional method as a doctrine of art)” (RGV, AA 6:12; see CCKR, p. 34). In other words, RGV seeks to leave untouched the question of precisely how any given religious faith will instruct its adherents on which theoretical truth-claims to believe.

The second issue or concern that Drogalis raises about my interpretation is that for Kant

there may be good reasons to believe that reasonable hope can be achieved by a moral agent without [such traditional religious] belief. Specifically, it seems that the command of the moral law alone could be sufficient for such hope.

I am actually quite sympathetic with this suggestion, especially given Kant’s own admission in the Fourth Piece that “illumined persons” typically feel as if they have no need of joining an empirical religious community (RGV, AA 6:158)—clearly an autobiographical confession of his own reluctance to take on the whole package of beliefs upheld by any of the religious communities that he had access to, and so perhaps also an attempt to justify his own decision not to attend church, at least during the second half of his life. Indeed, taken in the strictest sense, Kant’s “ought implies can” principle requires that even an atheist can be good—a view Kant explicitly affirms in his Third Critique discussion of Spinoza and the possibility of a righteous atheist (KU, AA 5:452). But in RGV Kant seems more uncomfortable with the potentially elitist overtones of such a position than with the need to apply this key principle in its strictest form. Instead, he emphasises that human beings are embodied, weak, and therefore inevitably imperfect—an emphasis that I highlight in CCKR and that Rossi’s essay affirms and develops in a helpful way (see below). Kant’s position is not that it is impossible for a righteous atheist to exist, but that such a person must live with such intense cognitive dissonance that despair is the almost certain result—a position he explicitly defends in response to the second difficulty discussed in Subsection C of Section One of the Second Piece. Right up to the end of his life, Kant recognises that, whatever his own biases may be, ordinary human beings need assistance, or at least they need a reasonable hope that assistance is possible. Indeed, as I read the passage at AA 6:158, he seems to be warning that those who, like Kant himself, choose to avoid affiliation with any specific religious congregation, may be fooling themselves when it comes to their own ability to do good independently of any organised religious community.

While Kant certainly recognises that it is possible for us to do good without believing in God’s help, his considered position is that it is not possible for human beings to understand how a person’s entire lifespan could be judged as good, without postulating a divine judge who looks into the depths of our heart’s convictions, and that having such an understanding is the best safeguard against existential despair; for this purpose, Kant on my reading does seriously believe that religious faith is required and that such faith is most effective when practiced in a community setting.

3. Response to Susan Shell

The first issue that Susan Shell raises relates to my understanding of the relationship between morality and religion. This issue is more complex than it might at first seem, primarily because Kant has two distinct meanings for the word ‘religion’: ordinary historical forms of religion (which I call ’empirical’ religion) and special rational religion (which Kant famously argues, as Shell rightly emphasises, must be moral in its orientation). According to my understanding of RGV’s title, empirical religion is always and inevitably ‘clothed’, and the question of whether or not it is a genuine expression of true religion rests primarily on the question of whether the ‘bare’ body underneath is moral. But a secondary reason for this complexity is that the issue cannot be resolved without also referring to the relation between morality and knowledge for Kant (or, as it is often expressed, between freedom and nature). Now, I readily concede that bare religion is grounded in the practical standpoint and therefore is essentially moral. My novel claim is that the clothing metaphor implied by Kant’s use of bloßen (‘bare’) in the title and throughout the book requires moral (or first-experiment) religion to expand to something else; that is, it must become historical (hence the need for what Kant calls his second experiment[11]). This does not mean that religion ceases to be moral, but only that it must have a moralising impact on people’s actions.

This is why I claim that RGV adopts the standpoint of the Third Critique rather than that of the Second Critique. By this I do not mean that religion is not essentially moral; rather, I mean that religion is necessarily manifested in history, and—like everything that adopts the judicial standpoint—it therefore poses to its practitioners a challenge: Will you prioritise the practical or the theoretical aspect of your faith? In a similar way one might ask: Is beauty, according to Kant, grounded in the practical? I think the correct answer would be yes; but only genuine beauty is so grounded, and this grounding (i.e. the form of beauty) exists through a symbol; a full-bodied experience of beauty therefore goes beyond the moral, just as do all genuinely religious objects, beliefs and actions. So by arguing that RGV adopts the judicial standpoint, I intend to make two main points: first, religion necessarily combines practical and theoretical aspects, so it cannot be reduced either to pure dogma or to pure morality; second, the key to authentic religion is to prioritise the moral over the theoretical, not to do away with theoretical commitments altogether. I don’t see how an interpreter could regard Kant’s view of religion as being more necessarily moral than this, without simply drawing an equal sign between the two, thus ignoring the many places where Kant says religion must go beyond bare morality by appealing to concrete symbols and rituals in order to succeed in overcoming the weakness of human embodiment.

Shell seems to portray Kant as presenting us with a stark, ‘either/or’ choice between pure morality that cannot be practiced and historical tradition that cannot be moral. I disagree quite profoundly with this way of reading the overall architectonic of Kant’s System.[12] For Kant, the formal and the material (like the subjective and objective) never appear separately in experience; they can be separated only in abstraction—i.e. in philosophical reflection. In the passage Shell cites from the Metaphysics of Morals Kant is therefore not setting apart two radically different kinds of religion, but is rather distinguishing between two perspectives on one and the same religious practice.[13] The matter of religion is indeed excluded from the form, but only because for Kant the business of philosophy is not to give the matter of anything, but always only the form. Yet this does not mean that a philosopher cannot or should not be religious! To draw such an inference from Kant’s statement would run contrary to crucial aspects of Kant’s own practical philosophy, including first and foremost the postulate of God’s existence. Shell herself confirms this by continuing the quote to reveal that for Kant ‘bare religion’ includes both the moral and the historical. This is the essence of my point, and since in the end Shell seems to affirm the importance of just such a proviso, I am not clear how her final position differs from my own.

The fact that, as Shell puts it, “Kant consistently emphasises that it is only from a practical (or moral) point of view, that belief in a personal God and final purpose of creation, has any workable meaning” fits right in with my reading—at least, what I intended to be arguing—though I do identify passages in RGV where Kant gives more credence to the historical than Shell lets on. But even if, for the sake of argument, I were to grant that there are no such passages, her comment misses my point in identifying Kant’s standpoint in RGV as judicial, which is that the believer in a historical religion (like the practicing artist or art critic or a person in the throes of a sublime experience or someone who reflects purposively upon nature) is adopting a standpoint other than the practical, but is nevertheless duty-bound to appeal to the practical for the purposes of explanation.[14] Because this seems to be precisely the way Shell interprets AA 6:14, I suspect that our disagreement is merely semantic.[15]

Another concern Shell expresses is that I

may push Kant closer to what Storr calls ‘supernaturalism’—a position that Kant in my [Shell’s] view gently but firmly disavowed—than can be textually justified.

I admit that Kant’s Critical Religion, especially Chapter IX (where I hypothetically adopt the stance of the biblical theologian to show how a robust Christian theology could be consistent with RGV’s philosophical theology), could be read as doing this, though I never mention Storr in that book. However, I believe Shell has applied this criticism inappropriately to CCKR. In the Preface, where I issue a call to Kant scholars (and to anyone interested in furthering the goals of genuine religion) to begin reading Kant less through the lens of Fichte’s interpretation and more through the lens of Storr’s, I might give readers the impression that I see Kant as backing Storr’s approach. But that is not my intent; rather, in my comments on the main passage where Kant discusses supernaturalism (RGV, AA 6:154–5), I peg Kant as a pure rationalist, and I claim that the other extremes, naturalism and supernaturalism, are meant as caricatures of Fichte and Storr, respectively. Kant scholars in general tend to read Kant as far closer to Fichte than to Storr, so if Kant’s intention was to promote a position midway between the two, then Kant scholarship needs to move in Storr’s direction in order to grasp what Kant had in mind. That is all I meant by the suggestion that in CCKR I am promoting an interpretation that puts Kant closer to Storr than he has generally been thought to be.

In discussing this issue Shell makes a point that I regard as an outright error—though one that she shares with the vast majority of interpretersincluding my previous account, in Kant’s Critical Religion. She claims that for Kant the supernaturalist “is not a rationalist at all”. But careful attention to the paragraph in question reveals that this is not what Kant says! (I owe this insight, incidentally, to Firestone and Jacobs’ In Defense of Kant’s Religion, which, for all its idiosyncrasies, does do an excellent job of parsing Kant’s taxonomy of different ways of thinking about religion in RGV, AA 6:154–5.) In short, Kant’s grammar requires that his use of “he”, just before introducing the position of the “pure supernaturalist” must refer to “the rationalist”. Kant’s point in this passage (the passage from the Fourth Piece that has received by far the most attention from commentators) is to determine how rationalists (i.e. those who agree that “merely [bloss] natural religion [is] morally necessary”) ought to view revelation. On this issue, the rationalist has three options: revelation may be definitely impossible (= the naturalist, such as Fichte), or possible and thus potentially genuine but not necessary (= the pure rationalist, such as Kant), or necessary together with the necessity of natural religion (= the pure supernaturalist, such as Storr).[16] Shell appears to be forgetting that before the passage she quotes, Kant already contrasts the supernaturalist with the naturalist. So the pure supernaturalist cannot hold the same position as the plain old, run-of-the-mill (philosophically naïve) supernaturalist—i.e. the one who denies natural religion. Note also that bloß in ‘merely natural religion’ does not mean ‘only natural religion’ so much as ‘naked natural religion’ and thus leaves open whether or not something else (namely, revelation) might also serve to clothe it. Shell goes on to claim that “precisely what rationalism as here defined explicitly denies” is the view that “‘universal religion’ of a sort adequate for moral purposes requires belief in the reality of divine revelation”. With all due respect, this claim simply cannot stand up to a close reading of the text: Kant unambiguously states that naturalism entails such an absolute denial, not rationalism per se. (Whether Kant’s use of these same terms, in the passage Shell quotes from Conflict of the Faculties, is identical to his use in RGV would require a deeper study of the text of Conflict than would be appropriate in this context. In any case, Shell’s quote refers only to ‘supernaturalism’, so it is possible that in Conflict Kant is no longer seeking to identify a less naïve, ‘pure’ form of supernaturalism, as he explicitly does in RGV.)

Shell tantalisingly claims that

the ‘unity’ of pure moral religion and historical faith (a unity encompassed by a ‘religion within the boundaries of bare reason’) may have a different character than [CCKR] […] suggests.

I would be curious to know more about how Shell sees that ‘unity’. In any case, I am glad she agrees that Kant was not a full-blown Fichtean in this respect. Again, I certainly agree that he did not want his position to be identified with Storr’s either. Along these lines, I suggest in CCKR that Kant’s much-neglected quip about avoiding a ‘Muslim’ in direct proportion to how many times he has visited Mecca (RGV, AA 6:189n) was probably directed against Storr, given that it is added in the second edition at exactly the point that Storr had referred to (in the first edition text) in the course of complaining about Kant’s unwillingness to take refuge in the sanctuary of the church! In other words, the point of Kant’s quip was not to denigrate Islam but to warn the reader, in direct response to Storr’s complaint, that a religious person who attends church for the wrong reason may be worse than one who avoids attending church altogether.

Another subtle error Shell makes comes in her claim that Kant’s use of the ‘oil and water’ metaphor in AA 6:13 refers to the mixing of the bare religion of reason with historical religion. This is not what Kant actually says. Rather, just before Kant introduces this metaphor, he stops talking about the contrast between moral religion and historical religion and begins a distinct argument, as follows: if one does not combine moral and historical religion properly, then the two will come together like one “genuine/actual religion” and one “ritual worship”. The latter is what can be held together with the moral only temporarily, not historical instantiations of moral religion as such. Kant actually argues (e.g. in Section V of Division One of the Third Piece) that such historical manifestations are absolutely necessary in order for any religion to be actual or genuine, because the form is nothing without the matter.

I shall end my response with a summary that follows the three-point conclusion of Shell’s essay:

1. I agree with Shell that Kant’s necessary link between morality and religion “never ceases to be contained within a grounding frame itself fundamentally moral”; yet Shell herself has stated that she understands Kant’s “religion within the bounds of bare reason” as requiring historical instantiation of moral aims. This is what I think Kant means when he says that morality expands itself to religion. Indeed, the grounding frame is fundamentally moral, just as the grounding frame for beauty and sublimity in the Third Critique is fundamentally moral. Nevertheless, the whole “path from morality to religion”, in its fullness, is not entirely moral, whereas the corresponding “path” in Kant’s strictly moral writings is entirely moral.

2. My “privileging of the ‘judicial’ over the ‘moral/practical’” is not meant to push Kant from a mediating position “toward ‘supernaturalism’”. Rather, the judicial is the home standpoint of critique itself, and its purpose is to ensure that, when we interpret experiences that go beyond mere scientific knowledge or moral action, we must give primacy to the practical, just as Shell rightly insists. My point is only that, to give primacy to the practical is not to reduce the religious to the moral (à la Fichte), any more than Kant’s theory of beauty as the symbol of the morally good reduces a judgement of beauty to a mere ought.

3. I certainly do not want to limit the reach of Kantian religion in the way Shell worries that I might be. Indeed, I do not see how such a limitation follows from my stance on RGV adopting the “judicial standpoint”. Kant has plenty to say about non-Christian religions, and while much of it may not be fair or accurate, especially from our modern point of view, in some places he does seem to bend over backwards to allow that historical instantiations of genuine religion do occur in historical faiths other than Christianity. Kant’s moral philosophy is sometimes criticised for being merely a translation of Christian ethics into philosophical terms, so I think the reading of RGV that pushes it more toward pure morality is far more limited in its reach—far less likely to make Kant attractive to those who associate themselves with religious traditions other than the Christian—than a reading like mine, which portrays Kantian religion as being open to a variety of historical instantiations, provided they serve as vehicles for the true matter of religion, which is moral.

4. Response to Philip Rossi

Philip Rossi’s essay consists mostly of a robust and detailed engagement with three topics that he sees arising out of the arguments I advance in CCKR. His goal is more to extend those arguments than to criticise or correct them. While I am very grateful that he saw fit to use his essay for this purpose, my agreement with nearly all of the points he makes leaves me with relatively less to say in response to his essay, compared to the foregoing three—an ironic result, as his is by far the longest essay and the one that delves into the issues in the most depth. If I were to rehearse his arguments and state my reasons for agreeing with them, I probably would not do justice to the nuanced insights he expresses. Instead, therefore, I shall limit my response to brief comments on each of the three topics he discusses, together with rejoinders to the several points of mild criticism that he advances. I readily agree at the outset with Rossi’s main criticism, that CCKR fails to develop “the full import” that many of my interpretations and arguments “have for Kant’s account of the philosophical and human significance of religion”. In order to bring out the import of such main themes I would have either had to make the book far less comprehensive, by not commenting on every sentence of Kant’s text, or else make the book two or three times longer than it already is. My hope, as expressed in the Preface, is that other insightful interpreters, such as Rossi and the other critics participating in the current exchange, will take up this challenge by developing such themes further than I was able to do—just as Rossi has begun to do in his essay.

I could not be more pleased that one of the commentators for Critique has drawn attention to my emphasis on the role of embodiment in Kant’s theory of religion. This emphasis is integrally bound up with my claim that RGV adopts the judicial standpoint—a point that Shell challenges (see above), despite the fact that in some of her own publications she has admirably highlighted the importance of embodiment for Kant. Rossi laments at one point that

Kant’s focus on the internal dynamics of an agent’s moral conversion [in the Second Piece] does not seem to provide conceptual resources for dealing adequately with […] what results […] from the performance of evil by human agents […].

While I fully agree that Kant’s arguments in the Second Piece provide no such resources, my response (in Kant’s defence) would be that the Second Piece is not the correct place to search for such resources. Instead, Kant provides the outlines of just the resources Rossi is looking for in the Third Piece, especially Section IV of Division One, where he puts forward four requirements for the true church. But this (Kant’s) answer to Rossi’s question is purely formal, and Rossi makes it clear that such an answer falls short of the material guidelines he seeks. Thus, at one point Rossi asks why Kant did not provide a set of guidelines that can be adopted by “a concrete human historical community, to overcome evil?” On my reading, this omission was very intentional: Kant avoided giving such guidelines because doing so is the task of the biblical theologian, not of the philosophical theologian. Some of Rossi’s further questions, however, could be read as relating more to the need for formal guidelines—i.e. guidelines that are philosophical in the sense of being applicable to any given historical religious tradition.

My response to the latter is that Kant does give formal guidelines, if we read the details of the Third and Fourth Pieces with sufficient care. Rossi poses the question:

[E]ven though Palmquist does not directly address questions of this kind in CCKR, do his discussions nonetheless at least suggest resources within Kant’s text that might be useful in response to them?

Rossi rightly answers in the affirmative, citing three helpful examples. My further response would be that I did intend to be addressing precisely the kinds of questions that Rossi raises, regarding the practical applicability (for real, empirical religious communities) of the formal guidelines Kant provides; so, even if Rossi did not think I meant to be doing so, I am pleased that he found three examples of how CCKR succeeded in that goal. My only comment here is that the attentive reader of CCKR will find many other practical guidelines in addition to the three that Rossi highlights. For example, in addition to constructive guidance on how to understand and conduct the ritual of communion, Kant offers similar guidelines regarding baptism, church attendance and public prayer. Still, if Rossi means that in such passages CCKR only scratches the surface of what is possible, then I fully agree. With this in mind, I would welcome the production of a book (perhaps an anthology of essays by like-minded interpreters?) that explores in greater depth a Kantian theology of ritual as embodied morality.

Rossi’s second topic, my portrayal of Kant’s way of engaging with theology and theologians as one of “respectful seriousness”, is a point that has recently been challenged by other commentators. In several publications Pasternack, for example, has accused Kant of prevarication by claiming to let biblical theologians have jurisdiction over their own field while actually (allegedly) devoting much of RGV to the destruction of various classical theological dogmas. Rossi’s affirmation of the alternative I propose is a refreshing confirmation of a tendency I have noticed: the more a commentator knows about the theological tradition, the more likely that commentator is to recognize in RGV an attitude of respectful engagement rather than elitist pontification, of the sort Firestone and Jacobs (2008:202) also impute to Kant. I fully agree with Rossi’s point in his third note, that “the reception of Kant’s Religion […] had a formative role in the articulation of […] the ‘secular’” that has shaped Western culture so radically over the past two centuries. Here again Rossi claims I do not deal with this theme in CCKR; while I would agree that my comments are quite sketchy and merit further development, I do not believe CCKR ignores the topic altogether. Rather, in discussing AA 6:154–5, for example, I have discussed in some detail the huge impact that the contrast between Fichte (the secular interpreter) and Storr (the religious interpreter) had on the reception of RGV and have claimed that it was through his desire to offer respectful clarifications of Storr’s feedback that Kant chose to prepare a second edition just one year after RGV’s initial publication. All in all, the point Rossi is rightly acknowledging could be expressed as follows: far from seeking to destroy historical religion or to sideline it (as Shell’s essay seems to suggest), Kant saw RGV as an attempt to bridge Lessing’s famous ‘ditch’ between historical and rational religion—this, again, being a good reason to see RGV as adopting the bridging standpoint of the Third Critique rather than the exclusively moral standpoint of the Second Critique. In this case Rossi’s extension of one of CCKR’s themes could be expressed by saying that interested readers of my book may wish to come together to produce an anthology called A Kantian Theology of Interfaith Dialogue.

I am profoundly appreciative of the fact that Rossi chose as his third topic the status of religious experience for Kant. Indeed, Rossi goes above and beyond the call of duty by responding not only to the views on Critical mysticism advanced in CCKR, but by also informing readers of the earlier work I did on the same topic, scattered throughout various chapters of Kant’s Critical Religion.[17] One gloss I especially appreciate in Rossi’s extension of my position is his observation that Critical mysticism is integrally bound up with the embodied nature of religion for Kant. Rossi could not be more correct to interpret my references to Critical mysticism as motivated, at least in part, by Kant’s emphasis on “the overall unitary intent of [his] critical project”. I fully agree that any given instance of Critical mysticism, as played out in the practice of an actual historical religious tradition, would need to be regarded “as an anthropological enterprise inescapably rooted in the finitude of embodied human reason”. That I said little about how this might work is a valid observation about CCKR, and my best rejoinder is to say that I hope to fill that gap in the projected fourth volume of my Kant’s System of Perspectives series, which (following the third volume, Kant’s Critical Science, which is currently in process) is tentatively entitled Kant’s Critical Anthropology. Here I would like to echo what I take to be one of Rossi’s most profound insights, that if the historical development of human religions follows Kant’s lead, then we will come to recognise that post-modern religiosity must be post-secular as well as (in certain traditional senses of the word) post-religious. That is, we must move beyond both Fichte and Storr.

Near the end of his introductory remarks, Rossi briefly mentions two disputes he has with the positions I defend, which he does not develop further in his essay. The first concerns “the social origin and sources of ‘radical evil’”, while the second concerns “the relationship between the ‘church’ and the political order as both participant [sic] in the human task of overcoming evil”. My rejoinder to the first point has already been presented in my response to Green above. On the second point, there is indeed much to say and I would have been happy to engage with one of the critics on this point. Many Kant interpreters—and as I read Rossi’s fine work on RGV, he is one of them—think that for Kant politics holds the key to the final destiny of the human race while religion is a temporary step on the way to that goal. In several publications prior to CCKR, by contrast, I have argued that Kant actually prioritises politics and religion in precisely the opposite order.[18] But as I have defended that position in some detail elsewhere and Rossi chose not to engage me on this key difference between our readings of Kant, I shall pass up this opportunity to supplement my previous arguments.

Received: 6 March 2017.


[1] A version of some parts of this paper were previously presented at an Author Meets Critics session at the APA Pacific Division Meeting held in San Francisco on March 31st, 2016. My thanks to those who provided feedback during that session.

[2] Although CCKR bears a 2016 copyright date, it was actually released as an eBook in October 2015 and appeared in print in mid-December 2015.

[3] I defend this novel translation at length in Palmquist (2015).

[4] A similar example worth noting is Kant’s claim in RGV, AA 6:62—in a paragraph that begins by referring to “the Son of God” as the person who “is presented [in the Bible] as having assumed human nature”, this being one of the rhetorical formulations Kant uses to refer to Jesus without actually naming him—that a person who has “practical faith” in Jesus “is entitled to regard himself as the One [denjenigen] who is an object not unworthy of divine satisfaction”. For a comparison of this reading with that of other English translators, see CCKR, p. 166n.60. The other translations obscure what I take to be Kant’s main point: religious believers are justified in hoping that God will see in them the perfection of Jesus (“the One” who is “the Son of God”) only if they adopt Jesus’s conviction as their own.

[5] See G. C. Storr, Annotationes quaedam theologicae ad philosophicam Kantii de religione doctrinam [Various Theological Notes on Kant’s Philosophical Doctrine of Religion] (Tübingen: Bornium, 1793). The full e-text is available online here.

[6] See Palmquist (2008).

[7] See especially Palmquist (2010a) and (2012).

[8] See e.g. RGV, AA 6:21–2, where Kant appears to claim that if we attempt to trace the temporal origin of a person’s good or evil character, then we must go all the way back to a time before birth (see CCKR, pp. 50–1).

[9] On Kierkegaard’s debt to Kant, see Ronald Green’s highly instructive books (Green 1988, 2011). For a detailed critique of these works and a defence of the claim that Kierkegaard borrowed even more from Kant than Green believes he did, see Palmquist (2016a). Whereas Green tends to read RGV with the assumption that Kierkegaard’s interpretation of Kant is accurate, I argue that Kierkegaard skews Kant’s intended meaning to fit his own purposes: he ends up attributing to Kant positions that are more extreme than the text of RGV merits, and then defending a position that is actually closer to Kant’s authentic position than the one Kierkegaard reads into the text.

[10] In the second edition (and so also in all previous English translations of RGV) the First Piece has only four numbered sections, followed by a lengthy General Comment, to which Kant appended a long additional paragraph in the second edition. In CCKR, pp. 120 and 144–5, I argue that this was very likely a printing error, for a variety of reasons that are too complex to rehearse here. I argue that the so-called first parergon consists only of the paragraph that was newly added, and that Kant never intended the passage that he had labeled Section V in the first edition to be regarded as the first parergon. As this makes a huge difference to the question of how important Kant’s argument in that section is to the book’s overall religious system, I refer to the section in question as Section V, not as the General Comment.

[11] RGV, AA 6:12. Lawrence Pasternack (2017) has recently called into question the whole notion that there are two “experiments” operating in RGV. In his article, he argues that the whole problem of how to distinguish and locate the two experiments that Kant refers to in the Prefaces to RGV is based on an unfortunate mistranslation of Kant’s term Versuch. He thinks Kant’s use of Versuch in RGV, AA 6:12 refers merely to the fact that he was now issuing a second edition, not that the book itself somehow contains two distinct experiments. Having been the most outspoken proponent of the usefulness of distinguishing between two experiments as a hermeneutic tool for interpreting Kant’s text, I plan to respond in detail to Pasternack’s provocative accusation that my use of this tool since the mid-1980s constitutes nothing but an “early misstep” (2017:107) in our contemporary attempts to interpret RGV and should therefore be entirely ignored from now on. In a nutshell, my response will proceed as follows. First, Kant’s use of Versuch is integrally bound up with an experimental method that Kant uses consistently, throughout his mature writings, so using “experiment” as a standard translation of Versuch is wholly justified. Second, in the same passage of the second Preface where Kant refers to the “second experiment”, he also employs an analogy of two concentric circles, which all who employ the “two experiments” terminology recognise as being intimately related to the latter. If Pasternack’s claim regarding the meaning of Versuch is correct, then the concentric circles should also relate in some way to the two editions rather than to the arguments contained in the book itself; but this is obviously not the case. Third, even if one were to de-link Kant’s use of Versuch from the concentric circles metaphor, so that the former could refer to the two editions even though the latter obviously does not, Kant’s account of the concentric circles metaphor on its own already implies that he intends to conduct a two-step inquiry that could adequately be described using the English word “experiment” for each step. Fourth, Pasternack’s argument exhibits a rather naïve approach to hermeneutics in general as well as to Kantian hermeneutics in particular. The key issue is not so much whether Kant intended to tell his readers that they should interpret his text through the use of two distinct experiments that can be mapped onto the concentric circles that he mentions in the same passage. Rather, what matters from a hermeneutic perspective is whether the use of this interpretative device has advanced our understanding of the text; and it is my contention that in this respect the two-experiments methodology is still alive and well, as Shell’s essay aptly illustrates.

[12] See Palmquist (1993a), especially Chs II and III.

[13] Throughout RGV Kant repeatedly adopts the same two-perspective approach to resolving the issues he discusses, always concluding that the solution is to give priority to the practical, not to exclude the theoretical altogether. Thus, for example, his approach to resolving the “remarkable antinomy” of “sanctifying faith” (RGV, AA 6:116–21) is to show that, viewed theoretically, the antinomy is irresolvable, while viewed practically, the two opposing approaches amount to the same thing. Along similar lines, when considering the relative importance of “statutory laws” and “moral conviction” in the Fourth Piece, he exclaims (p. 179): “So much depends, when one wants to bind two good things, on the order in which one binds them!”

[14] For one of the many examples of Kant’s emphasis in the Third Critique on establishing the proper priority between the theoretical and the practical, see KU §79, AA 5:417.

[15] Speaking of semantics, Shell follows di Giovanni’s translation of Kant’s “eigentliche Religion” (RGV, AA 6:12) as “genuine religion”; Pluhar has “religion proper”, which I revise as “actual religion”. While I agree with Shell that such religion “has meaning […] only from a moral-practical perspective,” I would say “has meaning only when interpreted from a moral-practical standpoint”. What Shell fails to mention here, but what is (rightly) assumed by her account of Kant’s position on the two experiments (cf. note 11, above), is that this “genuine religion” must be linked to some historical tradition. That is what makes it “actual”! Kant’s use of eigentlich emphasises that he is referring to an empirical religion that manifests the moral core.

[16] The latter is the position Storr adopts in his book-length review of the first edition of RGV (see note 5): he grants that Kant is correct about the absolute necessity of genuine religion to promote a moral lifestyle; but he thinks Kant fails to recognise that there are also certain historical elements of true religion that are equally necessary. Storr makes clear that he is a firm believer in Kant’s Critical philosophy—at least, the parts that had been published prior to RGV. So it is highly unlikely that Kant would have sought to alienate him by putting him in the naïve “supernaturalist” camp that he mentions before and in total contrast to the “rationalist” camp. The supernaturalist, unlike the pure supernaturalist, regards divine revelation as absolutely necessary without the need for any morally-based natural religion—a position that would not apply to Storr.

[17] After the appearance of CCKR I revised the relevant chapters of Kant’s Critical Religion, supplementing and updating that material, and have published it in the form of a new book (Palmquist 2017).

[18] For my interpretation of Kant’s position on the role of religion vis-à-vis politics, see Palmquist (1994, 2006, 2010b, 2013, 2016b). For a detailed account of my own position on religion and politics, see Palmquist (1993b).


Firestone, C. & N. Jacobs (2008), In Defense of Kant’s ‘Religion’ (Bloomington and Indianapolis: Indiana University Press).

Green, R. (1988), Kierkegaard and Kant: The Hidden Debt (Albany, NY: SUNY Press).

——— (2011), Kant and Kierkegaard on Time and Eternity (Macon, GA: Mercer University Press).

Palmquist, S. (1993a) Kant’s System of Perspectives: An Architectonic Interpretation of the Critical philosophy (Lanham, MD: University of America Press).

——— (1993b), Biblical Theocracy: A Vision of the Biblical Foundations for a Christian Political Philosophy (Hong Kong: Philopsychy Press).

——— (1994), ‘”The Kingdom of God is at Hand!” (Did Kant really say that?)’, History of Philosophy Quarterly 11(4): 421–37.

——— (2000), Kant’s Critical Religion: Volume Two of Kant’s System of Perspectives (Aldershot: Ashgate).

——— (2006), ‘Philosophers in the Public Square: A Religious Resolution of Kant’s Conflict of the Faculties‘, in S. Palmquist and C. Firestone (eds), Kant and the New Philosophy of Religion (Bloomington: Indiana University Press), pp. 230–54.

——— (2008), ‘Kant’s Quasi-Transcendental Argument for a Necessary and Universal Evil Propensity in Human Nature’, The Southern Journal of Philosophy 46(2): 261–97.

——— (2010a), ‘Kant’s Ethics of Grace: Perspectival Solutions to the Moral Difficulties with Divine Assistance’, The Journal of Religion 90(4): 530–53.

——— (2010b), ‘Theocratic Friendship as the Key to Kantian Church Government’, in M. Lee (ed.), Proceedings of the Twenty-Second World Congress of Philosophy, vol. 45 (Seoul: Korean Philosophical Association), pp. 251–60.

——— (2012), ‘Could Kant’s Jesus Be God?’, International Philosophical Quarterly 52(4), pp. 421–37.

——— (2013), ‘A Daoist Model for a Kantian Church’, Comparative Philosophy 4(1): 67–89.

——— (2015), ‘What is Kantian Gesinnung? On the Priority of Volition over Metaphysics and Psychology in Kant’s Religion’Kantian Review 20(2): 235–64.

——— (2016a), ‘The Paradox of Inwardness in Kant and Kierkegaard: Ronald Green’s Legacy in the Philosophy of Religion’, Journal of Religious Ethics 44(4): 738–51.

——— (2016b), ‘Kantian Theocracy as a Non-Political Path to the Politics of Peace’, Jian Dao 46: 155–75.

——— (2017), Baring All in Reason’s Light: An Exposition and Defense of Kant’s Critical Mysticism (Aurora, CO: Noesis Press).

Pasternack, L. (2017), ‘The “Two Experiments” of Kant’s Religion: Dismantling the Conundrum’, Kantian Review 22(1): 107–31.

© Stephen R. Palmquist, 2017.

Stephen Palmquist is Professor of Philosophy at Hong Kong Baptist University. He is the author of multiple books on Kant’s philosophy. His new book Baring All in Reason’s Light: An Exposition and Defense of Kant’s Critical Mysticism is forthcoming from Noesis Press.