I am grateful to Uygar Abaci and Peter Yong for the stimulating papers in this issue of Kantian Review. Although I will focus here on responding to their objections, it should be obvious that whenever one's work is carefully and constructively engaged in this kind of detail, the main response can only be one of thanks.
Abaci and Yong argue that the pre-Critical texts do not support what Yong refers to as my ‘Real Harmony Interpretation’ (RHI) of Kant's possibility proof; they also think that there are philosophical reasons to resist it. In brief, RHI (an acronym which I happily adopt here) says that, for Kant, there are three conditions on something's being ‘really’ possible – the existence of the basic content or ‘data’ of the predicates that constitute its concept, the logical consistency of those predicates, and the real or metaphysical harmony between those predicates (I call these ‘Content’, ‘Consistency’, and ‘Harmony’ for short).Footnote 1 I argued in the 2009 paper that Kant's proof of a unique, necessary and ‘most real’ ground of possibility only goes through if we interpret him as endorsing each of these conditions (Chignell Reference Chignell2009a). In a follow-up paper, I argued that endorsing Harmony also has a cost – it raises the ‘threat of Spinoza’ (Chignell Reference Chignell2012). By that I meant that the proof does indeed deliver a unique, necessary and ‘most real’ ground of possibility, but one that threatens to take the form of a ‘monstrous’ (ungeheur) being that exemplifies all possible predicates at once (for Spinoza's ‘monstrous’ idea, see Kant Reference Kant1993: The Only Possible Argument (OPA), 21: 50).
Abaci regards content and consistency alone as jointly sufficient, and so thinks he can avoid the Spinozistic threat from Harmony. Yong, by contrast, agrees with me that Kant's proof needs to appeal to the harmonious metaphysical relations between the predicates of possibilia in order to respond to ‘the plurality objection’ – i.e. the objection according to which the grounding work could be done by more than one necessary being. But Yong's reading of Harmony is different from mine, and this allows him to evade the threat of Spinoza.
My goals here are to discuss the textual concerns raised by Abaci and Yong together, and then to consider some of the most pressing philosophical concerns raised by each. Along the way I will note where these replies meet objections raised in other venues by Nicholas Stang and Anja Jauernig, although a full response to the latter two is a job for another time.
1. Is there Really ‘Real Modality’ in the Pre-Critical Texts?
Abaci begins by asserting that it is ‘not textually obvious that the Kant of the 1750s and early 1760s had a distinct notion of “real modality” as opposed to the “logical modality” of the Leibnizian-Wolffian school’ (3). The distinction Kant was using in that period, Abaci says, was between logical and real ‘elements or conditions’ of possibility, rather than logical and real ‘conceptions or kinds’ of possibility. He also notes that ‘[i]n fact, Kant does not use the term “real possibility” in any of his pre-Critical works’, though he concedes in n. 6 (22) that the phrase does turn up in a pre-Critical reflection from 1769–70 (Refl. 4196; 17: 452), and also that Kant clearly employs an ‘extra-logical conception of possibility’ in a reflection from the same period (Refl. 4004; 17: 382).
Abaci is right about the terminological fluctuation – in OPA Kant speaks more often of ‘the real’ (das Reale) in the ‘internal possibility’ of things rather than of ‘real possibility’ per se (1763a: OPA, 2: 78). But there is nothing in the concept of what he later calls ‘real possibility’ or ‘real necessity’ that seems particularly Critical; in fact, the idea of a modal status that attaches (in Abaci's words) to ‘the being or existence of a thing’ rather than to our way of cognizing it has a decidedly pre-Critical flavour. Abaci himself seems to recognize this: a few pages later he allows that ‘Kant assigns a substantial role to the notion of absolute real necessity in OPA’ and that ‘it can be said with fairness that the proof in OPA is concerned with (absolute) real modalities in distinction from logical modalities’ (5). So whatever the terminological fluctuations, all parties to the dispute agree, I think, that Kant invokes metaphysical-modal notions in his early writings, and that the most plausible interpretation of them is that they refer to the status of the ‘being or existence of things’.
2. What about Subject-Cancelling Real Repugnance?
2.1. Texts Against
More substantive textual issues come out in section 3 of Abaci's paper. He notes, first, that Kant's most extensive discussion of ‘real repugnance’ (Realrepugnanz) or ‘real opposition’ (reale Entgegensetzung) appears in the Negative Magnitudes essay of 1763, and that there Kant is focused on the way two opposed albeit ‘real’ motive forces can ‘annihilate’ (vernichtigen) or ‘cancel’ (aufheben) one another's effects. The Only Possible Argument was published that same year, and Kant uses many of the same examples – for instance, that of two opposed but equally powerful forces that ‘cancel’ one another such that a body is left at rest (1763b: NM, 2: 171; 1763a: OPA, 2: 86; compare Kant 1787: CPR, A272/B328).
Abaci takes this to be textual evidence for the claim that in all real repugnance relations ‘it is not the subject that is cancelled but only the opposing predicates or, more precisely, their effects’ (7). This is something that Yong also suggests, and that Anja Jauernig said in an unpublished commentary on these ideas delivered at the APA:
Although the discovery of real repugnance or real opposition, as he more often calls it, in the essay on Negative Magnitudes marked quite an important step in Kant's philosophical development, to my mind he never understood it as a threat to the real possibility of objects, be it sensible or supersensible ones. (Jauernig Reference Jauernig2006: 2)
My sense is that the critics have mistaken a species for the genus. I agree that Kant's presentations of the concept of real repugnance typically involve predicate-cancelling cases we can easily perceive or imagine – cases from mechanics, nautical science, empirical psychology, arithmetic and so forth. But I take this to be a result of the facts that (a) Kant himself seems to have become aware of the difference between logical and real opposition by reflecting on such cases and (b) Kant continued to think, throughout his career, that the best way to exhibit real repugnance (and lots of other philosophical notions) is by appeal to sensory, easy-to-imagine examples, even if the concept applies more broadly.
Thus in the first Critique Kant will say that Leibniz's thesis that positive predicates cannot be incompatible ‘is an entirely true proposition about the [logical] relations of concepts, but indicates nothing at all either in regard to nature or overall in regard to any thing in itself (of this we have no conception)’. This at least invokes the idea of non-logical, real oppositions amongst the properties of non-sensible things, but Kant immediately goes on to say that the clearest cases of such opposition are the predicate-cancelling ones ‘unceasingly placed before our eye by all hindrances and countereffects in nature’ (Kant Reference Kant1781/7: CPR, A273/B329). On the next page we find something similar: on the one hand, Kant condemns Leibniz and company for failing to ‘worry about opposition’ amongst the properties of a perfect ‘being that unites all reality’ but, on the other hand, admits that ‘the conditions for the representation of [such opposition] are only found in sensibility’ (Kant Reference Kant1781/7: CPR, A274/B330).
Kant's focus on predicate-cancelling examples in mechanics and psychology, then, seems like a methodological expedient in light of the fact that it is hard to grasp states or relations to which we cannot attach intuitional content.Footnote 2 But this does not imply that such states or relations cannot obtain outside of sensible nature. The very titles of the sections in Negative Magnitudes reflect this: Kant starts by elucidating the ‘concept in general’, then provides a series of empirical and psychological ‘examples’, and only later, in the third section, offers ‘some reflections which may serve to prepare for the application of the above concept to the objects of philosophy’ (1763b: NM, 2: 189). We find something similar in the third Critique with respect to concepts like ground, depend, flow from and even substance: Kant says we grasp these by way of empirical examples, but then go on to apply them, via analogical extension, in logical and philosophical contexts. This involves ‘the transportation of the reflection on one object of intuition to another, quite different concept, to which perhaps no intuition can ever directly correspond’ (Kant Reference Kant1790: CJ, 5: 352).
My suggestion is that Kant noticed the species of real repugnance that obtains between the predicates of some empirical objects, and then ‘transported his reflection’ to the concept of another species – absolute real repugnance, we might call it – that can obtain in non-sensible contexts too. He also noticed that such real repugnance, which is not relative to some set of empirical forces or dimensions, would be able to ‘cancel’ (aufheben) the subject as a whole. This is not an exotic idea: subject-cancellation is precisely what happens in familiar cases of logical contradiction. If the predicates <being A at t> and <not being A at t> are combined in the concept of something, the logical repugnance between them ‘cancels the subject’ altogether –such a thing cannot (logically speaking) exist. Similarly, if <being a human person> and <being made of ice> (this is Yong's Kripke-inspired example [29]) are contained in the concept of something, the real repugnance between them ‘cancels the subject’ altogether – such a thing cannot (metaphysically speaking) exist.
According to RHI, then, Kant's view is not that all real repugnance must be absent in order for something to be really possible; rather, all subject-cancelling repugnance must be absent. This is what Harmony says.
Abaci's other main piece of textual support is Kant's claim in Negative Magnitudes that the result of real repugnance is cogitabile and repraesentabile – i.e. something that can be consistently thought and represented. Abaci (7) takes this to mean that the subject of the repugnance remains and is not cancelled (Kant Reference Kant1763b: NM, 2: 171). But RHI openly acknowledges that the subjects of real repugnance can be represented in thought, even if some of them turn out to be impossible. Indeed, one of the central theses of RHI is that some real impossibilities are still logically possible and thus ‘thinkable’ in Kant's sense.
2.2 Texts in Favour
Having said all this, I concede that there is not as much as one might like by way of positive textual support for Kant's embrace of subject-cancelling real repugnance. I wish Kant had said more (or perhaps the same amount, but more clearly). I do offer a few passages in my earlier papers; let me now turn to Abaci's and Yong's assessments of those.
First, there is a passage in the third section of Negative Magnitudes where Kant posits the cryptic principle that ‘every passing-away (Vergehen) is a negative coming-to-be (Entstehen)’. He explicates it this way:
In other words, for something positive which exists to be cancelled, it is just as necessary that there should be a true real ground [of that cancellation] as it is necessary that a true real ground should exist in order to bring it into existence when it does not already exist. … In other words, only insofar as an equal but opposed real ground is combined with the ground of A is it possible for A to be cancelled. (Kant Reference Kant1763b: NM, 2: 190)
This looks on the face of it like an abstract principle that applies to things (‘something positive’) and their possible existence and non-existence. But, as Abaci points out, Kant goes on to illustrate the principle with cases of opposed predicates (thoughts, movements, temperatures, etc.) that come into and out of existence in a single persisting subject, rather than cases of opposed predicates that render their subject impossible. Abaci takes this to indicate that the concept of real repugnance itself is restricted to predicate-cancelling cases. But again, I think it is equally plausible to interpret these as empirical illustrations of a principle that may well have broader application.
It is worth pausing here to consider the ‘potential defence of Chignell’ that Abaci offers in this context, a defence that would allow us to generate a case of subject-cancelling real repugnance from any predicate-cancelling case (9). Consider the predicates <being acted on by force F> and <being acted on by force F*>, and assume that these forces are positive, opposed and equally powerful. As we have seen, Kant regards this as a case of predicate-cancelling real repugnance – the subject that has them can still exist. But now consider a concept of something that possesses the complex predicate <being acted on by equally powerful opposed forces F and F* and by no other forces> as well as the predicate <moving in the direction of force F>. This object is logically possible, but it also appears to be really impossible. So this is a subject-cancelling counterpart to Kant's predicate-cancelling example.
I think Abaci is right about this, and I happily accept the defence. Having offered it with one hand, however, Abaci seeks to take it away with the other by arguing that the first predicate – <being acted on by equally powerful opposed forces F and F* and by no other forces> – in fact entails the predicate <being at rest>. The latter predicate, he then points out, is logically incompatible with <moving in the direction of force F >. So the repugnance here turns out to be logical after all.
Needless to say, I want to resist this last move. I think we can do so by taking a closer look at the claim about entailment. What Abaci says is this: ‘F and F* cancel out each other's effects on A and since there is no other force acting on it A will remain at rest’ (10). But what is the force of ‘since’ here? Nothing in the rules of logic guarantees that a subject with the first predicate (i.e. <being acted on by equally powerful opposed forces F and F* and by no other forces>) will be at rest. Indeed, if there is any necessary connection between a body not being dynamically acted on and not moving, it is presumably a real or metaphysical one. But then we do have a subject-cancelling real repugnance at bottom after all: a being that is acted on by both F and F* and no other forces and still moves in the direction of F is impossible, but not for mere logical reasons. If this is right, then the ‘potential defence of Chignell’ may still have some potential.
With this in mind, let us turn to the second text, this time from OPA itself. I grant the critics’ point that there is a logical opposition between the predicates <being the supreme being> and <lacking understanding and will>, provided that <having understanding and will> picks out a fundamental good-making property that a supreme being must have.Footnote 3 Abaci, Yong and Jauernig all argue that in such discussions Kant is merely making the rather obvious logical point that a being that has all the best properties cannot also ‘be deprived of some of them’ (Abaci 9 and Yong 32–3; Jauernig Reference Jauernig2006: 2–3).
But note that the example that I was citing in this context (though not, I agree, with exemplary clarity) has to do with the repugnance that Kant spies within the complex predicate <emanating the actual universe> (see Chignell Reference Chignell2012: 646). In the compressed discussion of this in OPA, Kant seems to analyse that predicate into two more basic ones: <being a ‘blindly necessary ground of other things’> and <producing a universe full of ‘order, beauty, and perfection’ > (Kant Reference Kant1763a: 2: 88–9). He also seems to think that a blind emanator cannot produce a universe with the kind of ‘order, beauty, and perfection’ that we find in this one:
Now the necessary being is the sufficient real ground of everything else which is possible, apart from itself. It follows that the necessary being will possess that property, in virtue of which everything else, apart from itself, is able to become real in agreement with these relations. However it seems that the ground of the external possibility of order, beauty, and perfection is not sufficient unless a will in agreement with the understanding is presupposed. These properties [of will and understanding] must, therefore, be ascribed to the Supreme Being. (Kant Reference Kant1763a: OPA, 2: 88)
We might wonder about Kant's claim here – why think that a Neoplatonic One is not ‘sufficient’ to emanate a very pretty universe? The crucial point, however, is that, although it is not a logical repugnance that Kant has in mind, it is clearly subject-cancelling: an emanator of a beautiful, ordered universe, he suggests, cannot really exist.
Thirdly, Abaci admits that there is at least one passage in OPA ‘that does seem, however, to support [Chignell's] case for Kant's employment of the concept of “subject-cancelling” real repugnance’ (10). It comes just after the quoted passage above: Kant says that ‘the impenetrability of bodies, extension and such like, cannot be attributes of that which has understanding and will’ (Kant Reference Kant1763a: OPA, 2: 85). Abaci concedes that ‘the “cannot” here does not seem to express any logical contradiction but a metaphysical incompatibility … based on a Cartesian sort of dualist assumption’. Stang and Newlands also cite this passage as clear evidence in favour of ascribing to Kant a concept of real repugnance that is subject-cancelling.Footnote 4 Abaci, however, thinks that this reading of the passage is ‘impugned’ by the fact that Kant goes on to further ‘explain’ the concept of real repugnance by appeal to examples of physical forces. Similarly, Yong cites the passage briefly but dismisses it by saying that Kant goes on to ‘illustrate real repugnance through the example of motive forces’ (Yong 32).
By now it should be easy to anticipate how the response is going to go. Kant believed that some philosophical concepts are best exhibited by empirical examples, and so the fact that he reverts here to illustrating the concept of real repugnance with intuitive examples is no surprise. It certainly provides no grounds for denying (or impugning) the fact that Kant uses a modal notion (‘cannot be’) to refer to a subject that combines two really repugnant predicates (<being extended>, <having understanding and will>). Nor does it show that the repugnance between these predicates can be analysed in terms of opposed motive or psychological forces, on the one hand, or in terms of logical inconsistency on the other. And yet the repugnance here is subject-cancelling: Kant says that an extended, minded substance cannot really exist. We may not like the example, but it recurs throughout the Critical period too: ‘it is impossible to conceive a matter endowed with thought’ (Kant Reference Kant1793: R, 6: 128–9, n.).
3. Is Harmony a Distinct Condition on Real Possibility?
Abaci agrees with me (against Yong) that real possibility is grounded in the actuality of the fundamental predicates involved, rather than in the mere divine thought of them. I will say more about that debate in section 8 below. The main philosophical difference between Abaci and me is over whether Harmony is another, distinct condition on real possibility, one that goes beyond Consistency and Content. My claim is that it is: even if we have guaranteed that a concept is logically consistent and has positive content, we still need to make sure that there is no real repugnance between the constituent predicates in order to ‘prove’ that its object is really possible.
In my 2012 paper, I argued that a main advantage of this reading is that it gives Kant the resources to respond to ‘the plurality objection’ – the objection that his proof cannot rule out scenarios on which the ground of possibility consists in numerically distinct necessary beings. What is not so advantageous by Kant's lights, I said, is that harmony also seems to imply that only a certain sort of unique being – namely, a necessary being with some ‘monstrous’ Spinozistic attributes – could play the grounding role.
Abaci (16–17)agrees that the ‘plurality objection’ is a problem, and that RHI appears to solve it. But in light of the textual ambiguity and the Spinozistic threat, he prefers to conclude simply that the pre-Critical Kant ‘does not give any separate justification’ for holding that there is only one ground of real possibility.Footnote 5 He then suggests that Kant realized this later and tried to salvage something from the proof by appealing to a principle of theoretical economy and ‘downgrading’ the conclusion to a merely ‘subjectively valid’ one.
Note, though, that Abaci has given up on the proof in OPA altogether, since he thinks it cannot meet the plurality objection, whereas I was trying to find resources in Kant's text for reconstructing the OPA argument as a valid speculative proof of a single ens realissimum, albeit with some highly dubious premises and a lingering whiff of Spinozism. It is unclear to me which is the more charitable approach, but mine at least seems the more heroic (quixotic?).
4. Harmony as the Motivation for the Appeal to Intuition
Towards the end of the article (19), Abaci makes the following general remark:
The reasons that led the Critical Kant to the conclusion that we cannot know the real possibility of the ens realissimum are broader than a concern about real harmony. In a nutshell, the Critical Kant thinks that we cannot cognize the real modalities of supersensible or noumenal things because such things cannot be given to our intuition and our concepts of them are bound to remain without intuitive content and thus without objective reality.
This remark goes to the heart of another important disagreement. While I agree that Kant is worried about how we can establish the ‘objective reality’ of our concepts, I think it is also clear that what he means by this, most of the time anyway, is that we cannot be sure whether ‘an object corresponding to it is [really] possible’ (Kant Reference Kant1790: CJ, 5: 396; see also Kant Reference Kant1788: CPrR, 5: 134).
I also agree that appeals to intuition (or the field of sensible ‘givenness’) are a crucial aid in the project of securing objective reality for our concepts. If we can show that something is the object of an actual or possible experience, or is transcendentally required for experience in general, then we can prove to Kant's satisfaction that it is really possible. Further, since we have no other reliable way of establishing real possibility (mere conceivability is not a good guide, given the threat of real repugnance), Kant thinks we must restrict our knowledge-claims to objects and states that can be shown to be really possible in this empirical or transcendental way:
To cognize an object, it is required that we be able to prove its possibility (whether by the testimony of experience from its actuality, or a priori through reason). But I can think whatever I like, as long as I do not contradict myself … (Kant Reference Kant1781/87: CPR, Bxxvi, n.)
Elsewhere I refer to this as underwriting Kant's ‘Modal Condition’ on knowledge (Chignell Reference Chignell2009b). The difference between Abaci and me here has to do with the order of explanation. Kant's motivation for demanding the connection to intuition, I submit, is not fundamentally a semantic worry about conceptual content or an empiricist worry about justification. Given the carefree way in which Kant talks about free wills, God, monads and the like in his Critical writings and especially the lectures on metaphysics and rational theology, it seems clear that he thinks that the concepts involved have some determinate content, and can be applied in these contexts even without their objects being ‘given’.Footnote 6 The motivation for the appeal to intuition arises, rather, from Kant's new understanding of modal metaphysics and epistemology. We have no faculty of clear and distinct perception or any other way to prove that the objects referred to by our concepts are really possible or not. So we have to appeal to intuition. Put another way: establishing a link to what is ‘given’ in experience (via direct exhibition, causal connectedness or transcendental argument) is crucial, but only because of the way it responds to Kant's more fundamental concerns about Content and Harmony.Footnote 7
5. Charity
Peter Yong's piece includes some of the same textual challenges; I have tried to respond to those above. In what remains, I focus on Yong's philosophical objections. Yong says (34) that RHI is ‘uncharitable’ to Kant, since it ascribes to him a Spinozistic view, or at least claims ‘that his theory inadvertently commits him to it’. I agree that claiming that the historical Kant was an avowed Spinozist would be uncharitable: Kant regards the Spinozistic idea of God as ‘monstrous’ (ungeheur) throughout his career (Kant Reference Kant1993: OPA, 21: 50).Footnote 8 I also agree that claiming that Kant was an unavowed, closet Spinozist in the manner of the Pantheismusstreit is uncharitable (albeit scandalously fun). My claim, however, was merely that there is a ‘threat of Spinoza’ lurking in the details of Kant's proof, and that even our (or my) best efforts to reconstruct the proof in a way that avoids that threat may not succeed in overcoming it. So while I do think ‘his theory inadvertently commits him to it’, that conclusion comes at the end of a long discussion of other options. The principle of charity surely does not prohibit us from pointing out that a philosopher's background commitments entail a position that he or she officially eschews; otherwise, much that goes on in professional philosophy (and history of philosophy) is uncharitable in the extreme!
6. The Plurality Objection
Recall that RHI says that there must be a ground not only of all the real possibilities, but also of the harmony and repugnance relations between them. This is why we need a single being with a huge number of distinct attributes: the fact that some properties are really harmonious is grounded in their co-exemplification within a particular attribute of that being, while the fact that others are really repugnant is grounded in the fact that they are not so exemplified. Yong argues that these patterns of co-exemplification could obtain across numerically distinct necessary beings, and still succeed in grounding the domain of possibility. Thus, the ‘facts of real harmony do not require a single ground in an ens realissimum’ (35). This is a serious objection.
In response, note that the distinct beings in the alternative scenario Yong offers would each (by the logic of the proof) have to be absolutely necessary beings. Now suppose (for reductio) that X and Y are two such beings. What would explain the fact that Y is numerically distinct from X? The only explanation available seems to be that Y grounds a different collection of real possibilities than X does. But this explanation appeals to something about X in order to explain a feature of Y, and vice versa. So there is now a relation of explanatory dependence between two absolutely necessary beings. And that, according to Kant, is verboten.
To find out why, we have to turn to the puzzling passages in OPA where Kant considers and then sets aside the plurality objection on the grounds that ‘it is not possible for several beings to be absolutely necessary’ (Kant Reference Kant1763a: OPA, 2: 84–5). In my 2012 paper, I interpreted Kant's claim as one about causal dependency and took issue with it: it seems coherent to have two necessary beings, neither of which is causally dependent on the other or anything else, and each of which grounds two distinct subsets of the domain of real possibility (Chignell Reference Chignell2012: 652, n.). But reflection on Yong's objection now leads me to think that we should read Kant's claim as one about explanatory dependence more broadly. Kant sees that if the proof delivers two or more grounds of real possibility in the way just described, these beings would be explanatorily dependent on one another, and thus not absolutely necessary after all.Footnote 9
A great advantage of a monistic picture is that there is no numerical distinctness that needs explaining: the one, monstrous being is explanatorily independent. But what about the distinctness of its attributes – does explaining that violate some sort of rationalist principle? There are at least two interpretative options here. One is to follow Spinoza in saying that God's attributes are explanatorily independent of each other, and that this is somehow compatible with their being, at bottom, identical. This ‘parallelism’ doctrine obviates the need to explain the numerical distinctness of the X attribute from the Y attribute, because there is no such distinctness. But taking this approach also involves dragging the pre-Critical Kant very far down the Spinozistic rabbit hole indeed, and that is something (as Yong would surely remind us) that charitable interpreters should resist.
Fortunately, a second option is also available. In the 2012 paper I said that nothing in the proof requires that the necessary being have ‘parallel’ or ultimately identical attributes, though I did say it requires them to be explanatorily independent (Chignell Reference Chignell2012: 668, n.). But now I think we can see that Kant's proof provides positive reason to resist the explanatory independence of attributes doctrine as well. For the proof to go through (at least on RHI), we need to appeal to the exemplification of properties across distinct attributes in order to explain their real repugnance. So why is X attribute distinct from Y attribute? Because it exemplifies different properties than Y does. The threat to Kant then is not that of no-holds-barred Spinozism; rather, only the being that grounds possibility must be absolutely necessary and thus explanatorily independent of any other being; its attributes have some explanatory connections. That said, this slight modification to the Deus sive natura picture is unlikely to make Kant any happier to embrace it.
7. Failure to Explain?
Yong's final stated objection is that the picture I attribute to Kant ‘fails to explain’ real possibilities. This seems odd initially: what better explanation of something's possibility than the existence of its constituent predicates? Yong's real concern, I take it, is that RHI ‘eliminates the very phenomenon [i.e. non-actual possibilia] that it was created to account for in the first place’ (37).
I understand the objection, but I do not think it would be pressing for someone who is committed, in the way that Leibniz and Kant were, to grounding modal facts in facts about existence. This is often articulated as an ‘actualist’ commitment, but it does not have to be about the actual – David Lewis shares the commitment, but he construes the worlds that ground modal facts as existent but non-actual, at least from a this-worldly perspective. The point is just that it is supposed to be a virtue of such accounts that they can ground facts about possibility in what exists. If they do so by saying that all the possible properties are already exemplified in what exists, then so much the better by way of grounding. The real difficulties – for Kant as for Lewis – await in the other consequences of this move.Footnote 10
8. The Traditional Alternative
At the end of his piece, Yong develops his own version of the ‘more traditional’ account according to which God grounds real possibilities by representing them, rather than by exemplifying all of their fundamental properties. He recognizes, however, that the plurality problem arises for this interpretation too – why not divide the labour by having multiple minds representing different sub-sections of the domain? Yong's answer relies on the claim that there are additional facts, over and above the facts about individual possibilia, that still require grounding. These additional facts involve global relations between possibilia – for instance, ‘the fact that they together form a single domain of possibility’ (40), or the fact that this domain's ‘constituent possibilities are not fragmented, but stand together as a single space of thinkable content’ (43).Footnote 11
Yong's proposal, then, is that the only thing that can ground facts like this – explaining relations between individual possibilities as well between individuals and the whole – is their ‘schematization in the mind of their common cause’ (44). By this he presumably means something like their being thought together along with the relations between them by the being that actualizes some of them. Such mental schematizing ‘is precisely the role that God is supposed to play in Kant's account. … By appealing to divine schematization as the ground of the totality of all possibility Kant is thus able to provide an account of modality that minimizes bruteness and satisfies the demands of the principle of sufficient reason’ (44).
The principle of sufficient reason is an unremitting taskmaster, however, and in the end I am not sure its demands are really satisfied here, or that bruteness is really minimized. For Yong's proposal still leaves us with the question of why God thinks together these (where ‘these’ is a demonstrative gesture in the direction of all the real possibilities) rather than some others. There is a modal version of the Euthyphro problem here: does God think together Sherlock Holmes, my non-actual twin sister, and the Cubs-winning-ten-World-Series-in-a-row because these beings or states of affairs are antecedently possible? If so, then we still need an explanation of the latter fact, one which will presumably take us beyond the divine mind. But if not, then does God's choosing to think them make them possible? This runs headlong into the sort of Cartesian voluntarism about modality that both Kant and Leibniz clearly wanted to avoid.Footnote 12 Does God think them because he has the power to produce them? Here, too, the explanatory buck has merely been passed – we would still want an account of why these rather than those are the objects and states that God has the power to produce.Footnote 13
Here is a final option: does God's thinking them just make them possible, and itself have no further explanation? This, according to Samuel Newlands, is what Leibniz should have said, though he admits that he cannot find a passage in which Leibniz actually says it:
To put it cheekily, Leibniz could be more of a Humean about ‘modal laws.’ In virtue of what is p-and-not-p not possibly true? In virtue of the fact that God doesn't think that p-and-not-p. To some, that answer gets matters backward. But the promise of a reductive grounding account of modality should be attractive to an advocate of the principle of sufficient reason (PSR) like Leibniz. And we might well wonder, is a buck-stopping, table-pounding ‘God just can't!’ explanatorily better off than a buck-stopping, table-pounding ‘God just doesn't!’? Certainly the latter answer seems more in the spirit of Leibniz's general project of providing theistic grounds for modal truths: base what God (and creatures) can and cannot do on what God actually does and does not do. (Newlands Reference Newlands2013: 169)
The advantage of RHI over Yong's Leibnizian alternative, even construed in this Newlands-cum-Leibniz way, is that it seems clearly preferable for an explanatory regress to stop in the extra-mental, non-intentional predicates of a necessary being, rather than in its thoughts. To say that Holmes and Watson are possible because God represents them together, but that there is no explanation of why God represents them together (‘God just does!’) is not wholly satisfying, as Newlands himself admits. To say that Holmes and Watson are possible because all of their fundamental properties are already exemplified in the right way across the attributes of a necessary being is much more explanatorily satisfying. If you are looking for some bedrock, there is surely none so solid as the essential attributes of an absolutely necessary being. I take this thought to be part of what was motivating Kant.Footnote 14 But then, of course, the threat of Spinoza remains.Footnote 15
