OMRI BOEHM | Kant’s Critique of Spinoza | Oxford University Press 2014


By Nicholas Stang 

Spinoza’s monistic metaphysics is back on the agenda, both in historical scholarship and in contemporary metaphysics. The past decade has seen something of a renaissance in historically and contextually informed, analytically rigorous scholarship on Spinoza, primarily his metaphysics, epistemology and philosophy of mind. At the same time, the work of Jonathan Schaffer has put a version of Spinozistic monism on the map as a metaphysical view, while Michael Della Rocca (one of the main contributors to the renaissance of Spinoza scholarship), and others, have offered vigorous defences of the Principle of Sufficient Reason (PSR) very much in the spirit of Spinoza and other early modern rationalists. Simultaneously, Kant scholars have been paying far more attention to the influence of the eighteenth Century German rationalist tradition, exemplified by Leibniz, Wolff, Baumgarten and others.

The time is thus ripe for a consideration of the relation of Kant to Spinoza, a figure too often absent from the Kant literature. Spinoza and Spinozism, after all, were never very far from the minds of Kant’s immediate predecessors in German school metaphysics. Leibniz frequently claims that his metaphysics is the only way to avoid the atheism and fatalism taught by the Jewish heretic from Amsterdam. Christian Wolff was dismissed from his position at Halle on the accusation that his system led inevitably to Spinozistic fatalism. A reconsideration of the Kant-Spinoza relation is all the more relevant given the Pantheismusstreit that engulfed German intellectual life in the mid-1780s when Jacobi provocatively claimed that Spinozism is the true consequence of consistent Enlightenment philosophising and that we must, instead, make an irrational leap of faith to retain our religious and moral beliefs. Kant himself, as we shall see below, intervened directly in the ensuing debate, as did just about every single significant German intellectual of his day. Finally, Kant’s writings themselves abound in intriguing remarks about the philosophy of Spinoza, such as this one from an unpublished Reflexion: “Spinozism is the true consequence of dogmatic metaphysics” (Refl 6050, AA 18:436). Taken at face value, this remark means that our understanding of Kant’s critique of dogmatic metaphysics is incomplete without a systematic understanding of his critique of Spinoza.

Omri Boehm’s new book Kant’s Critique of Spinoza thus arrives at an auspicious moment and attempts to illuminate an important issue not only in Kant exegesis and the history of post-Kantian German philosophy, but, potentially, contemporary Spinoza-inspired metaphysics itself. The ambition of this book is, indeed, far greater than the typical volume in the Kant and Philosopher X genre. In the Preface, Boehm endorses a version of Jacobi’s claim: if we follow Spinoza in denying that nature has an inherent teleology, then all talk of value (especially moral value) is either meaningless or a mere fiction. In the rest of the book, he argues that Kant is right in thinking that the most consistent and defensible alternative to transcendental idealism is a specifically Spinozistic form of rationalism, which entails that we lack free will because all of our actions (indeed, everything whatsoever) are unavoidably necessary. Boehm is thus attempting nothing less than a reawakening of the Pantheismusstreit and a specifically Kantian solution to Jacobi’s problem. He ends the book by arguing that our moral awareness of the injustice of the world rationally grounds our belief in the contingency of the actions of rational agents and thus in restricting the PSR. “I had to deny knowledge in order to make room for faith”, Kant announces in the Preface to the B-edition of the Critique of Pure Reason and Boehm suggests that we must follow him if we want to avoid Spinozism.

What strengths the book does have lie in its analysis of Kant’s arguments in the Antinomial Conflicts of Pure Reason (‘Antinomies’ for short), which occurs in Chapters Two and Three. It makes an interesting case that, at least in the First and Third Antinomy, the Antithesis is meant to represent Spinoza and that Kant’s Thesis arguments are specifically anti-Spinozistic. It interprets those arguments, widely believed to depend on question-begging importations of Kant’s transcendental idealism, as, instead, rehearsing an eighteenth century objection to Spinoza that would have been familiar to Kant: Spinoza’s universe is a whole composed of infinitely many parts, which Kant follows Wolff and others in rejecting as incoherent.  Boehm points out that this objection is based on a misunderstanding of Spinoza: the world is not a whole composed of parts, but an infinite whole of which the parts are dependent modes. Boehm argues that the Kant-Spinoza dialectic should be reoriented around the more basic question of whether we can ‘adequately’ conceive of an infinite totum analyticum, a whole ontologically prior to its (infinitely many) parts. What will count as an ‘adequate’ conception of such a being is, of course, one of the things up for grabs between Kant and Spinoza.

I have organised my comments around the Preface, a brief discussion of the meaning of ‘Spinozism’, and then a chapter-by-chapter discussion of the book. I discuss the Introduction and the final chapter together in a short concluding section on the Pantheismusstreit.


The argument of the Preface is that a purely mechanistic conception of nature entails that talk of value is either fictional or meaningless. By a ‘mechanistic conception of nature’ Boehm means one that denies that there is real teleology in nature. According to his narrative, during the Enlightenment we came to reject the view that nature has a teleological structure, or, to use a handy formula, to reject the view that a natural product is the product of a process the purpose of which is to produce it.[1] This entails, supposedly, that there cannot be real or objective values. The only possible source of value is human desire, and there is no non-arbitrary target of human value, no goodness ‘out there’ that we can be incorrectly or correctly latching onto. In Boehm’s narrative, Spinoza has a special place. It is not so much Spinozistic necessitarianism or the denial of free will that leads to ‘nihilism’ (the term Boehm, following Jacobi, uses) about value but his denial of natural teleology. As Spinoza writes in the Appendix to Ethics I, it is a prejudice to suppose that “all natural things act, as men do, on account of an end”. This entails, according to Boehm, the further Spinozistic doctrine that “we judge something to be good” because we desire it, not vice versa. If nature is blind (does not act for purposes) then there is no source or ground for value other than our desires, and no question about what is good other than the contingent matter of what we happen to desire. This is Boehm’s ‘Jacobi moment’: the post-Enlightenment non-teleological mechanical conception of nature Spinoza did more than anyone to put on the intellectual map is incompatible with our moral commitments. So we should turn to Kant’s critique of Spinoza to see how we can retain both Enlightenment rationality and not-merely-fictional moral values.

Here is one of Boehm’s arguments that the denial of natural teleology entails nihilism about value:

Enlightenment rationalism entails nihilism to the extent that it deems appropriate only blind, mechanical conceptions of nature. If what exists is the result of what precedes it, and what precedes it has no relation to some separate (“transcendent”) non-accidental good, talk of value is relativized to some anchor within the world. However, if ex hypothesi everything within the world is an accidental consequence of blind causality, any anchor can only be as good as any other. Talk of value thus becomes either consciously fictional (a noble lie, perhaps) or meaningless. The point is this: if all value is arbitrarily fixed in relation to some anchor x, there is no reason not to fix value to non-x. […] From a Kantian point of view, the significant point to notice is that this conclusion doesn’t seem much affected if one substitutes x by ‘reason’, ‘rational beings’, or something of the sort. (p. xi)

The significant point to notice is that there is no real argument here.  What does it mean for value to be “anchored”? Why must value be “anchored” in the world? And assuming that value is “anchored”, why does the denial of teleology entail that its anchor is arbitrary? Why can’t value be necessarily anchored in whatever anchors it? Contrary to Boehm’s suggestion in the third sentence, the blindness of nature does not entail that its effects are “accidental”. Spinoza after all thinks that nature is “blind” in the relevant sense (does not act for purposes) but that everything that happens is necessary.

Just to bring out further how sloppy Boehm is being here, consider the range of meta-ethical positions he is ruling out by fiat, each of which is prima facie consistent with the denial of natural teleology: naturalist realism (moral properties are natural properties), Moore-style anti-naturalist realism (moral properties are primitive non-natural properties of entities in this world), and the view of many contemporary Kantians (who endorse a compatibilist conception of freedom, and eschew even Kant’s own ‘regulative’ role for teleology).[2] Showing that each of these views implicitly requires real natural teleology would require extensive argumentation, which Boehm does not supply. Nor does he ever clarify what it would be for morality to be ‘meaningful’. Boehm’s other arguments are no better; “even if we suppose that one acts on the basis of rational maxims, there is still little meaning to this if rationality is itself considered the product of blind causality” (p. xix) he writes at one point. But some Kantians would accept that the fact that homo sapiens possess practical reason is the product of “blind” evolution, while claiming that the morally good is what can be the object of a practically rational will. What is Boehm’s argument against that view?

A lot of names are dropped to support Boehm’s claim that ‘meaningful’ talk of values requires natural teleology—Jeremy Waldron, Wittgenstein, Leo Strauss, Bernard Williams, Horkheimer and Adorno, and others. But there is little in the way of argument. Boehm asserts the “meaninglessness” of natural rights talk, and his sole support for this claim (by way of Jeremy Waldron) is this: the only way to ground the principle that some beings have natural rights and others do not is through a teleological conception of nature. Why isn’t the fact that some beings have a capacity for rational end-setting (human beings, arguably some other animals) that others lack a difference that grounds a difference in natural right? Or, in a less Kantian spirit, the fact that some beings have a conception of themselves as existing over time that other beings lack (and can thus be harmed in ways that other animals cannot—by having their life projects cut short by death)? Neither of these views assumes that nature acts for ends and both of them are prima facie potential bases for a theory of natural right.

Boehm even tries to enlist Kant as an upholder of the idea that without natural teleology talk of value is “meaningless”. He quotes this wonderful passage about Spinoza from the Critique of Judgement:

We can thus assume a righteous man (like Spinoza) who takes himself to be firmly convinced there is no God and (since with regard to the object of morality it has a similar consequence) how will he judge his own inner purposive determination by the moral law, which he actively honors? He does not demand any advantage for himself from his conformity to this law, whether in this or in another world; rather, he would merely unselfishly establish the good to which the holy law directs all his powers. But his effort is limited; and from nature he can, to be sure, expect some contingent assistance here and there, but never a lawlike agreement in accordance with constant rules (like his internal maxims are and must be) with the ends to act on behalf of which he still feels himself bound and impelled. Deceit, violence, and envy will always surround him, even though he is himself honest, peaceable, and benevolent; and the righteous ones beside himself that he will still encounter will, in spite of all their worthiness to be happy, nevertheless be subject by nature, which pays no attention to that, to all the evils of poverty, illnesses, and untimely death, just like all the other animals on earth, and will always remain thus until the one wide grave engulfs them all together (whether honest or dishonest, it makes no difference here) and flings them, who were capable of believing themselves the final end of creation, back into the purposeless chaos of matter from which they were drawn.—The end, therefore, which this well-intentioned person had and should have before his eyes in conformity to the moral law, he would certainly have to give up as impossible. (KU, AA 5:453; quoted on p. xxi-xxii)[3]

But Kant’s point here is almost the opposite of Boehm’s.[4] The denier of a wise author of nature (Spinoza) is not forced to conclude that talk of morality and moral value is “meaningless”. As he makes clear throughout the discussion of moral teleology in the Third Critique, if there is no God morality remains just as binding as it was before.[5] Spinoza and other deniers of natural teleology are forced, according to Kant, into a conflict in practical reason: they are rationally obliged to set an end (the highest good) but they also believe this end is impossible. They must give up, not the validity of the moral law, but the highest end morality sets us (the highest good). The moral problem that Kant diagnoses for us inheritors of the post-Enlightenment ‘dis-enchanted’ conception of nature is not Jacobean nihilism, as Boehm claims, but despair: we are forced into the conclusion that our highest moral aims cannot possibly be achieved. This isn’t meant merely to be discouraging; Kant thinks it is a rational contradiction (but one that does not affect the validity of morality!). Avoidance of this contradiction warrants our belief (Glaube) in a wise author of nature and gives us grounds for interpreting nature teleologically as created for the purpose of achieving those very moral ends. Boehm is right that Kant thinks that modern moralists need to reintroduce teleology (albeit in a merely ‘regulative’ form) into their conception of nature, but is dramatically mistaken as to why. He seems to acknowledge this alternate Kantian route to teleology by the end of the Preface, but does not note that if this is Kant’s reason for reintroducing a ‘regulative’ conception of teleology it has no point of connection with his earlier argument that if value-talk is “meaningful” there must be real teleology in nature.

In fairness to Boehm, it is standard practice to begin a book with some dramatic claims, to get the reader’s attention. Further, all of these issues are apparently going to be dealt with in greater detail in a future book. But its effect on this reader was to make me question the author’s grasp of the issues.

What is “Spinozism”?

The Spinoza under discussion in this book is largely the metaphysician of Ethics I, with some consideration of the theory of ideas in Ethics II. Here are some Spinozistic doctrines that are relevant to Boehm’s discussion:

SP1. There is only one substance (God); everything else is a mode of this one substance. (Ip14–15)[6]

SP2. God is extended. (Id6, Ip15s)

SP3. Every finite being (mode) follows necessarily from a ‘preceding’ one. (Ip28)

SP4. Everything is a necessary consequence of the existence of a necessary being. (Ip29)

SP5. There is no incompatibilist free will, either in God or in finite beings. (Ip17s, Ip29)

SP6. Nothing happens for a purpose; there is no teleology in nature. (I, Appendix)

SP7. God does not ‘create’ the world, does not act for purposes, etc. (I, Appendix)

Boehm frequently identifies one of these doctrines as a target of Kantian criticism and then describes Kant as criticising the ‘Spinozistic’ position. But the problem is that, while all of these doctrines are upheld by Spinoza, some of them are upheld by non-Spinozists as well.[7] Arguably, Wolff and even Leibniz are committed to SP3-SP5, as is the pre-Critical Kant (who, pace Boehm, was never a Spinozist—see below). In fact, if we restrict ‘finite beings’ to ‘alterations (change in accident) in phenomenal substance’ then the Critical Kant is committed to SP3. So it muddies the waters somewhat for Boehm to describe Kant’s engagement with just some of these doctrines as his engagement with a ‘Spinozistic’ position. This is especially true with respect to SP1; many of Kant’s remarks about Spinoza identify this doctrine as the characteristic doctrine of ‘Spinozism’. But the doctrine of substance monism is barely discussed at all by Boehm. Instead, he focuses most of his attention on SP2–5. Since SP3–SP5 are upheld by Leibniz and Wolff, who reject SP1 and SP2, at times I wondered whether the focus on Spinoza had been lost. This is exacerbated by the fact that, for a book largely about Spinoza, there is remarkably little explicit discussion and exegesis of Spinoza’s writings. I count only one short section (pp. 134–9) in which Spinoza is discussed in detail.

I suspect the reason for this is that Boehm takes Spinoza as a stalking-horse for dogmatic rationalism in general. He quotes several passages in which Kant asserts that Spinozism is the true consequence of rationalism. For instance, in a passage from the Second Critique, Kant claims that transcendental idealism about space is the only way to avoid Spinozism:

Therefore I do not see how those who insist on regarding time and space as determinations belonging to the existence of things in themselves would avoid fatalism […]. Hence, if this ideality of time and space is not adopted, nothing remains but Spinozism, in which space and time are essential determinations of the original being itself, while the things dependent upon it (ourselves, therefore, included) are not substances but merely accidents inhering in it. (KpV, AA 5:101–2)

While Boehm quotes this passage several times he never fully explains or reconstructs Kant’s argument that if spatiality and temporality are properties of things in themselves then God is extended (SP2), everything else (ourselves included) is a mode of God (SP1), and every action is a necessary consequence of a preceding state (SP3).[8] Kant moves very quickly from the claim that things in themselves are in space and time to the conclusion that God is himself spatiotemporal, and what is more, the only substance, and Boehm does not fill in Kant’s argument (though he does acknowledge it needs to be filled in). In this passage and others, Kant asserts that once you accept transcendental realism (objects in space and time are things in themselves) and the principle of sufficient reason (PSR) you are committed to substance-monism and necessitarianism. Spinoza is, supposedly, the only one who drew the correct conclusions from his premises. Does Boehm think Kant is right about this? Why or why not? Boehm does not tell us. What about a Wolffian who embraces a transcendental realist view about space as well as the PSR, but denies substance monism? Or Crusius, who embraces transcendental realism about space but accepts only a restricted version of the PSR, thus making room for libertarian freedom? It might have been worthwhile for Boehm to investigate the coherence of these positions intermediate between Spinoza and Kant.[9]

What is more, given the absolutely central role that the PSR plays in this book, it is striking that Boehm never clearly articulates what this principle is, nor does he distinguish different versions of it and their relative strengths.[10] Nor does he, until relatively late in the game (p. 143) distinguish different relations of grounding, for instance, efficient causation (e.g. the relation of a body to a body it moves) from formal causation (e.g. the relation of the essence of a thing to its essential properties). Boehm’s unclarity about the exact meaning of the PSR leads him to invoke the principle in a bewildering number of ways in different arguments. Sometimes it is invoked to show that a being must have a ground of its existence, at other times to show that some possibility must have a ground, at yet other times to show that a conceptual truth must have an existing ground, etc. This makes it difficult to credit his reconstruction of any argument that makes a non-trivial use of the PSR.

Consider the following principles:

(i) All alterations in contingent beings have sufficient efficient (‘moving’) causes.

(ii) All alterations in contingent beings have determining efficient causes (defined as causes that necessarily bring about the same effect in the same circumstances).

(iii) All alterations in contingent beings—as well as such beings’ existence—have determining efficient causes.

(iv) All alterations in contingent beings—as well as such beings’ existence— have determining efficient causes and there is no infinite regress of causes.

(v) All properties of all contingent beings (as well as such beings’ existence) have determining grounds (where ground is a genus of which efficient cause is a species) and there is no infinite descending chain of grounds.

(vi) All properties of all beings (both contingent and necessary)—as well as these beings’ existence—have determining grounds and there is no infinite regress of grounds.

(vii) All truths within some circumscribed set S (e.g. the set of identities and definitions) lack grounds. All other truths have determinate grounds among the members of S, without infinite regress.

(viii) All truths whatsoever have determining grounds and there is no infinite regress of grounds.

Each of these principles has some claim to being called the PSR. But which of these is the Spinozistic PSR, the principle that (according to Boehm’s reading of Kant), once granted, if thought through to its consequences leads to Spinozistic necessitarianism and substance monism? There are passages that suggest that Boehm thinks that (viii) is the PSR and that endorsement of anything else is a failure to live up to the true requirements of rationalism.

But the identity of the PSR is a matter of significant controversy in the eighteenth century. In 1743 Christian August Crusius published a short book Dissertatio de usu et limitibus principii rationis determinantis, vulgo sufficientis, admired and praised by Kant himself in Nova dilucidatio (PND, AA 1:393), in which he distinguishes sufficient from determining grounds and argues that nothing stronger than (i) should be accepted. A sufficient ground makes its consequence true. A determining ground is one such that, necessarily, in the same circumstances, it makes that consequence true. Crusius denies that the free actions of rational agents have determining grounds (thus [ii] is false) while admitting that they have sufficient grounds. It is assumed throughout Boehm’s book that the unrestricted PSR is incompatible with libertarian freedom but this holds only if we assume that sufficient grounds are what Crusius calls ‘determining’ grounds. I am not asking Boehm to engage in Crusius exegesis (though this might have improved his book); I am merely pointing out that it is naïve to talk about the principle of sufficient reason in an eighteenth century context.

What is more, the most unrestricted form of the PSR, (viii), is arguably incoherent. 2 is the successor of 1. What is the ground of this truth? Perhaps, it is grounded in the essence of 2, the fact that to be 2 is to be the successor of 1. What grounds that truth? At this point, it is very plausible that we hit explanatory bedrock. The identities of things (e.g. the fact that 1=1) and their definitions (e.g. to be 2 is to be the successor of 1) do not have grounds. Everything else is grounded in those truths. Leibniz himself, in his more careful formulations of the PSR, exempts identities and definitions from the requirement of being grounded; they are the ‘primary’ truths in which everything else (including necessities) is grounded.[11] Even Spinoza, arguably, does not think that the definitions of things must be grounded (after all, the Ethics begins with definitions). What is grounded is whether these things are possible, whether they exist, and what other properties they have. It may be that to be a particular finite mode is to be a mode of God, so the essence of that mode involves God (God is the ‘cause’ of their essences), but the fact that that is what it is to be that mode may not require a further ground.[12] God is the unique being whose existence is grounded in his essence; all other beings exist in virtue of a ground distinct from themselves. All of God’s properties follow from his essence, while the properties of finite modes follow either from their essence (their essential properties) or from some external cause (e.g. the preceding finite mode that is their immediate cause). The reason for restricting principle (viii) (for adopting [vii] instead) is simple: (viii) entails that (viii) itself is grounded. What could be the ground of the truth of the principle of sufficient reason? (It cannot be God’s wise choice of the best, because the PSR is going to be invoked to prove God’s existence). All of this is very preliminary and sketchy, and would need to be filled out in greater detail. I bring it up merely to point out that invoking the PSR (as Boehm does) without clearly articulating which PSR is under discussion is philosophically problematic and unexplanatory.

One final note about the notion of ‘infinite regress of grounds’. Let us say that x has an infinite regress of grounds just in case (i) grounding is transitive and irreflexive (nothing grounds itself), (ii) x has a ground, and (iii) every ground of x has a ground. Whether the PSR allows infinite regress of grounds (an ‘infinite ascending series of conditions’) will be a crucial issue in Kant’s Antinomies. Boehm seems to assume that there cannot be infinite regresses of grounds for Spinoza (i.e. that we must go at least as far as [iv] above) but this is not the case. The series of causes of a given finite mode is an infinite regress of other finite modes (Ip28). Within the attribute of extension, the motions of my body are caused by the motions of the bodies that caused them, which are caused by etc. etc. etc. Now it is also true that each of these modes, and the infinite series of finite modes in the attribute of extension, are grounded in/caused by God. But this means there are two grounding relations: finite mode-finite mode, and God-finite mode (though Spinoza uses causa for both). Spinoza seems to allow infinite regresses in the former kind of relation, but not in the latter. Boehm should have more carefully distinguished different grounding relations in Spinoza (and Kant) and made clear the different status of infinite regresses in them. But this also means that we cannot simply assert that the PSR excludes infinite regresses of grounds (as Boehm does at various points—see below).

Chapter One: The One Possible Basis, The Ideal of Pure Reason, and Kant’s Regulative Spinozism

F.H. Jacobi, who inaugurated the Pantheismusstreit of the 1780s, claims that the work that first made him realise that consistent Enlightenment philosophy leads inevitably to Spinozism was Kant’s 1763 book The Only Possible Ground of Proof in Support of a Demonstration of the Existence of God (henceforth Beweisgrund). In order to reconstruct Kant’s critique of Spinoza, and with the eventual goal of understanding Kant’s position within the Streit itself, Boehm wisely begins with an examination of Beweisgrund.

Aside from containing the first published formulation of Kant’s famous objection to the ontological argument (“existence is not a predicate of a determination of a thing”—BDG, AA 2:72) Beweisgrund is primarily known for its innovative argument from the premise that something is really possible to the conclusion that there must be a unique absolutely necessary being that grounds all real possibility (God). The most basic premises of this argument are (1) not all logical possibilities are real (or material) possibilities, and (2) real possibilities are grounded in what exists. Kant’s idea is that the logical consistency of a set of predicates does not prove that a thing with those predicates is possible. The possibility of that thing requires that the predicates themselves are themselves possible. Kant calls this requirement the ‘material’ element in possibility (in later works he will call it the requirement that the predicates be ‘really’ possible and I shall borrow that terminology). He further requires (ii) that the material (real) possibility of some predicate be grounded in what exists. He seems to assume that the existing ground of a real possibility must be a substance, so let us grant him that requirement.

The question is: how does Kant use these premises to prove that there is a unique substance that exists necessarily and grounds all really possible predicates? Here is Boehm’s reconstruction of the argument:

D1: Internal possibility (a thing’s essence) depends not only on the formal element of possibility (the consistency of the predicates participating in the essence), but also on a real or “material element” (the predicates or properties participating in the essence).

D2: Formal possibility depends on material possibility. Contradiction is a relation posited between given predicates or things. This is no contradictory/consistent relation where nothing is pre-given that can enter into relations.

D3: Possibility is grounded in something actually existing. [By the PSR, if something is possible, there is something in virtue of which it is possible; further, by the PSR, ultimate grounds, existing unconditionally, must ground possibility; otherwise the fact that something is possible would remain inexplicable.]

D4: Necessarily, something is possible. Kant considers it impossible that absolutely nothing is possible. This claim can also be justified by the PSR. If nothing is possible, then nothing exists. But then, there can be no reason why nothing is possible. Therefore, something is possible.

D5: Necessarily, something exists. [From D3 and D4]

D6: All possibility is grounded in a single being. [By the PSR, all possibilities, including relations and possible relations have to be grounded. But this can be the case if and only if the same being grounds all possibilities; had different possibilities been grounded in two or more beings, the relations between these beings would have to be grounded as well by yet another being, and so on; by the PSR, this cannot regress ad indefinitum.]

C1: There is a being that exists necessarily. [From D6[13] and D5][14]

C2: There is only one necessary being. [From D6, a necessary being is a being on which all possibilities depend. Therefore, if two necessary beings existed, the possibility of each of these would have to be grounded in the other. But then, for each being there would be at least one possibility whose ground is external to it, namely its own possibility. Bu then, that being would not be necessary.] (p. 34; I have slightly modified Boehm’s numbering to avoid confusion)

One thing that is immediately obvious is that Boehm has not tried to make this argument logically valid. In principle, that is fine; informal arguments have a purpose. But the point of offering logically valid reconstructions of historical arguments is to bring out exactly what would be needed to get from the premises to the conclusion. Boehm’s manner of presentation obscures this. For instance, his parenthetical argument for D3 (claim [ii] from earlier) is that it follows from the PSR. But which PSR does it follow from? It is plausible that some version of the PSR lies behind Kant’s commitment to (ii), for (ii) is really equivalent to the claim that the fact that some really possible predicate is really possible cannot be a brute (ungrounded) fact. Since it is clear from the context of the rest of the book and by later invocations of the PSR in this argument (e.g. the derivation of D4 and D6) that by the PSR he understands something in the neighbourhood of (viii) from earlier, one has to ask: why does (ii) require Kant to be committed to the full-fledged PSR? Why can’t Kant hold something weaker: facts about the real possibility of predicates cannot be brute? This is especially problematic when we realise that the pre-Critical Kant explicitly denies the unrestricted PSR because he denies that God’s existence has a ground.[15] In fact, Kant had denied that God’s existence has a ground already in the 1755 Nova dilucidatio, a text in which he explicitly endorses the ‘Spinozistic’ doctrine that the acts of finite creatures are necessary (while maintaining that freedom is compatible with necessitarianism).[16] Kant never endorses the unrestricted version of the PSR that Boehm hastily attributes to him. Given this fact, Boehm owes us a more careful statement of the PSR and how it justifies the various claims in this argument. Just to take an example: why must the relations between distinct grounds of possibility be grounded (as stated in D6)? Kant argues in Nova dilucidatio that causal relations of mutual interaction between substances must be grounded in a distinct substance (God) but Boehm owes us an account of why, if there are multiple existing grounds of possibility, they stand in interaction-relations that require grounds. He also owes us account of why the PSR excludes infinite regresses of grounds (also invoked in D6). Does this apply to all grounding relations (does the PSR exclude an infinite descending chain of causes stretching back into the past) or only some of them (e.g. the non-causal relation between a ground of possibility and the real predicates it grounds)? Likewise, no explanation is forthcoming of how C1 follows from D6 and D5. Why can’t it be that the substance that actually grounds all real possibility does so only contingently, i.e. that role could have been played by a different substance?

Aside from its problems as a reconstruction of Kant’s argument, there is the even more pressing question of whether Boehm is right to concur with Jacobi’s claim that Kant is committed to Spinozism in Beweisgrund. On one point, we have already seen that Boehm’s argument is inadequate: requiring that real possibilities have grounds does not commit one to the unrestricted PSR, because it was a commonplace of eighteenth German metaphysics to adopt some restricted form of that principle (as Kant had been doing at least since 1755).

Furthermore, Boehm claims that Kant’s doctrine that all really possible things are either determinations or consequences of God (BDG, AA 2:79) commits him to a Spinozistic view on which all finite beings are mere modes of the One Substance. Compare two claims that might be attributed to Kant:

1. All really possible predicates inhere in God.

2. Everything other than God inheres in (is a mode of) God.

First of all, it is unquestionable that Kant did not think (1) is true, for it would entail immediately that really possible negative predicates like ‘is not omnipotent’ inhere in God. (This predicate is obviously really possible because various actual things, like us, instantiate it, and what is actual is, trivially, possible.) The question is whether all logically primitive really possible predicates inhere in God, where the logically primitive predicates are unlimited, positive, and non-complex. Following Leibniz and Wolff, we can think of the derivative predicates as derived from a stock of primitive ones by limitation, negation, and formation of arbitrary logical complexes.

Kant’s explicit statement is that all really possible predicates are grounded in God either as “determinations” or “consequences”.[17] Does the pre-Critical Kant think that all logically primitive really possible predicates inhere in God? Here is Boehm’s argument that he does:

To be sure, we’ve already established that the necessary being grounds all possibilities. The question, again, is how Kant understands the grounding relation. There are two possibilities. First, consequences (finite beings) are ontologically separate from the necessary being; they ground possibilities by existing (as required by D2); however, as finite beings, they must be created by a necessary being, and it is in this sense that all possibility depends on a single being. Second, consequences, like determinations, are properties of the necessary being; they inhere in it. (p. 30)

But notice that Boehm has just assumed that the only alternatives for the relation between really possible finite things and God are either efficient causation or inherence. Why isn’t there an alternative: a non-causal non-inherence grounding relation of making possible? Secondly, it does not follow from the fact that really possible finite things are “ontologically separate” in the relevant sense (they do not inhere in God) that they themselves ground any possibilities. Boehm seems to have overlooked, rather than argued against, a view on which God is the non-causal ground of the real possibility of finite things as well as the causal ground of those that exist (he creates them), but they do not inhere in him.

This is a dialectically relevant position, since it is precisely the one Kant articulates in his metaphysics lectures. He criticises Spinoza for mistakenly defining a substance as one that is not grounded in anything else, rather than as one that does not inhere in anything else.[18] And he distinguishes, within the genus ground, between causal grounds of existence (ratio fiendi) and non-causal grounds of the possibility of things (ratio essendi).[19] This should come as no surprise to any reader of the CPR since, when Kant claims that space, time, and the categories, are the grounds of the possibility of the objects experience, he surely does not mean they cause the existence of those objects, nor does he mean that objects of experience are accidents inherent in our subjective intellectual forms. Thus room is opened, in the context of Kant’s own discussions of Spinoza, for distinguishing the doctrine that all really possible predicates are grounded in (made possible by) God from the doctrine that they all inhere in God.

But this is merely to show that Boehm is too hasty in moving from grounding to inherence. Did Kant in 1763 make the same mistake? In a passage in Beweisgrund itself Kant makes abundantly clear that he does not think that all really possible predicates inhere in God:

The data of all possibility must be found in the necessary being either as determinations of it, or as consequences which are given through the necessary being as the ultimate real ground. It is thus apparent that all reality is, in one way or another, embraced by the ultimate real ground […] But this is not to be understood to mean that all possible reality is included among its determinations. This is a conceptual confusion  which has been unusually prevalent until now. All realities are attributed indiscriminately as predicates to God or to the necessary being. That all these predicates can by no means co-exist together as determinations in a single subject is not noticed. The impenetrability of bodies, extension and such like, cannot be attributes of that which has understanding and will. Nor does it help if one seeks to evade the issue by maintaining that the quality in question is not regarded as true reality. The thrust of a body or the force of cohesion is, without doubt, something truly positive. Similarly, in the sensations of the mind, pain is never merely a deprivation. (BDG, AA 2:85–6)

Kant here explicitly repudiates the early modern rationalist view that God possess all unlimited positive properties, i.e. that the properties of finite beings are limitations and negations of divine properties. In Spinozistic terms, Kant is denying that God possesses every attribute. Kant’s reasons for thinking this are not hard to find. It is one of the cornerstones of his pre-Critical metaphysics (retained within the Critical system) that evil is not a mere limitation or negation of good.[20] The will to do evil is a positive reality ontologically on a par with good (though neither evaluatively nor causally on a par, because the capacity to do evil is caused in us by a being that possesses no such capacity). If Kant held that all logically fundamental (un-limited, non-negated) predicates were inherent in God, he would have to conclude that God is both good and evil. He is adamant that not all possible properties are limitations or negations of divine properties, not only in this passage, but throughout various texts from the 1760s.[21]

Boehm not only attributes to Kant a view Kant could not have done more to distance himself from (the inherence of all really possible predicates in God), he also fails to distinguish (1) from (2). This is a telling mistake in this context because (1) is upheld by Leibniz and Wolff. There is nothing specifically Spinozistic about the view that the material for all possible predicates inheres in God, and the predicates of really possible finite things are limitations, negations, and complexes of them. Boehm assimilates that standard eighteenth century doctrine of school metaphysics to (2), the principle that all finite things are modes of God. Even if Kant were committed to (1), which he quite clearly is not, this would make his conception of God Leibnizian, not Spinozistic.

Finally, the attribution of Spinozism to the Kant of Beweisgrund faces an additional exegetical hurdle: Kant argues that God (the ground of all real possibility) possesses an  intellect and a will (BDG, AA 2:87) and claims that nothing extended can have an intellect (BDG, AA 2:85–6).[22] It follows that Kant’s pre-Critical God is not extended (contra Spinoza) and possesses both an intellect and will.[23] Boehm responds by claiming that Kant’s argument nevertheless has Spinozistic implications because he takes ‘extension’ to be a fundamental property, and all fundamental properties must inhere in God. He then quotes a long passage on extension, but leaves this out: “[I]f space did not exist, or if space was not at least given as a consequence through something existent, the word ‘space’ would signify nothing at all” (BDG, AA 2:81). Kant is here making clear that space is not a determination of God (does not inhere in him) but is merely a consequence of God. Boehm overlooks Kant’s pre-Critical cosmological theory of space (space is the order of mutually interacting substances), according to which God is precisely not spatial (because his causal influence on substance is one-directional), in order to pin the Spinozist label on him.[24] What is more, the majority of the text of Beweisgrund is devoted to the rather un-Spinozist project of giving a physico-teleological argument for the existence of a “rational Author with great wisdom, power, and goodness” (BDG, AA 2:159), a point conveniently not mentioned by Boehm.

I do agree with Boehm that Kant retains his pre-Critical Beweisgrund conception of God as the ground of all really possible predicates within the Critical period, but demotes it from the object of a demonstrative philosophical argument to the status of an idea of pure reason, a subjectively necessary hypothesis of pure reason we posit in order to bring a kind of explanatory unity to our modal thought without claiming any knowledge of it.[25] However, Boehm misreads Kant’s later invocations of this idea in the 1780s as a persistent commitment to a specifically Spinozistic conception of God (what Boehm calls a “regulative Spinozism”). To take just one example, he quotes this passage from the Transcendental Ideal:

If, therefore, reason employs in the complete determination of things a transcendental substrate that contains, as it were, the whole store of material from which all possible predicates of things must be taken, this substrate cannot be anything else than the idea of an All of reality (omnitudo realitatis). All true negations are nothing but limitations [Einschränkungen]—a title which would be inapplicable, were they not thus based upon the unlimited, that is, upon “the All”. (A575/B603)

But Kant is careful to go on to distinguish this Leibnizian idea that all really possible predicates are limitations of God’s infinite realities from the correct ‘regulative’ (to use Boehm’s term) view on which God is the ground of some predicates which are not limitations of his own infinite realities, (the very consequences which are not determinations in Beweisgrund):

The derivation of all other possibilities from this original being, strictly speaking, also cannot be regarded as a limitation of this highest reality and as a division, as it were, of it; for then the original being would be regarded as a mere aggregate of derivative beings […] Rather, the highest reality would ground the possibility of all things as a ground and not as a sum total; and the manifoldness of the former rests not on the limitation of the original being itself, but on its complete consequences; to which our whole sensibility, including all reality in appearance, would then belong, which cannot belong to the idea of a highest being as an ingredient (A579/B607; my underlining)

Boehm reads this passage as making merely the Spinozistic point that God is not an aggregate but a ground of finite beings, but misses the crucial point: not all really possible predicates of finite things are limitations of divine realities. Kant’s reason for stressing this anti-Spinozistic and anti-Leibnizian point is a familiar point from Beweisgrund: it would entail that negative predicates, like evil and pain, are mere limitations of divine perfections.[26] Kant retains within the Critical philosophy the rational idea of precisely the God he had argued for in Beweisgrund: he stands in a grounding relation, but not an instantiation relation, to all really possible predicates.

Boehm’s narrative renders inexplicable one of the best sources of evidence for Kant’s continuing retention of the Beweisgrund conception within the Critical philosophy, his remark in the Pölitz lectures on rational theology that:

On this point rests the only possible ground of proof for my demonstration of God’s existence, which was discussed in detail in a work I published some years ago. Here it was shown that of all possible proofs, the one which affords us the most satisfaction is the argument that if we remove an original being, we at the same time remove the substratum of the possibility of all things.—But even this proof is not apodictically certain; for it cannot establish the objective necessity of an original being, but establishes only the subjective necessity of assuming [annehmen] such a being. But this proof can in no way be refuted, because it has its ground in the nature of human reason. For my reason makes it absolutely necessary for me to assume a being which is the ground of everything possible, because otherwise I would be unable to know what in general the possibility of something consists in [worin etwas möglich sey]. (V-Phil-Th/Pölitz, AA 28:1034)

Kant here goes further than anywhere else in stating explicitly that the very idea of God he claimed to have proven in 1763 is a subjectively necessary hypothesis of pure reason, one we must accept in order to think about real possibility in general. If this is, as Boehm claims, a specifically Spinozist conception of God as the One Substance why do we find Kant railing against Spinoza in the very same lecture transcripts as follows:

Fundamentally one might just as well call Spinozism a great enthusiasm [Schwärmerei] as a form of atheism. For Spinoza affirms two predicates of God: extension and thinking. Every soul, he says, is only a modification of God’s thinking, and every body is a modification of his extension. Thus Spinoza assumed that everything that exists is to be found in God. But thereby fell into crude contradictions. For if only a single substance exists, then either I must be this substance, and consequently I must be God—but this contradicts my dependence—or I must be an accident—but this contradicts the concept of my ‘I’, through which I think myself as an ultimate subject, which is not the predicate of any other thing. (V-Phil-Th/Pölitz, AA 28:1052, my underlining)

If Kant, in the Critical period, thinks that Spinozism entails contradictions, how can he retain even a ‘regulative’ Spinozistic idea of God as the ground of all real possibility? The answer is that Kant never accepted a Spinozistic conception of God, as he repeatedly and adamantly tells us. The closest Kant came to Spinozism was in the 1755 Nova dilucidatio when he asserted the necessity of everything that happens, including the free actions of rational agents. But even in that text Kant was anti-Spinozist enough to deny that God is the only substance and that God is a causa sui. Boehm’s interpretations of Beweisgrund and of Kant’s continued Critical engagement with its themes are credible neither in detail nor in broad terms.[27]

Chapter Two: The First Antinomy and Spinoza

In Chapters Two and Three, the strongest in the book, Boehm turns to the Antinomial Conflicts of Pure Reason (‘Antinomies’ for short) in the Critique of Pure Reason. Chapter Two considers the First Antinomy:

1. Thesis: The world has a beginning in time, and in space it is enclosed in boundaries.

1. Antithesis: The world has no beginning and no bounds in space, but is infinite with regard to both time and space.

One persistent question about the Antinomies has been which historical figures are supposed to be represented by the Thesis and the Antithesis. Sadik Al-Azm’s highly influential 1972 book The Origin of Kant’s Arguments in the Antinomies argued that the Thesis represents the view of Samuel Clarke and the Antithesis represents the view of Leibniz in the famous Leibniz-Clarke correspondence. Boehm quite effectively defuses this interpretation by pointing out that neither the Antithesis of the First Antinomy, nor that of the Fourth, represents views held by Leibniz (the Second and Third are more complicated). Leibniz did not uphold the infinity of the world in space and time, but held on the contrary that the world has a beginning in time and that it is not infinite in space but indefinite: for any part of the world, there is a larger part of the world that contains the former. The distinction between an indefinite series and a completed infinite series will be crucial later, but let me just say something brief here. A complete infinity is a collection of infinitely many elements (for any finite N there are more than N elements) such that the collection ‘as a whole’ is an object that exists in its own right. A series of collections is said to be indefinite just in case, for any of the collections, there is a larger collection of which it is a sub-collection, but there is no ‘total collection’ of which every collection in the series is a sub-collection. From a contemporary point of view one might wonder whether this is a tenable distinction (can’t we just take the set of all the sub-collections?); I shall return to that point below.

Boehm’s argument that we should see the 1st Antithesis as representing Spinoza is a useful corrective to the tendency to too hastily identify Leibniz with the Antithesis position. However, this suggestion is problematic in its own right. Consider the Thesis/Antithesis pairs of the other three Antinomies:

2. Thesis: Every composite substance in the world consists of simple parts, and nothing exists anywhere except the simple or what is composed of simples.

2. Antithesis: No composite thing in the world consists of simple parts, and nowhere in it does there exist anything simple.

3. Thesis: Causality in accordance with laws of nature is not the only one from which all appearances of the world can be derived. It is also necessary to assume another causality in order to explain them.

3. Antithesis: There is no freedom, but everything in the world happens solely in accordance with the laws of nature.

4. Thesis: To the world there belongs something that, either as a part of it or as its cause, is an absolutely necessary being.

4. Antithesis: There is no absolutely necessary being existing anywhere, either in the world or outside the world as its cause.

The problem is that, while Spinoza does accept the 1st through 3rd Antitheses, he rejects the 4th Antithesis (as does Leibniz, of course). So there are reasons to question whether the Antithesis position is Spinoza’s throughout.

But there are more general reasons to question how useful it is to try to assign particular historical figures to positions within the Antinomies. The Antinomies, Kant tells us, are supposed to function as an indirect proof that Transcendental Realism (TR) is false. In particular, we assume TR and show that TR gives rise to ‘antinomial’ proofs: TR is committed to both the Thesis and the Antithesis of each conflict, and these cannot both be true. The solution consists in rejecting TR in favour of TI and realizing that on TI either both Thesis and Antithesis are false (in the case of the first two or ‘mathematical’ antinomies) or one of them is true of the phenomenal world (Antithesis) while the other may, for all we know, be true of the noumenal world (Thesis), in the case of the second two or ‘dynamical’ antinomies.

But this means that the Antinomies target TR as such, not any particular form of it. It would defeat Kant’s purposes in the Antinomies if he were to show merely that a specifically Spinozistic form of TR is committed to the incompatible First and Second Antinomy. Unless, of course, Kant thought that TR quite generally is committed to Spinozism. Boehm does quote a number of suggestive passages (discussed above) in which Kant claims that Spinozism is the most consistent form of TR. Earlier I raised doubts about what kind of ‘Spinozism’ Kant has in mind there and questioned whether the arguments in those passages can carry the weight of identifying Spinozism with TR tout court. What is more, they fall far short of a Kantian argument within the parameters of the Dialectic itself that TR as such can be equated with Spinozism. Even worse, there are reasons to think it is positively unhelpful in the Dialectic to identify TR too closely with Spinozism. For recall the argument structure of each of the Antinomies:

  1. Assume: TR.
  2. If TR then Thesis.
  3. If TR then Antithesis.[28]
  4. If TR then ~(Thesis & Antithesis).
  5. ∴ ~TR

This means that if TR=Spinozism then Kant is claiming that Spinozism is committed to the Theses, which include: the world has a beginning in time, there are absolutely simple constituents of matter, and there is freedom. But this is absurd. Or to put the point more dramatically, the fact that Kant took Spinozism to be the “most consistent” form of TR, within the context of the Dialectic, has to be read as faint praise: TR as such has inconsistent commitments, so being the most consistent transcendental realist is a bit like being the most rational young earth creationist.

To be clear: Boehm’s main point is that the 1st and 3rd Antithesis represents Spinoza’s view. My point is simply that Kant’s tendency to identify Spinozism as the most consistent form of TR, emphasised by Boehm, is of little interpretative help within the Antinomies. Similarly, the argumentative purpose of the Antinomies (to prove the falsity of TR tout court) is poorly served by too closely identifying either the Thesis or the Antithesis with any specific historical representative. Kant’s predecessors are inspirations for his formulations in the Antinomies; he cannot make argumentative hay out of the fact that a particular Transcendental Realist formulated things a particular way.

The difficult issue in reconstructing Kant’s proofs in the Antinomies has always been that it is deeply unclear why the Transcendental Realist should be committed to all of their premises. There are two issues here, both directly relevant to Boehm’s discussion. First of all, the PSR is appealed to in both the Thesis and the Antithesis proofs, which leads to the question: (i) why is Kant entitled to assume that TR is committed to the PSR? and (ii) to what form of the PSR is TR as such committed? Boehm leaves us without an answer to either question. He seems tempted to answer (i) by saying that TR here just is a stalking-horse for Spinozism but recall that this would deprive the Antinomies of its generality as a disproof of TR as such. Regarding (ii) I suspect Boehm’s answer is “the Spinozistic PSR” but without a clearer articulation of what this principle is, such an answer is not very informative.

The other, perhaps greater, difficulty with the Antinomies is how they can function as indirect proofs of TI when Kant appears to appeal to TI itself in the proofs themselves. Consider, for instance, the proof of the 1st Thesis, as reconstructed by Boehm:

Thesis: the world has a beginning

  1. Assume (for the sake of reductio) the Antithesis: the world has no beginning; it is infinite.
  2. It follows that up to any given moment, an eternity has elapsed. This means that an infinite number of successive changes (events) have actually taken place. That is, an infinite series has been completed.
  3. However, the concept of an infinity (Unendlichkeit) is just that which cannot be completed through a successive synthesis (sukzessive Synthesis).
  4. The notion that an infinite number of worldly events has passed, therefore, is contradictory.
  5. Therefore, there is a beginning in time, a first event. (p. 71)

The problem, of course, is that this argument is invalid. To make it valid, Kant would have to assume something like:

(*) An infinite series can be completed only if the successive synthesis of its parts can be completed.

More generally, on the assumption of TR, whether or not a completed infinity is possible should not be assumed to depend upon whether any act of synthesis or combination by any intellect is possible. To put the same point another way, the TR’ist could object to (2) that all that has been proved is that the series of past alteration is complete not that it is completed (by some act). Kant seems to have illegitimately smuggled in his own TI view that the possibility of a series of events in time depends upon the possibility of a successive synthesis of those events by a discursive intellect. In other words, Kant seems to have imported the ‘psychologistic’[29] implications of his own theory.

Boehm confronts the problem head-on, writing:

The charge of psychologism is ineffective, however. It overlooks Kant’s appeal to the notion of synthesis in this passage, which is not epistemological or psychologistic. As H. Allison points out, Kant’s argument relies on a conceptual, not a psychologistic, distinction between an analytic whole (totum analyticum) and a synthetic one (totum syntheticum). A totum analyticum is a whole whose parts are not independently conceived: they cannot be regarded as existing, pre-given entities but must be thought of as mere qualities, or limitations, of the whole. A totum syntheticum, by contrast, is a whole whose parts are pre-given entities, which is conceived as the product of its parts. An infinite and complete totum analyticum is possible since its “parts” are mere limitations of the whole, whose infinity is given as prior […] An infinite and complete totum syntheticum, however, is impossible: the whole is produced by its parts whose enumeration proceeds ad infinitum. The world is a totum syntheticum, since it is metaphysically constituted of pre-given parts (such as material bodies, minds, etc.). Therefore, if completed, it is not infinite. This is the reasoning applied by the Thesis’s claim that “completing an infinite successive synthesis” is impossible. The conclusion is that the world has a beginning. (p. 88)

Perhaps the major contribution of this book is Boehm’s pointing out that Kant’s objection (that the infinite Spinozistic series of finite modes requires a totum syntheticum) was a standard eighteenth century objection to Spinozism, raised by both Christian Wolff and Moses Mendelssohn (Boehm traces it back to Bayle’s Dictionaire). Unfortunately, though, even this objection may beg the question against TR. Why must the TR’ist accept that the possibility of a whole that depends on infinitely many parts depends on the possibility of the parts being synthesised, combined or having anything ‘done to them’? Even Boehm’s formulation of the idea of an infinite totum syntheticum might be accused of re-importing psychologism: its “enumeration proceeds ad infinitum”. But the resolute TR’ist will insist that the parts do not need to be enumerated by us or by any intellect for the whole to exist. There are infinitely many parts, and in virtue of this, there is a whole of which those parts are parts. No synthesis or combination needed.

In fairness to Boehm, this may simply be an oversight on Kant’s part (though I suspect it is not). Likewise, if it was standardly assumed in the period that an infinite totum syntheticum must be the product of a combination or synthesis then it would not be surprising for Kant to have made this assumption as well, without an antecedent proof that this is a consequence of TR as such. However, in addition, it renders the Thesis proofs (which, recall, are intended to show the falsity of the Antithesis and thus, on Boehm’s view, of Spinozism) inapplicable, for, on Boehm’s reading, Spinoza did not hold that the world is a totum syntheticum (a whole dependent on finite parts) but a totum analyticum, a whole ontologically prior to its parts.

Boehm engages in a long discussion of whether Spinoza’s theory in the Ethics has the grounds to account for how we can have an adequate idea of a totum analyticum but, dialectically, that is beside the point. Kant’s claim in the Antinomies is not that TR does not have the resources to explain how we, as finite discursive intellects, can cognise the world, given TR (that is the argument of the Transcendental Analytic). He argues that if TR is true, then it is committed both to the finitude and to the infinitude of the world. Whether or not TR has the resources to explain its own epistemic possibility is beside the point in the Antinomies (though interesting in its own right).

To return to the infinite/indefinite distinction, Boehm’s observation that Leibniz upheld the indefiniteness, not the infinity, of the world is a valuable corrective to the literature, but potentially disruptive of his own interpretation. On this reading, Leibniz already held the view that the world is not a completed totality, i.e. that there is no such thing as the ‘complete world whole,’ but this potentially does serious damage to Kant’s project in the Antinomies. Recall that Kant’s solution to the First Antinomy is a version of precisely this Leibnizian view: objects in space and time are possible objects of experience, no possible object of experience is the “complete world whole”, so there is no such object as the “complete world whole”. However, the world is not finite (contra the Thesis), since, for any (experience of any) finite world-part, there is (an experience of) a larger finite world-part that contains the former. To put it in modern set-theoretic terminology, (i) for any set S there is a set of which S is a proper sub-set, but (ii) there is no set that contains all of the sets. But the existence of such a solution in Leibniz threatens to jeopardise not only the originality of Transcendental Idealism but also its uniqueness as a solution to the Antinomial conflicts. To the Spinozistic totum analyticum we can add an additional Transcendental Realist way of navigating the minefield of at least the First Antinomy: objects in space and time are transcendentally real, but there is no transcendentally real complete totality of them.[30] To those who question the coherence of this view, I would merely point out that (i) and (ii) are true of Zermelo-Fraenkel (ZFC) set theory and many contemporary Platonist philosophers of mathematics take sets to be ‘transcendentally real’ in Kant’s sense: they exist and have their properties independently of us. It may be that Leibniz, rather than Spinoza, is the metaphysical opponent Kant should fear the most.

Before moving on, I would like to note how blithely Boehm deals with a major piece of evidence on Kant’s attitude towards Spinoza, his claim (repeated in various passages) that my awareness of myself as a thinking being entails that I am a substance not a mode of another substance.[31] Boehm dismisses these anti-Spinozistic arguments by saying “Kant’s argument won’t survive Kant’s own criticisms of rational psychology in the Paralogisms”. However, ever since the pioneering work of Karl Ameriks, we have known that Kant’s attitude towards the substantiality of the soul is more complex than a surface reading of the Paralogisms would suggest. For instance, it may be that Kant denies that I know myself as a persisting substance while upholding knowledge of myself as a substance in the barest ‘logical’ sense: I do not inhere in anything else. Boehm discusses none of the extensive scholarly literature on this subject.

Chapter Three: The Third Antinomy and Spinoza

In this chapter Boehm extends his analysis from the First Antinomy to the Third Antinomy:

3. Thesis: Causality in accordance with laws of nature is not the only one from which all appearances of the world can be derived. It is also necessary to assume another causality in order to explain them.

3. Antithesis: There is no freedom, but everything in the world happens solely in accordance with the laws of nature.

Boehm points out, contra Al-Azm, that Leibniz is not very plausible as the representative of the Antithesis because, far from claiming that freedom and the PSR are incompatible, he defends freedom and the PSR and even argues that freedom requires the PSR (contra incompatibilist conceptions of freedom). But the significance of this point is undermined when Boehm himself later points out that “Kant considers natural causality and causality of freedom (spontaneity) to be contradictories (A531/B561)” (p. 113). If, however, the “causality of freedom” in the third Antithesis is the contradictory opposite of natural causality then it is built into the concept of “causality of freedom” as Kant uses the term that the act of a free being cannot have a determining antecedent ground (or “contrary-excluding ground” to use the terminology of that period). But this means that Leibniz’s fully compatibilist conception of freedom, on which even the free acts of rational agents have fully determining antecedent grounds, is not a violation of the 3rd Antithesis because Leibniz rejects “causality of freedom” in Kant’s sense. So there is really no barrier to associating the 3rd Antithesis with Leibniz, as long we allow that the 3rd Antithesis is a Leibnizian doctrine expressed in Kantian terms.[32]

The rest of Boehm’s analysis in this chapter follows the basic model of chapter two. Here is Boehm’s reconstruction of the (allegedly anti-Spinozistic) argument of the Thesis:

To sufficiently explain all worldly phenomena it is necessary to assume both natural causality and causality of freedom

  1. Assume (for the sake of reduction) the Antithesis: there is no freedom; all worldly phenomena take place solely in accordance with laws of nature.
  2. It follows that every worldly event (E3) “presupposes a preceding state” (E2), from which it necessarily follows.
  3. Further, it follow that the preceding state (E2) also came into being “in time”. [If E2 always existed, E3 would also have always existed. But this contradicts the assumption that E3 came into existence subsequently to E2.]
  4. Thus every worldly cause (such as E2) presupposes a preceding worldly cause, which itself follows “according to the law of nature”, and so forth, ad infinitum.
  5. Therefore, on the assumption that “everything happens according to laws of nature”, there will always be a “deeper” cause but never an ultimate one. Because the regress continues ad infinitum, the series of causes remains incomplete.
  6. However, the “law of nature” consists in the claim that nothing happens without a cause “sufficiently determined a priori”.
  7. Therefore, when taken in an “unlimited universality”, the claim that all causality takes place only in accordance with the laws of nature is contradictory.
  8. Therefore, causality in accordance with the laws of nature is not the only kind of causality. There is also causality of freedom. (p. 112)

But notice that structurally identical arguments arise for premise (5) as arose for premise (3) in the reconstruction of the argument of the 1st Thesis in Chapter Two. Why assume that an infinite series of causes requires an infinite “synthesis” of them or that the series must be “completed”? Why isn’t it sufficient to assume that there simply is a complete infinite set of elements and the existence of that set does not require any intellect or act of synthesis or completion? Again, the anti-Spinozistic claim is that this would require an infinite totum syntheticum, widely taken to be incoherent in this period (although its incoherence can be denied). And, once again, the Spinozistic defence is that the infinite series of finite modes depends upon an infinite whole, a totum analyticum, so no synthesis is required.

Perhaps this is the point to discuss what I think is Boehm’s most intriguing idea in these two chapters: the real difference between Kant and Spinoza comes down to the basic question of whether we can have an ‘adequate’ idea of a totum analyticum. Earlier I argued that this is not dialectically relevant in the Antinomies. Kant is arguing that transcendental realism gives rise to contradictions—not that the epistemology of transcendental realism cannot explain one of its necessary presuppositions (our adequate grasp of a totum analyticum). Putting the details of the Antinomies aside, Boehm asks us to consider this basic difference between Kant’s and Spinoza’s philosophies and whether Spinoza can sufficiently justify his assumption that we understand the idea of a totum analyticum. However, I am not sure that Boehm deploys this interesting idea as effectively as he might. Kant does not deny that we possess the concept of a totum analyticum; after all, he himself uses the concept to illustrate the contrasting concept of totum syntheticum. What is more, we intuit two infinite tota analytica: space and time. While this is less often appreciated by readers of Kant, we also intuit finite analytic wholes: the objects we intuit in space and time are not synthesised from their parts, because they are infinitely divisible into parts, which themselves are infinitely divisible, etc. If we had to synthesise our intuitions of every single empirical object, we would have to perform an infinite task.[33] (Such an object is a whole but not a totality (totum) because a totality is not a part of any further whole.) We are presented with some basic finite analytic wholes which we then either synthesise into representations of larger finite wholes or analyse to find their parts. Kant’s claim is that we never intuit a real totum analyticum, an analytic whole that is not merely a subjective form in which objects appear to us (it is real) and which is not a part of any further whole (it is a totality). Since we are never presented with a real totum analyticum we can never know that such a being is really possible. We can abstractly think about the real totum analyticum but cannot cognize it, or know anything substantive and positive about it.

This does present a basic difference between Kant and Spinoza and other rationalist thinkers. But it is unclear why Boehm focuses so heavily on the indirect argument in the Dialectic, especially since he finds it (ultimately) wanting. Kant gives a direct argument for an anti-Spinozist conception of our cognitive faculties (in the Transcendental Aesthetic and Transcendental Analytic) that excludes the possibility of immediate acquaintance with (intuition of) a real totum analyticum. Its first key component is the concept/intuition distinction. Our mere possession of the concept <totum analyticum> does not entail that there are any corresponding objects given to us for thought. Its second key component is the thesis that our faculty of intuition is sensible: we can only intuit (be given) real objects that causally affect our senses. Once Kant has assumed that we have sensible faculties of intuition, it becomes very doubtful that we can intuit a real totum analyticum. So while Boehm may be right that this is the real crux of the Kant-Spinoza dialectic (a claim to which I am sympathetic) it might have been more profitable to pursue it by investigating these foundational assumptions of the Kantian system and Kant’s grounds for them. In other words, to get at the philosophical heart of the difference between Kant and Spinoza it might have been more profitable to focus on the parts of Kant’s philosophical system targeted by Hegel and post-Kantian thinkers most committed to pushing Kantian philosophy in a more Spinozistic direction: Why assume that there is a sharp concept/intuition distinction? Why assume that our faculty of intuition is exclusively sensible?

It is one of the basic tents of the Critique of Pure Reason that we can know that a concept is instantiated only by intuiting an object; likewise, we can know that a concept is really possibly instantiated only by knowing that it is really possible (compatible with our forms of intuition and understanding) to intuit an object that falls under it. But this is a downstream consequence of what is (arguably) an even more fundamental doctrine: there are no ontological arguments. Nothing exists solely in virtue of its concept. A metaphysical position that allowed ontological arguments could enable us to give existence-proofs on purely conceptual grounds and thus be able to prove the contentfulness of our concepts without any appeal to intuition, sensible or otherwise. This points to what might be an even deeper issue between Spinoza and Kant—whether existence is a ‘real predicate’—and it is to this issue which Boehm turns in the next chapter.

Chapter Four: The Causa Sui and the Ontological Argument, or, The Principle of Sufficient Reason and the Is-Ought Distinction

The Third and Fourth chapters leave us with an important result: there is a form of TR that appears unaffected by the arguments of the Antithesis. It allows that there is an infinite regress of past events, an infinite regress of causes of a given event, and infinite regress of parts of spaces. The world is infinite in space and in past time. (Leibniz’s indefinite view of space may also remain a viable contender—see above.) However, it denies that any of these transform the world into an infinite totum syntheticum  by claiming that they are dependent modes of an infinite totum analyticum: events in time and parts of space are ontologically dependent on the cosmos as a whole. In each case we defuse the argument of the Thesis by denying that the infinite series of finite elements constitutes an infinite composite whole, and we embrace the Antithesis by denying that there is ever an element within the series that lacks a ground of the relevant kind (a prior event in time, an enclosing space, a prior cause).

But what grounds the totum analyticum that grounds each of these infinite series? Why does this being exist? One route would be for the TR’ist to claim that it is a necessary being and there is thus no ground of its existence; its existence is a brute fact. A TR’ist committed to a stronger form of the PSR, on which all facts about existence have grounds, will reject this and demand a ground even of the necessary existence of the totum analyticum. What could ground its existence? One historically influential answer is its essence, what it is to be that being. Taking this direction will lead the TR’ist to defend the traditional ontological argument: the ground of the whole world exists in virtue of its essence. It is worth noting, though, that this is only a coherent move for the TR’ist to take assuming that the fact that the totum analyticum has the essence it does (its essence is to be x, say) does not itself require a further ground. Otherwise, the essence of the TA will require a ground, which will require a ground, etc. In other words, the TR’ist will not assume the strongest form of the PSR from earlier:

(viii) All truths whatsoever have determining grounds and there is no infinite regress of grounds.

Instead, the TR will want to assume something like the following restricted version of the PSR:

(ix) All facts about the existence of any being whatsoever have grounds, as do all facts about the properties of all beings. There is at least one fact about one essence that lacks a ground.

Otherwise, faced with the essence of the totum analyticum (e.g. to be the totum analyticum is to be the infinite whole of which everything else is a dependent mode), we shall have to ask: what is the ground of the fact that that is what it is to be the totum analyticum.

To summarise, the most consistent and coherent version of TR that appears to survive the arguments of the Antinomies unscathed is this broadly Spinozist view: the infinite series of finite beings in space and time is dependent upon an infinite substance that exists in virtue of its essence (the causa sui). In other words, faced with a very strong version of the PSR (something like [ix] above), the TR’ist is going to face significant pressure to resort to a position on which there is at least one being that exists in virtue of its essence (in virtue of what it is to be that being). Consequently, if we want to trace Kant’s arguments against TR in the Antinomies to their logical conclusions we should turn to the Ideal of Pure Reason, where he gives his famous criticism of the ontological argument: being is not a real predicate. As long as the ontological argument remains viable, there appears to be a viable form of TR, indeed, a broadly Spinozist one.

Consequently, Boehm now turns in Chapter Four to the ontological argument and Kant’s critique of it. However, the discussion in this chapter is deeply unclear and at several places (I shall argue) confused. Matters are not helped by Boehm’s announcement that “while I do not think this confrontation captures Kant’s and Spinoza’s actual historical positions, it does capture the spirit of the philosophical question at stake” (p. 152). This leads Boehm to write in terms of the “the rationalist” position, and what “the rationalist” is committed to. In doing so, however, he fails to capture either Spinoza’s specific position or the underlying philosophical issues. At many turns there are options available to “the rationalist” that Boehm fails to consider. They may not be Spinoza’s options but why should that stop one trying to “capture the spirit of the philosophical question at stake”? Boehm’s discussion would have merited from a wider survey of “rationalists” with whom Kant engaged, especially Leibniz and Wolff. I am going to comment closely on the many twists and turns of Boehm’s discussion before returning, eventually, to what I think he is getting at and how his points would have been more happily (and clearly) expressed.

After an introductory section, Boehm writes:

[W]e get a handle on the notion of a self-caused entity, a rationalist may argue, as soon as we come to affirm the following two propositions. (1) Existence is co-extensive with and follows from conceivability: a thing’s conceivability does not depend on representing the causes of its existence, for conceivability is prior both to ‘causality’ and existence. A thing exists if and only if it is conceivable. (p. 154)

This is an odd place to start, for Leibniz, Wolff, Baumgarten and Spinoza distinguish between conceiving a thing in itself and conceiving of it through its causes. Not every thing that can be consistently conceived in itself exists.[34] This corresponds in Wolff and Baumgarten to conceiving of something per se possible but not (in every case) actual, to conceiving of what is “possible in its own nature” in Leibniz, and to conceiving of a finite mode through its own nature, in Spinoza. We can distinguish this from conceiving of a thing fully or adequately: conceiving of it through the grounds that would be sufficient to bring it about. Call this conceiving of the thing in connection (in nexu). The second rationalist doctrine Boehm gives is:

(2) Concepts are conceived through themselves. (p. 154)

He immediately goes on to clarify, though, that is not the full position; many concepts (e.g. <bachelor>) are conceived through other concepts, the concepts of which they are composed (e.g. <unmarried> and <male>). In fact, there is only one concept, according to Boehm, that is conceived “through itself”. However, it is deeply uncertain why “the rationalist” is committed to this, since it is unclear whether Leibniz, Wolff, Baumgarten or Spinoza is committed to it. One standard way of defining God (the being that is the ground of its own existence) in Leibniz, Wolff, and Baumgarten is as the being possessed of every unlimited perfection. But this means that the concept <God> is conceived through his unlimited perfections. Now it might be claimed that God is not being conceived through something other than himself because these perfections are perfections he has (they are his properties). But then isn’t the same thing true of <bachelor> and bachelors? Consider Spinoza’s definition of God in Ethics I:

by God I understand a being absolutely infinite, i.e, a substance consists of an infinity of attributes, each of which expresses an eternal and infinite essence.

So conceiving of the concept <God> seems to occur by way of conceiving of infinitely many attributes. Likewise, the concept <causa sui > is conceived of through conceiving of the concept <cause>. Boehm’s point seems to be that the “rationalist” thinks that all conceptual content is ultimately derived from some primitive stock of unlimited perfections (attributes) and that there is a unique being that possesses all of these primitive perfections, so that conceiving of any other thing in connection (how it could be grounded) requires conceiving of it through its dependence on this ultimate being. But if this is Boehm’s point then it is a very familiar point about rationalist metaphysics expressed in a highly confusing fashion.

Boehm goes on to write:

Yet this assumption [that there are conceptual truths, truths made true wholly by the concepts they contain, and independent of what or whether anything exists] isn’t initially plausible at all. On a first pass, the truth of an assertion like “triangles have three sides” obviously depends on existence. If no triangles actually exist, triangles don’t have three sides. Thus, unless existing is something triangles do, so to speak, by definition, it isn’t conceptually true that triangles have three sides. Of course, one will here [have] recourse to interpreting this conceptual truth as a hypothetical, “if all triangles exist, it has three sides”. But this hypothetical truth isn’t conceptual, either. For despite being hypothetical, the assertion itself is supposedly actually true (to wit, it is actually true that if a triangle exists, it has three sides). This truth in turn depends on existence: if it is true regardless of the existence of triangles that if one exists it has this or that property, this truth depends on something that exists and isn’t a triangle that makes it the case. Here, this may be the nature of space (given that space exists); or my mind, or yours, or God’s (provided any of these exist); or another possible world existing (supposedly in David Lewis’s mind) in which triangles do exist, and so on. Truth, even hypothetical, cannot be divorced from existence. Therefore, unless one makes existence a first order predicate and builds it into a thing’s definition, there can be no conceptual truth at all. It is common to say that the ontological argument fails because existential truths cannot be conceptually true in the same way that, say, “triangles have three sides” is conceptually true. In fact, however, given that all truth depends on existence, if existential propositions aren’t conceptual there are no conceptual truths at all. (p. 156)

This sloppy argumentation is, unfortunately, typical of this book. What is Boehm’s argument that “unless existing is something triangles do, so to speak, by definition, it isn’t conceptually true that triangles have three sides”? He goes on to consider the possibility of reading “All triangles have three sides” as (x)(Tx → Sx) and claims “if it is true regardless of the existence of triangles that if one exists it has this or that property, this truth depends on something that exists and isn’t a triangle that makes it the case”. But why? Why can’t the rationalist claim that (x)(Tx → Sx) is made true by the inclusion of Sx as one of the constituents of Tx and claim that once you get to facts about the definitions of concepts you hit explanatory bedrock. There is no ground of why the concept of a triangle is the concept of a three-sided flat plane figure; that is just what it is to be the concept <triangle>. After all Leibniz himself claims that every proposition can be derived from identities and definitions, thus implicitly exempting that class of truths from the requirement of groundedness. And there is good reason for the rationalist to do this: if the definitions of concepts, and facts about what the essences of things are, need to be grounded then the fact that God’s existence is grounded in his essence just pushes the explanatory burden back a further step. Why does God have that essence? This is simply the point I raised above; it is another downstream consequence of Boehm’s failure to ever clearly articulate what the PSR is. Boehm’s idea might be that if “all triangles have three sides” is made true by the inclusion of the concept <three-sided> in the concept <triangle> then it is grounded in something that exists, namely, those very concepts. But this requires us to assume that existence is not only a first-order predicate but a first-order predicate of concepts. Why must the rationalist assume that?

Boehm comes close to appreciating this point when he quotes Michael Della Rocca:

What is it in virtue of which a is conceivable, and, more specifically, what is it in virtue of which a is conceivable in terms of such-and-such? The answer is this: a is conceivable in a certain way because otherwise it would not be a. That’s what it is to be a. Asking why a is conceivable as such-and-such a way is analogous to asking why bachelors are unmarried. In each case, the question betrays a misunderstanding of the very concepts at work. (quoted by Boehm, p. 159)

But Boehm glosses this passage as follows: “‘[C]onceivability’ is a primitive notion […] conceivabilty is conceived (or accounted for) through itself.” Notice, though, that he has conflated Della Rocca’s point (there is nothing in virtue of which a is to be conceived through b and c because that is just what it is to be a) with a different one: conceivability is conceived through itself. The issue is not what it is in terms of which conceivability is conceived; it is that facts about what is conceived through what may not, in some cases, require further grounds. The PSR’s demand for grounds hits explanatory bedrock when we get to the definitions of things (which state their essences).

Boehm then goes on to consider a rationalist argument for thesis 1:

Let us move on to consider (1), the claim that conceivability implies existence. A rationalist argument for this is the following. Assume the PSR: (a) It follows that a thing, x, is conceivable if and only if its existence involves no brute facts. Thus if x’s existence involves brute facts it is inconceivable. [This is just the meaning of the PSR]. (b) It follows that if x is conceivable, x exists. To see that this is the case, assume, for the sake of reduction, (c) that x is conceivable and that the existence of non-x is conceivable too. (d) State-of-affairs (c) implies that the existence of both x and non-x involves no brute facts [by (a)]. If this state of affairs were possible, x’s conceivability would not entail its existence (it would entail its possibility.) However, this state of affairs is impossible. For (e) if both x and non-x are conceivable (hence by [a] involve no brute facts) and say, x exists rather than non-x, there can be no reason that x exists and non-x does not [by (c) non-x is equally conceivable as x]. (f) However, this implies that x’s existence contradicts (c) [for its existence, involves, contrary to what (c) states, a brute fact. (g) Therefore, if x is conceivable, non-x is inconceivable. (h) Therefore, if x is conceivable, x exists. (p. 157)

First of all, the derivation of (a) from the PSR is rather confusing since the PSR is usually formulated as a claim about grounds, not as a claim about what is conceivable. Perhaps what Boehm has in mind is something like this: to conceive of x is to conceive of it through the PSR. Since the PSR states that everything has a ground, conceiving of x requires conceiving of the whole series of its grounds. If this is so, then conceiving of x in Boehm’s terminology is equivalent to conceiving of it “adequately” or “in connection” (in nexu), rather than in its own nature (per se). Secondly, in trying to show why (b) follows from (a) Boehm assumes for reductio that x is conceivable but does not exist, from which it follows that (c) x is conceivable and the non-existence of x is actual hence (because the PSR actually holds) conceivable. (He repeatedly talks about the existence of non-x, but this cannot be what he means for that refers to the existence of anything that isn’t x; I take it he means the non-existence of x.) He then argues that this state-of-affairs is impossible by claiming that if x and the non-existence of x are both conceivable then there can be no reason why x does not exist rather than does exist (we are assuming, for reduction, that x does not exist). But this does not follow. Conceiving of x involves conceiving of a (possibly) infinite series of grounds; conceiving of the non-existence of x involves conceiving of a distinct (possibly) infinite series of grounds of its non-existence. It needn’t involve conceiving of an infinite series of actual grounds. If it did, it would be trivial that if we conceive of x then x exists: we are conceiving of something actually grounded in what actually exists! So this argument is either invalid or establishes an immediate consequence of the definition of what it is to conceive of something (which Boehm never states). What is more, it would establish a claim that even Kant would not want to resist: when you conceive of a thing by conceiving of its sufficient grounds in what actually exists, you have conceived of an existing thing.

Perhaps what Boehm has in mind is this: if, in conceiving of the existence of x, we conceive of an infinitely regressing series of grounds, since we are conceiving of a state-of-affairs in which the PSR obtains (because the PSR determines the scope of what is conceivable), then there must be a ground of the fact that the alternate infinitely regressing series of grounds of the non-existence of x does not obtain. But this means that the PSR is being applied not just to individual elements in a series of grounds (finite modes) but to the entire infinitely regressing series, a point Boehm fails to make explicit.

Boehm eventually comes around to the point I opened this section with, that the upholder of a certain version of the PSR is going to want to endorse the ontological argument:

Assume that the PSR is true:

(a1) Then everything that exists admits of a full explanation [that is the meaning of the assumption.]

(a2) Then there exists a necessary being [without a necessary being not everything would ultimately be explained—(a1) would be false, on pains of an indefinite regress.]

(a3) But that necessary being has a reason for its existence [by (a1)], either in itself or in another being.

(a4) That reason can’t be found in another being [if it were, the necessary being wouldn’t enable—on pain of indefinite regress—the ultimate explanation for which it was posited in (a2)].

(a5) Therefore, the reason for the existence of the necessary being posited in (a2) is that necessary being itself [by (a3) and (a4)]. If the PSR is true, there exists an entity containing the ground of its own existence—a causa sui. (pp. 160-1)

One thing that is very notable about this argument is that Boehm has just spent Chapters Two and Three pointing out that a Spinozist can escape the Thesis arguments of Kant’s Antinomies by allowing an infinite regress of finite beings provided they are dependent parts of a totum analyticum. It is thus odd to see the rationalist here arguing that infinite regresses are impossible. It seems that Boehm’s rationalist must have in mind some version of the PSR as follows:

(x) The existence of every being (not just contingent ones) and every property of every being must be grounded. There can be infinite regresses of grounds provided those infinite series are themselves grounded in some further being without an infinite regress of grounds. At least some facts about essences lack grounds (e.g. facts about the essence of a necessary being).

This is all fine and good. What I would have liked is for Boehm to explicitly state the version of the PSR his rationalist is committed to, how it differs from more and less restricted versions (which, for instance, might deny that the existence of a necessary being requires a ground) and what reasons the rationalist might muster in favour of (x) rather than some more restricted version. After all, once we abandon (viii) (which, arguably, is incoherent) what non-ad hoc stopping place is there for the rationalist? It would also behove Boehm to clarify why infinite regresses are acceptable between finite modes of the One Substance, while the grounding relation between that infinite series (or the individual elements) cannot lead to such a regress. One plausible answer is that this is because there are multiple kinds of grounding relations going on here: efficient causation (between finite modes) and inherence (between God and the infinite series of modes). But, as we have seen, this is potentially damaging for Spinoza; Kant claims that once we distinguish inherence from grounding, the temptation of Spinozism will dissipate.

Back to the main line of Boehm’s discussion: the rationalist now has excellent reasons to think that there is a being that exists in virtue of what it is (its essence). Kant has a famous objection to this metaphysical view: existence is not a real predicate. What does this claim mean, and why does it entail that there cannot be a being that contains the ground of its own existence or (as is sometimes claimed) exists in virtue of its concept? Boehm never squarely answers these questions, but from various remarks I gather that he takes this to mean: (i) existence is not a predicate that some objects have while others lack (it is not a ‘discriminating’ predicate of objects) and (ii) it isn’t fundamentally a predicate of objects at all, but a predicate of concepts, applying to those concepts that are instantiated by at least one object. (We can talk about existence as a predicate of objects if we like, but it is merely a ‘logical’ predicate; it does not distinguish one class of objects from another). Why does this entail that there cannot be a being that exists in virtue of its essence? Why is rationalism committed to the claim that existence is a real predicate? Boehm never really explains.

He quotes the standard passages on existence and the ontological argument from Beweisgrund and the CPR and reconstructs their argument as follows:

(A1) It is possible to have a complete concept of a merely possible thing, z.

(A2) The merely possible can become actual. [Assumption]

(A3) Existence isn’t a predicate of a merely possible thing, z. [If existence were a predicate of the merely possible z, z wouldn’t be merely possible but actual.]

(A4) Therefore, if z were to come into existence, existence wouldn’t be among its predicates. (p. 163)

He goes on to claim that the argument begs the question because the rationalist would deny A1. But this confuses conceiving of a thing in its own nature and conceiving of it in connection with its grounds. Leibniz, Wolff, and Baumgarten (and Spinoza, I would argue) thought you could not conceive of a merely possible thing fully in connection with its grounds (if you did, you’d be conceiving of it as actual); nonetheless, we can consistently conceive of a thing in its own nature, even if it does not actually exist.[35] So this is not the point at which the rationalist should object to A1. Instead the rationalist should point out that this argument is invalid as stated. If we assume that z is merely possible then if z were to come into existence it would have a predicate it actually lacks; the fact that its actual complete concept does not contain this predicate is irrelevant, for it would have a different concept. Boehm is assuming that the rationalist is committed to the claim that nothing could exist with a different complete concept.[36] But this is straightforwardly inconsistent with two other commitments of the rationalist position Boehm is reconstructing: (1) all of the predicates of a thing are contained in its concept and (2) some things exist contingently. Why saddle such a rationalist with the inconsistent assumption that a finite thing could not have had a different complete concept, in particular, that its complete concept might have included existence even though it actually does not? After all, a version of argument (A1)–(A4) will arise if we replace existent/actual with any predicate that is contingently had by those things that have it. Compare: assume z is not in the room. The complete concept of z contains all of its predicates. Since z is not in the room, it does not contain the predicate is in the room. If it were to come into the room, it would have a predicate not contained in its concept. Is this sufficient to refute “the rationalist”? Boehm’s reconstruction of this argument raises more questions than it answers.

In fairness to Boehm, he does remark that many of these Kantian arguments beg the question against the rationalist who embraces the ontological argument. But he pairs this with the claim that the rationalist cannot ‘non-circularly’ establish that existence is a real predicate. But the dialectical situation here is one in which Kant is claiming that there cannot be a sound ontological argument. So if Kant begs the question he fails in his philosophical aim. It does not matter, for Kant’s purposes, if the rationalist cannot prove that existence is a real predicate. In general, when Boehm complains that rationalists cannot ‘non-circularly’ establish various claims, it is unclear what this criticism is supposed to amount to. After all, the rationalist has to assume something to get their arguments off the ground; why not let them assume the PSR and then reason that there must be an ens necessarium and, for that reason, embrace the ontological argument? If Boehm’s point is supposed to be the Kantian one that we cannot have sufficient epistemic warrant for asserting the unrestricted PSR as a claim about all reality überhaupt then the appropriate Kantian text to consider occurs much earlier in the Critique and the appropriate philosophical questions to be raised are fundamental ones about philosophical methodology and whether epistemic questions must be answered before metaphysical ones.

Which is all to say that Boehm sheds little light on the ontological argument, Kant’s critique of it, and how the rationalist might respond. What is worse, he seems confused on basic points. Of Kant’s critique of philosophers who model philosophy on geometry (which Boehm, convincingly, reads as a condemnation of Spinoza) he writes:

Kant can dismiss the metaphysical use of definitions, axioms, definitions, and demonstrations only on the assumption that existence isn’t a first order predicate. For if it is a predicate, the geometrical method’s attempt to derive metaphysical truths from definitions would seem most appropriate. (p. 168)

No. Even if existence is a real predicate many of Kant’s objections to the “geometrical method” still stand: in philosophy we do not start with arbitrary definitions of concepts we have synthesised but with given concepts we must then analyse (A730–1/B758). Even if existence is contained in some of those concepts this by itself gives us no reason to think that philosophy can proceed as (Kant thinks) geometry does: by construction of its objects in intuition (A734/B762). Later on we read:

If we believe there is a distinction between the merely possible and the actually existing (if we assume that necessitarianism is false), we are committed to believing that existence isn’t a predicate. (p. 170)

Again, no. We can deny necessitarianism and hold that existence is a real predicate by claiming that there are objects that do not exist (existence is a discriminating first-order predicate of objects) while denying the PSR holds true at all. In fact, Boehm’s claim is almost the opposite of the truth: the most flat-footed way of distinguishing possibility from actuality is to claim there are merely possible objects that do not possess the predicate exists.[37] Anti-necessitarianism with a first-order existence predicate remains, for all Boehm has argued, a prima facie consistent position. Later on we read:

[W]e saw above that believing that necessitarianism is true is necessary for believing the ontological argument. (p. 171)

Nope. It is consistent to accept the ontological argument for God while rejecting necessitarianism because one holds, for instance, that God has incompatibilist free will and there is no determining ground why he chooses to create the world he does (rather than some other possible world). I quote these sentences not because they are crucial to Boehm’s line of argument, but because they are typical of the unseriousness and sloppiness of this chapter.

Now let us return to the main narrative of this chapter. The rationalist position that best escapes the arguments of the Antinomies is a broadly Spinozist one: an infinite series of finite beings standing in ground-consequence relations which are dependent upon a single substance whose essence is to exist (the causa sui). Kant objects that the idea of a causa sui is incoherent, but his objections beg the question against the rationalist. The rationalist, meanwhile, motivates the idea that there is a causa sui partly through appeal to a relatively unrestricted form of the PSR (see [x] above). How do we break the stand-off? In the rest of the chapter, which is significantly better than the initial sections, Boehm considers the source of our intuitive motivation for the PSR and then argues that the very motivation for this undercuts the strong rationalist application in Spinozistic metaphysics. Before continuing, a caveat: I am not sure why the rationalist position that emerges from the Antinomies and the ontological argument should be considered specifically Spinozist. While Leibniz and Wolff may have wanted to retain an initial moment of creation, there seems to be room for this kind of compromise view: the indefinite series (infinitely continuable though not a completed infinity) of past events depend upon God (so they are not a mere totum syntheticum) but they do not inhere in him. Since finite things do not inhere in God, there is room for more than one substance (ultimate subject of inherence). Likewise, we adopt a compatibilist conception of freedom: freedom does not require contingency in connection with a thing’s grounds, but merely that the omission of that act is consistently conceivable in its own nature. So far, this is a Wolffian view with the modification that the past series continues ad indefinitum but does not constitute a totum syntheticum, because finite beings all have a common ground.

In a short section, Boehm considers our philosophical motivations for accepting the PSR. He claims that we apply the PSR when we intuitively judge that some situation is contingent, that it could have been false: “[W]e don’t demand reasons for things we take to be necessary” (p. 175). But notice that this is belied by his own previous arguments. The rationalist does not rest content with a necessary being but proceeds to ask about the ground of its existence. He then goes on to point something out that would have benefited his earlier discussions: we do not ask for a ground of conceptual truths (e.g. why are bachelors unmarried?). (Note, though, he did earlier argue that these very conceptual truths must be grounded in something that exists—see above.) So even the rationalist has to admit the demand for grounds ends somewhere. Contra Boehm, the PSR does not rest on the anti-necessitarian distinction between what actually exists and what is possible. It rests on the distinction between what must be grounded and what does not require a ground, a distinction that holds as much within the scope of necessary truths (e.g. that the internal angles of a triangle sum to 180 degrees, to be a bachelor is to be an unmarried man) as within the scope of contingent truths. This should be especially clear in the context of Spinoza’s metaphysics, according to which everything exists necessarily, even though some things are grounded in others.

Boehm ends the book with a discussion of why, from a Kantian point of view, we at the same time both want to apply the PSR and to restrict it. His Kantian solution is that the contingency of events in the world is most manifest when we judge that they ought not to be the case. But this likewise leads us to restrict the PSR; an unrestricted PSR, Boehm claims, leads to necessitarianism, which is incompatible with its being the case that some things ought not be the case (assuming that ought implies can). But this only highlights one of the curious features of the book, his assumption that the PSR must take the Spinozistic/Wolffian form of demanding a determining ground rather than a merely sufficient ground. The free actions of rational agents are not uncaused, according to Kant; they are the activations of a capacity that could have produced the opposite (an omission) under the same circumstances. Why can’t the Kantian claim that this Crusian form of the PSR (even free actions have sufficient but non-determining grounds) is what lies behind our demand for explanations of free actions? Or to put it in the dramatic terms in which this book is couched, why isn’t the dilemma: Kant or Spinoza or Crusius?  What is more, the entire drama of the book (the PSR leads to fatalism) only holds if we assume by freedom we mean incompatibilist free will. If we accept the Leibnizian-Wolffian analysis of free will, freedom requires only (i) intelligence, (ii) spontaneity (being the causal source of the action, and that (iii) the action is contingent in its own nature (its omission is possible per se)? Why, then, isn’t the dilemma: Kant, Spinoza, or Leibniz?

Finally, there is a curious disconnect between the argument of the Preface and the rest of the book. As we saw earlier, the Preface argues that it is Spinoza’s denial of natural teleology that leads to what Boehm, following Jacobi, calls “nihilism”: all talk of value is meaningless or fictional. At one point, Boehm explicitly denies that the nihilistic import of Spinozism comes from necessitarianism (p. x). But the body of the book focuses on Spinozistic necessitarianism. According to the Preface, the kind of compatibilism Kant defends in Nova dilucidatio (1755) would seem to be enough to fend off what Boehm calls (in one of the more groan-inducing passages in this book) “our Nietzschean predicament”. Likewise, if we could combine the claim that everything that happens in nature is a necessary consequence of an antecedent state (all of which are necessarily grounded in God) with the claim that there is a teleological structure in nature, we could fend off the threat of meaninglessness.

Chapter Five: The Pantheismusstreit

There has been an apparent danger to Boehm’s project all along. In a letter to Reinhold, Hamann reports that Kant claimed he had “never been able to understand Spinoza’s philosophy”. Boehm’s response to what might be thought devastating to his project is that Kant was not being serious; in the same letter Hamann reports that Kant, having received a copy of Jacobi’s Letters on the Doctrine of Spinoza, was “very pleased with the presentation”. Boehm draws from this the conclusion that Kant cannot be taken at face value in claiming never to have understood Spinoza because at the same time he is “quite pleased with the presentation” of a book that claims that his 1763 work Beweisgrund was Spinozistic. This seems like very questionable hermeneutics. Can’t Kant be sincerely praising a book, even if he rejects one of its (patently false, as I argued earlier) conclusions (that Kant was a Spinozist in 1763), while confessing he himself never really understood the philosopher in question (Spinoza)? But Boehm needn’t have gone to such interpretative lengths because the Hamann letter is not really a problem for his project at all. What Boehm should have said instead is that even if Kant never understood Spinoza, the topic of this book is the relation of the content of Kantian Critical philosophy to the content of Spinozistic transcendental realism not the psychological states of Immanuel Kant, human being and resident of Königsberg.

My comments have gotten inordinately long so I am not going to discuss the final chapter, which is mainly a detailed blow-by-blow of the Pantheismusstreit. Boehm argues, convincingly, that the standard narrative, according to which Spinoza was a “dead letter” in German philosophy until Jacobi revived interest in him in 1783, needs at least to be seriously revised. He further argues that Kant’s essay ‘What does it mean to orient oneself in thinking?’ is not his only direct foray into the Streit; the Preface to the B-edition is intended to point out that the Critical philosophy showed how to be an Enlightenment philosopher without being a Spinozist before Jacobi ever raised the issue. These are valuable contributions to the growing literature on the Streit.

However, if the ambition of this book is not merely to contribute to our understanding of the Streit, and the relative positions of Kant and Spinoza, but to reawaken the Streit, as Boehm attempts to do in the Preface by arguing that a non-teleological conception of nature entails that “all talk of value is fictional or meaningless”, then he needs to provide some reason for thinking that Spinozistic monistic metaphysics (which has been the focus of the book) entails that nature is non-teleological. The PSR is prima facie compatible with a teleological account of nature: every alteration has a sufficient reason, both a mechanistic-causal one and a teleological one. This can be elaborated in various ways, but one historically relevant one is Leibniz’s: finite things strive for perfection and God has pre-arranged the order of efficient causes and the order of final causes so that efficient causes produce the very end for which finite things strive. It is not hard to give this a Spinozistic spin: God necessarily emanates a world of finite modes with this precise efficient and final-causal structure. The Pantheismusstreit arose because Jacobi assumed that Spinozism entails fatalism and nihilism. I just don’t understand why Boehm follows him in this assumption.

Acknowledgements: Thanks to Catharine Diehl and Lucy Allais for reading an earlier draft of this essay, and to Karl Schafer for discussing it with me.

Invited: 22 October 2014; received: 24 October 2015.


[1] This formula is due to Calvin Normore, in conversation.

[2] Not to mention the fact that Boehm owes us an argument that sophisticated versions of meta-ethical expressivism are committed to first-order nihilism.

[3] I have quoted the Guyer–Matthews translation, rather than Boehm’s somewhat looser, though more readable, rendition.

[4] Boehm says of this passage: “We may say that the situation facing most modern ethicists—current most-metaphysical Kantian ethicists included—is the situation that Kant here ascribes to Spinoza. Their theories formulate version of the moral law, but their position on matters of metaphysics (in the broadest sense of the term) forces them, in the final analysis, to give up the meaning of their theories.” That is precisely not Kant’s point in this passage, as I explain the main text.

[5] Cf. “This proof [the moral proof of the existence of God] […] is not mean to say that it is just as necessary to assume to existence of God as to acknowledge the validity of the moral law, hence that whoever cannot convince himself of the former can judge himself to be free from the obligations of the latter. No! All that would have to be surrendered in that case would be the aim of realizing the final end in the world” (MS, AA 4:451).

[6] I cite the Ethics as follows: roman numerals indicate the five parts, p refers to propositions, d refers to definitions, Dem refers to demonstrations, and s refers to scholia. E.g IIp40s2 refers to part II, proposition 7, scholium 2.

[7] Boehm acknowledges (p. 6) that it is often unclear which Spinozistic doctrine Kant is criticising when he criticises ‘Spinozism’. But then careful scholarship would require distinguishing different doctrines, considering which is the most likely target of Kant’s criticism, etc. Boehm does not do this.

[8] He does, however, argue that this passage in the Second Critique is a reference back to the proof of the Thesis of the Fourth Antinomy (A458/B486), that there is a necessary being in space and time. There are reasons not to be satisfied with this, though. First, both the proof of the Thesis and the Antithesis appeal without further argument to the problematic claim in this Second Critique passage that if God is the cause of space and time then he is in space and time (A455/B483; A455/B483); this is precisely what Wolff, Leibniz, and the pre-Critical Kant denied, and Boehm ought to give us the Critical Kantian argument against it. Secondly, Kant claims in the Second Critique that if God is in space and time then we are modes of him, something he does not argue for in the Fourth Antinomy. Thirdly, Kant’s argument in the Thesis of the Fourth Antinomy is from the PSR to the spatiotemporality of God, not from the latter to fatalism (as it is in the Second Critique passage).

[9] Boehm does discuss Kant’s remark in the Second Critique that “the dogmatic teachers of metaphysics have shown more shrewdness than sincerity in keeping this difficult point out of sight” (KpV, AA 5:103), namely, that transcendental realism about space entails Spinozism.

[10] The closest he comes is distinguishing what Kant calls the logical principle of reason (for every conditioned object, find its condition) from the pure principle of reason (for every conditioned object, the unconditioned series of its conditions is given). The latter is supposed to represent the rationalist PSR. But there are many things ‘the rationalist PSR’ could be (given in the main text) and each of these principles demands further explication. What does ‘condition’ (ground) mean here? And what falls within the scope of the ‘conditioned’? And does this principle allow for infinite regresses of grounds?

[11] See Leibniz (1969:226; 1989:28–31). For critical discussion, see Wilson (1980).

[12] Though, of course, the correct interpretation of Ethics IP8d is going to remain controversial.

[13] Boehm writes ‘A6’ which I think is a typo; I think he meant D6. I flag it just to show that I might be misreading his argument.

[14] It remains unclear why C2 follows from D6 and D5 on Boehm’s reading, i.e. why, if all possibilities are grounded in a single being, that being exists necessarily. Presumably, Boehm has something like this in mind: if this being did not exist, nothing would be really possible, so it exists with absolute necessity. But this was worth spelling out in the formal presentation of the argument itself.

[15] In the Herder metaphysics lecture from the 1760s: “[T]hat which absolutely necessarily exists does not exist because of a cause [Ursache], but rather because its non-existence cannot be thought. But this is merely a ground of knowledge [Erkenntnisgrund] and not an antecedently determining ground [ratio anticedenter determinans]. In short: it exists” (V-Met/Herder, AA 28:55). Cf. BDG, AA 2:86, 88.

[16] PND, AA 1:396.

[17] BDG, AA 2:79; see also AA 2:86 (“it follows from this . . . ”), 89, 125; see also the Herder metaphysics lectures (V-Met/Herder, AA 28:134).

[18] See V-Met-L2/Pölitz,AA 28:562; V-Met-K2/Heinze, 28:779; V-Met/Dohna, AA 28:638–9; V-Phil-Th/Pölitz, 28:1041, 1104–5, 1308.

[19] Cf. PND, AA 1: 392; V-Met/Herder, AA 28:37, 54; V-Met/Mron, AA 29:809, 810; V-Met-L2/Pölitz, AA 28:571; V-Met/Dohna, AA 28:648; V-Met-K2/Heinze, AA 28:802; Refl 5182, AA 18:111.

[20] See A273/B329.

[21] See the Herder metaphysics lectures, V-Met/Herder, AA 28:128, 132–3, 150.

[22] The structure of Kant’s argument at BDG, AA 2:87–9 is enough to show that Boehm’s Spinozistic interpretation is wrong-headed. In this section Kant argues that intellect and will are not merely consequences of God they are determinations of him. The problem he is concerned with is this: does God have a mind, or is he merely the ground of minds? If, as Boehm claims, real possibilities that are consequences of God inhere in God, this question would not even arise for Kant.

[23] Boehm notes the apparently anti-Spinozistic import of Kant’s attribution of intellect and will to God (pp. 42–3) but attempts to ameliorate this by citing Kant’s remark that “this leaves undecided the question whether the properties of understanding and will are to be found in the Supreme Being as determinations inhering in it, or whether they are to be regarded merely as consequences produced by it in other things”. Boehm fails to note, however, that this is a remark about what “the third Reflection establishes” (Reflections being the parts of the Beweisgrund), a remark that itself is contained in the Fourth Reflection, which itself contains the anti-Spinozist argument that God has an intellect and will (BDG, AA 2:87–9). Kant is very clearly saying: the intellect and will of God were left unproven earlier, that is why I have just argued for them.

[24] Found in the Nova dilucidatio and the Herder metaphysics lectures (contemporaneous with Beweisgrund).

[25] Boehm’s description of this as a ‘regulative’ use of the Beweisgrund idea of God is somewhat unfortunate because it does not, like the regulative ideas, play a role in regulating our empirical inquiry into the natural world. I prefer to describe it as a postulate of pure theoretical reason, but this is a relatively small difference in our interpretations.

[26] Cf. Amphiboly, A273–4/B229–30.

[27] For further criticisms of the Spinozist reading of Beweisgrund see Wyrwich (2014) and Hoffer (2016).

[28] It should also be noted that the Thesis and the Antithesis proofs themselves are indirect. In the case of the ‘mathematical’ antinomies (#1 and #2) they have the form: 1. Assume: Thesis (or Antithesis); 2. If Thesis (Antithesis) then p (e.g. that there is an uncaused first event in time); 3. ~p; 4. ∴ ~Thesis (Antithesis); 5. Thesis or Antithesis; 6. ∴ Antithesis (Thesis). Kant’s solution in the mathematical antinomies is to point out that the disjunctive premise (5) is true only if TR is assumed. In Transcendental Idealism, both the Thesis and the Antithesis can be false. The dynamical antinomies (#3 and #4) work differently.

[29] By ‘psychologism’ I mean merely that possibilities for object in space and time are grounded in possibilities of mental acts by discursive intellect.

[30] As well as the position I suggested earlier, that the world is a completed infinity, but that this isn’t the product of a synthesis (the view that Kant, like others in the eighteenth century, might have overlooked).

[31] E.g. V-Phil-Th/Pölitz, AA 28:1041, 1052.

[32] Boehm further argues against identifying the 3rd Antithesis with Leibniz by claiming that Leibniz holds that the grounds of the actions of finite agents are ‘incomplete’, e.g that there is no complete set of grounds for why Caesar crossed the rubicon. But Leibniz’s consistent position is that there is a complete set of grounds but it is infinite and involves perfection, so such acts remain contingent (unlike the truths of geometry). Boehm identifies, without argument, Leibniz’s ‘infinite analysis’ defence of contingency with his view that the world is indefinite but not infinite.

[33] Cf. McLear (2015).

[34] See especially Ethics IP33s1.

[35] In the case of Spinoza, see Ethics Ip33s1.

[36] This may remind readers of what is sometimes referred to as Leibniz’s doctrine of ‘counterfactual non-identity’. But it isn’t clear that Wolff or Baumgarten held this doctrine, and it is not clear that it is vulnerable to the objection as reconstructed by Boehm. Why can’t Leibniz claim that if the thing comes into existence at time t, this predicate is part of its complete concept? Presumably, Boehm’s idea then would be to run the argument (A1)–(A4) for a possible thing that never comes into existence. Then why would the rationalist accept premise (A2)? The rationalist might admit a version of (A2): the merely possible thing is consistent in its own nature and is in this minimal internal sense ‘possibly actual’.

[37] Rather than, say, expressing the difference between what is actual and what is possible in terms of what is the case, rather than what there is, i.e. deny that ◇p → p.



Hoffer, N. (2016), ‘The Relation between God and the World in the pre-Critical Kant: was Kant a Spinozist?’, Kantian Review 21(2):185–210.

McLear, C. (2015), ‘Two Kinds of Unity in the Critique of Pure Reason’, Journal of the History of Philosophy 53 (1):79–110.

Leibniz, G. W. (1969), Philosophical Papers and Letters, trans. and ed. L. Loemker (Dordrecht: D. Reidel).

——— (1989), Philosophical Essays, trans. and ed. R. Ariew and D. Garber (Indianapolis: Hackett).

Wilson, M. (1980), Leibniz’s Doctrine of Necessary Truth (New York: Garland Press).

Wyrwich, T. (2014), ‘Kants Anti-Spinozismus—Eine Antwort auf Omri Boehm’, Kant-Studien 105(1):113–24. 

© Nicholas Stang, 2017.

Nick Stang is Assistant Professor of Philosophy at the University of Toronto, Canada. His primary research interests are metaphysics and its history, mainly German philosophy. He has published in, among other places, Philosophy and Phenomenological Research, Kantian Review, Philosophers’ Imprint, Noûs and the British Journal for the History of Philosophy. His monograph Kant’s Modal Metaphysics has recently been published with Oxford University Press (2016), and is the subject of a forthcoming Author-Meets-Critics session on this site, with Jessica Leech and Andrew Stephenson as discussants.