2. Interpreting Kant on Miracles

Nuyen (Reference Nuyen2002: 319) argues that with respect to miracles Kant is an ‘empirical sceptic’, in the sense that these have no explanatory power at the empirical level, but that he is equally a ‘transcendental liberal’, in the sense that the concept of a miracle is a necessary concept of reason. As Nuyen puts it, ‘something that cannot be known can nevertheless be thought’, and ‘reason could well find it necessary to think [miracles], just as reason thinks it necessary to think the things-in-themselves’ (2002: 319). What warrants reason to ‘think’ miracles is Kant’s teleology, which ‘dictates that reason must suppose the biggest miracle of all’, namely, the miracle of the ‘creation of a world suitable for human existence’ (Nuyen Reference Nuyen2002: 319). Nuyen points to Kant’s claim in the third Critique that ‘if it is granted that cattle, sheep, horses, and the like, were to be in the world, then there had to be grass on the earth’ to argue that nature is purposive – as Nuyen puts it, Kant holds that ‘life is fine tuned for the existence of all things’ (2002: 320). Nuyen then argues that the question of the origin of this fine-tuned universe leads Kant to conclude that a fine-tuner, namely, God, has created the world according to a finality of ends (321). It is in this sense that reason in its regulative capacity is motivated to suppose a ‘supernatural basis for nature’ and, consequently, ‘to believe in miracles’ (321).Footnote ² Nuyen thus appeals to Kant’s teleology in order to account for the warrant for thinking miracles as well as for the form that miracles take. In the case of the former, because the teleological structure of nature warrants us in considering nature as the product of a creator we are also warranted in taking further miraculous acts of the creator within nature to be possible. In terms of the latter – though Nuyen does not spell this out explicitly – miracles are presumably to be understood as modelled on the act of creation in that God brings about a given effect in nature through teleological causation.

However, there are at least three problems with a straightforward appeal to teleology as a justification for miracles. First, the quotations that Nuyen cites refer to what Kant calls merely ‘external’ or ‘relative’ purposiveness (CPJ, 5: 367). An inference from this kind of purposiveness to an ‘absolute’ purposiveness established by a creator would be justified only if the beings for whom these purposes exist themselves existed for a purpose (CPJ, 5: 369). In other words, in the example cited above one would be able to assume that the grass exists for the purposes of feeding cattle only if the cattle themselves existed as an absolute purpose. However, Kant denies this is the case: to claim that such apparently purposive relations exist in order to provide an ‘advantage for certain miserable creatures’ is a ‘very bold and arbitrary judgment’, one which moreover would be ‘presumptuous and ill-considered’ (CPJ, 5: 369). For this reason, relative purposiveness ultimately ‘justifies no absolute teleological judgments’ (CPJ, 5: 369), and so Kant would deny that the inference from the apparently purposive relations we seem to find in the world to an absolute purpose in the form of a creator is warranted.

There is a second and equally significant problem with Nuyen’s teleological account, namely, the inference from creation to the possibility of miracles. Nuyen argues that since reason is compelled to accept the actuality of one class of miracles in the form of creation, it is further compelled to accept the possibility of miracles in general. However, Kant at various places explicitly denies that creation itself is a miracle: for example, in a lecture he is reported to have stated that ‘what happens outside the world contrary to the order of nature is not a miracle, e.g. creation is no miracle’ (L-Met, 29: 870). Thus, even if we were to accept the event of the creation of the world, this would in itself not warrant the inference that miracles as events within the world are possible. In this sense, Nuyen’s inference from creation, which would be a miracle outside of the world (if it is a miracle at all), to miracles operating within the world also seems unwarranted.

A third but associated problem is that the definition of miracles as ‘events in the world the operating laws of whose causes are in principle unknown to us’ that Nuyen (Reference Nuyen2002: 313) ascribes to Kant is not sufficiently comprehensive. This definition is drawn from the Religion, where Kant emphasizes that his discussion is restricted to the practical viewpoint – specifically, this definition is said to be only ‘for us’ and merely ‘for our practical employment of reason’ (Rel, 6: 86). Relying on a merely practical definition, however, threatens to miss some of the fine-grained distinctions and further qualifications Kant makes to the definition of miracles from the perspective of theoretical reason. In this latter context, he relies on a distinction between ‘course of nature’ and ‘order of nature’, where the former is understood as ‘the series of the alterations of events’ and the latter as this same series of alterations ‘but only to the extent that they stand under a general rule’ (L-Met, 28: 216). Miracles at various stages are said to be ‘exceptions’ to (L-Met 29: 872), ‘contrary’ to (L-Met 29: 870) or simply to ‘not happen according’ to the order of nature (L-Met, 28: 217). This implies that miracles are not merely ‘unknowable’ as Nuyen states but also somehow stand in opposition to the order of nature and thus, presumably, to natural laws.Footnote ³

In the context of the above discussion Kant further distinguishes between various kinds of miracles, including between what he calls ‘comparative’ and ‘rigorous’ miracles or between ‘pre-established’ and ‘occasional’ ones, and it is not obvious that the possibility of the one entails the possibility of the other, even within the world. So far we have seen that neither is ‘external’ teleology strong enough to warrant the positing of an absolute teleology nor would an absolute teleology acting from outside the world necessarily warrant an inference to the possibility of miracles within the world. We can now add that any attempt to account for Kant’s argument for the possibility or impossibility of miracles would need to take into consideration both the definition of miracles from the perspective of theoretical reason (namely, as contrary to the order of nature) as well as the above distinctions between the various kinds of miracles Kant distinguishes.

Chignell’s (Reference Chignell2014) account of Kant’s conception of miracles is able to bypass the first set problems by dropping all reference to teleology, instead providing what I shall call a ‘mechanistic’ account of miracles. Furthermore, Chignell agrees that creation is not a miracle, and paraphrases Kant’s definition in a way that takes into account the exceptional status of miracles with respect to natural laws. Namely, in Chignell’s words, miracles are said to obtain ‘when something that is not a part of nature purposively intervenes to produce an event in time that counts as an exception to a particular empirical law’ (2014: 107). Finally, he teases out some of the distinctions between kinds of miracles, arguing that Kant considers only one particular class of miracles to be possible.

Especially significant for Chignell’s account on this last point is Kant’s distinction between ‘material’ and ‘formal’ miracles, where Kant defines the former as events in which ‘even the cause is outside of nature’ and the latter as events in which ‘the cause is indeed in nature, but the determination of its action does not happen according to the order of nature’ (LM, 28: 217). He illustrates this distinction with the following example:

The difference seems to be that in the case of material miracles God himself operates as the immediate cause (in this case of the drying of the Red Sea) whereas in the case of formal miracles God’s causal efficacy is mediated by a natural cause that is nonetheless supernaturally ‘determined’, in some sense of the term. Chignell argues that only formal miracles are possible and that material miracles are not possible. The latter are ruled out on the grounds that their occurrence would constitute a violation of the third law of mechanics (i.e. the law of effect and counter-effect), which according to Chignell is a necessary law derivable from the Third Analogy. Specifically, a material miracle would ‘involve the direct introduction of new motion (force)’ that would alter the centrum gravitatis of the world (Chignell Reference Chignell2014: 111), which Kant claims is impossible (specifically with respect to miracles in Refl, 18: 321 and 419, and more generally MFNS, 4: 563 and Refl, 17: 682).

Chignell interprets the nature of the causal relationship involved in formal miracles, on the other hand, as consisting in what he calls a supernatural ‘complement’ to natural laws, which is deployed when these laws are on their own insufficient to determine an event according to divine purpose. Thus, for instance, ‘a wind that would normally cause a few whitecaps can be “complemented” in such a way that the entire sea parts’ (Chignell Reference Chignell2014: 112). According to Chignell, such miracles are possible within nature because they retain the ‘form’ of lawfulness in that ‘all alterations do have (partial) empirical causes’, which in turn makes them ‘consistent with the letter of the Second Analogy Principle’ (112). What makes such events miraculous, however, is that the above-mentioned empirical determinations are ‘only part of the total cause’, where the other part is the ‘complementary determination’ from outside of nature (112).

However, material miracles were ruled out due to their violation not of the Second but of the Third Analogy specifically, especially of the law of effect and counter-effect associated with the latter (this is Kant’s third law of mechanics; see MFNS, 4: 558). In light of this, it is necessary to ask whether formal miracles are also consistent with this law. Chignell is aware of the problem and, as he puts it, it is initially ‘unclear how such a miracle could occur without introducing action that has no reaction into the system’ (2014: 113). He finds an ingenious though (as he admits) rather ‘speculative’ solution to this problem, according to which God would introduce not only the desired effect into nature but also the requisite counter-effect, thereby satisfying Kant’s third mechanical law of the equality of effect and counter-effect. As Chignell puts it, on this model ‘a reaction equivalent to the quantity of motion or force contributed by the complement is also simultaneously added’, which avoids the aforementioned shift in the centrum gravitatis of the universe by maintaining a balance in the quantity of motion or force in the system (113). Formal miracles, then, are possible because under this model, as Chignell writes, ‘neither the transcendental laws of the Critique nor their application to matter in the “metaphysics of corporeal nature” would be broken’.

It is in the above sense that I take Chignell’s justification of the possibility of miracles to be a ‘mechanistic’ one. Unlike Nuyen’s teleological account modelled on creation, Chignell’s account proceeds by way of the mechanical laws of motion in that one species of miracles is made possible through its conformity to Kant’s third law of mechanics (though the added counter-effect is supernaturally determined). On this picture, Chignell seems correct in insisting on the inviolability of the third law of mechanics (namely, the law of effect and counter-effect) and the subsequent need for a supernatural supplementation in order to make miraculous events compatible with the order of nature as underpinned by the above law. However, in my view further conditions with respect to space in relation to the application of the third law pose a problem for the manner in which Chignell envisions such a complement. While he is right to say that the divine provision of the quantity of force contributed by the supernatural causal complement would avoid a shift of the centrum gravitatis, the same issue seen specifically from the perspective of Kant’s third law takes on a slightly different tinge. Specifically, as I shall develop in more detail below, this law seems to require the quantity of force to be localized with respect to the interaction of appearances in space rather than merely abstractly supplied into the system. In other words, the question that arises is how (or if) the source of the counter-effect itself is able to enter into empirical space.

To see why the question of space becomes an issue, it is useful to recall in what sense material miracles violate the law of effect and counter-effect. In the case of these miracles, the effect upon some A (e.g. the parting of the Red Sea) is caused immediately by B (God). However, because A cannot exercise a counter-effect on B (since B, God, is not an appearance), Kant’s third mechanical law would be violated. In the case of formal miracles, on the other hand, A is affected by a cause B (the source of the wind) that is itself a natural substance and is thus an appearance in empirical space. For this reason, A is able to exercise a counter-effect on B (through the interaction of forces between the wind and the waves, for example). Now B, however, as a causally active substance, would itself need to exercise a counter-effect on its own cause, C. On Chignell’s mechanical explanation, God simply supplies C, thus preserving the efficacy of the third mechanical law. According to Chignell it may be that a cause C operates in conjunction with B in order to bring about A, but this would not obviate the need for some D as the cause of C, upon which C would then exercise a counter-effect. In order to preserve the third law with respect to D, then, another causally active appearance in empirical space would need to be posited, and we can see that a regress arises, one reminiscent of the chain of condition-conditioned in the Antinomies. If the requirement that all correlates (A to D in this case) need to be in empirical space in order to preserve the possibility of applying the third mechanical law, then the question of how a supernatural cause ‘from outside of the world’ is able to enter into the empirical space in which mutual causal interactions between substances occur becomes pressing.

Significantly, the requirement that mutually causally interacting substances be located in space was added to Kant’s second-edition formulation of the Third Analogy (B256), with which Chignell associates the third mechanical law. Furthermore, in Kant’s most developed proof relating to the possibility of miracles, as mentioned above, the argument revolves not on the causal principle or on natural laws, but on the notion of space (Refl, 18: 320–2). The argument begins with an apparently conditional premise: ‘If miracles in space were possible, then it would be possible for appearances to occur in which effect and counter-effect are not equally large.’ From this, together with the subsequent premise that it is not possible for appearances to occur in which effect and counter-effect are not equally large, it would follow that miracles in space are not possible. But Kant does not proceed this way; instead, he continues that ‘all alterations are motions’ and that all motions are ‘represented in relative space’ (Refl, 18: 320–1). He then argues that since miracles would not stand under the law of effect and counter-effect and would hence not be represented in relative space, they would operate in ‘empty space’. This, then – namely, motion in empty space – rather than the direct violation of any putative natural laws, is what generates a contradiction establishing the impossibility of miracles. As Kant puts it, ‘a motion in empty space is a contradiction, namely, it would be the relation of a thing to a nothing, for empty space is a mere idea’ (Refl, 18: 321).

The specific nature of spatiality seems to play a significant role in Kant’s denial of the possibility of miracles.Footnote ⁴ Thus, while Chignell appears correct to argue that compatibility with the third law of motion appears to be a condition on the possibility of miraculous events, in the above passages Kant appears to set further conditions related to space – among others, that any alterations be represented in relative space and that no motion in empty space occur. However, while Chignell’s solution manages to satisfy the former condition it is unclear whether it is also able to address the latter. In the following section I examine Kant’s conceptions of relative and empty space and their role in the application of the law of effect and counter-effect with a view to determining precisely what kind of events are and are not possible within the order of nature.

3. Space, Relativity and Mechanical Laws

The relativity of space is most widely discussed in the MFNS, and is first introduced in the opening definition (Erklärung) of the Phoronomy section, which states that ‘matter is the movable in space’ and ‘that space which is itself movable is called material, or also relative space’ (MFNS, 4: 480). Because material bodies are determined based on their spatial relations to other material bodies, including to (the material body of) the perceiver, a given body must be located within an empirical space for such relations to be determined. Based on these considerations, Kant provides the following argument for the relativity of space. If the aforementioned empirical space is to provide a framework of spatial relations for locating a given body it must also be ‘an object of experience’, and so must itself be material. In this sense, Kant reasons, this material space ‘presupposes in turn an enlarged material space, in which it is movable’, leading to an ever larger progression of material spaces ‘to infinity’ (MFNS, 4: 481). He concludes that all space, as material, is relative to another given material space, and so ‘material space’ can also be called ‘relative space’ (MFNS, 4: 480).

As Friedman (Reference Friedman2013) stresses, material space serves as a frame of reference initially established relative to our bodily position that is always in principle capable of being modified, be it through expansion, reduction or shift in location. What is thus established in such modifications, according to Friedman, is not only a ‘frame of reference or coordinate system’ for locating bodies in space, but also ‘the first term of one of Kant’s “infinite progressions” of ever more inclusive relative spaces’ (2013: 44). As he puts it:

The relativity of space thus consists in the relational nature of space, the arbitrariness of the initial frame of reference for locating bodies within these relations and the subsequent possibility of establishing an ever new and distinct spatial frame from a different but equally arbitrary position.

From the relativity of space as described above, two significant consequences follow. The first is that, because motion is defined as changes in spatial relations, all motion is likewise relative to a specific spatial frame (with a significant caveat to come, as we shall see). In the MFNS, Kant formulates this important point as a ‘principle’, which reads:

In other words, it is arbitrary whether changes in spatial relations (i.e. motions) are ascribed to a body or to the relative space enclosing that body. This principle results from the above consideration that every spatial frame of reference within which the motion of a body is determined is relative, and thus arbitrary and subject to revision. As Kant puts it in an earlier text, the relative spaces in which determinations of motion are made are based on the spatial framework established by the perceiver, and in this sense a mere change in perspective on a given body in space leads to changes in the spatial relations of the body:

Thus, when I change my point of view on a body I impart motion to it. This claim must be read in the context of Kant’s conception of motion as change in outer relations: any new perspective on body’s outer relations will entail changes in those outer relations, and in this sense motion will be ‘imparted’ to the body, so to speak. As mentioned previously, it is always possible to take up such a new perspective and in this way to set a given body in motion: as Kant puts it, ‘I am completely free to expand my point of view as much as I wish, and to consider my body in relation to ever more distant surroundings’ (FGDS, 2: 16). From this it follows that all motion is relative to an arbitrary spatial framework established by a given point of view: ‘I grasp that my judgement about the motion and rest of this body is never constant, but can always rather be changed by new points of view’ (FGDS, 2: 16).

The second consequence that follows from the relativity of space is that no absolute space is possible. Because the motion of a body is determined through changes in its relations to the relative space in which the perceiver is located, any such determinations of motion will depend on the particular relative space the perceiver chooses to occupy. From the potentially infinite shift or expansion in frames of reference in the form of relative spaces it follows that the perceiver is in principle unable to take up a perspective from a non-relative or absolute space from which to determine changes in the external relations of matter:

Though such a conclusion recalls the arguments of the Antinomies according to which it is impossible to convert a totality of conditions into an object of experience, there is a subtle difference in Kant’s present argument that will become significant in the context of his refutation of miracles. It is true that experience of an absolute space containing all possible relative spaces is precluded by the finitude of our cognitive apparatus in the context of transcendental idealism. However, in MFNS Kant provides further reasons for the impossibility of absolute space that stem directly from the above considerations pertaining to the nature of relative space. Specifically, Kant argues that it would be impossible even to identify a non-relative, absolute space within experience. In order to determine the relations of a given body two correlates must be given in experience, the body and another space, and since both these correlates must be material in order to be given it follows that they are both relative. As Kant puts it, because a spatial relation ‘can be an object of experience only insofar as both correlates are objects of experience’, a non-relative, or what he calls an absolute or an empty space, is impossible for the reason that it would entail a spatial correlate that was not in relative space (MFNS, 4: 556). For this reason, absolute space ‘signifies only any other relative space’ (MFNS, 4: 481). This is so because in order to locate the putatively absolute space within experience, a second sensible (material) spatial correlate in relation to which the first space can be located would be required. But the establishment of such a relation would convert the putatively absolute space into a relative one because its location would be determined relative to a second space rather than given absolutely.

The conjunction of these three points – the relativity of space, the relativity of motion and the impossibility of absolute space – seems to leave Kant’s metaphysics of nature with a serious problem. Namely, if a non-subjective, absolute space from which to determine motion is unavailable then it becomes difficult to distinguish relative from true motion. For example, in such a framework it seems prima facie impossible to distinguish motions resulting from (objective) causal processes and motions resulting from (merely subjective) shifts in spatial frames of reference. However, Kant emphasizes that the radical perspectivalism that plagues our apprehension of the motion of bodies in space occurs at the level of the subjective apprehension of the empirical manifold, and further argues that at the level of connections between objects it is possible to determine true motion of bodies through the application of rules. What becomes the issue here is how we can move from the subject-centred, perspectival syntheses of spatial relations to objective, rule-governed ones, without appealing to a non-relative, absolute perspective. In other words, the question is how we can move from the course of nature, as the succession of motions of bodies, to the order of nature, where such succession is connected by rules into a law-governed unity. In the case of miracles, it is precisely this move that Kant rules out, and a brief look at how Kant resolves this question will allow us to determine the nature and scope of Kant’s refutation of miracles.

Kant’s solution is to argue that it is by way of the application of the third law of mechanics that we are able to fix the relations of two or more bodies by regarding changes in the spatial relations holding between these bodies not as the result of changes in the arbitrary spatial context in which they are intuited but rather as the result of mutually acting forces. Kant’s procedure is thus to construct a privileged reference frame in which the motion of two (or more) given bodies can be objectively determined by reducing these motions not to ‘the empirical space surrounding’ them but the space in which they interact (MFNS, 4: 545). In Kant’s words, when two bodies are subsumed under the law of effect and counter-effect they ‘are considered simply in relation to one another, in accordance with the influence that the motion of the one can have on the change of state of the other, abstracting from all relation to the empirical space’ (MFNS, 4: 545). In this manner, we determine the spatial location of two bodies and any changes in those relations (i.e. motions) not in relation to an arbitrary spatial frame but to one another, which in turn allows us to determine motions in the object. As Carrier has put this point, ‘if the [causal] interaction of bodies is taken into account it is no longer true that all frames of reference have equal status’ (Carrier Reference Carrier1992: 405); rather, in such interaction one body causes another to move, thereby determining motion objectively and in abstraction from relative space. In other words, the third law of mechanics enables us to determine that motion is imparted not by a shift in perspective but by a causal interaction between bodies. For this reason, as Carrier puts it, ‘recourse to causality allows for distinguishing privileged reference frame’ (405). This privileged frame of reference becomes the centre of mass frame, and it is this frame that Kant refers to as ‘absolute space’. As Friedman puts this same point, Kant ‘takes the centre of mass of any interacting system of bodies to define a privileged frame of reference for reducing observable motions within [a] system to absolute space’ (2013: 429).

In this sense, furthermore, absolute space takes on a merely regulative function in the fixing of spatial frames: it is, in Kant’s words, ‘a necessary concept of reason, and thus nothing more than a mere idea’ (MFNS, 4: 559). Despite its role in the determination of motion for the sake of a possible experience, as with the other ideas of reason Kant continues to emphasize the danger of hypostatization, where to treat absolute space as an object of experience would be ‘to transform the logical universality of any space into a physical universality of actual extent’ and so to ‘misunderstand reason in its idea’ (MFNS, 4: 482). Absolute space can be expanded beyond the centre of mass of a simple two-body system to the centre of mass of, say, the entire universe, but it nonetheless retains its merely regulative status. As Carrier puts it, absolute space in the above case becomes ‘the imaginary limit of the embedding procedure’ (1992: 403) of ever more inclusive relative spaces, but one which, nevertheless, can never take on a corresponding object that would itself be anything but merely another relative space. In this sense, furthermore, absolute space as the counter-concept of relative space in the phoronomical sense (that is, pertaining to mere motion in abstraction from any further properties) is merely another name for the empty space of the miracle-proof passage: as Kant puts it, ‘empty space in the phoronomical sense is also called absolute space’ (MFNS, 4: 563; see also 4: 556). The reason that absolute space is also called empty space in this sense is that absolute space cannot contain any matter, since matter is defined as the movable in space, and is thus empty – if absolute space contained matter, as we have seen, it would itself be movable, and would thus again be only one more relative space.

4. Miracles – What is and is Not Possible

If this reading is correct, the application of Kant’s third law of mechanics effects the transition from the merely possible succession of appearances in the course of nature to the necessary connection of appearances that makes up the order of nature. It does so by adding the particular rule that allows us to connect A and B in a manner that is constitutive of experience. For example, I push a ball (appearance B) down the length of the table with my hand (appearance A), and I can judge that the motion (change of relations) I impart to the ball (B) is the necessary effect of the force generated by my hand (A), together with any other requisite forces, such as friction. I thus subsume A and B under the third law which determines the motion of A (and B, reciprocally) in an objectively valid manner. In the case of a motion imparted by change of perspective I cannot effect such a subsumption and thus the predicates of motion I ascribe to the ball either remain at the level of mere possibility or become a Schein-judgement. The subsumption of two appearances under the third law is what allows me to fix a given set of relations, and convert them into the necessary relations that constitute an experience of the order of nature.

We can now also see what cases of motions it is not possible to subsume under the above law in a manner that would be constitutive of experience, namely, cases in which the source of the cause of a motion in appearance A is not located in a material (i.e. relative) space. Thus, if I moved appearance A with my mind (or by any other non-spatialized entity), such a motion could not be integrated into the order of nature because the application of the rule under which the two appearances could be subsumed would be blocked – such a motion would either remain at the level of appearance (and so of possible motions) or would be a Schein-judgement because one of its correlates would not be in relative space.

This impossibility is the result of Kant’s arguments, analysed above, according to which the application of the third law requires two correlates in material space, and all material space is relative space. To the extent that it ‘does not hold of the causes in the case of miracles’ that they are represented in relative space (Refl, 18: 320), it follows that miracles would be a case of precisely this kind of motion, namely, one that is impossible to integrate into experience. Consequently, any miracle that requires a relation between a non-spatial correlate and an appearance in relative space is for the same reason apodictically ruled out. To assume otherwise would be to hypostatize the idea of absolute space, which has a merely regulative function for coordinating the relation between relative spaces – in Kant’s terms it is ‘nothing’. This is the upshot of the argument in the passage on miracles where Kant concludes that ‘a motion in empty space is a contradiction, namely, it would be the relation of a thing to a nothing, for empty space is a mere idea’ (Refl, 18: 321).

In all three versions of the proof, Kant mentions a shift in the centre of gravity of the cosmic system in connection with the impossibility of motion in absolute space (Refl, 18: 321, 419; OP, 21: 431). This reference points back, first, to Kant’s discussion in the MFNS. There Kant argues that since any motion of a body is always relative to another body, the only absolute motion would be that of a body that did not have any relation to any other matter. Because all matter within the cosmos stands in relation to other matter, the only such absolute motion would be the motion of the cosmos itself. However, any ‘deviation’ from the third mechanical law would shift the centre of gravity of the cosmos as a whole, which would in turn lead to an absolute motion of the cosmos. Kant takes the ‘utter impossibility’ of such a motion to constitute an ‘apodictic proof’ for the third mechanical law.Footnote ⁵

This impossibility points, secondly, to Leibniz’s assertion against Clarke that the motion of the universe as a whole in an empty space is a ‘chimerical supposition’ (Leibniz Reference Leibniz2015: 328–9). According to Leibniz’s relational theory of space, the motion of a body can only be perceived in relation to another body; lacking such a relation, the motion would not be capable of being perceived, and hence would not constitute a case of motion at all. Clarke’s idea of an absolute space in which non-relative motions (such as the rectilinear motion of the universe as whole) can occur is, according to Leibniz, a mere ‘fiction’ (2015: 334). Though Kant agrees with Leibniz that the concept of absolute space lacks any reality, he nonetheless classifies it as an ‘idea’ that has a regulative function with respect to determining true motion, in the manner we saw above. This function does not, however, warrant us in ascribing objective reality to a mere idea – as Kant puts it in the first Critique, to do so would be to ‘hypostatize what exists merely in thoughts’ (A384). Thus, if one assumes the real possibility of a motion of the universe as a whole, one must also assume the reality of an absolute space in which such a motion could occur, which in turn amounts to a hypostatization of the idea of absolute space because one thereby ascribes objective reality to a mere thought. Consequently, Kant can conclude that because a miracle requires just such an illicit hypostatization, ‘no miracle occurs in the material world’ (OP, 21: 321).

In the passage on miracles, Kant also refers to a ‘similar’ proof that operates by way of time. The discussion is prefaced with a distinction between outer and inner miracles, where the former are said to occur in space and the latter in time (Refl, 18: 320). In the Religion Kant calls the latter ‘grace’, which he glosses as ‘supposed inner experience’ (in distinction from ‘alleged outer experience’) (Rel, 6: 53). The discussion of the proof that I analyse above begins with the words ‘If miracles in space were possible …’ (Refl, 18: 320), whereas the discussion of the temporal proof begins with the claim that ‘in a similar way it is proven that there can be no miracles with regard to appearances in time’ (Refl, 18: 321). The reference to ‘appearances in time’ is potentially ambiguous because it could refer to all appearances (given that all appearances are temporal) or it could refer to appearances in inner sense only, that is, acts of grace. Due to the structure of Kant’s argument, and the distinction that precedes it, I take each proof to target a distinct class of miracles, those that pertain to alterations in space and those that pertain to alterations that do not have a spatial correlate (i.e. grace). Furthermore, in Kant’s most condensed version of the spatial argument, one which makes no reference to the temporal argument, he limits his conclusion to the impossibility of miracles in the material world (in der körperlichen Welt, OP, 21: 321), implying both that the spatial argument targets miracles in outer sense specifically and that miracles of inner sense potentially require a distinct argument. Because my focus is on miracles within nature, I forgo here a more detailed analysis of the temporal argument against miracles of inner sense.

As we have seen, Chignell argues that Kant does allow the possibility of a certain class of miracles, namely, those that complement the natural causes of a given motion. Such miracles are possible on the condition they do not violate Kant’s third law of mechanics, which would lead to a shift in the centrum gravitatis. This condition can indeed be satisfied by Chignell’s account of miracles because the requisite measure of force can be supplied by God to prevent such a shift. However, as I have argued, there are further, spatial conditions on the integration of motions into experience, where experience is understood as the connection of appearances according to laws. These conditions preclude the possibility of motions in empty space, because such motions would require that empty space enter into the law-governed succession of appearances that constitutes the order of nature. In Kant’s words, cited above, this would represent the relation of a thing to a nothing.

The specific problem is thus relating a thing to a nothing – that is, relating an appearance in relative space to an empty space – rather than straightforwardly violating conservation laws (though this too is a problem). While in my view Chignell’s strategy would address the problem of the potential imbalance in the total quantity of motion within a system, it does not seem to apply to the more specific problem of the impossibility of causal interventions into relative space from empty space, which is what Kant states miracles would require. On Chignell’s model of formal miracles an effect A is jointly caused by appearance B in relative space and complementary cause C located in empty space. We can now see that any intervention by C into the causal chain connecting appearances in empirical space is apodictically ruled out by Kant, for the reasons given above.

Kant’s proof, furthermore, rules out both material and formal miracles as construed by Chignell, for both would require the intervention into relative space from an empty space, the former immediately and the latter mediately. This squares with Kant’s formulations in both versions of the proof in the unpublished notes (Refl, 18: 320–2, 419) as well as with a similar formulation from the Opus Postumum in what is probably Kant’s last word on the subject of miracles, where he writes simply: ‘No miracles occur in the material world … for that the world itself as a whole moves is not an object of possible experience because absolute space is not an appearance’ (OP, 21: 431). In my view, then, Kant’s disenchantment of nature is much more complete than was implied by Chignell’s and Nuyen’s accounts, because all miracles that would entail the above kind of intervention into relative space are strictly ruled out.

Kant’s proof is characteristically critical, in that miracles are events that are impossible to integrate into experience in a rule-governed manner. There is a strong epistemic element in Kant’s proof that nonetheless has metaphysical import. In a late essay he defines metaphysics as ‘the science of progressing by reason from knowledge of the sensible to that of the supersensible’ (PM, 20: 260), and analyses various attempts to reach the supersensible that inevitably miscarry due to their transgression of the limits of reason. Any attempt to establish the possibility of miracles, to the extent that this involves the hypostatization of an idea in order to make room for the requisite supersensible causal correlate, would miscarry for analogous reasons. Nonetheless, Kant seems to be open to the possibility of some kinds of miracles, and the final remaining puzzle relates to the character of these miracles and how they fit more generally into the Kantian framework, especially given the above proof. I will argue that while we cannot have knowledge of the supersensible, we can nonetheless have an oblique ‘indication’ (CPJ, 5: 196) of it, which opens up a space for one specific kind of miracle.

We have seen that miracles can be accounted for neither by teleological causation because this does not provide a sufficient warrant for accepting their possibility, nor can they be reduced to mechanical causation because an intervention into relative space seems to require a causality distinct from mechanistic causality. Prima facie, there seems to be no place to locate the possibility of miraculous causation within Kant’s architectonic. I want to suggest, however, that Kant faces a structurally analogous dilemma with respect to another class of appearances that are inexplicable solely by mechanical or teleological causality, namely, organized matter, and that his solution is in both cases similar. At the relevant juncture of the third Critique, Kant appeals to (at least) the possibility of a supersensible ground that would unite mechanical and teleological laws so that both would ‘flow’ from a common source. In this sense, both sets of laws would be able to ‘subsist alongside one another’ (CPJ, 5: 412). Specifically, mechanical laws would be ‘subordinated’ to teleological laws without a violation of the former as on Chignell’s account and without the latter having to bear an unwarranted explanatory burden as on Nuyen’s account. However, I am in this sense in agreement with Nuyen that teleology is a necessary ingredient in Kant’s account of miracles (though not with the model Nuyen provides) and with Chignell that the integrity of the order of nature must be preserved, two insights which the above model integrates.

Apart from finding a home in Kant’s architectonic, such a solution seems to be borne out by his discussions of particular cases of miracles. Kant provides the following example:

In other words, what is miraculous in the above event is not a direct intervention into chains of natural causes – in this case, the wind is a sufficient cause to bring about the given effect. What is miraculous, rather, is that natural laws were coordinated to bring about this effect at a particular time, i.e. in the moment that it would effect the overthrow of Sennacherib. Kant provides a similar analysis of the ‘passage of the children of Israel through the Red Sea’:

Again, what is miraculous is not the violation of natural laws, but their apparent subordination to a ‘special direction’ by which their effect would operate in the right moment, so to speak. Finally, Kant claims that naturalistic explanations of the biblical flood are in a sense correct but that the event is a miracle nonetheless ‘for it is still no law of nature that a comet should have flooded over the earth just then, and for the reason that its inhabitants were godless’ (L-Met, 29: 870–1). Again, we have an apparent subordination of natural causes to teleological ‘direction’ based on which the natural laws fulfil an apparently divine intention without themselves being violated.

In sum, we thus have a case of miraculous causation that is modelled neither on the teleological creation of appearances and their motions nor on a mechanical intervention into the law-governed succession of appearances in material space. Rather, we have a subordination of the latter to the former closer to the model of the supersensible substratum of nature of the third Critique. Kant introduces the supersensible substratum of nature in the context of his solution to the antinomy between two forms of causality, mechanical and teleological. Though it is unclear how precisely the appeal to the supersensible is supposed to resolve this antinomy (Watkins Reference Watkins2009: 218–19), Kant takes himself to have ‘secured’ at least the possibility that both causal orders may be ‘objectively unifiable in one principle’ (CPJ, 5: 413). His thought, presumably, is that while a given effect of nature must be viewed as embedded in a chain of mechanical causes, it is nonetheless possible that this causal chain itself responds to a teleological order. The teleological ground is used, in Kant’s words, ‘as if it were the tool of an intentionally acting cause to whose ends nature is subordinated, even in its mechanical laws’ (CPJ, 5: 422). In the context of miracles, the subordination of mechanical laws to a telos would be used as a tool enabling a given event to be the result both of mechanical causes and a divine intention. Thus, in the example above, it is possible that the causal chain leading up to the flood responds to God’s intention to punish its godless victims. However, it must be pointed out that Kant’s argument does not license us to ascribe objective validity to teleological judgements, in a manner that, for example, Nuyen’s conception of miracles requires. For this reason, any interventions into the mechanical causal order, and with it miracles based on teleological causation, are ruled out. Rather, the only license Kant’s conception of the supersensible substratum provides is to keep open the (mere) possibility that the entire mechanical causal chain itself responds to divine intention.

If this is correct, then we can in conclusion answer the question of which kinds of miracles are possible. Chignell argues that formal miracles are possible while material ones are impossible. However, Kant further distinguishes two kinds of formal miracles, occasional and pre-established, and on the model I have been arguing for only the latter would be possible. Kant defines an occasional miracle as ‘that which God effects immediately at each occasion and thereby interrupts the natural order’ (L-Met, 29: 870) – I have argued at length that such interruptions are impossible if they entail an intervention into the law-governed succession of appearances in relative space. A pre-established miracle, however, is defined as ‘what God effects mediately through natural causes, but which is used by God for special purposes, or what God already determined from the beginning and whose execution requires natural causes’ (L-Met, 29: 870). This kind of miracle, on the model I suggest, is possible because it entails a supernatural determination of events through the coordination of natural laws and divine intentions ‘from the beginning’ rather than direct violation of the former through an intervention into nature at a given point in time.

In Kant’s illustration, a miracle is pre-established ‘if already from the beginning God had made such an arrangement that the wind would have to blow in the single case of the passage of the children of Israel through the Red Sea’ (L-Met, 28: 218). In this example, there would be ‘a wind determined since eternity, which was to blow exactly at the time, where everything is arranged like a clockwork’ (L-Met, 28: 669). In my view, then, what is arranged ‘like a clockwork’ in the case of pre-established formal miracles is the harmony of teleological and mechanical causes in a manner that does not require intervention into the order of nature modelled on divine creation. Rather, God coordinates, in the supersensible realm, an event to result from mechanical causes in the phenomenal realm at a particular point in time.

In line with the above, Kant concludes one of the versions of his proof of the impossibility of miracles with the following claim:

The first part of the claim should be familiar: motions cannot begin by anything that is not itself antecedently moved because the correlate in relation to which a body is set in motion must itself be movable, and hence be located in relative (material) space. The second part of Kant’s argument comprises the claim that occasional miracles are not ‘to be found’ in the world of phenomena but pre-established miracles are. Because an occasional miracle would require a teleological act of creation or mechanistic intervention into the spatiotemporal realm of phenomena, Kant is able to rule out any such interferences with the order of nature. To the extent, however, that it is possible that the order of nature has been coordinated with a divine order beyond our cognitive ken in a supersensible ‘world that itself has no beginning’, it is nonetheless no contradiction that miracles are possible and equally that the order of nature allows no exceptions. Because, however, the principle under which the two causal orders are unified is supersensible, it can have no explanatory power, and so an appeal to divine causation to explain a given event would be illegitimate – the most one could say is that it is possible that that event is in line with divine intention. Though such a result would most likely not please a supporter of miracles, more robust forms of miracles are ruled out by Kant’s proof. It is, however, precisely this proof which at the same time allows the integrity of the order to nature to be maintained and so prevents reason from losing itself in an enchanted world.