A Kantian Response to the Problem of Evil: Living in the Moral World

Abstract

James Sterba has presented a powerful and existentially sincere form of the problem of evil, arguing that it is logically impossible for God to exist, given that there are powerful moral requirements to prevent evil, where one can, and that these requirements would bind an all-powerful and good God, who would indeed be able to prevent such evil. The ‘Kantian’ argument that I set out, if accepted, would undermine the following stage of Sterba’s argument: Significant and especially horrendous evil consequences of immoral actions do obtain all around us, which, if God exists, would have to be through his permission. The Kantian argument will hold that we are able to believe that, in some sense, such horrendous evil consequences do not really obtain, although they appear to. The claim is not that the Kantian argument is ‘persuasive’, but that if some Kantian assumptions are granted, we do have a response to Sterba, which throws open a different way of looking at things. I conclude with some more informal reflections on what we might take away from the Kantian argument, even if we do not accept the deep assumptions, or the progression of the argument. I will not worry too much about demonstrating that this is a ‘correct reading’ of Kant, although I think it is.

1. Overview

A standard line of critique of analytical philosophy of religion, from various Wittgensteinian and post-Kantian traditions, is that no one has ever really been brought to religious conviction, or dissuaded from it, because of a formal argument. Sterba presents himself as a striking counter-example, insisting that he would give up his atheism, if his argument can be demonstrated to be faulty. Having formerly been religious, indeed, in a religious order, Sterba writes:

My commitment to atheism is only as strong as the soundness and validity of my argument. Undercut my argument and proof, at least in my case, no more atheist.1

As set out by Sterba, this is his argument:

• There is an all good, all powerful God. (This is assumed for the sake of argument by both Mackie and Plantinga).
• If there is an all good, all powerful God, then necessarily he would be adhering to Moral Evil Prevention Requirements I–III.
• If God were adhering to Moral Evil Prevention Requirements I–III, then necessarily significant and especially horrendous evil consequences of immoral actions would not be obtaining through what would have to be his permission.
• Significant and especially horrendous evil consequences of immoral actions do obtain all around us, which, if God exists, would have to be through his permission. (This is assumed by both Mackie and Plantinga).
• Therefore, it is not the case that there is an all good, all powerful God.
  The three Moral Evil Prevention requirements in turn, quoting Sterba, are as follows:
  • Moral evil prevention requirement I
    Prevent rather than permit significant and especially horrendous evil consequences of immoral actions without violating anyone’s rights (a good to which we have a right) when that can easily be done.
  • Moral evil prevention requirement II
    Do not permit rather than prevent significant and especially horrendous evil consequences of immoral actions simply to provide other rational beings with goods they would morally prefer not to have.
  • Moral evil prevention requirement III
    Do not permit rather than prevent significant and especially horrendous evil consequences of immoral actions (which would violate someone’s rights) in order to provide such goods when there are countless morally unobjectionable ways of providing those goods.2

The ‘Kantian’ argument that I am about to unfurl agrees with stages 1, 2 and 3 of Sterba’s argument above, and accepts the Moral Evil Prevention Requirements. Where the disagreement occurs is at stage 4, specifically the words I have italicised below:

Significant and especially horrendous evil consequences of immoral actions do obtain all around us, which, if God exists, would have to be through his permission

The Kantian argument will hold that we are able to believe that, in some sense, such horrendous evil consequences do not really obtain, although they appear to. I will now give a skeleton overview of this argument, which I will then fill-out.

• It ought to be the case that being moral is met with proportionate happiness.
• Such a world is one where the Moral Evil Prevention Requirements I–III hold.
• Where knowledge is limited, practical reason enjoys freedom to hold-things-for-true.
• In the world that appears, being moral is not met with proportionate happiness, such that in the world that appears, the Moral Evil Prevention Requirements I–III do not obtain.
• What appears is not fundamental.
• Practical reason is entitled to believe in a moral realm (from 3), wherein the Moral Evil Prevention Requirements I–III do obtain. Such a fundamental moral realm is one where God does adhere to the Moral Evil Prevention Requirements I–III.
• Therefore, the argument from the Moral Evil Prevention Requirements I–III does not provide a necessary argument for the non-existence of God.

I do not pretend the argument is ‘persuasive’. Certainly, I do not ‘believe’ it. Let me say now: I am confident that this is not the counter-argument that will shift Sterba away from his atheism. But, it does seem to me that if some Kantian assumptions are granted, this does provide an interesting response to Sterba, which throws open a different way of looking at things. Perhaps, a strange and quixotic and utterly uncompelling way, but, a way nonetheless.

The question arises, here, of where the burden of proof lies. A natural reaction is to affirm that it lies with the Kantian (on my interpretation), to demonstrate that the rather outlandish assumptions apply, or, that they are at all plausible. Something interesting happens, though, when we consider the strength, scope, and ambition of Sterba’s proposed argument. The point is this: Sterba does not just say that the existence of God is extremely unlikely, or improbable, given the existence of evil. He affirms that it is logically impossible. The strength of this claim is what may shift the burden of proof, away from the Kantian, back to Sterba, or his defender. There seem to me to be two main lines of response available to Sterba, or to a philosopher defending Sterba’s position:

(i)

To show that the required Kantian ‘outlandish’ assumptions involve affirming something that is indeed logically impossible. In this way, the claim that the existence of God is logically impossible will still stand.

(ii)

To qualify and nuance the scope of the argument for the logical impossibility of the existence of God, by specifying that it applies only to the world on a more or less ‘common—sense’ conception of what the world consists of, where apparent spatio-temporal facts and events are, more—or-less, as they appear to be.

Perhaps we might embody the second-approach by adding a further requirement, over and above the three ‘moral evil prevention’ requirements specified by Sterba. We could call this the ‘common sense preservation requirement’:

The world as it appears is more-or-less the world as it fundamentally is.

The addition of something like this ‘common-sense preservation requirement’ is interesting, as it would explicitly limit the scope of the ‘logical impossibility’ argument, in a way that is particularly germane in the context of religious belief. This is because it is a pronounced and distinctive feature of much religious belief to claim that the world is indeed not as it straightforwardly appears to be.

To insist on the ‘common-sense preservation requirement’ is, therefore, not a neutral thing to do, in relation to vast swathes of religious belief. It is to stack matters against the religious believer, and to denature the holistic nature of some religious belief, by lopping-off a vital element and dimension of how believers describe the world that presents itself to them. In this article, I explore how this works in relation to a Kantian world-view (on one interpretation), but a similar structural issue may arise in relation to other worldviews and fundamental metaphysical pictures: perhaps in some strands of Hinduism and Buddhism, which regard the world as it presents itself as an appearance of something more fundamental; something like this question may also arise in relation to elements of Platonism, which are present in Christian thinkers such as Augustine and Aquinas, where spatio-temporal phenomena are not the definitive ‘version’ of what is really going on.

I suspect that those more versed in the details and depth of Sterba’s arguments, across a number of works,3 and in the literature arising from Sterba’s work, will have plenty to say in relation to this structural challenge, if not, so much, in relation to the Kantian specifics. The structural challenge, to summarise, is this:

To what extent does (i) Sterba’s logical argument for the impossibility of the existence of God, and (ii) similar logical arguments, rely upon the ‘common-sense preservation requirement’, in particular in relation to stage 4 of Sterba’s argument? Stage 4 of the argument, we recall involves the following claim:

Significant and especially horrendous evil consequences of immoral actions do obtain all around us, which, if God exists, would have to be through his permission.

The possible answers, as I understand the matter, fall into two camps:

(a)

Such logical arguments do assume the common-sense preservation requirement.

(b)

Such logical arguments do not assume the common-sense preservation requirement.

In either case, we have an interesting result. If (a), we face the problem that this is not a neutral requirement, particularly in relation to religious belief. If (b), how are we to go forward? Do we need, in each particular case, to show that the violation of the common-sense preservation requirement amounts to a logical impossibility? This is quite a different project from the one we started out with. The other alternative might be to abandon the claim that this is so austerely an argument about logical impossibility, but that we need not insist on ‘common-sense’ in a way that so flatly excludes much religious belief and instinct. Perhaps we could frame something along the following lines:

(c)

Such logical arguments work alongside most plausible and non-extravagant ontologies, although may not work with more ‘extreme’ or ‘outlandish’ metaphysical positions.

With (c), of course, the hard-work has only just begun, of specifying the bounds and limits of a plausible and non-extravagant ontology.

An even more radical response would be to give up on the claim to demonstrate the ‘logical impossibility’ of the existence of God, and ‘merely’ to argue that the existence of God is impossible, given widely held assumptions about the epistemological status of statements about reality that are based upon how the world appears to us, which, although plausible, are not logically indubitable. The burden of proof is then thrown back onto the Kantian (or to whoever is defending a perspective that goes beyond ‘common-sense’). But, this would come at some cost to anyone, including Sterba, who is eager to insist on the logically impossibility of the existence of God. In relation to this distinctive claim, response (d) amounts to a significant concession, and a retreat, albeit a dignified one.

With these wider considerations behind us, I will now set out the specifics of one possible version of this wider challenge, arising from an interpretation of Kant, which I have set out extensively elsewhere, and draw upon here. After setting out some of the wider Kantian framework in a bit more detail, I will ask what the main assumptions are, if one is to accept the Kantian argument. On the surface, it might seem that the most important assumption is a metaphysically extravagant world-view (noumena and phenomena, and so on). I will suggest that this is not the case: that the main assumptions are more epistemological, about the status of a certain type of transcendental argument, and the freedom of practical reason to make its own moves, when theoretical reason is in the dark. This complexity may make the various possible ‘responses’ outlined above even more complicated: of demonstrating that outlandish metaphysical positions are logically impossible (strategy b), or of articulating what the limits are for a ‘non-outlandish’ metaphysical (strategies b and c). This is because the metaphysically ‘outlandish’ viewpoints are better motivated than they might initially appear. They are not constructed upon some claims to supernatural metaphysical intuition or inspiration, but are grounded upon a type of epistemic humility, combined with a particular understanding of what constitutes the purposes and possibilities of belief-formation. I will conclude with some more informal reflections on what we might take away from the Kantian argument, even if we do not accept the deep assumptions, or the progression of the argument. I will not worry too much about demonstrating that this is a ‘correct reading’ of Kant, although I think it is. I will offer some grounds for finding such an argument in Kant, with footnotes gesturing to more evidence, for those who care enough.

2. The Kantian Picture

In this short section, I will set out a sweeping interpretative picture of how I read Kant, or, the Kant needed to give us the argument sketched out above. This is not the place to offer an extensive defence of this account. This I have attempted in various publications.4 But it might be helpful, here, to locate my claims in the wider realm of ‘Kant studies’: I associate myself with a recently resurgent ‘metaphysical’ reading of Kant, which understands Kant as having more substantive commitments—ethical, ontological and theological—than more deflationary commentators thought possible or proper for Kant.5 Within this movement, I have a particular interest in Kant’s theological commitments, especially with respect to human freedom as it relates to divine action. A number of recent commentators have also been interested in Kant’s theological convictions. My most distinctive claim, perhaps, is to affirm that although Kant believes in God and in a meaningful conception of transcendence, he consciously diverges from Christianity as he would have received it. My grounds for saying this have been that Kant avoids the categories of revelation, tradition, and authority, as well as denying that God can be the final or efficient cause of human action. Kant also rejects the traditional claim that loving and knowing God is our highest good. I have found this to be a more productive lens for appreciating Kant than approaches which judge Kant to be a more-or-less lousy Christian of some stripe (with different emphases on Lutheranism or a more Platonically infused theological rationalism).6

Having marked out the terrain a little, I will now move at a bracing pace through my main interpretative headlines.

For all the undoubted difficulty of Kant’s texts, a firm grasp of four principles serves to illuminate the fundamental contours of his ‘critical’ thinking (broadly speaking, Kant’s thought after 1770):

(1)

The ‘inner value’ of the world is freedom, and nothing else. Freedom means: setting ends for yourself, without being impacted upon by anything external to you. Other things may be admirable, or impressive, but they lack this value.

(2)

Reason is a larger category than knowledge. There is far more that we can have rational beliefs about, than we can know about. This means that Kant would not recognise the ultimate validity of a debate between ‘faith and reason’: because faith, religious belief (Glaube), is entirely within the stretch of reason, even though it goes beyond the bounds of knowledge.

(3)

Thinking about the ‘conditions of possibility’ of something can expand your knowledge, and your set of rational beliefs. Consider: if you know something, or have a rational belief about something, you can then ask, ‘what else must be the case, or, what else do I need to believe, in order to make this possible?’. You then have warrant for affirming, for ‘holding-for-true’, whatever comes out of this conceptual investigation. You might not know it, but, as we have seen, from the second principle, knowledge is not everything. There is a caveat here: anything you come up with must not contradict something that you know. But that is a fairly minimal test, precisely because we do not know very much.

(4)

Kant thinks in a way that is big and binary. His philosophy tends to lead us to a crossroads, where he finds that everything (created and uncreated) is either this way or that way, where what is offered is an entire package, a whole and encompassing world-view. In relation to the question of morality and freedom the options are these: we either live in a ‘moral world’ where freedom is possible, or, we live in a world of mechanistic determinism, where freedom and morality are impossible. The former world has value, the latter world is a ‘mere desert’, entirely without value. Kant finds that we can, indeed, must rationally believe that the entire and whole world is undergirded by freedom, and not mechanism, and so, that it is a world with value.

Kant’s ‘transcendental idealism’ arises from Kant’s ability to affirm such a world undergirded by freedom. Putting it briskly, the idea is this: if space and time are features of the world in itself, and directly created by God, they go ‘all the way down’ into reality, and we are contained within them. This is bad news for freedom, because, Kant believes, space and time are through and through deterministic in ways described by Newtonian mechanism. If, then, space and time are features of our reception of the world, and not in the world in itself, this is good news for the possibility of freedom. It enables us to believe in freedom and morality. There is a conceptual space for fundamental reality to be quite other than it appears to be. Here, we can recall the third principle: something providing the ‘conditions of possibility’ of something is itself permitted to provide warrant for a belief, if it does not contradict what we know, because, reason is a larger category than knowledge (the second principle).

For Kant, the ‘noumenal realm’ is the ground of the world of appearances (‘phenomena’—‘that which appears’), whereby ‘noumenal objects’ affect us. These noumenal objects bring about our experience, which experience is always mediated through our forms of intuition, space and time. Although we understand that all our experience is always on this side of this mediation, coming downstream of how we receive the world, we also understand that it is dependent upon the world as it is in itself, even though we cannot know anything substantial about this world, except that it does indeed ground our experience. This interpretation of transcendental idealism is known in the literature as the ‘noumenal-affection’ account.

‘Transcendental idealism’, on this interpretation, has three dimensions. First of all, it sets the limits to knowledge (which, as we have seen, is a more constrained category than reason). Secondly, within those limits, knowledge is made secure. Thirdly, it opens up possibilities for rational thinking beyond the limits of knowledge. That is to say, transcendental idealism retains epistemic humility about what we can know, whilst opening up the possibility that the way things are is fundamentally different from the way things appear to be. Things appear to be determined, but this is just an appearance. We can believe in freedom without epistemic irresponsibility, precisely because belief in freedom is a ‘condition of possibility’ of morality.

Once we are properly equipped with such an interpretation of transcendental idealism and noumenal freedom, we are ready to understand the type of transcendence that is really at work in Kant’s philosophical religiosity. The history of human actions, as with everything that appears, is the appearance of that which is fundamentally non-spatial and non-temporal, where there is no sense in which we move towards or further away from the noumenal dimension wherein morality and freedom resides. For this reason, Kant emphasises the invisibility of moral action, stating, for example in the Groundwork, we can never recognise whether an action is actually grounded on conformity with the moral law, rather than happening to coincide with it.

This position can be understood as delivered by a combination of the four principles set out above. Kant asks, ‘what sort of entire world is the condition of possibility of the freedom that is the inner-value of the world?’ (principle four). Transcendental idealism is part of Kant’s answer. Because it does not contradict what we know, but only goes beyond it, it is rational to believe it (principle two). Because, Kant thinks, it is the only entire world, the only way the world can be, that can sustain such an ambitious conception of freedom (principle one), we should believe in it (principle three). Furthermore, it is not rational to believe more than is required, in order to sustain the possibility of the freedom that is the inner value of the world. In these two constraints (‘we should believe’ and ‘believe no more than is required’) lie Kant’s epistemic discipline and humility.

With this in place, we are able to understand Kant’s conception of autonomy. Autonomy never appears, for Kant, in space and time. Autonomy is only possible if there is a realm of noumenal freedom, where rational will is able to will itself, in its activity of end-setting. We note, then, that the possibility of autonomy is itself the great philosophical-religious hope of Kant’s whole system. Only if there is a dimension of reality beyond mechanism, is end-setting, and so autonomy possible. The alternative to a moral world with freedom, is a universe without end-setting, and without freedom, which Kant tells us repeatedly, would be a sort of ‘desert’ with no ‘inner value’. Believing in the possibility of autonomy already, and in itself, leans into what we might call religious hope: the hope that things are not as they seem, and that there is a dimension to reality which is saturated with reason, wherein which we find our ‘proper selves’. The result of a fully autonomous Kingdom of Ends, where everyone acts harmoniously and universally, would be the ‘happiness’ of the highest good, where ‘everything goes according to the wish and will’ of every ‘rational being in the [moral] world’ (CPrR, 5: 124).

A way to inflect all this in relation to Sterba’s argument would be to say that (‘for all we know’) it is only an appearance that we live in a reality where the Moral Evil Prevention Requirements I–III do not obtain. I realise there is quite a lot going on in this claim. In what sense might the Moral Evil Prevention Requirements not obtain in fundamental reality, given that the world-that-appears is nothing else than an appearance (a well-founded phenomenon) of fundamental reality? I will sketch two possible Kantian responses: one apophatic, the other brave (with the potential to cause offence). Again, I am not ‘promoting’ the Kantian response to Sterba, and I state here my clear preference for the apophatic line of response. I call the response ‘apophatic’, rather than ‘agnostic’, because it arises from a rational meditation on the in principle limits of our knowledge, rather than on the insight that we simply do not know whether something is the case or not.

The apophatic response is simply to say that we have no idea how, or in what way, the Moral Evil Prevention Requirements are being fundamentally met, but that knowing that the violations of these requirements only occur in the world-as-it-appears gives us grounds for belief that things may be otherwise fundamentally, in a way, perhaps, that ‘we’ may encounter and receive full consolation for. We cannot even really try to address the issue eschatologically, that things will be ‘made right’ in time, in the eschaton, because time is a feature of the way we receive the world, and is not in the world as it fundamentally is. In the same way, Kant thinks that space and time are features not of the world as it fundamentally is, but of how we receive the world: this we know, but how the world is in itself, we do not. That is the whole point of ‘transcendental idealism’.

I think it is wise to leave things there, but, in fact, Kant does sketch out some braver speculative thoughts on this matter. These braver thoughts begin with the thread of thought in Kant’s writing about God not being the creator of appearances. God, for Kant, is the author of ‘nature’, but not of appearances, the realm where natural evil presents itself and occurs. Kant makes this clear at a number of points:

Just as it would be a contradiction to say that God is the creator of appearances, so it is also a contradiction to say that as creator he is the cause of actions in the sensible world and thus of actions as appearances, even though he is the cause of the existence of the acting beings (as noumena). (CPrR, 5: 102)

Whatever God did is good, but it does not lie in the sensible world as a mere schema of the intelligible world. Thus space is nothing in itself and is not a thing as a divine work, but rather lies in us and can only obtain in us […] The appearances are not actually creations, thus neither is the human being; rather he is merely the appearance of a divine creation. His condition of acting and being acted upon is an appearance and depends on him as bodies depend on space. The human being is the principium originarium of appearances (R 6057).

These initially perplexing claims are, in fact, an implication of Kant’s claim that space and time do not ‘go all the way down’ in the universe, but are features of our reception of the world, such that we are directly and immediately, the creators of space and time. God, for Kant, is the creator of noumenal substances, outside of space and time. Human beings are the direct and immediate source of space and time, and all the appearances in space and time. Given this, it is unclear what, precisely, an earthquake relates to at the ‘noumenal’ level. It is at least conceptually possible that an earthquake is the appearance of a disturbance at the level of free noumenal decisions.

This seems such an extravagant claim, that it has been used as evidence, in the form of a reductio ad absurdum, that Kant cannot possibly ascribe to a notion of the noumenal realm and noumenal freedom.7 But the textual evidence does not support this reductio. This is because Kant himself seems to experiment, approvingly, with precisely this implication, in a passage dating either from the late 1770s, or the 1790s:

The actions here in the world are mere Schemata of the intelligible [actions]; yet these appearances (this word already signifies “schema”) are still interconnected in accordance with empirical laws, even if one regards reason itself, in accordance with its expressons, as a phaenomenon (of the character). But what the cause of this may be we do not discover in phaenomenis. Insofar as one cognizes one’s own character only from the phaenomenis, one imputes it to oneself, although it is, to be sure, itself determined by external causes. If one knew it in itself, then all good and evil would not be ascribed to external causes but only to the subject alone, together with the good and the disadvantageous consequences. In the intelligible world nothing happens and nothing changes, and there the rule of causal connection disappears.

(R 5612)

The extraordinary, but consistent, thought here is that ‘if one knew it in itself’, then all good and evil, and we might include ‘natural evil’, would ‘not be ascribed to external’, that is ‘natural’ causes, but ‘only to the subject alone’. This resonates with another pregnant reflection from the 1770s, where Kant reflects that:

Between nature and chance, there is a third thing, namely freedom. All appearances are in nature, but the cause of the appearance is not contained in the appearance, therefore also not [in] nature. Our understanding is such a cause of the actions of the power of choice, which as appearances are certainly natural but which as a whole of appearances stand under freedom. (R 5369)

If this claim, about all evils depending upon freedom, is an implication of noumenal freedom, and Kant makes this claim, this provides support for the noumenal freedom interpretation of Kant. We would have to add that Kant will always be parsimonious, in such a way that he will not make any specific claims, about any specific events. There will be no ‘moralising’, crude or otherwise, about the causes of earthquakes, or the individual springs of the tragedies suffered by individuals. There could not possibly be, given the inaccessibility of the noumenal realm. But we could know that somehow everything that appears is in some way the appearance of underlying freedom, such that tragedies, pain, suffering and natural evil are an appearance of an underlying moral disturbance. This, of course, is a ‘strange thought’, but it is a strangeness that attaches to Kant’s entire picture, and not particularly a problem in this specific case. The strange thought can quickly become an upsetting or offensive thought, if it is moralised into a thought that we ‘somehow’ are freely responsible for our own suffering. I offer no defence of the thought, but we might also note that it is an idea that is not without precedent in classical Christian theology, albeit against a different metaphysic. So, for example, Augustine, himself drawing upon St Paul, makes the suggestion that the travails and suffering of the entire creation are expressive of a deep moral fissure that enters the creation, with our free fall away from our state of original justice, into original sin.8

In the light of all this, we can revisit the summary version of my argument above, and inscribe against relevant phases which of the four Kantian principles set out above are operative:

• It ought to be the case that being moral is met with proportionate happiness.(Principle 1, but slightly developed—acting freely is the inner value of the world, and the highest expression of freedom is autonomy, which involves acting according to the moral law).
• Such a world is one where the Moral Evil Prevention Requirements I–III hold.
• Where knowledge is limited, practical reason enjoys freedom to hold-things-for-true. (Principles 2, 3 and 4)
• It is not the case that in the world that appears, being moral is met with proportionate happiness, such that in the world that appears, the Moral Evil Prevention Requirements I–III do not obtain.
• What appears is not fundamental. (Principle 4)
• Practical reason is entitled to believe in a moral realm (from 3 above), wherein the Moral Evil Prevention Requirements I–III do obtain. Such a fundamental moral realm is one where God does adhere to the Moral Evil Prevention Requirements I–III. (Principles 2 and 3).
• Therefore, the argument from the Moral Evil Prevention Requirements I–III does not provide a necessary argument for the non-existence of God.

3. Kant’s ‘Moral Proof’

Is anything like this argument actually set out by Kant? It may not matter much, for our constructive purposes. But, I believe Kant does set out something like this arc of thought, and will say something briefly about this. The position I think we find in Kant goes against readings that construe Kant’s ‘moral argument’ as some sort of gratuitous ‘add-on’ to an ethical system, onto a picture that should have no place for an eschatological reward. Beck9, Auxter10 and Murphy11 all find that the concept of the highest good, and the subsequent moral proof, is unimportant, even pernicious, introducing ‘extra-moral theological purposes’12. Other influential commentators agree (Rawls13, O’Neill14, Velkley15, Reath16 and Pogge17), finding that Kant’s moral proof violates the purity of the moral law, as well as being metaphysically extravagant, at least, until the hope for the highest good has been deflated and secularised. Those who insist on the redundancy of concept of the highest good, and the moral proof that it gives rise to, tend to circle around the following set of reasons: that the highest good is never cited when Kant discusses the categorical imperative; that Kant could not (or should not) countenance a non-moral material end (happiness) when framing the moral law, and that as we cannot know the extent of another’s virtue, we are simply unable to promote the highest good (the proportionality of virtue and happiness).

On the other hand, when Kant’s notion of the highest good is supported, the argument typically made is that the highest good, and, perhaps, some conception of God, is required in order to provide extra content to the moral law,18 or by providing some sort of supportive motivation to obey it, if only by removing obstacles to our hope that the highest good is at least possible.19 What we notice here is a shared presupposition held to by both sides: that the concept of the highest good, and the subsequent moral proof, can only be important inasmuch as it either provides extra content to the moral law, or insofar as it supports our motivation to obey it. So, where it can be shown that the highest good adds no such content or motivation, it cannot be important. My suggestion is that the concept of the highest good can remain important, and leans into the moral proof, even where its function is not that of providing the content or supporting the motivational force of the moral law20.

I would suggest that the picture is more this: the most significant move into something like hope, for Kant, is to believe in freedom at all. Once this move is made, we are already in a transcendent space of reasons, and not so very far away from the possibility of God, where God does not violate Moral Evil Prevention Requirements. I show how central it is to Kant, when thinking about God in relation to morality, that the fate of morality is closely bound up with the concept of this realm of reality beyond, and prior to, space and time: only if (deterministic) space and time do not go ‘all the way down’, is freedom, and autonomy, possible. As Kant puts it, if space and time are ‘things-in-themselves’, ‘then freedom cannot be saved’ (A536/B564). Only if there is a dimension of reality beyond mechanism, is end-setting, and so autonomy possible. The alternative is a universe without end-setting, and without freedom, which Kant tells us repeatedly, would be a sort of ‘desert’ with no ‘inner value’. First of all, I set out the broad shape of the argument, as found in the second Critique (and I think, elsewhere, but for our purposes here, this is enough): from a need, faced with a problem, moving to a solution.

4. A ‘Need, a Problem, and a Solution’ in the Second Critique

Kant’s overall argument towards the highest good has the following shape: there is a need of practical reason; there is also a problem in meeting this need, upon a certain conception of the shape of reality. The need and the problem seen side by side constitute Kant’s ‘practical antinomy’. There is then a solution that involves modifying our understanding of the shape of reality, in such a way that the problem is dissolved. I set each of these stages out below, following Kant’s paradigmatic statement of the ‘antinomy of practical reason’ in the second Critique.21

4.1. The Need of Practical Reason

The moral law ought to have a certain end-point and completeness, such that being moral is met with proportionate happiness. If this ought to be the case, it can be the case:

In the highest good which is practical for us, that is, to be made real through our will, virtue and happiness are thought as necessarily combined, so that the one cannot be assumed by pure practical reason without the other also belonging to it.

(CPrR, 5: 114)

In the practical task of pure reason, that is, in the necessary pursuit of the highest good, such a connection is postulated as necessary: we ought to strive to promote the highest good (which must therefore be possible).

(CPrR, 5: 125)

There is a lot one might say, here, about the nature of the relationship between this contestible ‘ought’, and the, also contestable, claim that it ‘must therefore be possible’. It is hardly satisfactory to cite here the principle extracted from Kant that ‘ought implies can’, where Kant tells us that ‘duty commands nothing but what we can do’ (Rel., 6: 47). This simply repeats the controversial claim. In fact, I think Kant has quite a lot to say here that is persuasive, if we have accepted deeper premises in his philosophy in relation to the possibility of freedom.22 At this point, though, our concern is to bring out the centrality of freedom and transcendental idealism in the moral proof, in relation to Sterba’s argument.

4.2. The Problem

If, in truth, we inhabited an entirely mechanistic space and time, with no immortality of the soul, and with no God, there seem to be two ways in which happiness and virtue might be ‘combined’. First of all, following the ‘Epicurean’ model, we can make the ‘desire for happiness’ the ‘motive to maxims of virtue’, or, following the Stoic model, we can make the ‘maxim of virtue’ the ‘efficient cause of happiness’, such that simply being virtuous (whatever befalls us) is identical with happiness. The Epicurean model, Kant declares, is ‘absolutely impossible’, because ‘maxims that put the determining ground of the will in the desire for one’s happiness are not moral at all and can be the ground of no virtue’ (CPrR, 5: 114). In such a case, there is nothing like morality in the world, and no ‘inner value’ to the world. The Stoic conception is ‘also impossible’, because any practical connection of causes and effects in the world, as a result of the determination of the will, does not depend upon the moral dispositions of the will but upon knowledge of the laws of nature and the physical ability to use them for one’s purposes. (CPrR, 5: 114)

Here, once again, Kant has in mind the world viewed as sheerly mechanistic. Upon this conception of the world, there is no necessary connection, or, indeed, even a remote possibility of a connection, between virtue and happiness, where virtue could be a cause of happiness. This is Kant’s ‘practical antinomy’, whereby, with either the Epicurean or Stoic conception:

No necessary connection of happiness with virtue in the world, adequate to the highest good, can be expected from the most meticulous observance of moral laws.

(CPrR, 5: 114)

What we might note here is that Kant upholds Sterba’s Moral Evil Prevention Requirements, in a way that neither Epicureanism nor Stoicism do: Kant does not permit the thought that anything like the ‘highest good’ is manifest in a world in which virtue is met with anything other than true happiness. I take it here that Kant’s commitment to a moral world involves moral actions being met with proportionate happiness, and moral agents being protected, ultimately, against the immoral actions of others, and against other types of undeserved suffering (where ‘natural evil’, for Kant, may, in the end, be a consequence of immoral actions). As I understand Sterba’s Moral Evil Prevention Requirements, such a moral world would meet these requirements, although I anticipate that this might be a point of contention. Both Stoicism and Epicureanism in different ways encourage us to reconcile ourselves with a world in which the Moral Evil Prevention Requirements are clearly not met.

4.3. The Solution

Therefore, given the need of practical reason, which must be met (we grant, for the moment), and given that the antinomy arises because of our conception of what the ‘present conditions’ are, it follows that there must be something wrong with our conception of what the present conditions in fact are. The problematic conception of the ‘present conditions’ involves construing them as being exhaustively constituted by mechanistic space and time, with no immortality of the soul, and with no God. The solution, then, is to alter our conception of what the fundamental structure of this world really is (the ‘present conditions’). The first realisation is that at a more fundamental level of reality, the world is not spatial or temporal, and so not deterministic. The noumenal world is a conceptual space within which moral freedom is possible. Kant explicitly draws a parallel between the way in which transcendental idealism offers a solution to the practical antinomy, and the way in which it similarly solves theoretical antinomies in the first Critique. In the first Critique, Kant solves, at least to his own satisfaction, the antinomy whereby it seems that we must both affirm a first cause, from which all other causes and effects follow, and also affirm that there is no first cause, because the idea of a first cause is itself incoherent, given that everything has a cause. Kant’s solution is to affirm both as true in different ways: in the realm of spatial and temporal appearances, every effect has a determining cause; but in the world in itself, prior to our spatial and temporal reception of it, there are genuine first causes, which are not the effects of predetermining causes. This is how Kant draws the parallel between the solutions to the theoretical and the practical antinomies:

In the antinomy of pure speculative reason there is a similar conflict [to the practical antinomy] between natural necessity and freedom in the causality of events in the world. It was resolved by showing that there is no true conflict if the events and even the world in which they occur are regarded (and they should also be so regarded) merely as appearances; for, one and the same acting being as appearance (even to his own inner sense) has a causality in the world of sense that always conforms to the mechanism of nature, but with respect to the same event, insofar as the acting person regards himself at the same time as nouemenon (as pure intelligence, in his existence that cannot be temporally determined), he can contain a determining ground of that causality in accordance with laws of nature which is itself free from all laws of nature.

(CPrR, 5: 114)

Transcendental idealism, which generates the possibility of a noumenal, and therefore a moral world, removes the ‘seeming conflict of a practical reason with itself’, whereby ‘the highest good is the necessary highest end of a morally determined will and is a true object of that will’:

For it is practically possible, and the maxims of such a will, which refer to it as regards their matter, have objective reality, which at first was threatened by that antinomy in the combination of morality with happiness in accordance with a universal law, but only from a misinterpretation, because the relation between appearances was held to be a relation of things in themselves to those appearances.

(CPrR, 5: 115)

The way that Kant’s argument for the highest good works then, is to show, with respect to the ‘problem’ that faces our need, that things (the ‘present conditions’) are indeed not as they seem: that mechanistic space and time are features of our reception of the world, and not fundamentally the way things are; that the soul is capable of an ‘endless progress’, and that there is a God. This way of putting the sequence is significant, because it fills in a frequently missing stage in Kant’s movement towards the highest good. Typically, even in thinkers broadly sympathetic to Kant’s notion of the highest good, there is a tendency to go straight from the need for virtue to be met with happiness, to the notions of immortality and God, with scant attention paid to the crucial middle-term, which is belief in the non-spatial and non-temporal intelligible/moral world, which fundamental reality may be one where the Moral Evil Prevention Requirements are (somehow) upheld. Sometimes this omission is philosophically well-motivated, although not, perhaps, exegetically, in that the commentator wishes to abstract Kant’s philosophical theology from a metaphysically committed interpretation of the noumenal realm, and to read Kant’s transcendental idealism in a more deflationary way. This is a curious feature of some theologically sympathetic interpretations of Kant, in that one might think that a willingness to consider the concept of God (classically, non-spatial and non-temporal) as meaningful, might help a thinker to be less allergic to metaphysical commitment. As it happens, though, theologically sympathetic readers of Kant have often tended instinctively to gravitate towards more deflationary readings of Kant23: perhaps not to surround themselves with even more metaphysical extravagance, or, perhaps, out of a sense that such a non-spatial and non-temporal space should be reserved for God alone. In any case, and for whatever reason, this intermediate move, from the possibility of virtue being met proportionately with happiness, to the noumenal/intelligible/moral realm, is often overlooked. It ought not to be, for both exegetical and philosophical reasons: for exegetical reasons, because this step is a central feature of all Kant’s discussions of the highest good; and for philosophical reasons, because this intermediate step opens up a way in which a commitment to morality and freedom already has an intrinsic momentum, in Kant’s thought, given Kant’s wider arguments and assumptions, towards hope and belief in God. If morality and freedom are possible, and we must think that they are, the step towards the possibility of happiness, I will argue, is not a large one at all, and nor therefore, is the step towards God and immortality, to the extent that these are connected with the possibility of happiness. Indeed, the more difficult thing might be to stop the progression of ideas moving in this direction (from freedom to happiness and divinity).

5. Concluding Reflections

As I said in my opening remarks, I do not ‘believe’ or ‘promote’ this whole Kantian picture and argument. But it is interesting to reflect on what would make someone more likely to look on it with some sympathy. I doubt that the key thing is to be drawn to grand inflationary metaphysics (freedom, noumena) and so on. The more likely draw to the argument may be one’s attitudes to the Kantian principles set out earlier, especially, perhaps principles 2, 3, and 4. In turn, the crucial commitment in each of these principles could be summarised as follows:

• the success of transcendental style arguments at generating justified beliefs.
• the ability of practical reason to move beyond the limits of knowledge.
• the ability of practical reason, when moving beyond such limits, to think in entire systems and world-views.

None of these epistemological commitments are directly concerned with the existence or otherwise of God.

Apart from the (perhaps implausible) moves in the Kantian argument set out above, what might we take away in a ‘big picture’ sense from the possibility of this rather alternative perspective on the problem of evil, apart from the wider ‘structural challenge’ to logical-impossibility arguments that I set out at the beginning of this article? Sterba is generous and transparent in offering some personal and biographical framing for his argument, and I would like to reciprocate here by offering a more personal reflection here.

When I was teenager, I remember always having the same reaction to any sort of formulation of the ‘problem of evil’. Rather than the existence of ‘evil’ and suffering being a reason not to believe in God, it seemed to me ‘obvious’ that it was the best sort of reason to believe in God, as an expression of a type of yearning. God was the concept that should be reached for on the other side of a lament. No other concept would suffice, because it would be inadequate to the task of engaging with evil. Anything else would be unrealistic: no amount of historical or political action, or human self-improvement, or economic management, or whatever, could possibly address the evil and suffering that had already happened, let alone prevent or ameliorate what was to come. At that stage in my life, a sense of the problem of evil and suffering was my main ‘motivation’ or reason, if I had to give one, for believing in God. I think I probably tried to express this thought in school, and soon learned that it was not one of the grooves down which one’s mind was supposed to travel. At least, the formal problem went the other way round, in a way that is expressed with rigour and existential-commitment by Sterba: the existence of evil and suffering is a problem for belief in God, not a good motivation for it. I learned to keep (mostly) quiet about this. Some thirty years later, coming back to the thought, I have a few reflections about it.

First of all, something like this movement of thought expresses the momentum of Kant’s own ‘moral argument’ for the existence of God: that we live in a realm of antinomies, where things are not as they ‘should be’, and where belief in God is the only way through this situation. Secondly, I am struck by the thought that much religious and theological thinking is really quite skeptical and nihilistic, at least about the sufficiency of most purported and suggested ‘solutions’ to our various predicaments, which might be expressed in terms of antinomies (for example, between how things are, and how they ought to be). At least one striking feature of some variants of religiosity is not credulity, optimism, or confident knowledge claims about the absolute, but a sense of how partial, broken, and fragmented our condition is, where a sense of ‘wholeness’ or ‘healing’ is only gestured to in the faintest way, but where this gesturing constitutes (for some people) one of our most important ‘ecstatic’ moments. This is a very different instinct than the one underlying Sterba’s argument, although we share an instinct that the problem of evil is of central significance in relation to belief in God. I offer the personal reflection not as a rival ‘argument’, but as an alternative perspective. It would be fascinating to explore what is ‘at stake’ in occupying these different perspectives: both in terms of the explicit arguments and reasons that might be given, but, also, the underlying intuitions, hopes and fears. Some of this exploring might be beyond the bounds of mere philosophy.