CHRISTOPHER INSOLE | Kant and the Creation of Freedom: A Theological Problem | Oxford University Press, 2013 


By Wolfgang Ertl

In the wake of Mendelssohn’s infamous dictum, Kant is very often described as “all crushing” with regard to traditional metaphysics, in particular with regard to its theological dimension. As far as Kant’s own positive claims about God are concerned, Schopenhauer, in his Kritik der Kantischen Philosophie, argued that Kant, “when tearing down old errors and knowing the danger of the matter, only tried to temporarily push in a couple of weak support pillars through moral theology so that the collapse would not hit him and he would win time to move away” (1992:684ff.; trans. mine). With remarkable accuracy Schopenhauer’s bon mot sums up what actually happened in large parts of the reception history of Kant’s thought. While this is true mainly for philosophy, Kant’s thought was a huge challenge for theology which intersects with philosophy in the field of natural theology—a constellation similar to the case of the doctrine of practical natural law, which has traditionally been covered by jurisprudence as well.

Kant’s perceived anti-theological stance has led to an unhelpful construal of dichotomies of which Neo-Thomism and its rival, Liberal Protestantism, are perhaps particularly good examples. These movements were both heavily involved in the so-called Kulturkampf regarding cultural independence and autarchy, for example in late nineteenth-century Germany. In Neo-Thomism especially, Kant and Aquinas were seen as being at opposite ends of the spectrum, and this—albeit not exclusively—particularly as far as the feasibility of philosophical theology is concerned. To this day, it is not rare to detect something like outright hostility towards Kant among Aquinas scholars, while the attitude towards Aquinas among many Kantians is little different. Sadly, this often ideologically charged assumption of a Kant-Aquinas duopoly has turned out to be rather detrimental to intellectual historiography, since it tended to obfuscate the importance of other scholastics, such as Duns Scotus[1], and to ignore the towering influence of Aquinas (along with Scotus) in the philosophical discourse across the various Christian denominations. This influence was also palpable at the German universities well into the eighteenth century, and well beyond the theological faculties of the universities.

Christopher Insole—along with a number of other recent Kant scholars defying the orthodoxy—is swimming against the strong current of an anti-theological reading of Kant in philosophy and a strong anti-Kantian tide within theology. He tries rather to unlock some of the mysteries of transcendental idealism (TI) by taking it to be concerned with an essentially philosophico-theological question. This question is about the relationship between divine and human actions and their respective freedoms.

As Insole himself points out, he started off on the assumption that Kant’s approach amounted to what he calls “Titanism” (p. 11). According to this view, Kant held that the world needs to be accounted for in terms of the human mind, so that man could be regarded as having taken over this fundamental role traditionally assigned to God in Kant’s dogmatic predecessors. Insole emphasises that he came to reject this perhaps not altogether uncommon view, and that he now sees a substantial theistic thread in Kant’s mature theoretical philosophy (including rather strong claims about the noumenal realm), and also that he takes this thread to lead us to a better understanding of it altogether. Moreover, he thinks that, conversely, transcendental idealism can also provide important insights into the structure of the relation of divine and human freedom, in particular the idea of God somehow being involved in bringing about a free action on the part of humans.

Neither the freedom of God, nor the freedom of man, which both differs from and is similar to divine freedom, are as such a matter of course. Historically, doubts regarding God’s freedom arose perhaps most prominently in the wake of the adoption of Neoplatonist doctrines, for example in Avicenna, according to which there is a certain ‘automatism’ leading from the possibility of the world to its actualisation. Man’s freedom, by contrast, even if possible as such, appears to be under threat by certain actions and features of God, for example the very actualisation of the world even if this act itself could legitimately be called free.[2]

Insole’s project is tremendously important and provides an equally comprehensive, focused and illuminating take on familiar Kantian issues from what for many readers still is a thoroughly unfamiliar perspective. It is rare to see philosophers like Aquinas, Suárez and others along this line discussed in the context of Kant, and Insole’s book shows that invoking this context has a clarifying effect on a large number of passages in Kant’s œuvre which sometimes do have an air of obscurity when approached from a contemporary perspective.[3]

However, as much as I fully endorse invoking Aquinas’s thought as a relevant context for understanding Kant, I have reservations about Insole having retained the restricted focus on Aquinas and Kant in his attempt to overcome the perception of a wholly adversary nature of this relation. In this respect, Insole’s strategy is perhaps similar to that of the so-called ‘transcendental Thomists’ who tried to undercut the mainstream view by trying to uncover similarities in Aquinas and Kant. While Suárez, on whom Insole draws heavily, can of course not at all be reduced to being a mere Thomist, there are other scholastics—doubtless relevant to the topic at issue—who should have perhaps been considered as well. As I shall try to point out, certain rival models of the interaction of divine and human rational agency in scholasticism could be used to bring out the intricacies of Kant’s stance on this matter in a perhaps more nuanced manner.

In this respect, the work of Luis de Molina (1535–1600), one of the greatest theologians and philosophers of the sixteenth century, is particularly important. Molina’s Concordia (Molina 1953) exerted a profound and long-lasting inter-confessional influence[4] and raised a variety of issues related to human and divine freedom both in philosophy and theology. Only gradually is this influence being unearthed again in our time, whereas only a couple of centuries ago this work occupied centre stage in intellectual discourse (see Aichele & Kaufmann 2014).

That said, Insole uses his approach to shed light on Kant’s complex philosophical development, i.e. his Entwicklungsgeschichte (mainly in Chapter 4), with great success. Moreover, Insole claims that the Critical Kant is by and large a mere conservationist, transcendental-idealistically modified through the distinction between things in themselves and appearances (see p. 223). ‘Mere conservationism’, as I shall explain in more detail further below, is a position within the debate about the interplay of God as the first cause and the created entities as secondary causes and belongs to the doctrine of divine concursus. Finally, for Insole, it is by virtue of this mere conservationism with regard to things in themselves as opposed to appearances, that transcendental freedom of man, required in turn for Kant’s moral theory to prevail, can be upheld.[5]

Insole also suggests that there is a strand of ideas in Kant’s later works which, perhaps not fully explored by Kant himself, has affinities to rival conceptions in this very debate, and which may point to a different, perhaps weaker conception of human freedom.

Somewhat reluctantly, but for the sake of continuing and perhaps even expanding this discussion, I shall be focusing on two points of disagreement. These concern the so-called compatibility question regarding the threats emerging from God to human freedom. In my view, it is doubtful (i) whether, in this regard, Kant really attempted to establish that human freedom in the strong form required can be upheld by subscribing to mere conservationism regarding things in themselves in the manner Insole suggests, and whether Kant is successful in this attempt. Moreover, (ii) it is not obvious that a so-called ‘concurrentist’ view, i.e. a view according to which God’s causal activities with regard to the world exceed creation and conservation, would commit Kant to a compatibilist view in a more conventional sense.

In order to establish my points, I need to resort to a somewhat lengthy preparation, not least because the pertinent doctrines do not figure prominently in contemporary research on Kant. I shall first try and explain the different models of concursus, then lay out the various ‘facets’ of the problem of human freedom and what these models imply regarding possible threats to this freedom, and then I shall discuss a couple of key passages in Kant to substantiate my worries about Insole’s view on Kant’s stance on the compatibility question at issue.

1. Concursus

As indicated and simply put, concursus is the cooperation of a number of causes to bring about an effect. This is first and foremost a general account, similar perhaps to ideas underlying John Mackie’s (1974) theory of causality according to which the cause of an effect under scrutiny “is an insufficient, but necessary component in a bundle of factors that was unnecessary, but sufficient for the occurrence” of it (Loux 2002:197). This similarity is not explored by Insole himself, but as he rightly emphasises (pp. 214–18), concursus should not be thought of in terms of a number of different actions, but in terms of causes working together to produce one action. In contrast to Mackie, therefore, all, or at least several, factors of the bundle can count as causes in concursus.[6] In line with Mackie, though, causes are not by themselves sufficient conditions for an effect to occur. Also, concursus is usually meant to apply to efficient causality and this also holds for what follows, but of course in principle nothing stands in the way of applying it, for example, to the other three in the Aristotelian list. In any event, within classical theism this general account of concursus can and has been applied to the special case of divine concursus: Here, the issue is the interplay of God as the first cause, and created causes as secondary causes to bring about an effect.

For the sake of clarification, I shall use ‘cooperation’ as the wider concept and ‘concurrence’ as the narrower concept in the following. Mere creationists restrict divine cooperation to creation and conservation of the secondary causes (and of the true substances whose states can count as effects of secondary causes), while concurrentists think that divine causal cooperation involves more than this.

The doctrine of concursus has been investigated in rich detail by Alfred Freddoso (1988a, 1988b, 1991, 1994, 2002), on whom Insole basically relies in his account, and I shall more or less follow his lead too. As Freddoso has shown, almost everyone involved in the pertinent mediæval and early modern discussions subscribed to the creation and conservation thesis, namely that the substances in the universe, which might qualify as causes[7] apart from God, have been created by God ex nihilo and are sustained in existence or conserved at any moment in time through God. It is important and illuminating to see that the generally perhaps more well-known doctrine of occasionalism can at least in part be accounted for against the background of the problem of cooperation. In a sense, occasionalism is at one extreme end of the spectrum of possible positions, simply by denying cooperation on the part of created entities. The occasionalists (such as Al-Ghazali, Nicolas of Autrecourt[8] and, later, Malebranche) think that God is the only (genuine) efficient cause in the universe. While being the creator of the entities which make up the world and while sustaining them in existence, none of these entities contribute causally (in the sense of a [genuine] efficient cause) to what is going on in the world.

However, in the mediæval and early modern accounts of cooperation it was generally assumed that the created causes do contribute causally (as efficient causes) to changes in the world. There are, however, a number of very different views as to how this contribution works. These are mere conservationism on the one hand and concurrentism on the other. In a sense, mere conservationism is diametrically opposed to occasionalism and hence at the opposite end of the spectrum of possible positions. A mere conservationist claims that divine cooperation is restricted to creation and conservation of the created causes, while the created causes themselves do all the work of bringing about change in the world.

The classic exponent of mere conservationism is Durandus,[9] or more precisely, Durandus de Sancto Porciano (ca. 1275–1334), the Doctor modernus or Doctor resolutissimus, as he is sometimes called. Durandus is discussed at length by both Suárez in his Disputationes metaphysicae (XX–XXII) and Leibniz in his Essais de Theodicée (I. 27, III. 330, 361, 381).

Durandus’s mere conservationism, however, as Insole points out (p. 202), was a fringe position in the mediæval and early modern debate about cooperation. Most of the scholastics were concurrentists and assumed that God does indeed play a more active role than this, i.e. being involved in bringing about the changes in the world as well, adding something to the activity of created causes in this respect.

There are, however, at least two very different versions of concurrentism developed by philosophers and theologians who claim that God’s cooperation with secondary causes goes beyond mere creating and conserving them. These versions are built on different models of how God’s activity exceeds the creation and conservation of the substances.

This is on the one hand the Aquinian, or at any rate the model used by Thomists in the ensuing debates.[10] This model involves the idea that God acts on, or perhaps better, provides a causal contribution to the secondary causes, and that this activity is required to release their causal power. When it comes to rational agents, this divine contribution to secondary or created causes has somewhat polemically been called praedeterminatio or praemotio (see Hübener 1989).[11] In short: God premoves the will of a rational agent to render it active.

A quite different account has been provided by Luis de Molina. According to Molina, God acts, or perhaps better, provides a causal contribution, along the finite causes or together with them. As we shall see, Insole does not really discuss this alternative model of concursus and this does have some impact on his overall account of Kant’s stance on the compatibility question.

Most importantly, neither for the Thomists nor for the Molinists can created causes on their own ever be sufficient to bring about an effect in the world.[12]

As Insole points out (pp. 211–14), an account of miracles within these frameworks is instructive to show how concurrentism works. Along these lines, a miracle does not amount to a violation of a natural law requiring an intervention to somehow block the natural flow of events. Rather, by God withholding concurrence the natural causes on their own cannot achieve the effect they usually achieve when combined with a divine contribution. By itself the fire of the furnace, to take one prominent example discussed also by Hogan (forthcoming), cannot burn the young men in Dan 3:23.

Insole indicates, though, that while Kant himself accounts for miracles in the now-standard way of an interference, Kant—ironically enough—does that by employing the term concursus which runs counter to scholastic opinion.

One could add that the two models of concurrentism under consideration just differ in how they account for this inability of the fire to wreak havoc. According to the Thomist model God does not activate the causal powers of the created entities, while according the Molinist model Kant does not add the extra factor required for the bundle of factors to be sufficient for the effect in question.

Insole perhaps also underexposes the importance of a denial of concurrentism for a core feature of Kant’s theoretical philosophy (pp. 205–8). If one indeed wishes to use these doctrines to shed light on Kant’s thought, an important point to make is that, assuming Kant to opt for a theist stance in one way or another, mere conservationism is required in connection with what one could call the ‘methodologically naturalist’ dimension of TI. Kant is adamant that events in the world of appearances can and indeed must be accounted for by natural science, and this, plainly, presupposes that natural causes can be sufficient for bringing about an effect in this realm.[13] In any event, one reason for subscribing to mere conservationism concerns something which in and of itself poses a threat to human freedom, namely by the determinism thesis underlying this methodological naturalism.

What we have here is rather typical of the way Kant engages with doctrines of the philosophical tradition in that he often reconfigures and reassembles some of those doctrines in quite an unexpected manner. Kant’s creative engagement with doctrines of his philosophical predecessors can also be observed in the solution he provides—in my opinion at least[14]—for defusing this threat to human freedom occurring via methodological naturalism. Kant takes divine agency with regard to the world of appearances as a whole to underwrite a version of a nowadays so-called ‘altered-law compatibilism’,[15] but this is a topic for another discussion.

Of course, methodological naturalism could do well without any recourse to theological claims in the first place, but here Kant’s texts are quite clear. There is the regulative idea of the whole of the world of appearances being a divine creation and there is the idea of creation with regard to things in themselves. Although these claims transcend what can be known by us, Kant in one way or another wishes to retain the view that there is a God along the lines of the theistic conception of it.

I shall return to this issue below when talking about Insole’s account of how various—what I shall call—‘facets’ of the freedom problem are interconnected. To repeat: In Insole’s opinion, Kant thinks that the freedom thesis about human agents can be upheld if and when the creation (and conservation) thesis is applied to the agent as a thing in itself, and not as an appearance.[16]

Let me get back for a moment to the different models of concursus.  It is fair to say that Insole focuses on the Aquinian or, at any rate, Thomist account. I am not trying to press my point about whether Kant was a Molinist, but as we shall hopefully see below, Molina’s view of concursus (and in this respect, Concordia II and III are pertinent) is an important element of the very context within which Insole himself chose to assess Kant’s thoughts.

We must also distinguish various fields of this interplay of divine and created causality: The question of cooperation arises with regard to

(i) ‘ordinary’ natural causality, as we could loosely call it, not involving human agents and their will.

It arises with regard to

(ii) moral acts (in the conventional sense) of human agents.

Finally, it arises also

(iii) in the special field of human moral acts in their relation to divine acts of grace (and this, obviously, is a rather specifically theological context).

With regard to point iii, one important qualification needs to be added regarding the two main models of concurrentist positions: In the moral context concerning grace, Molina (1953, Concordia III. 41) switches to the Thomist model but disagrees with the Thomists about the scope of divine sovereignty. When it comes to grace, God’s causal contribution indeed relates to the human agent qua potential cause, but for Molina, human cooperation is needed for divine grace to unfold its potential, which a fortiori means that lack of human cooperation ‘blocks’ the efficacy[17] of divine grace. In other words, while the granting of grace through God is necessary for salvation, it is not sufficient. In Molina’s account, moreover, neither human cooperation nor the lack of it can be overwritten or compensated for by God. Put somewhat dramatically, the human capacity of cooperation is beyond the reach of divine power, and this in turn is precisely where human freedom predominantly manifests itself. In this manner, Molina places humans and God on the same level with regard to their freedom—a truly revolutionary claim, from a most unlikely source. This has some affinity to Kant’s core position in moral philosophy, namely that all rational beings equipped with a will have equal, unconditional worth. I return to this issue below.

It is with regard to the field of divine concursus concerning grace that Molina’s celebrated theory of scientia media or middle knowledge (e.g. Molina 1953, Concordia IV, 52, 9–19) has its main application, although it is not restricted to this. Molina thought that by virtue of his middle knowledge, God knows how a free agent would act in every possible circumstance, hence in particular, God by virtue of his middle knowledge knows how each agent would ‘react’ to acts of grace on the part of God.

For the majority of the Thomists, most prominently perhaps Domingo Bañez (1528–1604),[18] human agents cannot ‘prevent’ divine grace from becoming effective. Expressed in terms of language taken from Mackie this idea can be rendered as follows: God’s causal contribution in grace renders the bundle of the human and divine causes sufficient to bring about an action on the part of humans. In this sense, divine grace cannot be ‘resisted’ according to the Thomists. As we shall see, through his opposite claim, Molina can defuse a problem which arises in the Thomist model in general. This is the threat to human freedom by virtue of this very irresistibility. In Molina, God’s causal contribution does not render this bundle sufficient to bring about an action involving the human will. Clearly, in the face of the irresistibility of divine acts of grace, such an approach to human freedom on the part of the Thomists cannot essentially involve the ability to do otherwise. For other reasons, such as the externality of God as a partial cause, or the possible involvement of coercion, human freedom can nonetheless still be under threat.

Let us return to the problem of concursus. One of the likely sources of Kant is Baumgarten. In his Metaphysica, Baumgarten first defines concursus in general and then applies it to the case of divine cooperation with the activities of created causes. (It is noteworthy in this regard that, in line with Leibniz, for Baumgarten efficient causality of created monads occurs only internally to them). The pertinent paragraph is §954 of the Metaphysica, which deserves to be quoted in full:

Since the efficient causes, with the exception of God, and all the substances of this world (§319, 846) are subordinated to God, he is simply the first efficient cause, and all the rest are secondary (§315, 28). Now, all the actions of finite substances are, at the same time, sufferings caused by the other finite substances that influence them (§451). Therefore, in a mediated way, God concurs with all the actions of finite substances as an efficient cause (§314, 320). However, seeing that all the sufferings of finite substances caused by other finite substances are at the same time their very own actions (§463), not only when they are chiefly conceived of as acting, but also when they are observed to be suffering, God, who actualizes the sufficient ground of this alteration and thus their power through conservation (§953) at the very moment in which they are altered, concurs immediately with the actions of all finite substances as an efficient cause (§320). For, his action pertains to the present existence of a finite substance (§210, 55). (Baumgarten 2013:317–18)

With regard to Kant’s criticism of Baumgarten, I see things slightly differently than Insole. Insole takes Baumgarten to mean, in line with concurrentists, that the divine causal cooperation exceeds mere creation and conservation (pp. 203–8). Of course, concerning grace, this is certainly true, but it does not necessarily hold for divine cooperation tout court. In my opinion, Kant understands Baumgarten as maintaining that the divine activity of cooperation in terms of creation and conservation already amounts to concursus (in the sense of a contribution to the causal activity of the created causes, which is perhaps not implausible given Baumgarten’s view of existence as reality) and Kant tries to argue against this position, keeping mere conservationism and concurrentism strictly apart.

In this reading, Baumgarten would be a Durandist, but at the same time insists that this amounts to concursus proper, while Kant would be a Durandist (with regard to issues i and ii, to be sure) denying that conservation and creation amount to concursus. Several passages in the so-called Lectures on the Philosophical Doctrine of Religion seem to point in this direction:

For if the causes are subordinated one to another and constitute a chain or series of causes in which each is a particular link, then each link in the chain is the complete causes of the next, even if all together they have a common ground in the first cause. (V-Phil-Th/Pölitz, AA 28:1105)

Kant’s criticism may, in this sense, only be that there is some sort of conceptual confusion in Baumgarten.

2. The Multifaceted Problem of Human Freedom

Having laid out the different models of the interplay between divine and created causes, in particular human agents, we can now turn to the multifaceted problem of human freedom. I skip Insole’s important and illuminating investigations of the possibility of divine freedom and its role as a paradigm for human freedom, and instead move on directly to the threats human free actions are facing.

In this respect, we need to distinguish between threats emerging from God and threats emerging from elsewhere, for example from the activity of causes in the world of nature, which—from a theological perspective—is the realm of created causes.

On the part of God, we can distinguish the following three potential threats (overlapping with points ii and iii in the concursus discussion above):

1. His knowledge about our (future) acts might undercut the freedom of these acts. We can call this the (fore)knowledge facet of the problem of freedom.

2. His causal activity of creation and conservation might undercut the freedom of human acts due to the radical ontological dependency of human agents this may be seen to involve. Accordingly, we can speak of an ontological dependency facet of the problem of freedom.

3. His causal activity in the world in addition to creation and conservation (as the concurrentists have it) might undercut the freedom of our acts. This primarily concerns points ii and iii in the account of concurrentism sketched above, namely conventional moral acts and those acts which have an extra quality making the agent fit for redemption. With regard to such acts the overwhelming agreement of the scholastic tradition is that a divine contribution is required and is indispensable for acts of this quality to ensue. Hence, the main question under consideration has traditionally been as to whether this special form of a divine contribution is compatible with human freedom. In line with these ideas, there is something like a divine contribution, involvement, or indeed a concurrence facet of the problem of freedom, which covers point ii as well. Plainly, if one really thinks that God does more than merely create and sustain the free causes in existence, then the question arises as to whether this doing more is compatible with the freedom of these causes and their actions. To repeat, however, the issue of divine acts of grace and their compatibility with human freedom used to be of primary concern. As we have seen above, the Molinists and the Thomists differ with regard to the question as to whether acts of divine grace on the human agent can be resisted.

In Aquinas at least, we also need to distinguish two modes in which God can be regarded as a cause pertinent to human action. God is a final cause as the object of happiness in the form of the beatific vision of the essence of God (visio beatifica). Within the praemotio theory, God is moreover also an efficient cause on the will or perhaps other capacities of the free agent.

Amazingly, Aquinas sees no fatal threat to human freedom with regard to either. One might perhaps find this understandable with regard to final causality,[19] but the really problematic case is that of efficient causality. In Insole’s treatment of Aquinas (pp. 86–90), it is the fact that no coercion is involved on the part of God which makes it possible to uphold human freedom, but coercion (or its absence, respectively) is a somewhat surprising category in this respect, because coercion can arguably be absent also in the case of causation through created or finite causes while still undermining human freedom. This will hopefully become clear once we turn to the next facet of the freedom problem.

On the part of creaturely causes the following threats to human freedom emerge:

4. The causal activity of causes in the natural world different from God might be undercutting the freedom of the human agent. These activities could be both internal and external to the agent. This threat can come in at least two variants, namely (a) in the form of the natural causal determinism thesis discussed widely in the contemporary debate about the freedom problem, with van Inwagen (1983) providing perhaps the standard definition or account of this thesis. (b) This threat can come in the form of coercion through natural causes, and, clearly, (a) and (b) do not amount to one and the same thing. While coercion is perhaps obviously incompatible with freedom, the situation is very different with regard to the determinism thesis. The discussion of this compatibility question is perhaps one of the most intensely discussed issues in the contemporary philosophy of freedom. True, some have argued that one might mistakenly take the determinism thesis to be incompatible with freedom because of an unwarranted identification of determinism and coercion. Nonetheless, the incompatibility could of course arise for a different reason. The difference between determinism and coercion notwithstanding, we can speak of a natural causal facet of the freedom problem.

There are even further possible threats to human freedom, which are not necessarily linked to causality, for example a threat arising from the nature of truth:

5. One might argue that the mere truth-aptness of propositions about future human acts is incompatible with their freedom. This is of course a different issue from the one arising in the context of perfect divine knowledge. One can call this the logical or truth-theoretical facet of the problem of human freedom.

3. Divine Cooperation and Compatibilism vs. Incompatibilism

It is fair to say that Insole focuses, on the one hand, on the ontological dependency and divine contribution facet as far as theological threats to freedom are concerned, and on the other hand on the natural causal facet of it. With regard to these facets of the freedom problem under consideration, Insole provides something like a matrix of possible ways of addressing them in terms of taking either a compatibilist or incompatibilist stance. He distinguishes four different positions:

1. Theological Compatibilism A: compatibilist account of the relation between God and the creature, but incompatibilist with respect to our freedom in relation to other created causes.

2. Theological Compatibilism B: compatibilist account of the relation between God and the creature, and compatibilist with respect to our freedom in relation to other created causes.

3. Theological Incompatibilism A: incompatibilist account of the relation between God and the creature, and incompatibilist with respect to our freedom in relation to other created causes.

4. Theological Incompatibilism B: incompatibilist account of the relation between God and the creature, but compatibilist with respect to our freedom in relation to other created causes. (p. 221)

Insole then claims that in around 1755 Kant’s position was position 2, i.e. theological compatibilism B and that, in the Critical period, Kant moved to position 3, i.e. theological incompatibilism A, because he ultimately could not accept a concurrentist account of God’s relation to the world. With regard to 1755 this is, I think, right, as the dialogue between Caius and Titius in the Nova Dilucidatio (PND, AA 1:401–5) illustrates.[20] It is also true to say that, in a sense, Kant’s Critical position can rightly be labelled as 3, ignoring, for the moment, possible developments of Kant’s theory of freedom within the Critical period itself here. However, Insole appears to go further: When investigating the question as to whether this amounts to ascribing a concurrentist position to the pre-Critical Kant, a suggestion which Insole denies, he also says:

There is a conservationist version of ‘Theological Compatibilism B’: God determines human actions, so that human beings do not have the ability to do other than they do, and are not ultimately responsible for their action, although, when the human being acts, it is the human being (deterministically) acting, rather than God. (p. 222)

In this vein, there is a case to be made that conservationism on its own undercuts the alternative possibility condition (AP). However, what Insole says seems to suggest—though he himself does not put it this way—that the Critical Kant’s position can also be described as a conservationist variant of (2), i.e. theological compatibilism A, albeit—and this is a crucial qualification—in a different sense of ‘compatibilism’. This is indicated by Insole’s repeated claim (Chapter 8, passim, e.g. p. 180) that—at least in a number of important passages—Kant does not find the ontological dependency of noumenal substances on God problematic any longer, while requiring transcendental freedom for his moral theory to hold. Apparently, then, Insole takes the Critical Kant to regard the ontological dependency facet of the problem of human freedom as manageable, even though human freedom must in some manner include AP, i.e. the ability to do otherwise. For, Insole, as we shall see, AP is a key ingredient of transcendental freedom.

In my view, by contrast, (α) concurrentism does not coincide with a classical compatibilist reading, (β) mere conservationism cannot by itself deliver compatibilism in a more unconventional sense, and (γ) Kant’s own position about (β) is actually slightly different and more complex than Insole suggests. As already indicated, Insole takes the Critical Kant to hold that mere conservationism with regard to things in themselves does not undercut human freedom.

The point regarding β I am trying to make is not that mere conservationism is not concurrentism, but that—as we have seen in the outline of the different facets of the freedom problem—not all the threats emerging from God’s causal activities in the world stem from the specifically concurrentist aspects of divine cooperation. As Insole himself suggests, God might not cooperate further than creating and conserving the agents, but still determine the allegedly free causality in a way that undercuts this freedom.

In order to understand what is at issue here we need to look at the account of compatibilism and incompatibilism (e.g. p. 30). The crucial question in these discussions is indeed the importance of AP.

As we shall see, up to a point my worries are merely terminological, concerning the label to be attached to Kant’s stance on the compatibility question at issue, and do not indicate a deep disagreement. However, there is a further aspect of my objection which, as I shall try to develop in detail, concerns a more fundamental level of disagreement.

Let us look at the terminological issues first. It has often been discussed how to properly classify Kant’s stance on the question regarding the compatibility of human freedom and natural causal determinism. There is a general agreement that Kant’s position does not fit easily or neatly into the established taxonomy. It is obvious that Kant is not a typical compatibilist, given the concept of freedom he has, but on the other hand it is also obvious that he thinks the freedom thesis and the determinism thesis can, albeit with regard to different frameworks, realms or perspectives, both be upheld. A crucial question regarding Kant’s concept of human freedom is indeed the issue of AP. I completely agree with Insole that for Kant AP is basically an expression of imperfection not afflicting divine freedom (p. 184). It needs pointing out, however, that Kant’s account of human freedom involves a rather special form of AP, which one could call an ‘asymmetrical variant’ of it: It must be possible to perform a morally correct or even good action instead of a morally bad one. This corresponds with Kant’s refusal to define human freedom in terms of libertas indifferentiae, which amounts to the same as AP (see MS, AA 6: 226), operating across the basic moral categories of being able to act in accordance with and in opposition to the moral law. That said, this limited form of AP is in fact present in Kant’s conception of human freedom.

According to Insole, the two core elements of transcendental freedom are AP and the function of transcendental freedom of underwriting universal responsibility (UR) (e.g. p. 59).

For Insole, a compatibilist typically does not regard AP as necessary for human freedom, while an incompatibilist does (p. 30). This claim could, as nearly everything in the freedom debate, of course be questioned, since a number of compatibilists have tried to accommodate AP, but this is not what I am trying to get at. At any rate, in this special sense Insole has in mind (and with all of the due qualifications regarding AP) Kant can indeed be called an ‘incompatibilist’. Clearly though, as we have seen earlier, Kant does indeed think that the freedom and the determinism thesis with regard to natural causality can be upheld which, judging from the literal meaning of term ‘compatibilism’, turns him into some sort of compatibilist. Insole, drawing largely on Simon Shengjian Xie’s (2009) paper, discusses suggestions to indicate the special form of compatibilism endorsed by Kant, such as Allen Wood’s “higher order compatibilism” (1984: 74). Elsewhere (Ertl 2014), I have suggested Vivhelin’s (2000) ‘libertarian compatibilism’ as the appropriate tag, but whatever the case may be, Kant is a compatibilist of sorts. Given the established conventions, it would also make sense to call Kant an unconventional ‘soft determinist’, since in some way he seems to at least believe in the truth of the determinism thesis and in the truth of the freedom thesis.[21]

Hence, returning to the matrix proposed by Insole, Kant is rather an unconventional, indeed perhaps ‘libertarian’ compatibilist with regard to the natural causal facet of the freedom problem. How about the theological dimension? Is Kant really a compatibilist (in the unconventional sense) about the divine actions of creation and conservation, i.e. the ontological dependency facet, as Insole seems to acknowledge implicitly? This is precisely the point at which we leave merely terminological issues behind.

Put bluntly, can we secure a sufficiently strong variant of human freedom in spite of God’s creation and conservation activity with regard to the moral agents and does the Critical Kant think we can?

First of all, it is not clear whether the problem really hinges on AP, since ontological dependency might be seen to clash with UR. In his attempt to dismiss UR as a metaphysical impossibility, Galen Strawson (2010, 2011), for example, to cut a very long story short, claimed that UR involves the idea of self-creation, which is incoherent when taken literally. In a weaker sense of self-creation, it means that we must be able to turn ourselves into the way we are (mentally speaking) prior to acting on the grounds of how we are (mentally speaking). Arguably, Strawson thought that Kant’s conception of a timeless choice of character was meant to make this requirement work, although at a price too high or in the end in an unsuccessful manner for Strawson. Ironically, Strawson himself did not deal with the ontological dependency facet of the problem of human freedom nor with Kant’s treatment of it.

So, could UR be upheld in the face of ontological dependency and does Kant think it can? Insole’s claim is that for Kant TI makes the higher order compatibilism with regard to the threat emerging from creation and conservation possible.

On the face of it and independently of Kant’s stance on this matter, the odds for such a position to succeed are not good and this precisely because actually at least two facets of the problem of freedom come into play, but unlike Insole I take these to be the (fore)knowlege and the ontological dependency facet.

Can we say that a being who is rendered actual by God who knows exactly what this being will do is ultimately responsible for his or her actions? At least part, if not all of the blame, should go to God in this case, shouldn’t it?[22]

Regarding γ, Insole, however, as we shall see, treats the ontological dependency facet in the light of the natural causal facet of the problem of human freedom, and at the end of the day argues that cutting the link—prominent in the pre-Critical Kant—between these two facets through TI removes the threat to freedom stemming from ontological dependency altogether.

Here I have a problem, but I need to elucidate where exactly the problem is located in order to avoid misunderstanding. Insole’s strategy (Chapter 6) is, on the one hand, to defend Kant’s claims which are unpopular in today’s, perhaps broadly naturalist, philosophical climate, most notably the timelessness claim of noumenal causation and the claim that there can be something like noumenal first causation in the first place.

This is a formidable task and I do not take issue with it at all (although I would perhaps suggest a somewhat different line of approach), but for the sake of argument let us assume that this can be done successfully.

Moreover, he also tries to reconcile Kant’s claims with the strictures of Kant’s own demands for humility or what Insole calls ‘epistemic discipline’ (p. 6, see Chapter 7), and rightly so, since it is not obvious in a Kantian framework that we can legitimately talk about noumenal substances and noumenal first or free causation at all, even if this happened in a practical context only. Here the distinction between mere thinking (and its legitimacy) vs. gaining knowledge in a strong sense of this term is at issue. Moreover, Kant’s general strategy about the reach of pertinent metaphysical principles comes into play at this point; he seems to assume that since certain metaphysical principles can be proven to be valid only with regard to the world of appearances, conceptual space opens up for these principles not to be valid in the world of things in themselves. In this respect, not just the distinction between things in themselves and appearances requires a lot of work in order to be rendered convincing. One also needs to account for certain asymmetries in Kant’s approach, insofar as he rules out spatio-temporality for things in themselves, allows a limited applicability of pure concepts, but retains the overall architecture of things being constituted by realities for things in general, i.e. straddling the distinction between things in themselves and appearances.

Here, metaphysically austere (typically of a Neo-Kantian background) and metaphysically more liberal minded Kantians usually clash, but again let us assume the metaphysically more liberal minded Kantian can find a way of justifying these more substantial claims about noumenal first causation and the like.

Still, it is difficult to see why and how ascribing the creation thesis to things in themselves only should be able to establish a theological form of unconventional compatibilism. To repeat: The systematic problems are formidable in the light of the pressure generated by divine knowledge in combination with creation and conservation.

Why should restricting the creation and conservation thesis to noumena help, given that it is the noumenal self and/or its acts which are supposed to be free? Insole is perfectly aware of the problem arising with regard to this restriction here, when he says:

The difficulty is this: as God is still the source of the reciprocal relations between created noumenal selves, the question about how we can both be free and created seems to re-instate itself, albeit at a noumenal level. (p. 178, emphasis added).

So why does the problem not re-instate itself in TI after all? Insole thinks that it is Kant’s retraction of a principle (more precisely, two principles) that ties interaction of substances to their common dependence on God which is crucial here (pp. 178–86). According to him, Kant sees the main problem not in God creating the human noumenal substances, but in the impact other created substances may have on the allegedly free human noumenal substance. In this context Kant’s claim that God does not create appearances is important for Insole, since in the Transcendental Analytic, causal interaction is shown to happen on the level of appearances. Moreover, the community of created rational beings and God constitute a realm in which the ontological dependency is not freedom threatening according to Insole. The community of rational beings (qua things in themselves) does not form a series, as Insole is putting it.

However, whether the recourse to the community of rational beings is sufficient to remove the threat to freedom on the level of noumenal substances remains to be seen. Other rational noumenal substances might not endanger freedom, but how about other, as it were, conventional noumena or things in themselves? Again, we might want to say that Kant’s strategy of claiming that the principle of interaction can only be established on the level of phenomena together with epistemic discipline with regard to the level of noumena creates conceptual space for the claim that there are no causal threats on this level, since their might not be any interaction at all on this level. Even if all this is conceded, though, there still is a different problem here, and a problem Kant himself is investigating.

In all this, at any rate, Insole implicitly distinguishes an indirect threat arising from ontological dependency, through fellow created substances, and something like a direct threat. I think this is an important, illuminating move. With this in mind, let us look at one of the pertinent passages Insole is discussing. He is to be commended for making ample use of lecture transcripts and the Reflexionen, but I will focus on a passage from his published works, which should, however, in no way be regarded as a dismissal of the other sources. In the Critique of Practical Reason Kant says, and the somewhat lengthy passage deserves to be quoted in full:

That is to say: if it is granted us that the intelligible subject can still be free with respect to a given action, although as a subject also belonging to the sensible world, he is mechanically conditioned with respect to the same action, it nevertheless seems that, as soon as one admits that God as a universal original being is the cause also of the existence of substance (a proposition that can never be given up without also giving up the concept of God as the being of all beings and with it his all-sufficiency, on which everything in theology depends), one must admit that a human being’s actions have their determining ground in something altogether beyond his control, namely in the causality of a supreme being which is distinct from him and upon which his own existence and the entire determination of his causality absolutely depend. In fact, if a human being’s actions insofar as they belong to his determinations in time were not merely determinations of him as appearance but as a thing in itself, freedom could not be saved. A human being would be a marionette or an automaton, like Vaucanson’s, built and wound up by the supreme artist […].

The difficulty mentioned above is resolved briefly and clearly as follows. If existence in time is only a sensible way of representing things which belongs to thinking beings in the world and consequently does not apply to them as things in themselves, then the creation of these beings is a creation of things in themselves, since the concept of a creation does not belong to the sensible way of representing existence or causality, but can only be referred to noumena. Consequently, if I say of beings in the sensible world that they are created, I so far regard them as noumena. Just as it would thus be a contradiction to say that God is a 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). If it is now possible to affirm freedom without compromising the natural mechanism of actions as appearances (by taking existence in time to be something that holds only of appearances, not things in themselves), then it cannot make the slightest difference that the acting beings are creatures, since creation has to do with their intelligible but not their sensible existence and therefore cannot be regarded as the determining ground of appearances; but it would turn out quite differently, if the beings in the world as things in themselves existed in time, since the creator of substance would also be the author of the entire mechanism in this substance. (KpV, AA 5:100–2)

Insole’s distinction of an indirect and a direct threat emerging from dependency is indeed an important device for understanding this passage, but let us proceed step by step: For Kant, in transcendental realism (TR) the natural causal and the ontological dependency facet of the freedom problem arise for the same reason, namely falsely taking time to be a feature of things in themselves. In Kant’s opinion, the correct understanding of time can solve, or at least contribute to resolving and defusing the threat of the natural causal facet of the problem of freedom. Kant also says that in the light of a correct understanding of time, the creation thesis poses no threat for this facet of the problem, but he does not rule out that the creation and the conservation thesis can be problematic for other reasons.

Kant seems to be maintaining that the createdness of the noumenal self makes no difference with regard to the freedom of the agent insofar as that agent’s actions are appearances, or even with regard to the natural causal facet of the freedom problem altogether. This sounds astonishing indeed, since one of the key pillars of Kant’s strategy is precisely that the agent who performs the free action is a noumenon or thing in itself.

But then, Kant is not ascribing freedom to actions qua appearances. Rather, he says that provided that we can secure freedom (presumably in the sense of a legitimate assumption well short of genuine knowledge) for the noumenal agent, her being created cannot make a difference in this respect. In a sense, this manœuvre is quite weak and has a whiff of tautology. Of course, provided the freedom thesis can be secured, it can be secured. At any rate, Kant seems to indicate here that, in his opinion, it is not through the creation thesis that the freedom thesis is under fire.

In this sense, Insole is quite right in maintaining that Kant does not regard the creation and conservation thesis as fatal to freedom. It is quite a different issue, though, to argue whether Kant is correct here and to assess the reasons he provides why this should be the case.

Moreover, from the KpV passage quoted above we do not learn why it is so clear that in a TR environment we could not only not solve the natural causal facet of the problem, but that we would have to concede that the creator configures the agent (KpV, AA 5:102) and therefore undercuts her freedom. One way of reading Kant here could run like this: The configuration problem, as we could call it, arises because, along with the creation thesis, it seems as though we have to concede that the way a created entity acts is due to the creator. This is precisely what the simile to a Vaucanson automat is supposed to stress. Using the terminology Kant employs elsewhere (namely mainly in the Critique of Pure Reason and the Religion), we might wish to say the character of the agent would not be self-chosen, but assigned.

One plausible suggestion as to why this problem is fatal in TR might be that in a TR environment we are committed to the position that natural laws precede every action, whereas in TI the possibility opens up for natural laws to supervene on certain human actions. How this can be worked out in detail need not and cannot be discussed here. Presumably, though, Kant thinks that the priority of the laws is both temporal and metaphysical and that this must have something to do with creation itself being an act in time in TR.

Part of the reason for this priority might look like this: Within TR the act of creation, as an act in time, stands itself under laws which transfer a form of necessity holding for the past to the laws enacted in creation, and it is this necessity which underwrites the priority of these laws over the actions covered by them.

In any event, it seems to me that for Kant the problem in TR is that the direct ontological dependency of finite agents is the main issue here and not—as Insole suggests—or at any rate not exclusively that in such a scenario freedom is undercut by the causal activity of other created substances. The causality immanent to the agent whose configuration, moreover, is due to the creator leaves no space for freedom on the in-itself-level. It would seem as though Kant has Leibniz in mind here, in particular Leibniz’s contentious doctrine of monad-internal time and laws regulating the sequence of states of substances[23].

Furthermore, Kant—in the passage still under consideration—surprisingly does not elaborate on why in TI the configuration problem is manageable, a point to which we shall have to return to in a moment.

Clearly, there is indeed a thread of unconventional compatibilism with regard to the ontological dependency facet, and this unconventional compatibilism is combined with agnosticism about its justification, but there are passages in later works where Kant seems to have doubts about the feasibility of this unconventional compatibilism, which brings us to (α) in the list of disagreements above.

We can see this in a passage which Insole discusses in the last chapter of the book in which he detects a possible veering of Kant towards what he thinks is a different stance altogether, namely a more conventional compatibilist position in the light of concurrentist intuitions (as opposed to mere conservationism). In my opinion, this indeed crucial passage does not so much point to a shift towards concurrentism as that it possibly indicates an even stronger commitment to the strong conception of human freedom.

In a passage from the ‘General Remark’ to the third piece of the Religion Kant says:

It is, however, totally incomprehensible to our reason how beings can be created to use their powers freely, for according to the principle of causality we cannot attribute any other inner ground of action to a being, which we assume to have been produced, except that which the producing cause has placed in it. And since through this ground (hence through an external cause) the being’s every action is determined as well, the being itself cannot be free. So through our rational insight we cannot reconcile the divine and holy legislation, which only applies to free beings, with the concept of the creation of these beings, but must simply presuppose the latter as already existing free beings who are determined to citizenship in the divine state, not in virtue of their creation, but because of a purely moral necessitation, only possible according to the laws of freedom, i.e. through a call. So the call to this end is morally quite clear; for speculation, however, the possibility of beings who are thus called is an impenetrable mystery. (RGV, AA 6:142–3)

To begin with, two classic Kantian positions are pertinent for this passage which, however, pull in opposite directions: On the one hand, the way we think or necessarily think does not always indicate truth, as Kant’s doctrine of transcendental illusion shows. On the other hand, the way we must think does at least sometimes or in some cases coincide with truth according to Kant, as the very issue of the freedom thesis in Groundwork III seems to suggest. Kant claims there that a being who must think itself as free can, in certain contexts at least, count as free (GMS, AA 5:448). What does this mean for our unconventional compatibilism that is under consideration?

True, the evidence this text provides is not straightforward. Kant claims that it is “incomprehensible” (unbegreiflich) for our reason how the freedom thesis and the creation (and conservation) thesis can be true at the same time. This of course opens up the possibility that Kant wishes to retain the claim that they are compatible and remain agnostic simply with regard to an account or explanation of this compatibility. Still, it is also at least possible that our inability to understand compatibility here points to genuine incompatibility, so that the creation thesis indeed needs to be given up for the sake of the freedom thesis. In short, the passage allows both a reading along the lines of an unconventional compatibilism with regard to the ontological dependency facet of the problem of freedom, but at the same time a libertarian reading along the lines of what, for want of a readily available term, one might wish to call something like ‘metaphysical existentialism’. According to this position, free agents are uncreated, their existence is something like a bedrock fact.

Incidentally, one possible reason for Kant’s doubts is precisely the pressure generated by the combination of the ontological dependency and the foreknowledge facet of the problem of human freedom.

How plausible is it to assume that Kant held or at least seriously considered metaphysical existentialism? Kant—for example in his critique of the cosmological argument for the existence of God—deals with the alleged metaphysical principle that contingent substances need to have a cause for their existence and maintains that this principle cannot be established for the realm of things in themselves (A609/B637). Again, this does not amount to denying that contingent things in themselves are created, but at least it opens up the possibility of denying it. In fact, Kant’s phrasing or way of putting it in the Religion passage under consideration may simply be an expression of caution and an attempt to soften the impact of such an indeed radical claim, which the denial of the creation thesis would amount to. If he really does deny the creation thesis, Kant’s stance is remarkable in at least three respects. On the one hand, and relating to the history of ideas, such a move can be seen as both indebted to and expanding on ideas found in the broadly Scotist tradition of scholasticism. As Jacob Schmutz (2002) has shown, there were tendencies within this tradition of turning God into a mere contemplateur des idées which existed in some sense independently of God. To be sure, these idées first and foremost had to do with what is metaphysically possible, but Kant could be seen as extending this doctrine to maintaining the God-independent actuality of at least some entities, namely free agents. Secondly, this move would amount to repairing a notorious problem within a Molinist version of concurrentism since the creation thesis threatens to undercut Molina’s efforts to put the full blame on finite rational agents in case they act against the moral law (which God, by virtue of his middle knowledge, would foresee while rendering them actual nonetheless). Thirdly, understood this way, Kant could be seen as providing a possible answer to Galen Strawson’s (2010, 2011) ‘Basic Argument’ against human freedom. The existence of the agent being something like a basic fact could underwrite the so-called agent theoretical objection against Strawson according to which conceptual space can be found for free agency independently of the requirement to choose the way we are beforehand.

At any rate, Insole should perhaps have at least considered the alternative reading of Kant’s position regarding the ontological dependency facet of the problem of human freedom, namely the possibility of metaphysical existentialism.

As indicated above (see α), a further point of disagreement regarding compatibilism and Kant’s stance on it stems from Insole’s association of concurrentism and compatibilism in the conventional sense. This is true, or at least plausible, only with regard to the Thomist model. Insole’s identification, or at least close association of concurrentism and compatibilism in the conventional sense, has important consequences: If it is indeed correct that Kant is toying with the idea of concurrentism (with regard to grace) or if there is something like a concurrentist strand in some of the texts of the later Kant, as Insole suggests in the final chapter, this does not necessarily amount to Kant toying with compatibilism in the classical sense with regard to grace.

Insole seems to think that for the Critical Kant the main threat to human freedom comes from the concurrentist dimension of God’s activities, not from the creation and conservation activities which the concurrentist also assumes. I presume that it is at least partly for this reason that Insole construes a relation between Kant’s denial of concurrentism and his acceptance of Kant’s unconventional theological compatibilism.

I am of course not saying that the distinguishing mark (of God’s activities exceeding creation and conservation) in concurrentism does not pose any threat, but I think here the question of the different models turns out to be important: Arguably the Thomist model has a steeper hill to climb and it is precisely here that the question about the impact of divine actions becomes pertinent.

As I have tried to point out above, Aquinas, or at any rate many Thomists, take praemotio of the will to be compatible with freedom, even though it is irresistible. Assuming such an irresistible praemotio, however, seems to be a rather special feature of the Aquinian variant of concurrentism and does not apply to concurrentism as such, as Molina’s account shows.[24]

At any rate, it appears that something along the lines of an irresistible praemotio is ruled out in the first attempt to establish the freedom thesis in Groundwork III. Simply put, in this section Kant attempts to establish that we are under the moral law by first establishing the so-called reciprocity thesis according to which a free will and a will under the moral law is one and the same. Subsequently, Kant attempts to establish the freedom thesis for human beings (in some way short of its known truth). In this first attempt, Kant says that it is inconceivable that reason, with regard to its function of generating judgements (presumably judgments involving value), receives guidance (Lenkung) from outside, and this fits strikingly well as a rejection of an irresistible praemotio in the sense of the Thomists.

In this sense, Insole is right: there are compatibilist tendencies in this model of concurrence with compatibilism taken in the conventional sense. However, this does not link concurrentism and compatibilism in the traditional sense altogether. The Molinist model of concurrence with regard to grace does not have these compatibilist tendencies, with compatibilism taken in the conventional sense. As we have seen, it is up to the human agent to accept grace and render it efficient. The question even arises, as to whether Kant has endorsed or at least should have endorsed this model, since it fits rather well with his overall framework and the emphasis he puts on freedom. The answer to this question requires lengthy investigations in their own right and therefore needs to be postponed to a different occasion.

I should like to close by reiterating how deeply illuminating Insole’s approach is and how he manages to shed light on a number of as yet largely unexplored aspects of Kant’s thought. Moreover, he provides the resources for reassessing quite a few allegedly familiar claims by placing them in a context thoroughly unfamiliar and perhaps even alien to many quarters of contemporary Kant scholarship. He has done us all a great service by linking Kant to the Thomist tradition of scholasticism, even though the Thomist tradition is not the only one which matters in this respect.

Acknowledgements: I should like to thank Dennis Schulting and Maximilian Forschner for their very helpful comments on an earlier draft of this paper.

Invited: 21 May 2015. Received: 27 February 2017.


[1] With regard to the Scotist tradition and its relevance for Kant’s thought the work of Honnefelder (e.g. 1990, 2004), Möhle (e.g. 2003) and Kaufmann (2011) is pertinent.

[2] Various threats emerging from God to human freedom have recently been examined by Eric Watkins and Kimberly Brewer (2014), Desmond Hogan (2015), and Nick Stang (2016, esp. ch. 10) in his comprehensive account of the theological dimension at the heart of Kant’s metaphysics altogether, while Patrick Kain (forthcoming) provides an account of Kant’s conception of divine freedom and its development. As I shall try to point out in detail below, it is indeed crucial to distinguish different respects in which God threatens to undermine human freedom. When speaking about ‘threats’ emerging from God I simply take this as a short form for saying that certain truths or, more cautiously, claims about God, e.g. that he is the creator ex nihilo of agents or that he is an omniscient being, render it difficult to see how the freedom thesis can be upheld with regard to human beings. Kant’s conception of God has been the subject of Allen Wood’s pioneering work (1978), Theis (1994, 2012), Michalson (1999), Byrne (2007) and Forschner & Fischer (2012). For an approach to Kant’s relevant doctrines from a perhaps more theological perspective see Firestone (2009), Anderson & Bell (2010), as well as Essen & Striet (2005). Kant and his relation to scholasticism altogether is a somewhat neglected topic, but the topic of God, in particular with regard to questions regarding the controversy about the so-called analogy versus the univocity of being, should be a valuable vehicle for detailed comparisons which can highlight important features of Kant’s transcendental approach and his metaphysics. Kant and scholasticism has been investigated, besides by Honnefelder, in the work of, for example, Lotz (1955), Hinske (1970), Courtine (1990), Boulnois (1999), Tommasi (2008), and di Vona (2009). To some extent, there is a continuity between the scholastic tradition and German school philosophy, examined, in particular with regard to Suárez, for example by Sgarbi (2011, 2016). In the US, Allen Wood, Robert Adams and their former students have contributed strongly to rekindling and sustaining philosophical research on the theological dimension of Kant’s thought.

[3] Ironically, there are several studies comparing, for example, Kant and Aristotle or Kant and the Stoics, but there is still some kind of reluctance with regard to connecting Kant to scholastic authors, the grounds of which point to deep-seated cultural obstacles, which are themselves worthy of more thorough investigations. See Ertl (2013) for a short sketch of these cultural obstacles.

[4] See Piro (2014) for a comprehensive account of Molina’s impact on a large number of philosophers after him and Ertl (2014) for the relevance of Molina’s thoughts for Kant’s stance on the compatibility question regarding freedom and natural causality. As I shall reiterate in a forthcoming article, my claim is not that Kant is following Molina with regard to reconciling human free will and divine foreknowledge, as one might perhaps think.

[5] Desmond Hogan (forthcoming) pursues a similar line of investigation but maintains that there remains an unresolved tension between Kant’s adherence to mere conservationism and his commitment to transcendental freedom. I basically agree with Hogan, but as I shall try to show below, Kant, in my opinion, may have been prepared to give up the claim that human free agents are created entities.

[6] Although for Mackie different factors of the bundle can count as the cause of the occurrence, depending on the context of the enquiry. In any case, the similarity of Mackie’s account and concursus is surprising since Mackie usually counts as a Neo-Humean theoretician of causality.

[7] Generally speaking, efficient causes are considered to be substances by the participants of this debate, which may be seen as being at odds with Kant’s approach usually regarded as taking events to be the relata of causality. Eric Watkins (2005) has challenged this view and argued that Kant’s account of causality is far more Aristotelian by also assuming causes to be substances. I largely agree with Watkins, but cannot discuss this matter in detail here.

[8] Of course, I need to gloss over many important details here.

[9] Durandus’s account can be found in his commentary on Petrus Lombardus’s twelfth-century Quatuor Libri Sententiarum (The Four Books of Sentences), i.e. one of the most important textbooks in the history of mediæval and early modern thought channeling patristic ideas to subsequent eras and shaping the context and even the way many fundamental theological as well as philosophical questions have been addressed. Important passages of this commentary have been translated and discussed by Freddoso (1994).

[10] I leave it open here what Aquinas’s own position is on this matter.

[11] This term, admittedly, by virtue of its having polemical overtones needs to be handled with care.

[12] Accordingly, for concurrentists a human agent cannot be the complete or sole cause of an action and the question is whether this by itself undercuts the freedom of actions, which I think is not the case. There are a number of conditions required to hold for an agent to perform an action. Moreover, even if some of these conditions might be external to the agent, this by itself does not rule out that the action in question is the agent’s. As we shall see, much depends on the specific features of the conditions involved.

[13] Although Hogan (forthcoming) claims that for Suárez, for example, the point at issue here is a metaphysical condition of secondary causality and hence something altogether different from questions of methodology in science, I still think the sufficiency problem matters here.

[14] See Ertl (2004, 2014).

[15] According to altered-law compatibilism there is a dependency relationship between the laws of nature and human acts of freedom, that is, had the agent acted differently, different laws would have been valid.

[16] I ignore the question here as to what this means with regard to treating the world of appearances as a whole as a divine creation.

[17] I omit a discussion of the original terminology regarding the sufficiency and efficiency of grace.

[18] For a concise account of Bañez’s position and the Molina-Bañez-controversy, see Freddoso (1988a, esp. pp. 36–42).

[19] Aquinas’s position with regard to final causality was the topic of intense debates in the Middle Ages. For example, final causality was not really causality proper for Scotus (see, e.g. Möhle 2003:324–6), and therefore, in his opinion, an external final cause presumably cannot be a threat to human freedom. Strikingly, Kant appears to see this differently. At least it seems plausible to assume that in his short argument for the so-called reciprocity thesis in Groundwork III (GMS, AA 4:446), he wishes to rule out determination through foreign final causes as well—which we can plausibly take as being the same as ends there—in his negative account of freedom of will.

[20] See Byrd (2008) for an important critical discussion of Kant’s position at that time.

[21] The background assumption, at any rate, must not be that AP and determinism are incompatible, as it is sometimes the case, since this is precisely one of the key contentious points in the debate about freedom.

[22] As I shall try to explain on a later occasion, this difficulty points to a major weakness in Molina’s account.

[23] See Ertl (2010) for a more detailed account of this line of thought.

[24] Hogan (forthcoming) suggests that Molina’s (and Suárez’s) pertinent doctrine is “not explored by Kant and it is unclear how much he understands of it”. He maintains, though that this doctrine was meant to reconcile ontological dependency and freedom, but I think it is doubtful whether Molina and Suárez are successful in this regard. Kant’s possible endorsement of metaphysical existentialism in my view amounts to an attempt at rectifying these difficulties while at the same time remaining true to the spirit of Molina’s original insights, namely that human freedom is the source of the truth of fundamental contingent truths beyond the reach of the divine will. At any rate, even if Kant had indeed been unfamiliar with these doctrines, it is a different matter altogether as to whether they are crucial for assessing the relationship of concurrentism on the one hand and compatibilism as well as incompatibilism on the other.


Primary sources

Baumgarten, A. G. (2013), Metaphysica. A Critical Translation with Kant’s Elucidations, Selected Notes, and Related Materials, trans. and ed. C. Fugate & J. Hymers (London & New York: Bloomsbury).

Molina, L. (1953) Liberi arbitrii cum gratiae donis, divina praescientia, providentia, praedestinatione et reprobatione Concordia, ed. I. Rabeneck (Oniae, Matriti).

Schopenhauer, A. (1998), ‘Kritik der Kantischen Philosophie’, in Die Welt als Wille und Vorstellung, vol. I (Frankfurt/M.: Insel), pp. 559–715.

Secondary literature

Aichele, A. & M. Kaufman (2014), ‘Introduction’, in Aichele & Kaufman (eds), A Companion to Luis de Molina (Boston & Leiden: Brill), pp. viii–xxxiii.

Anderson, P. & J. Bell (eds) (2010), Kant and Theology (London & New York: T&T Clark).

Boulnois, O. (1999), Être et représentation (Paris: PUF).

Brewer, K. & E. Watkins (2012), ‘A Difficulty Still Awaits: Kant, Spinoza, and the Threat of Theological Determinism’, Kant-Studien 103(2): 163–87.

Byrd, J. (2008), ‘Kant’s Compatibilism in the New Elucidation of the First Principles of Metaphysical Cognition‘, Kant-Studien 99:68–79.

Byrne, P. (2007), Kant on God (Aldershot: Ashgate).

Courtine, J.-F. (1990), Suarez et le système de la métaphysique (Paris: PUF).

Di Vona, Piero (2009), L’ontologia dimenticata. Dall’ ontologia spagnola alla Critica della ragion pura (Napoli: La città del sole).

Ertl, W. (2004), ‘Schöpfung und Freiheit. Ein kosmologischer Schlüssel zu Kants Kompatibilismus’, in N. Fischer (ed.), Kants Metaphysik und Religionsphilosophie (Hamburg: Meiner), 43–76.

——— (2010), ‘Persons as Causes in Kant’, in S. Palmquist (ed.), The Unity of Human Personhood (Berlin & New York: de Gruyter), pp. 217–30.

——— (2013), ‘”Kant und die Scholastik heute”. Vorüberlegungen zu einer Neueinschätzung’, The Geibun-Kenkyu 105(2): 20–40.

——— (2014), ‘”Ludewig” Molina and Kant’s Libertarian Compatibilism’, in M. Kaufmann & A. Aichele (eds), A Companion to Luis de Molina (Boston & Leiden: Brill), pp. 405–45.

Essen, G. & M. Striet (eds) (2005), Kant und die Theologie (Darmstadt: Wissenschaftliche Buchgesellschaft).

Firestone, C (2009), Kant and Theology at the Boundaries of Reason (Farnham: Ashgate).

Fischer, N. and M. Forschner (eds) (2012), Die Gottesfrage in der Philosophie Immanuel Kants (Freiburg: Herder).

Freddoso, A. (1988a), ‘Introduction’, in L. de Molina, On Divine Foreknowledge (Part IV of the Concordia), trans. A. Freddoso (Ithaca & London: Cornell University Press), pp. 1–81.

——— (1988b), ‘Medieval Aristotelianism and the Case against Secondary Causation in Nature’, in T. Morris (ed.), Divine and Human Action. Essays in the Metaphysics of Theism (Ithaca & London: Cornell University Press), pp. 74–118.

——— (1991), ‘God’s General Concurrence with Secondary Causes: Why Conservation is not Enough?’, Philosophical Perspectives 5: 553–85.

——— (1994), ‘God’s General Concurrence with Secondary Causes: Pitfalls and Prospects’, American Catholic Philosophical Quarterly 67: 131–56.

——— (2002), ‘Introduction’, in F. Suarez, On Creation, Conservation and Concurrence, trans. A. Freddoso. (South Bend: St Augustine’s Press), pp. xi–cxxiii.

Hinske, N. (1970), Kants Weg zur Transzendentalphilosophie. Der dreißigjährige Kant (Stuttgart: Kohlhammer).

Hogan, D. (2014), ‘Kant on Foreknowledge of Contingent Truths’, Res Philosophica 91(1): 47–70.

——— (forthcoming), ‘Kant’s Theory of Divine and Secondary Causation’, in B. Look (ed.), Leibniz and Kant (Oxford: Oxford University Press).

Honnefelder, L. (1990), Scientia transcendens. Die formale Bestimmung der Seiendheit und der Realität in der Metaphysik des Mittelalters und der Neuzeit (Hamburg: Meiner).

——— (2004), ‘Metaphysics as a Discipline: From the “Transcendental Philosophy of the Ancients” to Kant’s Notion of Transcendental Philosophy’, in R. Friedman & L. Nielsen (eds), The Medieval Heritage in Early Modern Metaphysics and Modal Theory, 1400-1700 (Dordrecht: Kluwer), pp. 53–74.

Hübener, W. (1989), ‘Praedeterminatio physica’, in J. Ritter & K. Gründer (eds), Historisches Wörterbuch der Philosophie, vol. 7 (Basel: Schwabe), pp. 1216–26.

Kain, P. (forthcoming), ‘The Development of Kant’s Conception of Divine Freedom’, in B. Look (ed.), Leibniz and Kant (Oxford: Oxford University Press).

Kaufmann, M. (2011), ‘Die Willensfreiheit, das moralisch Gute und das Ziel des Menschen bei Duns Scotus, Ockham und Molina’, in H. Busche (ed.), Departure for Modern Europe. A Handbook of Early Modern Philosophy (1400-1700) (Hamburg: Meiner), pp. 167–79.

Lotz, J. (ed.) (1955), Kant und die Scholastik heute (Pullach: Berchmannskolleg).

Loux, M. (2006), Metaphysics. A Contemporary Introduction (London & New York: Routledge).

Mackie, J. (1974), The Cement of the Universe. A Study of Causation (Oxford: Oxford University Press).

Michalson, G. (1999), Kant and the Problem of God (Oxford & Malden, MA: Blackwell).

Möhle, H. (2003), ‘Scotus’s Theory of Natural Law’, in T. Williams (ed.), The Cambridge Companion to Duns Scotus (Cambridge: Cambridge University Press), pp. 312–31.

Piro, F. (2014), ‘The Philosophical Impact of Molinism in the 17th Century’, in M. Kaufmann & A. Aichele (eds), A Companion to Luis de Molina (Leiden & Boston: Brill), pp. 365–403.

Schmutz, J. (2002), ‘Un Dieu indifférent. La crise de la science divine durant la scolastique moderne’, in O. Boulnois et al. (eds), Le contemplateur et les idées. Modèles de la science divine, du Néoplatonisme au XVIIIe siècle (Paris: Vrin), pp. 185–221.

Sgarbi, M. (2011), Logica e metafisica nel Kant precritico. L’ambiente intellettuale di Königsberg e la formazione della filosofia kantiana (Frankfurt: Peter Lang).

——— (2016), Aristotle and Kant: Epistemology, Logic, and Method (Albany, NY: SUNY Press).

Shengjian Xie, S. (2009), ‘What is Kant: A Compatibilist or an Incompatibilist? A New Interpretation of Kant’s Solution to the Free Will Problem’, Kant-Studien 100: 53–76.

Stang, N. (2016), Kant’s Modal Metaphysics (New York: Oxford University Press).

Strawson, G. (2010), Freedom and Belief, revised ed. (New York: Oxford University Press).

——— (2011), ‘The impossibility of ultimate responsibility?’, in R. Swinburne (ed.), Free Will and Modern Science (Oxford: Oxford University Press), pp. 127–41.

Theis, R. (1994), Gott. Untersuchung zur Entwicklung des theologischen Diskurses in Kants Schriften zur theoretischen Philosophie bis hin zum Erscheinen der Kritik der reinen Vernunft (Stuttgart-Bad Cannstatt: Frommann-Holzboog).

——— (2012), La raison et son Dieu. Étude sur la théologie kantienne (Paris: Vrin).

Tommasi, F. (2008), Philosophia transcendentalis. La questione antepredicativa e l’analogia tra la Scolastica e Kant (Firenze: Olschki).

Van Inwagen, P. (1983), An Essay on Free Will (Oxford: Clarendon).

Vihvelin, K. (2000), ‘Libertarian compatibilism’, Philosophical Perspectives 14: 139–66.

Watkins, E. (2005), Kant and the Metaphysics of Causality (Cambridge: Cambridge University Press).

Wood, A. (1978), Kant’s Rational Theology (Ithaca & London: Cornell University Press).

——— (1984), ‘Kant’s Compatibilism’, in A. Wood (ed.), Self and Nature in Kant’s Philosophy (Ithaca & London: Cornell University Press), pp. 73–101.

© Wolfgang Ertl, 2017.

Wolfgang Ertl is Professor of Moral Philosophy at Keio University, Tokyo. He has held visiting positions at University College London, Humboldt University, Berlin, and the University of St Andrews, UK. He is the author of Kants Auflösung der ‘dritten Antinomie’: Zur Bedeutung des Schöpfungskonzepts für die Freiheitslehre (Karl Alber, 1998) and of David Hume und die Dissertation von 1770. Eine Untersuchung zur Entwicklungsgeschichte der Philosophie Immanuel Kants (Peter Lang, 1999). His research interests focus on topics at the intersection of metaphysics and ethics, and he is currently working on a book on the relevance of scholastic thought for an account of the core features of Kant’s theoretical and practical philosophy.