2.1 The Ontological Theory of the Best Possible World

Kant’s most well-known endorsement of the view that God creates the best possible world is found in the pre-Critical ‘An Attempt at Some Reflections on Optimism’ (1759). Here Kant explicitly associates the view that God creates the best possible world with Leibniz and thinks in broadly Leibnizian terms about what a best possible world looks like (2: 29).

Kant identifies the notion of ‘perfection’ that he uses to rank worlds with the concept of ‘reality’: ‘First of all, permit me to equate the absolute perfection of a thing with its degree of reality’ (2: 30–1). This suggests that ‘perfection’ is an ontological notion. In an associated Reflexion, Kant likewise advances a conception of the perfection of worlds that seems ontological, insofar as it suggests that the more beings or types of being a world contains, the more perfect that world is:

Both the terminology of ‘reality’ and the identification of the level of perfection with the number of existent things or types of existent things suggest that Kant ranks worlds using an ontological metric that assigns value to beings on the basis of the sheer fact that they are. The metric would be ontological insofar as it affords value to being as such, without regard for whether the being is itself ‘good’ in some further, more demanding sense.

However, neither the terminology of ‘reality’ nor the affordance of value to being as such is sufficient to imply that Kant uses a purely or fundamentally ontological metric. There is a long tradition in both ancient and medieval philosophy of treating being and goodness as either interchangeable terms for one property or as closely related (see MacDonald Reference MacDonald1991). Kant’s own use of the terms ‘perfection’ and ‘reality’ indicates the ongoing influence of this tradition, which remained important in Leibniz’s work and in early modern Germanophone rationalist philosophy in general, although it finds little echo in contemporary Anglophone philosophy.

It might seem that Kant’s notion of ‘perfection’ completely equates axiological and ontological considerations. For instance, Kant talks about worlds that are ‘equal in respect of perfection’ (which he identifies with ‘reality’) as those which have ‘the same degree of goodness (Güte)’ (2: 31), thereby either identifying or correlating perfection, reality and goodness. But in spite of Kant’s use of axiologically loaded terminology such as ‘goodness’ and ‘excellence’ (Trefflichkeit), the use that he makes of ‘reality’ in the argument itself indicates that the term must be understood as irreducibly ontological – that is, as appealing to the structure of being or reality where ‘being’ and ‘reality’ are understood in value-neutral or value-indifferent terms. This is because Kant’s argument for the uniqueness of the best possible world requires such an understanding of ‘reality’.

In order to establish that there is only one best possible world, Kant appeals to a principle about how realities differ from one another: ‘reality and reality differ from one another, not in respect of their quality (qualitate), but in respect of their magnitude (gradu). Accordingly, if things differ from one another, they always do so through the degree of their reality’ (2: 31). Kant here commits himself to the view that any two things that have the same overall degree of reality are numerically identical. In order for this view to be minimally compelling, reality cannot be understood as an axiological or value-theoretic notion. It is very hard to render plausible the view that two things that are equivalent in value or goodness are eo ipso the same thing, and therefore identical with respect to all properties. In order for the degree of ‘reality’ to determine whether things are metaphysically distinct, ‘reality’ ought to be understood as an ontological notion. Otherwise, it seems that one is forced to assign value-theoretic significance to every difference in properties, although plausibly there can be variation in properties without an accompanying value-theoretical variation (consider two things that differ merely with respect to colour). Although one might hold that there are more ‘properties’ than ‘realities’, and perhaps it is only variation with respect to ‘realities’ that implies value-theoretic difference, it is then hard to see how it would be plausible to hold that ‘if things differ from one another, they always do so through the degree of their reality’ (2: 31) – the plausibility of that principle requires a broad understanding of what qualitative properties count as ‘realities’. Regardless of whether Kant’s argument for the uniqueness of the best possible world is successful, the important point for present purposes is that his use of ‘reality’ in the uniqueness argument requires a non-axiological understanding of the concept.

Now Kant does attempt to treat ‘reality’ as an axiologically significant concept that is equivalent to or correlated with ‘goodness’ and ‘excellence’. But in doing so, he risks an equivocation that consists in appealing to the ontological concept of reality to make the uniqueness argument, and then redescribing his results using axiologically loaded synonyms for ‘reality’ that fail to share the formal structure that is supposed to make the uniqueness argument work.

This brief treatment of Kant’s early views on the best possible world is not intended to conclusively establish that the notion of ‘perfection’ that Kant uses to rank worlds as of 1759 was exclusively ontological, with absolutely no value-theoretical component or implications. Although the uniqueness argument strongly suggests that the ontological dimension is more fundamental, it might be the case that Kant treats ‘reality’ and ‘perfection’ as more ontological at some points and more axiological at others. For present purposes it suffices to point out that the notion of ‘perfection’ to which Kant appeals in the 1759 essay contains an irreducibly ontological component.

2.2 The Axiological Theory of the Best Possible World

Kant’s development of his mature axiology alters what it means for Kant to consider a world to be ‘best’ or the most ‘perfect’. In the Critical period Kant adopts a dualistic axiology according to which there are two types of good, the moral and the natural, which in turn are grounded in the fundamental dualism of sensibility and understanding:

Kant begins to develop this view as soon as the early 1760s, and he is recorded in the Herder lecture notes as holding that, ‘Blessed happiness (Glückseligkeit) is constituted out of happiness (Glück); non-moral good: physical well being … and blessedness (Seligkeit); moral good’ (27: 46, see Bader Reference Badermanuscript a). The moral good or ‘perfection’ that makes the ‘worth of the person’ is to be understood as the good will. As Kant famously states at the beginning of the Groundwork of the Metaphysics of Morals (1785), ‘It is impossible to think of anything at all in the world, or indeed even beyond it, that could be considered good without limitation except a good will’ (4: 393). The good will, or a virtuous will, is the moral good. The natural good is happiness.

There is controversy regarding whether these two types of goodness, moral and natural, exhaust the types of things that the Critical Kant regards as valuable. Some commentators hold a much more expansive view of what Kant thinks is valuable, appealing to instances in which Kant speaks in ways that seem to impute value to humanity, freedom, reason and such (Korsgaard Reference Korsgaard1986, Reference Korsgaard1996; Guyer Reference Guyer1998). By way of contrast, others hold that the sense in which something like humanity can be said to have ‘value’ is much broader than the sense in which the good will and happiness have value (Bader Reference Badermanuscripts a, b). Proponents of this view note that what Kant refers to in the second Critique as ‘the whole and complete good’ or the summum bonum consummatum is made up of virtue and proportional happiness (5: 110). Given that the highest good is supposed to be the totality of that which is good, the status of complete virtue and happiness as the ‘complete’ highest good suggests that virtue and happiness exhaust that which has ‘value’ in the most technical or strict sense of the term.

Regardless of one’s position on the broader issue of how many kinds of value or valuable things the Critical Kant recognizes in general, it is clear that Kant is using the strict dualistic account of value or goodness in his Critical theory of the best possible world. In both the moral philosophy and the rational theology lecture transcripts, he offers an account of the best world that appeals to the two basic categories of natural and moral value:

Here Kant articulates a very different view from the broadly Leibnizian approach of the 1759 essay, according to which ontological features of worlds (e.g. the number of beings or types of beings) add to the value of worlds. But Kant is not completely overthrowing Leibnizian value theory so much as adapting it to the basic features of his own philosophy. In the Theodicy (1710) Leibniz identifies three categories of evil: metaphysical evil (‘mere imperfection’), moral evil (‘sin’) and physical evil (‘suffering’) (Leibniz Reference Leibniz and Huggard1951: §21). This threefold taxonomy in turn grounds a threefold categorization of the types of good as natural, moral and metaphysical. Kant’s own terminology of perfection shows the influence of Leibniz insofar as Kant refers to physical perfection or good (e.g. R5511, 18: 216; 28: 1101), moral perfection or good (e.g. R5511, 18: 216; 28: 1101) and metaphysical perfection (e.g. R3889, 17: 328; R4028, 17: 390). But only moral perfection and physical perfection are ultimately admitted into Kant’s mature value theory.

Kant revises Leibniz’s trifold account of value to accord with his own dualism of theoretical and practical reason. In a Reflexion that Adickes dates to the 1780s, Kant explicitly distinguishes between perfection ‘in a theoretical sense’, as either ‘transcendental’ or ‘metaphysical’ perfection (the latter of which is identified with the ‘reality’ of a thing), and perfection in a ‘practical sense’ as ‘the good’, which is either good ‘in itself (the good will)’ or what is good as a means to an end (18: 344). Here Kant signals his own break with the tradition of treating being and goodness as somehow deeply intertwined. Being is theoretical, goodness is practical. Metaphysical perfection or ‘reality’ lacks the practical significance that would enable that type of perfection to be classified along with moral and physical perfection.

The separation between the theoretical and the practical is even clearer in the Mrongovius metaphysics lecture transcripts, in which Kant distinguishes the ‘good’, as the ‘objective ground of our choice’, from ‘perfection’, as ‘the ground of the greatest objective unity of cognitions’ (29: 936). In fact, here Kant even goes so far as to criticize accounts of an ‘optimal’ world that fail to make that distinction, and that therefore treat attributes associated with the theoretical notion of ‘perfection’ (Vollkommenheit) as features that make a world more ‘good’ (Gut, bonum) or ‘optimal’ (optimum) (ibid.). While there may be structural resemblances between the concepts, the different terminology underscores the way that they correspond to different notions and are proper to different faculties: ‘the good is for the will what the perfect is for the understanding’ (ibid.). Now this terminology is not consistent across Kant’s corpus, insofar as in the rational theology lecture transcripts he is recorded as referring to the combination of ‘moral and physical perfection (Vollkommenheit)’ as accounting for why the world is ‘best’ (Th-Volckmann, 28: 1202; Th-Pölitz, 28: 1101). However, here the context makes it clear that Kant is using the language of ‘perfection’ to invoke his dualistic account of value, rather than any sort of notion that combines ontological and axiological features.

Moral and physical goodness in turn acquire a distinctive Kantian significance insofar as they become associated with the dualism of the understanding and sensibility within the domain of practical reason. Moral perfection or goodness (virtue) is aligned with the understanding, while physical goodness (happiness) is aligned with sensibility. The types of perfection that Leibniz includes in his more expansive value theory become regimented and gain systematic import in accordance with the fundamental dualisms of Kant’s Critical philosophy. Although a reconstruction of the trajectory of this shift exceeds the scope of this article, the contrast between the foundations of the best possible world theory presented in the 1759 essay on optimism and that which is presented in the 1783/4 rational theology lecture transcripts is clear. In the latter, Kant is recorded as giving up on the effort to associate axiology and ontology in his metric for the best possible world, in favour of a purely dualistic axiology that finds its foundation in his dualistic conception of practical reason.Footnote ⁴

2.3 Accounting for Non-Axiological Features of Worlds

The Critical account of the best possible world excludes from its metric many of the features that add to the goodness of a world on a Leibnizian account. According to Kant’s dualistic axiology, the only way for features of the world such as harmony to add to the value of the world is indirectly, insofar as they increase the virtue or happiness of rational creatures. But they are not themselves value-bearing or value-increasing features of worlds.

Kant nonetheless retains a place for order, beauty and harmony in his account of the kind of world that God creates. He grounds these features in the non-volitional aspects of God. Kant identifies non-axiological but (what Leibniz would call) metaphysically good features of ‘order, beauty, [and] harmony’ with God’s essence, ‘rather than the will of God’ (Th-Danzig, 28: 1261; Th-Pölitz, 28: 1034). There are different strategies one might employ in order to explain why the world that God creates exhibits order, beauty and harmony. Perhaps there are features that influence or inform the creation of worlds that are not the value of worlds, even though they might at least partially explain why this world is created. Or perhaps the set of worlds from which God chooses would be restricted by the non-volitional properties of God. Or a third way of accounting for features of worlds such as order, beauty and harmony is to argue that, in choosing the world that maximizes moral and natural value, God eo ipso chooses a world that maximizes other properties.

Features such as justice are also excluded from the Critical Kant’s axiology. According to Kant’s narrow dualistic axiology, more just worlds are not eo ipso better worlds. They are simply more just. This implication of Kant’s narrow axiology might seem strange insofar as the requirement that the world exhibit proportionality between virtue and happiness (i.e. that this world is just) supplies the Critical Kant with his justification for positing the existence of God (A815/B843; Lec-Th, 28: 1012, 1153, 1242). But while the world that God creates is just, the justice of such a world need not be part of what makes that world the best world possible.

Justice is not itself a type of value, but rather concerns the conditioning relationship between the two fundamental types of value that Kant identifies, moral and natural value. Kant presents God’s justice as a limitation (Einschränkung) of God’s goodness: God only distributes as much happiness as rational beings have made themselves worthy of through their virtuous activity (Lec-Th, 28: 1074, 1185, 1292). As a principle concerning the distribution of what is good, rather than a principle concerning what kinds of things are good, justice has a structurally distinct role from value, and Kant correctly excludes it from his value theory. But because God is just, a world that God creates exhibits proportionality (or perhaps, even more strongly, necessary proportionality) between virtue and happiness.

2.4 Why the Critical Philosophy Needs a Best Possible World

The Critical Kant’s doctrine of God’s creation of a best possible world is indirectly integrated into his account of the highest good. Indeed, one can derive the justification for that doctrine from the considerations that Kant uses to support the postulation of the existence of God. In the first Critique and rational theology lecture note transcripts, Kant argues that we are justified in postulating the existence of God for the sake of moral motivation (A815/B843; Lec-Th, 28: 1012, 1153, 1242). In particular, moral motivation requires that I rationally believe in the achievability of my individual highest good, which in turn requires rational belief in a necessary proportionality between my virtue and my happiness, and (as Kant argues) since this proportionality can only obtain if God exists, I am justified in believing that God exists. Kant changes his view in the second Critique, arguing instead that something like the rationality of morality is what requires rational belief in the aforementioned necessary proportionality, and therefore in God (see e.g. Pasternack Reference Pasternack2014: 44–5). But in both cases, there is some way in which the viability of my moral activity requires postulating the existence of the sort of being who can ground this proportionality.

The link between Kant’s Critical moral philosophy and the doctrine of the best possible world consists in the requirement that I rationally believe in the existence of a being who has the traditional divine attributes. More specifically, according to Kant, rationally believing in the necessary proportionality of individual virtue and individual happiness requires believing that the world is created by a being with the traditional divine attributes, including omnipotence, omnibenevolence and omniscience (A815/B843; Lec-Th, 28: 1012, 1153, 1242; CPrR, 5: 140). Each of these traditional attributes is supposed to have a role in explaining why God brings about the necessary proportionality. For instance, according to Kant, I can only expect that my happiness will be (necessarily) proportioned to my virtue if the world is created by a being who, at minimum, has the capacity to ensure that I am granted happiness, and who (according to Kant) is correspondingly omnipotent. Now Kant holds that a being like God who is omnibenevolent and omnipotent has a perfect will, and that a perfect will is perfect insofar as it chooses the best object (Lec-Th, 28: 1097, 1199, 1301). Thus the Critical concept of God, as a being who is both omnibenevolent and omnipotent, is the concept of a being who creates a best possible world.

There is therefore an indirect connection between moral motivation (and/or the rationality of morality) and the belief that this world is the best possible. Namely, in order to be morally motivated, I have to rationally believe that this world is one in which my individual virtue will be rewarded with proportional happiness, and therefore that this world was created by the sort of being who would create a best possible world. The belief that this world is the best possible is indirectly connected to individual moral motivation through a chain of multiple inferences. The doctrine of God’s creation of the best possible world supplies a link between the individual highest good and what one might call ‘collective considerations’ about the virtue and happiness of finite rational beings as a whole.

2.5 Did Kant Really Believe This?

Kant’s Critical doctrine of God’s creation of the best possible world has been widely neglected in the secondary literature. One possible reason for this neglect may be the fact that the lecture transcripts on rational theology are the major (although not the sole) source of textual evidence for his Critical views on the matter. Since Kant never authorized for publication the notes taken during these lectures, and moreover was lecturing on the set texts of Baumgarten’s Metaphysica (1743) and Eberhard’s Vorbereitung zur natürlichen Theologie zum Gebrauch akademischer Vorlesungen (1781), one might worry that the views on the best possible world expressed in the lecture transcripts are Kant’s rehearsals of the views of Baumgarten and Eberhard, or the regurgitated wisdom of his day, rather than views of his own.

However, there is strong reason to hold that Kant’s Critical account of the best possible world reflects the outcome of his own serious reflection on the topic. Kant is recorded as updating his account of the perfection of the best possible world in the 1783/4 transcripts in a way that brings that account into line with his own mature axiology, according to which virtue and happiness are what have value. It is hard to see why he would have done this if he did not take seriously the view that God creates the best possible world.

Furthermore, in the transcripts Kant is recorded as providing two arguments in favour of holding that God creates the best possible world that follow from his own distinctive commitments. Kant holds that ‘it is possible to recognize the doctrine of the best world from maxims of reason alone, independently of all theology’ (Th-Pölitz, 28: 1098; Th-Volckmann, 28: 1199; Th-Danzig, 28: 1301). The first argument proposes that the hypothesis that this world is the best possible is analogous to a necessary maxim for human reason in the course of scientific discovery, and that we should ‘expect things to be similar’ in ‘organized, albeit non-rational nature’ on the one hand, and ‘the nobler part of the world’ on the other (ibid.). Regardless of what one thinks about the strength of this argument, the consideration regarding necessary maxims of human reason that are used in scientific discovery is a distinctively Kantian one that could not have been taken from Baumgarten or Eberhard.

Kant provides another argument in favour of the doctrine which appeals to the fact that ‘in a best world it is impossible for good conduct to exist apart from well-being’, and under such conditions the ‘incentives’ of morality ‘retain their strength’ (Th-Pölitz, 28:1098; Th-Volckmann, 28: 1200, Th-Danzig, 28: 1302). Here Kant repeats the considerations that justify postulating the existence of God in the first Critique, which appeal to the need for a proportionality of ‘good conduct’ and ‘well-being’ for the sake of moral motivation (A815/B843). Once more, regardless of how plausible one finds this sketch of an argument, this argument for the doctrine of a best possible world is a distinctively Kantian one. It is not regurgitated from Baumgarten or Eberhard, and Kant would not be going to the trouble of finding arguments from within his own practical philosophy in favour of the claim that this world is the best possible one if he did not genuinely affirm it.

3.1 Maximalism

The value of this world is what makes it a fitting object of divine choice. But then how does one account for the fact that it seems that there are worlds that could be better than this world, and that this world therefore seems not to be a fitting object of divine choice? One strategy is to hold that any world God creates must have defects – perhaps very many defects. Defective though this world is, God could not have done any better. Leibniz endorsed some variant of this view, although the details are controversial (see e.g. Rescher Reference Rescher1979, Rutherford Reference Rutherford1995, Strickland Reference Strickland2006).

Leibniz’s strategy for excusing this world for its defects becomes problematic when goodness is understood axiologically rather than ontologically or quasi-ontologically. This section argues that Kant’s axiological theory of the best possible world is not plausibly understood as a theory that incorporates a defective best possible world, whose defects are then excused using Leibniz’s strategy of holding that any world must have at least some defects. For Kant, there is no world that is better than this world (i.e. the world God creates), but this is not because it is necessary that any world that God creates has defects or is defective. This is what it means to say that Kant is a maximalist about the best possible world.

The strategy of excusing this world by holding that some defects are necessary is a plausible strategy when goodness is measured using an ontological or quasi-ontological metric. For Leibniz, the necessity of defects follows from the non-compossibility of essences. That is, there are some essences such that the instantiation of one of them in a world rules out the instantiation of other essences (Look Reference Look2013). However, the appeal to non-compossibility becomes implausible within the context of a Kantian axiology combined with a Kantian understanding of morality. The most direct argument for this point proceeds as follows: the moral good and, arguably, the natural good are intrinsic goods (see Bader Reference Bader, Hirose and Olson2015b). Intrinsic goods are by definition compossible, insofar as intrinsicality implies that something’s status as good is independent from anything apart from the bearer of goodness itself.Footnote ⁵ One can also mount additional arguments for this point by examining and rejecting various strategies for appealing to non-compossibility when using a Kantian axiology.

One way to argue that there must be value-theoretic defects in a world due to non-compossibility would be to hold that there is a trade-off between the virtue of agents, such that some agents’ virtue somehow comes at the expense of the virtue of others. If we understand virtue as consisting in strength of will in conforming to the moral law, then this view is difficult to sustain. How can one person’s increasingly strong conformity to the moral law rule out an increase in another person’s conformity to the moral law? For if strengthening one’s will is a purely inward process, which does not require that the agent perform any action in a world shared by other agents, it seems highly mysterious how the strengthening of one agent’s will might negatively constrain the strengthening of another agent’s will.

One might attempt to make the argument for necessary moral defects more plausible by identifying a connection between virtue and moral action, and then appealing to the non-compossibility of certain moral actions. The connection between virtue and moral action can be established in multiple ways. There are some interpretations of virtue according to which virtue consists not in strength of will, but rather in the proportion of actions performed that are moral (Bader Reference Bader, Aufderheide and Bader2015a). On this interpretation, moral action constitutes virtue. Or, if one chooses to interpret virtue as consisting in strength of will, one might hold that improving the strength of will requires performing moral actions. Given either a constitutive or causal connection between moral action and virtue, one might try to argue that certain combinations of moral actions are impossible. It is not prima facie implausible that one agent’s acting in a certain way could rule out another agent acting in a certain way, and thus that non-compossibility of moral actions could imply the necessity of moral defects in a world.

However, such a position is at odds with a Kantian view of moral action. For Kant, what it is for an action to be moral is for it to be compossible with the actions of other rational beings. That is, the version of the moral law expressed by the Formula of the Law of Nature can be understood as treating compossibility with the actions of other rational beings as a defining feature of moral action (Grundlegung, 4: 421). Furthermore, the kingdom of ends can be understood as the union of beings who form a harmony insofar as they are interconnected by their obedience to the moral law (4: 438, see also A812/B840). The idea that moral actions could somehow be in tension with one another, such that the performance of some moral actions necessarily rules out the performance of other moral actions, and such that there is no way for all actions to be moral, thus arguably stands in significant tension with how Kant conceives of morality itself.

One might instead try to argue that there is non-compossibility with respect to the existences of sensibly affected wills, and that consequent limits on the number of sensibly affected wills render value-theoretic defects unavoidable. This line of thought is unattractive, since according to the postulate of immortality, each finite rational being has an infinite amount of time in which to morally improve. Moral improvement brings about an increase in moral value. This suggests that a limit on the number of moral agents is not enough to bring about defects with respect to moral value.

Thus there does not seem to be a plausible Kantian notion of non-compossibility with respect to moral value or its sources. An appeal to non-compossibility would therefore not seem to provide Kant with an adequate explanation of defects in the moral value of the best possible world. And given that God creates a world in which justice obtains and in which happiness is therefore necessarily proportional with virtue (i.e. moral value), the same holds of appealing to non-compossibility to explain defects in the natural value of the best possible world.

A second consideration in favour of maximalism is that it establishes a structural parallel between individuals and worlds. In the second Critique, Kant argues that the achievement of the complete highest good for the individual requires ‘the complete conformity of dispositions with the moral law’, which is a kind of ‘holiness, a perfection of which no rational being of the sensible world is capable at any moment of his existence … [and] can only be found in an endless progress toward that complete conformity’ (CPrR, 5: 122). For an individual, it is thus clear that the achievement of the complete highest good requires a maximalist conception of conformity. The individual who achieves the complete highest good is in a state such that no state is better for it, and this fact in turn is not explained via the necessity of there being some defects in that individual’s state (or any individual’s state). The question is then whether the highest good of worlds should also be understood in a maximalist manner.

Kant presents the complete highest good of the individual and of a world as structurally similar notions:

Complete virtue and complete happiness in proportion (or perhaps necessary proportion) make up the complete highest good for both individuals and worlds. As noted above, Kant understands the complete highest good of an individual (as opposed to the supreme highest good, which is virtue alone) in a maximalist manner. The general parallelism that Kant advances between the complete highest good of individuals and that of worlds suggests that the complete highest good of worlds should also be understood in a maximalist manner.

This does not suffice to establish that the best possible world should likewise be understood in a maximalist manner. The highest good of a world is a ‘collective’ notion of the highest good. Collective notions of the highest good consist in the complete virtue and happiness of all finite rational beings, as well as the (perhaps necessary) proportionality of virtue and happiness (see e.g. Yovel Reference Yovel1980: 64, 69; Reath Reference Reath1988: 603; Bader Reference Bader, Aufderheide and Bader2015a: 202).Footnote ⁶ But while there are some resemblances between the notions of the best possible world and of a collective highest good or the highest good of a world, these notions are not identical.Footnote ⁷

Nonetheless, the maximalist conception of the highest good of a world lends support to a maximalist construal of the best possible world. For each individual, the highest good of that individual is possible. Parallelism between individuals and worlds with respect to the highest good would then suggest that the highest good of each world is possible, in the sense that it can be achieved. If the highest good of each world can be achieved, then it is hard to see how a putative best possible world could fail to achieve its highest good, yet be best possible. For then a better world than that putative best possible world would be possible – namely, a world that is just like it, except that the highest good of that very world is achieved. This, in turn, suggests that the best possible world should be understood in a maximalist manner.

A third consideration in favour of maximalism is that there are variants of the maximalist view, such as the one this article will present, which resolve an open problem with Kant’s theology. Since it will be easier to demonstrate this point after developing the variant in question, the presentation of the problem-solving advantages of (one variant of) maximalism will be postponed until the fourth section.

However, there is an obvious problem with maximalism. This problem arises from the fact that bad things have already happened – both morally bad actions and instances of undeserved unhappiness. The maximalist view is that this world is such that no world is better than it, and that this ‘no-world-is-better’ or ‘best possible’ claim is not explained via the necessity of defects. But this world quite clearly seems to be defective. So how can this world be such that no world is better than it, given that this world has witnessed evil and ill? What explains the truth of the claim that no world is better than this world, if not the necessity of defects?

3.2 The Infinite Value of the Best Possible World

The solution that the remainder of this article will explore is the view that the best possible world is infinite in value. This is a view that has recently been proposed within the contemporary philosophy of religion (Climenhaga Reference Climenhaga2018). While there are a number of ways of developing this view, the basic idea is simple: if the value of the world is understood as a sum of goodness (with badness subtracted), then if there is infinite goodness and only finite badness, the overall value of the world is infinite. Different variants of this approach would measure the overall goodness of worlds in terms of the proportion of goodness to badness. Within the latter such group of such views, depending on how one aggregates the value of individual lives to calculate the value of worlds, it may not be apt to describe the overall value of the world itself as infinite, but the availability of infinite value-bearing moral actions and episodes of happiness within lives would be the key to explaining the truth of maximalism. For the sake of simplicity, I will assume the summation framework and speak in what follows about infinitely valuable worlds, but this is not intended to imply endorsement of the summation framework at the expense of a proportionality framework.

If the world is infinite in value, then the explanatory challenge to maximalism can be addressed by pointing out that a world with finitely many bad actions and finite unhappiness, but infinitely many good actions and infinite happiness, has precisely the same overall value as a world in which there are no bad actions or unhappiness, and infinitely many good actions and infinite happiness. The former kind of world is no worse than a world that has infinitely many exclusively good actions and infinite happiness with no unhappiness.

One can describe a world with finitely many bad actions and finite unhappiness, and infinitely many good actions and infinite happiness, as a world in which the trajectory of improvement ultimately reaches a point at which rational beings become sufficiently morally developed to stop committing bad actions entirely, and experience only happiness (perhaps due to technological advancement, or to happiness granted by God after death).

In general, there are many ways to envision the precise way in which infinite value is distributed across time, space and species. It does not matter whether it can be shown a priori which of the various options obtains. What matters is simply that it is compatible with everything that has been witnessed on earth (i.e. finitely many bad actions and finitely much unhappiness) that this world is such that no world is better than it. Thus we do not need to appeal to the necessity of defects in any world God creates in order to hold both that evil and ill obtain in this world, and that no world is better than this world.

An infinite value view provides a structurally parallel resolution at the level of worlds to the resolution of the problem that the maximalist requirement poses at the level of individuals. The first part of the antinomy of pure practical reason arises because individuals apparently cannot bring themselves into complete conformity with the moral law. The resolution is the postulate of immortality, the function of which is to afford individuals endless time with which to bring about ‘endless progress from lower to higher stages of moral perfection’ (CPrR, 5: 123). Here Kant appeals to considerations related to infinity in order to explain how the exacting demands of morality might be satisfied. The infinite value view provides a parallel kind of appeal at the level of worlds and is therefore consistent with the overall spirit of Kant’s views.

3.3 How Bad can you Get?

Admitting infinitely valuable worlds resolves the problem with maximalism by establishing the in-principle compatibility of the finite quantities of evil and ill (such as have been observed in this world) with the status of this world as best possible (that is, no world is better than it), and without appealing to the putative necessity that any world have defects. The key to this compatibility is that arbitrary finite quantities of evil and/or ill are compatible with the status of the world as best possible, so long as there is also infinite virtue and happiness. But this way of establishing the in-principle compatibility of evil and ill with perfection has the counterintuitive consequence that a world with massive quantities of evil and horror can be best possible. Thus, for instance, a best possible world might contain billions of murders and trillions of episodes of unhappiness inflicted upon the virtuous. This seems incorrect.

However, this is precisely the result that one should want on the basis of the fundamental principles of Kant’s practical philosophy. Kant’s practical justification for belief in the existence of God is that such belief is required for moral motivation and/or the rationality of morality. Since the concept of God is that of a being that (among other features) has a perfect will and therefore creates the best possible world, moral motivation and/or the rationality of morality requires belief that this world is the best possible world. This in turn means that, if morality is to be rational and/or capable of motivating moral action even in terrible-seeming worlds, then rational beings in such a world are committed to believing that their world is such that no world is better than it, or best possible.

Furthermore, on a Kantian theory, the rational status of morality and the capacity of the moral law to motivate moral action are both supposed to be in some sense knowable a priori. This means that, no matter how bad the empirical world seems according to any finite amount of observation, one is committed to the motivational and/or rational status of morality in this world, and therefore to maintaining belief in all those articles in which one must have belief in order to maintain that motivational and/or rational status. So whatever a Kantian theory of the best possible world looks like, it has to be compatible with arbitrarily much observed wretchedness. This is precisely what one gets according to a theory that admits infinitely valuable worlds that are valuable in virtue of containing infinitely much virtue and happiness. No amount of observed evil and ill can rule out this world’s being the best possible one.

The apparently counter-intuitive view that a wretched(-seeming) world with billions of murders and trillions of episodes of unhappiness suffered by the virtuous can be perfect is therefore precisely the view to which moral agents are committed on the basis of the fundamental principles of Kant’s practical philosophy. The infinite value view explains how moral faith, and by extension the rational and/or motivational status of morality, can be rational even in terrible-seeming worlds. And while the counter-intuitive consequence may seem too counter-intuitive to countenance, it is worth bearing in mind that, if the world conforms to the requirement of exhibiting perfect justice (which is suggested in turn by the attribution of justice to God), the virtuous sufferers ultimately will be rewarded with proportionate happiness (which might take place in an afterlife). Thus even extremely large quantities of suffering need not cause the lives of the virtuous to be unhappy on the whole.

3.4 Infinite Value and the Rejection of Theodicy

Although the infinite value view explains how Kant can endorse maximalism in a world that seems very bad (such as this world), one might worry that this interpretation is inconsistent with Kant’s rejection of theodicy in the 1791 essay ‘On the Miscarriage of All Philosophical Trials in Theodicy’. As the title would suggest, this essay subjects theodicy arguments to stringent criticism.

However, Kant’s rejection of ‘theodicy’ in the 1791 essay does not entail a rejection of the view that this world is the best possible world, or of the infinite value view. This is because Kant has a quite specific understanding of what a theodicy is. According to Kant, a theodicy is an ‘interpretation of nature insofar as God announces his will through it’ (8: 264, emphasis in original). A theodicy therefore requires providing an ameliorative account of particular ‘counterpurposive’ events that are observed in nature (8: 255). Crucially, it works by providing a kind of gloss or spin on putatively bad events. Kant holds that there are three possible strategies: either the ‘counterpurposive’ events observed are not in fact counterpurposive, or they are non-intended and unavoidable, or they are to be imputed to human beings rather than to God (8: 255). However, the infinite value view does not require providing any sort of gloss or spin on bad events. Rather, it stipulates the inclusion of enough good events to make the world best possible, without providing any account of the nature or function of the bad events, or of the reason the bad events occur. Even if, implausibly, a finite number of bad events were genuinely ‘counterpurposive’, and they were intended by God, and avoidable, and imputable to God instead of to human beings, it could still be the case that the world is infinite in value, and therefore that God creates the best possible world.

Theodicies are often used to defend the view that this world is the best possible world from the threat posed by the reality of evil and ill. But this does not mean that every such defence of the view that the world is best possible is eo ipso a theodicy. According to Kant’s account of theodicies in the 1791 essay, the infinite value view is not a theodicy.