1

The tenth proposition of Spinoza’s Ethics reads: ‘Each attribute of substance must be conceived through itself.’¹ Developing and defending the argument for this single proposition, it turns out, is vital to Spinoza’s philosophical project. Indeed, it’s virtually impossible to overstate its importance. Spinoza and his interpreters have used EIp10 to prove central claims in his metaphysics and philosophy of mind (i.e., substance monism, mind-body parallelism, mind-body identity, and finite subject individuation).² It’s crucial for making sense of his epistemology (i.e., Spinoza’s account of knowledge and response to skepticism) and in resolving puzzles within the Ethics (i.e., explaining human ignorance of all but two attributes).³ Even those who do not attribute some of the above claims to Spinoza need EIp10 to defend much of what they believe about Spinoza’s system. This paper locates a previously unnoticed argument for this proposition in Spinoza’s Short Treatise on God, Man, and His Well Being. There, Spinoza shows himself concerned with a powerful and underappreciated form of philosophical skepticism, one with surprising echoes in the work of his contemporary Leibniz as well as in the later Wittgenstein.⁴ Spinoza’s introduction of EIp10 in the Ethics circumvents this form of skepticism, solving the problem the Short Treatise envisions while also explaining that text’s argument’s absence from the explicit justificatory structure of the Ethics.

As commonly interpreted, EIp10 expresses Spinoza’s surprising claim that each attribute—that is, each fundamental way of being an independent substance—is conceptually and causally isolated from every other. Crucially, here I take all non-causal forms of dependence to be conceptual, including what contemporary philosophers call ontological or explanatory dependence. Conceptual talk, though perhaps not the most natural term for non-causal dependence relations in the contemporary context, best captures Spinoza’s own usage and thus suits my purposes here. And ultimately I side with those interpreters who think Spinoza identifies causal and non-causal forms of dependence.⁵ That said, for the two attributes Spinoza thinks humans can know,⁶ Thought and Extension, as well as their modes, particular minds and bodies, EIp10 entails:

Barrier: Nothing mental can be caused by nor conceived through anything physical, and nothing physical can be caused by nor conceived through anything mental.⁷

Spinoza’s predecessors, principally Descartes, accept the strict conceptual independence of Thought and Extension, such that, in Descartes’s words, ‘the concept of one is not contained in the concept of another’ (CSM I 298). In this paper, I assume that the conceptual, that is, non-causal, independence of attributes is common ground between Spinoza and his interlocutors.⁸ Given this shared presupposition, Descartes’s rejection of Barrier can be framed in two equivalent ways:

Interactionism: It is not the case that every attribute is causally and conceptually independent.

Interactionism’s functional upshot, for Descartes, is that that mental things (e.g., substances or modes) can cause physical things (e.g., substances or modes), and physical things can cause mental things, though of course only God, rather than any finite substance, can actually create other substances.⁹ But it’s also useful to see Descartes and Spinoza as disagreeing about two types of dependence relations. On this framing, the Spinozist defender of Barrier affirms, while the Cartesian Interactionist denies, the following:

Correlation: Conceptual independence entails causal independence.¹⁰

The heart of Spinoza’s explanation for the causal isolation of physical and mental things, on this reading, is the absence of any sort of conceptual, non-causal dependence between the two.

Two further facts about Barrier make this paper’s attempt to locate a textually grounded, genuinely Spinozistic argument for EIp10 essential. First, rejecting Barrier courts widespread inconsistency in the Ethics. But there is nonetheless substantial interpretive pressure, as historical and recent debates about either ‘idealist’ or ‘materialist’ readings of Spinoza make clear, to weaken or otherwise amend Barrier.¹¹ This might involve seeing minds as in some way ontologically dependent on, and thus pace Barrier causally or conceptually dependent on, bodies, as in materialist readings, or vice versa for idealist ones. Whether billed as critical correctives to or friendly clarifications of Spinoza’s metaphysics, such readings typically presuppose that much of Spinozism can survive the change. As Karolina Hübner forcefully reminds us in criticizing materialist readings of the Ethics, this is a mistake: ‘What is also at stake is the internal consistency of Spinoza’s philosophical system in the Ethics.”¹² EIp10 is an ineliminable part of the justification for substance monism and the other theses cashing out God’s nature (especially EIp14–16), from which virtually all the other central claims of Spinozism are argued to follow.¹³ Almost nothing of Spinozism survives its rejection.

Second, almost none of Spinoza’s most sympathetic interpreters have worked to elucidate his key argument or primary motivations for believing Barrier. Given its centrality, it might seem odd that so many of Spinoza’s most charitable readers have tried to make Barrier a negotiable part of his overall system. Della Rocca expresses this paper’s primary motivation in explaining why this occurs:

It is actually not clear why Spinoza holds that there is no conceptual connection between the mental and the physical, and thus it is not clear why he holds that any dependence between them would be inexplicable. This is a sore point between me and Spinoza. But given that he does regard mental-physical dependence as inexplicable, for this reason he invokes the PSR to deny that there is any such dependence. (Della Rocca 2007: 854, emphasis added)

In later work, Della Rocca reverses himself somewhat and gives the only extant argument for Barrier I am aware of, a derivation from the PSR.¹⁴ I think this argument fails, as does Spinoza’s own demonstration of EIp10, taken in isolation. To be sure, there are oft-cited texts from the Ethics, in EIp10 and elsewhere, which explain, given the totality of Spinoza’s other views, why Barrier is a natural position to have. I’ll discuss these in the final section. But given that Barrier figures as a premise in the proofs for nearly all members of this totality, these other views cannot justify it on pains of circularity.¹⁵

Since this point is essential for what follows, it bears repeating: Spinoza requires an argument for Barrier that precedes almost all his well-known commitments—substance monism, mind-body parallelism or identity, and necessitarianism—in the order of justification. The kernel of Spinoza’s previously unnoticed Short Treatise argument I give here both meets this criterion and answers Della Rocca’s question. There is no conceptual connection between the mental and the physical because conceptual connections imply causal interactions, causal interactions between attributes lead to skepticism, and skepticism is false. More specifically, if Interactionism is affirmed against a background of attribute conceptual independence, nearly all agents, nearly all the time, should have well-grounded doubts about their tacit or explicit assumption that certain incredibly implausible causal relations between minds and bodies do not obtain. But Spinoza thinks this version of skepticism is false, and so Interactionism must be rejected. Thus EIp10, which states the conceptual and causal independence of the attributes and implies Barrier, is true.

The next section gives the key passage from the Short Treatise, in which representative skeptical scenarios from this class are described, and identifies features of the context necessary for constructing Spinoza’s argument. Section 3 uses parallels with Leibniz and Wittgenstein, as well as recent work on Spinoza’s methodology and epistemology, to formalize Spinoza’s Short Treatise argument itself and to clarify the sort of rule-following skepticism at play. In the final section, I uncover hints of the Short Treatise argument in Spinoza’s Ethics and examine recent work tackling issues related to the truth of Barrier and the broader disagreements between Spinoza and Descartes. My reading helps explain why Spinoza says what he does in EIp10’s favor in the Ethics, shows why those remarks are insufficient in isolation but powerful when read in the light of the Short Treatise on the interpretation offered here, and identifies outstanding questions for the Spinozistic argument for Barrier.

2

Spinoza’s Short Treatise is a confusing text.¹⁶ Thankfully, for present purposes we do not require a coherent and comprehensive interpretation of the text as a whole. To extricate Spinoza’s previously unnoticed argument for EIp10 from it, it’s simply important to note that a) in the Short Treatise Spinoza only inconsistently endorses Barrier—that is, he occasionally accepts Interactionism—and that b) this presents him with a sort of skeptical problem.¹⁷ This problem then provides the raw materials for the Spinozistic argument for Barrier, in the precise sense that only by rejecting the Short Treatise’s occasional interactionism, in favor of Barrier strictly speaking, can Spinoza avoid the skeptical problem he discusses in the earlier text.

Now, that Spinoza does occasionally violate Barrier in the Short Treatise should not be controversial. For while in that early text he often hews quite closely to the position of strict causal and conceptual independence of the attributes,¹⁸ he also at times envisions a sort of war for causal control of the body’s motion, suggesting that this conflict results in anxiety:

And because these spirits can also be moved by the body, and so determined [in their direction], it can often happen that having their motion in one direction because of the body, and in another because of the soul, they bring about those anxieties which we often perceive in ourselves, without knowing the reasons why we have them. (KV I/92, 19–25, emphasis added)¹⁹

Spinoza is certainly sensitive to the apparent contradiction between those passages where he almost adopts the Ethics’s Barrier and those where he sounds interactionist. Like Descartes before him, Spinoza initially attempts to resolve the tension by appeal to the tight union between mind and body (KV I/91, 29–31): ‘The cause of this is, and can only be, that the soul, being an Idea of this body, is so united with it, that it and this body, so constituted, together make a whole.’²⁰ A few pages later, Spinoza suggests that the body’s only effect on the soul is to cause the soul to perceive its own body, and through its mediation, other external bodies. He then explains this union by connecting love with knowledge, saying: ‘So whatever else apart from this perception happens to the soul cannot be produced through the body. And because the first thing the soul comes to know is the body, the result is that the soul loves the body and is united to it’ (KV I/93, 10–13, emphasis added).

Wolf has tried to use passages like these to argue for strict, continuous endorsement of Barrier from the Short Treatise to the Ethics, suggesting: ‘Similarly, mind and body are really one whole, merely a double-faced mode of substance; mind does not affect body, nor body mind; the one simply is the other—that is, a parallel aspect of the same reality’ (1910: 228–29, original emphasis).²¹ This interpretation fails for two reasons. First, it still cannot explain the anxiety passage given above. Second, wholes have questionable ontological status in the Short Treatise. Spinoza characterizes them in this way: ‘Part and whole are not true or real beings, but only beings of reason; consequently in Nature there are neither whole nor parts’ (KV I/24, 19–21). So either the whole composed of mind and body does not exist to do the explanatory work required of it, or mind and body form a whole in some different, as yet unexplained sense. Therefore, on these two grounds I argue that Spinoza’s own attempts in the Short Treatise to resolve the apparent contradiction, while continuing to affirm a limited Interactionism, fail.

But what exactly is the problem that Spinoza’s Short Treatise Interactionism causes for his overall views, the problem that Spinoza’s own appeals to the union of mind and body as one whole fails to solve? Spinoza states the issue as a decidedly skeptical challenge, in a passage that I’ll refer to henceforth as Deviance:²²

The second possible objection is this: [A] we see that the soul, though it has nothing in common with the body, nevertheless can bring it about that the spirits, which would have moved in one direction, now however move in another direction—why, then, [B] could they not also make a body which is completely at rest begin to move? Similarly, why could [C] it not also move wherever it will all other bodies that already have motion? (KV I/97, 1–8, emphases added)²³

Even if souls and bodies interact causally, as Spinoza in his interactionist moods proposes, there must nonetheless be principled limits on the scope of souls’ causal powers over bodies. One plausible such limit is that each soul is naturally connected to a particular body with which it interacts preferentially. It seems that souls cannot control the motions of just any old body, or multiple bodies. But what could explain such limits and justify our confidence that souls cannot, for example, ‘[C] move wherever [they] will all other bodies’, given the complete conceptual independence of the mental and the physical?

Spinoza’s response to his own objection is twofold. On one hand, he sometimes argues for a tight mind-body unity as one whole, and suggests that this unity constrains causal interactions across attributes to those between a given soul and its single body (KV I/91, 29–31). I have already argued that this strategy fails on the Short Treatise’s own terms.²⁴ On the other hand, Spinoza’s response also includes the passage cited above, where he sounds most like the mind-body parallelist of the Ethics (KV I/98, 9–15, 24–36).²⁵ There, Spinoza sketches a view that closely approximates that given by Barrier, before tying his comments back to the objection imagined in Deviance and concluding: ‘[D] So Paul’s soul can indeed move his own body, but not that of someone else, such as Peter’ (KV I/98, 14–15).

The initial thought in the Short Treatise seems to be this: Spinoza can use the Barrier-like response to the possibility envisioned by [C]—that the soul could move wherever it will all other bodies that already have motion—to block the skeptical consequences of [C]. But this Barrier-like response, the thought continues, is no obstacle to the sorts of minimally interactionist claims expressed in [D]; namely, that ‘Paul’s soul can indeed move his own body, but not Peter’s,’ perhaps because ‘no change can occur in the object [Paul’s body] unless the same thing actually occurs in the Idea [Paul’s soul].’ To foreshadow: I do not think this initial thought, which seeks to move towards the view Spinoza defends in the Ethics while allowing that souls can move their own bodies, is tenable. Only embracing Barrier itself does the trick. I focus on developing this insight through further examination of the Deviance passage, along with [D], in what follows.

Interpreting the terse, rather generalized objection in Deviance requires some context. The first and crucial thing to note is this: Spinoza’s worry here is not just the traditional interaction problem! Princess Elisabeth formulates the traditional interaction problem in her 1643 letter to Descartes: ‘I ask you to tell me how man’s soul, being only a thinking substance, can determine animal spirits so as to cause voluntary actions’ (Descartes 2007: 22).²⁶ Spinoza echoes this traditional formulation when he argues: ‘Since there is no common measure between will and motion … the forces of the Body cannot in any way be determined by those of the Mind’ (EVPref, emphasis added).²⁷ But in Deviance Spinoza just assumes that the traditional problem is solved, at least for the sake of argument [A]: ‘… we see that the soul, though it has nothing in common with the body, nevertheless can bring it about …’ So Spinoza’s problem cannot just be the unintelligibility of Descartes’s view, pace several interpreters.²⁸ In this paper’s terms, Spinoza’s imagined objection presupposes Interactionism’s truth and overall intelligibility, denies Correlation, and evaluates what follows from these presuppositions.

Two very different sorts of problems arise for Interactionism on these assumptions. The first, expressed in [B], is more esoteric, an internal tension Spinoza identifies between Cartesian interactionism and Cartesian physics. Arguably, Descartes thinks that minds can redirect preexisting motions but are incapable of causing new motion in motionless bodies, believing this the only way to preserve the physical conservation laws.²⁹ The imagined objector questions the legitimacy of this restriction in the context of Interactionism. The second sort, in [C]–[D], is more general, more skeptical-sounding, but also more nebulous. I do not claim that mine is the only plausible interpretation, but at least this much seems clear. In [C]–[D], the objector imagines, again on the assumption of Interactionism and the denial of Correlation, a mind’s having limitless causal capability. This limitless capability covers any body ([C] ‘all other bodies that already have motion’), including that of another person ([D] ‘Paul’s soul can indeed move his own body, but not that [body] of someone else, such as Peter’ only if Spinoza’s Barrier-like response to the objection succeeds), and any possible motions (‘why could [the soul] not also move [them] wherever it will’). If minds and bodies can interact, Spinoza worries, isn’t any pattern of events, any combination of mental acts and extended motions, possible? If so, what explains why such deviant patterns do not obtain in general and what justifies the widespread assumption that they do not?

3

Spinoza, it appears, identifies a rule following paradox for interactionism. Wittgenstein’s §201 of the Philosophical Investigations is the most famous example: ‘This was our paradox: no course of action could be determined by a rule, because every course of action can be made out to accord with the rule.’ Kripke, whose development of Wittgenstein’s argument will be most relevant to my purposes here, initially envisions the rule-following paradox in the guise of a previously unconsidered mathematics problem, ‘68 + 57’.³⁰ Two possible rules are compared, the ‘plus’ rule which adds in the predictable way and gives ‘125’ and the ‘quus’ rule, which adds in the predictable way only if both numbers to be summed are less than 57, but simply produces ‘5’ if either summand exceeds 57. Wittgenstein’s rule-following skeptic (Kripke 1982: 11) then questions both whether there is any fact that grounds my having meant ‘plus’ rather than ‘quus’ (the constitutive question) and whether I am now epistemically justified in applying the ‘plus’ rule to answer ‘125’ rather than the ‘quus’ rule to obtain ‘5’ (the justification question). The problem, though Kripke first states it in mathematical terms, ramifies throughout the philosophy of language and metaphysics, in discussions of meaning or causation or anywhere rules or laws purport to connect certain groups of phenomena but not others. Spinoza’s favored version begins not with mathematics but with the rules or laws that purport to connect minds to one particular body but not to any old body indiscriminately.

Three parallels between Spinoza’s Short Treatise and Wittgenstein’s rule following paradox are especially helpful for appreciating further avenues for development of this argument in the Spinoza literature. First, Kripke (1982: viii, 60) sees Wittgenstein as having identified a quite general paradox about rule following, an entirely new form of skepticism, which recurs in tackling particular species of the issue in the philosophy of mathematics and the philosophy of mind. Regarding the latter, the rule following paradox recurs in Wittgenstein’s discussions of the ‘private language argument,’ the ‘problem of other minds,’ and what is known in the contemporary literature as the ‘pairing problem’ for dualism.³¹ Similarly, Spinoza’s problem actually extends far more generally, beyond mind-body interaction, to deviant interactions between modes of any of God’s infinitely many attributes.³² Prima facie, the infinite attribute version of the problem is far worse than its specific application to mind-body interaction in Deviance.³³ Nonetheless, those looking to defend Spinoza’s Barrier along the lines suggested here might find his argument sharpened by more explicit engagement with the pairing problem debate.³⁴

Second, on Kripke’s reading Wittgenstein gives a sort of transcendental deduction of §580: ‘An “inner process” stands in need of outward criteria.’ These outward criteria provide necessary evidence governing the correct application of an agent’s inner rule for speaking meaningfully or acting purposefully. Now, one important aspect of Wittgenstein’s discussion of his own problem is that he in some sense embraces the skeptical upshot of his paradox, before proposing a ‘skeptical’ rather than ‘straight’ solution to it.³⁵ This solution acknowledges that the skeptic’s questions are in a sense unanswerable. It then relies on a language game essential to our form of life to undermine skepticism that might prohibit the ordinary business of getting by:

[This language game] licenses, under certain conditions, assertions that someone ‘means such-and-such’ and that his present application of a word ‘accords’ with what he ‘meant’ in the past. It turns out that this role and these conditions involve reference to a community. They are thus inapplicable to a single person considered in isolation. (Kripke 1982: 79)

Wittgenstein’s skeptical solution requires that correct rule following attributions generally, and pain or meaning ascriptions more specifically, always involve reference to a community. One common criticism of this move is that Wittgenstein presupposes verificationism or behaviorism as a premise.³⁶ Kripke (1982: 100, 100n81, 120n5, original emphasis) rebuts this charge, arguing that the need for outward criteria, ‘circumstances, observable in the behavior of an individual, which, when present, would lead others to agree with his avowals,’ are transcendentally deduced in a Kantian sense, not simply assumed. These outward criteria are among the necessary conditions for the possibility of warranted assertions of rule following, meaning, and pain attribution required to avoid interminable entrapment in the skeptical paradox.

On my interpretation, Spinoza’s argument for Barrier has a similarly Kantian flavor.³⁷ Interactionism takes the place that transcendental realism occupies in the Critique, in that both are seen as engendering skepticism or antinomies not absolutely, but only on the (optional) assumption of their truth.³⁸ For Spinoza, embracing Barrier completes a key step in his arguments for the main conclusions of the Ethics, which helps avoid the skeptical results to which Interactionism gives rise. But the parallels between Spinoza and Wittgenstein are even sharper, I think. For the conditions Spinoza places on the proper investigation of the mind quite clearly echo Wittgenstein’s demand that inner processes stand in need of outer criteria.³⁹ Spinoza writes:

And so to determine what is the difference between the human Mind and the others, and how it surpasses them, it is necessary for us, as we have said, to know the nature of its object, i.e., of the human Body. I cannot explain this here, nor is it necessary for the things I wish to demonstrate. (EIIp13s, emphasis added)

For Spinoza, natural scientific investigation of God or Nature, which results from Extension’s publicly observable character, is the best way to understand the mind, even though nothing mental directly explains anything extended.⁴⁰

Third and finally, Kripke argues that, despite the importance of outward physical and behavioral criteria for licensing warranted assertions of rule following, meaning attributions, and being in pain, Wittgenstein still believes minds and the mental are ineliminable. Kripke interprets Wittgenstein this way for two reasons that emerge most clearly in his discussion of a body’s ‘having’ a ‘mind’ that is in pain. Kripke envisions someone who attempts to ‘bypass these [mental] intermediaries and deal directly with the connection of the sensation and physical object,’ by trying to locate pains in three senses, thereby avoiding any reference to minds (Kripke 1982: 129, original emphasis). Pains, one might think, are located in the sense that damage to a specific area produces pain, relief in that same area often eliminates it, and pains are subjectively felt in specific areas, like the feet or arms (Kripke 1982: 128). But, Kripke thinks, a wholly physical or behavioral account of pain, on the model sketched here, cannot succeed by Wittgenstein’s lights. This is ‘precisely because I cannot then define what it means for “another mind” to have the sensation in a given physical object, as opposed to “my” having it there’ (Kripke 1982: 129–30)

This discussion neatly parallels recent and historical debates about materialist readings of Spinoza and relatedly about why Spinoza thinks there are multiple attributes of God, including Thought.⁴¹ Indeed, Spinoza’s Ethics-style response to Deviance in the Short Treatise parallels in striking ways Wittgenstein’s argument that a purely physical account of pain sensations cannot account for the distinction between my having a pain at that location and some other mind doing so:

We said then that although Nature has different attributes, it is nevertheless only one unique being, of which all these attributes are predicated. We added that the thinking thing is also unique in Nature and that it is expressed in infinite Ideas, according to the infinite things that are in Nature. For if the body receives one mode, such as, for example, Peter’s body, and again another, such as Paul’s body, the result of this is that there are two different ideas in the thinking thing: One Idea of Peter’s body, which makes the soul of Peter, and another of Paul’s body, which makes Paul’s soul. So then, the thinking thing can indeed move Peter’s body, through the Idea of Peter’s body, but not through the Idea of Paul’s body. So Paul’s soul can indeed move his own body, but not that of someone else, such as Peter. (KV I/98, 9–15, emphases added)⁴²

That is, defining which mind, Peter’s or Paul’s, is having a given thought involves seeing whose body, which corresponding set of neural activity, is at a particular location in Extension. But Peter and Paul are distinguished both by the different locations of their bodies, and by the two different ideas those bodies make in the thinking thing. And this is why Peter’s soul can only move—Spinoza will say ‘is only of’ rather than ‘can move’ in the Ethics—Peter’s body, not Paul’s or anyone or anything else’s. The mind-body parallelism Spinoza develops fully in the Ethics and foreshadows in the Short Treatise requires both Thought and Extension to dissolve the skeptical paradox. So too, Wittgenstein requires both outward physical criteria and ineliminable reference to ‘the elusive conception of another “self” and its relation to material objects and sensation’ to achieve the same (Kripke 1982: 130).

Even if one finds the foregoing parallels somewhat thought provoking, one might nevertheless think that applying Wittgensteinian rule-following considerations to the early modern period to understand Spinoza’s texts constitutes gross anachronism.⁴³ On the contrary, quite apart from Spinoza, rule-following considerations are native to the early modern period, as Leibniz’s Discourse on Metaphysics makes clear:

… Not only does nothing completely irregular occur in the world, but we would not even be able to imagine such a thing. … if someone traced a continuous line which is sometimes straight, sometimes circular, and sometimes of another nature, it is possible to find a notion, or rule, or equation common to all the points of this line, in virtue of which these very changes must occur. … But, when a rule is extremely complex, what is in conformity with it passes for irregular. (Leibniz 1989: 39, emphases added)

Leibniz’s invocation of a notion or rule or equation in virtue of which any changes, no matter how irregular, must occur should help forestall a natural misunderstanding of Spinoza’s Deviance passage from the Short Treatise. One might think that the issue with denying Barrier and affirming Interactionism is that the latter violates the PSR, but this is a mistake.⁴⁴ For Leibniz’s point is precisely that, for any pattern of events, there is a notion or rule that explains how and why they occur exactly as they do. Any pattern of laws connecting minds to bodies would satisfy the PSR, or would, as Spinoza puts it in Ethics EIp11s, have ‘a cause or reason, as much for its existence as for its nonexistence.’⁴⁵ The issues raised by Deviance extend beyond both the traditional interaction problem and any putative costs of violating the PSR.

Moreover, while Descartes does not echo Leibniz’s explicit invocation of rule following, he too appears deeply concerned with issues contained in each of the three parallels drawn above. This, at least, appears to be one upshot of Chantal Jaquet’s insightful recent book on the affects and passions in Spinoza. Jaquet begins by describing how Descartes and Spinoza differ from their predecessors:

Although the passions of the soul bring into play the union and are in reality of a psychophysical nature, Descartes seeks to explain them as a natural philosopher. … [But] Spinoza also calls for the naturalization and rationalization of the passional phenomena when he proposes to deal with them geometrically. This approach involves breaking with those who considered the passions as things outside nature and contrary to reason. … It must be noted that Spinoza also deals with the affects as a natural philosopher, to the extent that he contemplates considering them as bodies. (2018: 35–37)⁴⁶

Jaquet’s comparison between Descartes and Spinoza here, both of whom examine the affects as natural scientists in contradistinction to their predecessors, echoes one feature of the second parallel above. Namely, for both of them scientific investigation of the body is in some sense the best avenue for ultimately understanding the mind, despite Spinoza’s causal and explanatory isolation of the mental and the physical.⁴⁷ This task, as Jaquet outlines it, involves a rigorous natural scientific investigation of the mind and the body that avoids two temptations. Namely, it must not stray either into dualism (hence the connection to the causal pairing problem in the first parallel) or into materialism (hence the third parallel’s insistence on the ineliminability of the mental).⁴⁸ Taken together, these parallels help organize the connections between Spinoza’s early work, that of his contemporaries, and contemporary rule following discussions. The foregoing shows how they outline a task (inquiry into the mind through the natural scientific investigation of body) and ward off potential pitfalls (dualist or materialist models). Moreover, it shows that we must resist easy solutions (among other reasons, by showing that the PSR alone cannot play the outsized role in Spinoza’s system that some recent interpreters have taken it to play). Further consideration of these connections thus potentially reorients purely interpretive discussions in the Spinoza literature, but also provides rich raw materials for crafting an alternative to currently dominant views in the philosophy of mind.⁴⁹

Having situated Spinoza’s skeptical challenge within contemporary discussions of rule following, especially as they relate to the metaphysics of mind, we should now try to extract a formal sketch of the argument for Barrier from the materials in Deviance.⁵⁰

1. Every attribute is conceptually independent.

2. Interactionism is true. That is, Correlation—conceptual independence entails causal independence—is false.

3. If (1) and (2), there are good grounds for doubting: [C] souls cannot impart any motion to any body wherever they will and [D] Paul’s soul cannot move Peter’s body.

4. But there are no good grounds for doubting: [C] souls cannot impart any motion to any body wherever they will or [D] Paul’s soul cannot move Peter’s body.

5. If there are no good grounds for doubting that Paul’s soul cannot move Peter’s body, then there are no good grounds for doubting that Paul’s soul cannot move Paul’s own body.

6. Therefore, given (1), not (2). Interactionism is false and Correlation is true.

7. Therefore, Barrier is true.

I take premise (1), as stated before, to be common ground between Spinoza and his interlocutors, most notably Descartes. Interactionism, introduced for reductio, is given by (2).⁵¹ Premise (3) turns the questions posed in the Deviance passage into concrete skeptical possibilities—that souls can impart any motions to just any body, and in particular, that Paul’s soul could move Peter’s body. It then denies that an interactionist who affirms the total non-causal conceptual independence of the mental and physical can justifiably deny these deviant causal possibilities. Premise (4) expresses anti-skepticism, as it seems we do have good grounds for denying that souls can impart motions to just any body, or that Paul’s soul can move Peter’s body. Premise (5) claims that whatever anti-skeptical grounds preclude Interactionism in general work equally well against the specific interactionist case of a soul and its own body. But given that the skeptical causal possibilities can be rejected, Interactionism is false and thus Barrier is true.

With Barrier defended, Spinoza can assert (let us simply suppose), all his major conclusions from the Ethics, including substance monism, mind-body parallelism and mind-body identity, and necessitarianism.⁵² Spinoza’s mature system precludes the possibility that any soul might impart just any motion to any body, or indeed to any body whatsoever. There is exactly one necessary set of finite modes of each attribute, and thus one set of connections among the modes of Extension, which is causally isolated from but paralleled by the corresponding modes of Thought to which they are identical. In sum, Interactionism gives rise to the rule following paradox and cannot solve it. Barrier solves the rule following paradox, gives rise to the totality of Spinoza’s views in the Ethics, and Spinoza’s mature system prevents the rule following paradox from reemerging.

Fully defending an argument that both results from the materials in Deviance and whose formal presentation jibes perfectly with discussions of skepticism in contemporary epistemology lies well beyond the scope of this paper. For one, Spinoza does not have, and probably would not accept, many of the epistemological concepts such discussions typically presuppose, though his thinking would be far more at home in current debates in the philosophy of mind. But I can say more about the crucial premises (3)–(5), drawing on Peterman’s insightful recent interpretation of Spinoza on skepticism. Based on her analysis of Spinoza’s most explicit discussion of these issues in the Principles of Cartesian Philosophy, Peterman suggests that, according to Spinoza:

S’s belief that p is certain if and only if all three of the following are true:

1. p is a “mathematical truth.” A “mathematical truth” in the PPC seems to be a truth that we are “compelled to infer” from a clear and distinct idea, and not necessarily one that concerns mathematics (C 238).

2. S cannot discover a “ground for doubting” p. (According to Spinoza, a “ground for doubting” p is an idea from which a doubt that p can be inferred—so ultimately we will want to know: what is a doubt that p, for Spinoza?)

3. S can infer p (or, S can “infer that p is most certain” (C 237)) from every other idea that S can attend to (Peterman 2018: 343–44).

Criterion (b) requires that a ground for doubting p be an idea from which a doubt that p can be inferred. Premise (3) of my reconstructed Spinozist argument suggests that, on the assumption of Interactionism, we have good grounds for doubting: [C] souls cannot impart any motion to any body wherever it will or [D] Paul’s soul cannot move Peter’s body. That is, we have good reason to think that skeptical interactionist possibilities cannot be ruled out. Premise (4) then suggests that anti-skeptical grounds do exist, such that Interactionism is the assumption to reject, and premise (5) applies those same anti-skeptical reasons to the limited, intuitive case of a soul’s interaction with its own body.

Obviously, the interactionist would like to exclude the more skeptical interactionist possibilities while preserving the special case of interaction between a soul and its own body. But given criterion (b) above, I think we can see why Spinoza thinks this desire is a quixotic one. The problem is that the idea from which a doubt can be inferred about our ability to rule out skeptical options—a soul moving any body at will, or Paul’s soul moving Peter’s body—just is the non-causal conceptual independence of the attributes. This shared Spinozist-interactionist premise denies that there are any conceptual connections between the attributes of Thought and Extension.⁵³ But now consider the interactionist who suggests preserving the limited case of causal interaction between a soul and its own body while denying that other deviant causal relations can arise between a soul and another’s body, or between souls and any distant body whatsoever.

In making this proposal but denying the deviant causal possibilities, Spinoza thinks, one thereby asserts a set of conceptual connections between what might be called the active powers of a soul and the receptive powers of its own particular body. One correspondingly denies any such receptivity in another’s body or in far-flung distant bodies. But the impossibility of such conceptual connections—those that definitively countenance some possibilities but definitely preclude others—can be straightforwardly inferred from attribute conceptual independence.⁵⁴ This entails not only the weaker claim that [C] and [D] are not certain, on the assumption of Interactionism, in virtue of failing criteria (b) (i.e., since a ground for doubting [C] and [D] exists). It further entails premise (5), which claims that all positive claims about a soul’s causal connections to its particular body will have the same grounds for doubt, which are rooted in the complete conceptual independence of Thought and Extension.

Of course, Spinoza does not himself think that [C] or [D] is doubtful, but his reasons for thinking them immune to doubt, using Peterman’s framework, are far stronger than those of the interactionist. For once Interactionism is rejected and Barrier affirmed, Spinoza’s other arguments in the first two books of the Ethics, those which employ Barrier as a premise, can help him establish that EIIp47: ‘The human mind has an adequate knowledge of God’s eternal and infinite essence.’ From this idea, as Peterman reads Spinoza, ‘If I begin with a C&D idea of God’s essence, carefully deduce what follows from it in the proper order, and populate my mind only with the resultant ideas, I will never generate a ground for doubting that I know any of these ideas’ (Peterman 2018: 345). Spinoza, but not the interactionist, thus ends up affirming [C] and [D] with certainty, in a way that respects premise (4) by making clear that there is no ground for doubting them, contrary to the position in which the interactionist finds himself.

To be clear, this does not necessarily entail that the interactionist has many, or even any, false beliefs about these far flung possibilities. Both Spinoza and Descartes, for instance, would accept [C] and [D]. Rather, here Spinoza simply turns the well-known Cartesian point about the atheist geometer against interactionism itself. Whereas Descartes acknowledges that the atheist geometer can perform all sorts of technically correct proofs, the atheist geometer’s results are always open to doubt if the evil demon scenario happens to occur to him.⁵⁵ Similarly, Spinoza admits that the interactionist truly believes that skeptical scenarios of the sort alluded to in [C] and [D], such as Paul’s soul moving Peter’s body, do not obtain. But he argues that his interactionism, combined with his belief in attribute conceptual independence, makes such beliefs forever vulnerable to well-grounded, reasonable doubts. For by the interactionist’s own lights, given conceptual independence, he should not be able to make any firm claims about which bodies a soul can (its own) and cannot (all other bodies out there) control.

Since there are zero conceptual connections between the powers of any soul and the receptivity of any given body, the interactionist cannot guarantee himself exclusively true beliefs that attain some other reasonable epistemic standard while simply failing to meet the standards of certainty Spinoza requires. That is, interactionism’s problem is not just that Spinoza deploys tacitly infallibilist standards of knowledge. Every claim about causal relations between particular minds and bodies is doubtful in exactly the same way and to the same degree. The conceptual independence of the attributes prevents the interactionist from having any justification at all, not just infallible ones, for limiting the scope of possible causal relations between Thought and Extension.

Much remains to be said about this argument. One natural way of pushing the conversation further would be to bring Spinoza’s arguments into dialogue with the vast literature on Cartesian skepticism. Those few interpreters who have written at length about Spinoza’s relationship to skepticism have taken this tack; the Short Treatise argument would simply be another instance of this general trend.⁵⁶ But Spinoza’s brand of skepticism is actually not all that similar to the familiar Cartesian fare. Rather, Spinoza envisions sets of competing causal relations between minds and bodies, ordinary on one hand and deviant, skepticism-engendering on the other, all of which are necessitated and explained just as the PSR requires. Unlike Cartesian skepticism, Spinoza’s problem arises when the unexpected implications of a widespread, ordinary view of mind-body interaction come to light. By contrast, Descartes knows full well that evil demon doubts are artificially constructed for a particular purpose in a specific, limited philosophical context. Spinoza thinks, however, that the mere dual assumptions of Interactionism and attribute conceptual independence simply do not permit us to say that there is any fact that grounds the obtaining of ordinary but not deviant causal relations, or that epistemically justifies our assumption that they do.

4

The preceding discussion of the Deviance passage gives us good reason to question Della Rocca’s (2007: 854) contention that Spinoza’s reasons for Barrier are opaque and undefended. As we have seen, allowing causal or conceptual connections between Thought and Extension, as Interactionism does, gives rise to a set of skeptical scenarios in [C]–[D] whose resolution requires Barrier. While in the Short Treatise itself Spinoza adopts a view quite close to, but not yet exactly Barrier that still licenses,⁵⁷ he thinks, claims like ‘Paul’s soul can indeed move his own body, but not that [body] of someone else, such as Peter,’ by the Ethics this exception has been removed. Spinoza’s Ethics affirms the complete conceptual and causal isolation of each attribute from every other, including in the case of the mind and body of one individual human being. I have been suggesting that Spinoza’s Barrier in its mature form springs from continued meditation on the inadequacy of Spinoza’s own proposed solutions in the Short Treatise, whose failures by Spinoza’s own lights I’ve already discussed. Granted, I admit to lacking biographical evidence that would fully substantiate the claim that meditation on this problem leads Spinoza directly to his mature views. I’ve simply argued that it is a problem for Interactionism, one the Spinozism of the Ethics lacks thanks to Barrier.

This final section does three things. First, it confirms what has thus far been presupposed, namely that Barrier requires additional justification beyond what it receives in the Ethics. In so doing, we can also see remnants of the Short Treatise’s Deviance passage rear their heads in the Ethics, albeit outside its explicit justificatory structure. Second, it discusses possible Cartesian interactionist attempts to preserve mind-body interaction in the special case of a soul and its own body in greater detail, through discussion of recent work seeking to identify and adjudicate the deepest disagreements between Spinoza and Descartes. Finally, it also identifies some major outstanding questions for the argument developed thus far.

Having set out an alternative history of the development of Barrier, I now consider what Spinoza says on its behalf in the Ethics, beginning with EIp10’s demonstration, which reads: ‘For an attribute is what the intellect perceives concerning substance, as constituting its essence (by D4); so (by D3) it must be conceived through itself, q.e.d.’ That is, Spinoza appears to suggest that Barrier follows simply from his definitions of substance and attribute. But this demonstration is not confidence inspiring. For instance, EId3 defines substance as self-inhering, self-conceived, and conceptually self-contained. As Spinoza puts it: ‘By substance I understand what is in itself and is conceived through itself, that is, that whose concept does not require the concept of another thing, from which it must be formed.’ And EId1 connects self-conception explicitly to self-causation—‘by cause of itself I understand that whose essence involves existence, or that whose nature cannot be conceived except as existing.’ But why should constraints on self-conception, self-inherence, and self-causation apply directly to the attributes of substance, rather than only to substance itself?

One too quick answer might be: the same constraints apply to substances and attributes because Spinoza thinks, as he says in EIp4d, that substances and attributes are ‘the same.’ Eisenberg (1990: 4, 9) discusses this response before rejecting it, arguing:

If (e.g. Thought), as an attribute of God, is supposed to be identical to the essence of substance but also Extension is identical to the essence of substance, then two things which, according to Spinoza himself, are supposed to be irreducibly different from one another (namely, Thought and Extension) turn out to be identical to one and the same thing and, hence, identical to each other … [Furthermore] it is hard (for me) to see why Spinoza gives the emphasis that he does to the notion of substance if in fact he believes it has the same content as does the notion of attribute(s).

Eisenberg’s description of the problem with treating ‘substance’ and ‘attribute’ as the same appeals to an apparent failure of the Transitivity of Identity. For if both Thought and Extension were identical to Substance, simple substitution would seem to garner that Thought=Extension, contra Spinoza’s explicit insistence that each attribute is independent, distinct from the others, and conceivable only through itself.

Recently, Garrett has proposed a reading of Spinoza’s metaphysics that makes sense of the failure of the Transitivity of Identity Eisenberg judges fatal to the sameness proposal. The stakes here, as Garrett rightly notes, are high, since ‘given their fundamentality [the principles of the Transitivity of Identity and Indiscernibility of Identicals] to our own understanding of identity, we cannot fully understand Spinoza’s philosophy unless we understand how this [Spinoza’s apparent denial of each of the preceding] is possible’ (Garrett 2017: 13). Garrett’s argument is rich and complex, and I cannot do justice to its full scope here. But I do think that his solution is a satisfying one as regards the limited question of whether Spinoza’s assertion—that substance and attribute are in some sense the same—courts inconsistency. However, I simply deny that it helps Spinoza with the particular problem we are concerned with most immediately, namely, justifying EIp10.

Garrett’s basic strategy is to argue that attributes are not items within Spinoza’s ontology, as are substances and modes, but structures of it.⁵⁸ Claims about the ‘sameness’ of substance and attributes are thus genuine truth claims, for Spinoza, but they are not statements of numerical identity, since ‘substance’ names an item within the ontology while ‘attributes’ are structures of those items, substances and modes, within it. As Garrett puts the point:

The particular sameness expressed by G3 and G4—Substance-Attribute Identity—concerns a sameness of truths concerning substance and the manners of existence that structure the ontology [attributes], and does not constitute an identification of substances or affections within the ontology itself. If this is correct, then Spinoza can accept a version of the Transitivity of Identity understood as a principle governing the numerical identity of items—namely, substances and modes—within his ontology. (Garrett 2017: 40)

I think this is exactly right, and I think that Garrett is correct to look for a way of understanding Spinoza’s philosophy that does not problematically reify attributes as just another item in Spinoza’s ontology, alongside substances and modes. His way of understanding so-called “attribute neutral” properties, as disjunctive, derivative properties of attribute-specific ones, is similarly salutary (Garrett 2017: 30). Taken together, with these two moves Garrett helps us to avoid many problems that I think are ultimately specious, but might initially seem quite serious, in understanding Spinoza’s mature system.⁵⁹

But the value of Garrett’s interpretation should not mislead us about its starting points. Indeed, Garrett is quite clear that he is trying to show how we might see Spinoza as preserving, in some sense, fundamental properties of our contemporary understanding of the identity relation, if we presuppose the basic claims of Spinoza’s system—EIp10 itself, but also substance monism, mind-body identity theory, etc. Thus Garrett uses EIp10 and its consequences ubiquitously to generate the very puzzles concerning the Indiscernibility of Identicals and the Transitivity of Identity he then attempts, successfully in my view, to solve. So his interpretation does constitute an additional reason for finding this mature Spinozist system plausible. He does show that we need not see Spinoza as radically at odds with our contemporary understanding of the identity relation. But he does not give an independent argument for Spinoza’s basic conclusions.

This all makes sense, since it would seem rather surprising if Barrier followed directly from the definitions of substance and attribute, a terminology Spinoza ostensibly shares with Descartes, who nevertheless accepts Interactionism and denies Correlation. If his idiosyncratic restrictions on attributes and their modes (namely, Barrier) did result merely from his own particular definitions of attribute and substance, they would be woefully dialectically ineffective against interactionists like Descartes.⁶⁰ For as we have seen, Descartes accepts a strong conceptual independence of Thought from Extension, as Spinoza does. Descartes just denies Correlation, and thus Barrier, in part because he thinks attributing multiple, conceptually independent attributes to a single, simple subject is incoherent:

As for the attributes which so constitute the natures of things, it cannot be said that those which are different, and such that the concept of the one is not contained in the concept of the other, are present together in one and the same subject; for that would be equivalent to saying that one and the same subject has two different natures – a statement that implies a contradiction, at least when it is a question of a simple subject (as in the present case) rather than a composite one. (CSM I 298, emphases added)

Fundamentally then, it seems to me that Spinoza’s demonstration in EIp10d and EIp10s cannot justify Barrier, since his Cartesian interlocutors simply deny that their shared terminology of substances and attributes remains coherent if used as Spinoza intends.

However, recently Stetter (2019) has argued that the ostensible overlap in substance-attribute terminology between Descartes and Spinoza is merely apparent, and that their deeper definitional differences are central to evaluating their respective positions. Stetter develops Francois Lamy’s argument against Spinoza’s Ethics on Descartes’s behalf in his recent work, one that runs, if I understand it correctly, as follows.⁶¹ First, Lamy argues that Spinoza faces the problem of God’s unity (the Unity Objection), namely, that of explaining how many conceptually independent attributes can pertain to a single substance. Spinoza’s response, which Stetter deems successful, is to argue that the idea of God as infinitely expressive through infinite attributes is central to Spinoza’s very concepts of ‘substance’ and ‘attribute’. These concepts, therefore, are God-specific and must be understood with reference to this Spinozistic definition of God. Lamy then retorts that Spinoza’s definition of God simply begs the question against Descartes (the Bedrock Objection). But, Stetter thinks, Spinoza would simply respond first, that we do all do implicitly already have his concept of God, rather than Descartes’s, and second, that the ‘onus of proof should fall on the theologically-minded Cartesian to explain how God could be the most real being without possessing infinitely many attributes.’⁶² While Stetter seems broadly sympathetic to both lines of reasoning, he ultimately concludes that Spinoza can meet the Cartesian challenge in the way he describes.

I think that there are two related problems with Stetter’s characterization here. The first concerns Stetter’s attempt, on Spinoza’s behalf, to shift the burden of proof to Descartes, on the assumption that he wants to argue that God is the most real being ‘without possessing infinitely many attributes.’ Whereas before Stetter spoke, quite correctly, of the problem of God’s unity as the problem of how ‘several, conceptually independent attributes can pertain to one substance,’ here he shifts to talking of infinitely many attributes simpliciter, dropping reference to their conceptual independence.⁶³ This is a crucial qualification, however, since as many scholars of Descartes have noted, and as Stetter appears to recognize at many points throughout his article, the Cartesian God has infinitely many (albeit conceptually interrelated) attributes.⁶⁴

Relatedly, once we reinstate the conceptual independence requirement for attributes, then we can see that Stetter’s attempt to shift the burden of proof in just this way fails. For Descartes (CSM I 298) already concedes that if a subject is composite then it can have multiple conceptually independent essences. Hoffman, for example, relies on this point extensively in his discussion of the unity of the human being as a composite subject in Descartes’s thought. But this just returns us, in the terms of Stetter’s argument, to Lamy’s Unity Objection, which was that Spinoza’s God is a composite of infinitely many other substances, rather than a single one. Stetter’s argument has not advanced to bedrock so much as returned to the beginning.⁶⁵ So whatever the merits of this burden-of-proof shifting strategy are, I think that Spinoza’s position is significantly stronger when seen in the light of the Short Treatise argument I develop here, one that develops concrete problems for the Cartesian that arise on the assumption of Interactionism.

Spinoza’s interpreters, both friendly and critical, have not simply sought to justify Barrier by appeal simply to Spinoza’s demonstration of EIp10. So I want to consider two final recent attempts to adjudicate matters between Descartes and Spinoza. The first, due to Anat Schechtman, begins with the ostensible overlap in basic definitions between Descartes and Spinoza, though she ends up focusing on a particular disagreement over the classification of dependence relations that is especially relevant to our concerns here. The second, Karolina Hübner’s recent work on the mind-body relationship in Spinoza, tries to justify Barrier more directly.

Schechtman begins with what she calls the familiar story, according to which Spinoza borrows Cartesian fundamental concepts of ‘substance’ and ‘attribute’, pursues the implications of these concepts more rigorously than Descartes himself does, and ends up with his substance monism. Schechtman argues that this familiar story, though in some ways grounded in the texts, is not true. Specifically, while Descartes and Spinoza both accept that substance must be independent in some way, Spinoza requires, while Descartes denies, that this substantial independence requires causal independence. As Schechtman puts it:

The crucial disagreement between Descartes and Spinoza, it will turn out, concerns whether causation is what I call an ontological dependence relation—the kind of dependence relation, of which inherence is the paradigm, that a substance does not (indeed, cannot) bear to any other entity. (Schechtman 2018: 1)

Translating Schechtman’s terms to those of this paper, their key conflict concerns Correlation—that conceptual independence, which explicitly includes non-causal relations of ontological dependence, entails causal independence.

Schechtman’s basic idea is that denying Correlation, in particular by arguing that causation is not an ontological dependence relation and thus that causal independence is not entailed by ontological independence, allows Descartes to parry the charge that he inconsistently applies his own basic concepts, as in the familiar story. On this reading, minds and bodies, as Cartesian substances, are ontologically independent, both from each other and from God, while nonetheless causally dependent on God. This position is only inconsistent if Spinoza can successfully argue that the first (ontological) sort of independence entails the (causal) second sort. Rather than discuss all aspects of her sustained argument in detail, I want to accept much of what she says about Spinoza for the sake of argument and focus on her defense of Descartes.⁶⁶

Schechtman (2018: 12) argues persuasively in my view that both Descartes and Spinoza require that ‘modes, which ontologically depend on a substance, both inhere in and are conceived through a substance.’ However, Schechtman thinks their agreement ends at this point, since for Descartes, ‘from the perspective of intelligibility—with respect to understanding or conceiving what an entity is—causal relations are not always relevant’ (Schechtman 2018: 12). While Descartes preserves the link between causal and conceptual independence in some cases, most notably God’s, he does not require it in general. Schechtman’s preferred understanding of the Cartesian position is that a Cartesian substance ‘may (and generally does) stand in a causal relation to, and hence may be causally dependent on, other entities; it just cannot do so by its nature’ (Schechtman 2018: 14). As Schechtman reads Descartes, whereas God’s nature involves necessary existence, and hence demands consideration of God as self-caused, minds and bodies are substances that causally depend on God, albeit not in virtue of the natures of mind and body alone. Thus, this causal dependence does not threaten their substantiality.

In response to Schechtman’s defense of Descartes, I would like to make two subsidiary points, and one major one. First, distinguishing ontological (i.e., non-causal conceptual, in this paper’s terms) independence from causal independence, as Schechtman does, is a theologically risky route for Descartes to take. It relies on our simply presupposing that Descartes can show, on other grounds Schechtman leaves unspecified, that the two senses of independence coincide for God. But for all she says, Descartes’s ontological argument might be able to show at most that God is a necessarily (ontologically independent) existent deity that is nonetheless Himself causally dependent. This, of course, would be a theologically disastrous result. At several points she does appeal to Descartes’s notion of causa sui, but this notion, it is well-known, was controversial when used by anyone, including but not limited to, Descartes and Spinoza.⁶⁷ Indeed, much of the preceding Christian tradition explicitly denies that God requires any cause, or that is it at all coherent to speak of God as cause of Himself. Saving Descartes from one inconsistency by (potentially) introducing another might be dialectically effective against an interlocutor like Spinoza, but it is not an ideal strategy.

The second challenge for Schechtman’s interpretation, I think, is that there are good reasons for thinking that minds and bodies are, by their natures, causally dependent in general, and causally dependent on God in particular. To her credit, Schechtman considers two objections along these lines. I will focus on the question of whether minds and bodies depend by their nature on God specifically. In a long footnote, she considers the question of whether minds and bodies, as finite, depend by their natures on God as the infinite being. Schechtman (2018: fn. 53) offers three reasons for rejecting this possibility in the Cartesian case: first, finitude is a privation, and thus an accident rather than an essence; second, minds and bodies are not necessarily finite, since they might have been indefinite (albeit not infinite); finally, that even if minds and bodies are necessarily finite, this is arguably only one of their propria, rather than part of their essence. But I think this might be all somewhat beside the point, since arguably the Cartesian meditator does not argue indirectly, through the finitude of mind, to its necessary dependence on God. Rather, the meditator argues directly that being made by God, and hence being causally dependent, is at least a clear a part of his nature as thinking is. This comes out clearly in the Third Meditation, where Descartes writes:

And indeed it is no surprise that God, in creating me, should have placed this idea in me to be, as it were, the mark of the craftsman stamped on his work—not that the mark need be anything distinct from the work itself. But the mere fact that God created me is a very strong basis for believing that I am somehow made in his image and likeness, and that I perceive that likeness, which includes the idea of God, by the same faculty which enables me to perceive myself. That is, when I turn my mind’s eye upon myself, I understand that I am a thing which is incomplete and dependent upon another. … The whole force of the argument lies in this: I recognize that it would be impossible for me to exist with the kind of nature I have—that is, having within me the idea of God—were it not the case that God really existed. (CSM II 35, emphases added)

Descartes here not only strongly implies that the divine mark—the mind’s created nature—might be indistinguishable from the mind itself, and thus that the mind’s nature just is the mark of being made. He also uses almost precisely the same language when he claims that his nature is that he is a thinking thing as when he states that his nature is to be dependent on another, namely God.⁶⁸ So not only does Schechtman’s interpretation endanger the causal independence of God, per my first response, but it also seems to ignore Descartes’s own direct appeals to his nature as created, stamped by the craftsman and constantly dependent on him.

Finally, and most centrally, even if we ignore these first two issues, I actually think that Schechtman raises and concedes the central issue of this paper. That is, in at least two places she notes that one consequence of severing any necessary connection between ontological and causal independence is that the causal relations between minds and bodies need not be intelligible. The first case occurs when she observes in a long footnote: ‘A similar point is well-documented in the literature on specific types of Cartesian interaction (body-body, mind-to-body, and body-to-mind) each of which has been argued to hold without the relation of conception or intelligibility between the cause and effect’ (Schechtman 2018: fn. 40). The second occurs in her conclusion, when she notes that the basic disagreement between Descartes and Spinoza might concern what she calls a genetic view of natures, namely, that causal relations belong to the nature of their effects (Schechtman 2018: 19–20).

But if what I have argued in the first sections of this paper is correct, then Spinoza’s Short Treatise argument is precisely that the denial of this genetic view of natures, this too easy acquiescence to the absence of any intelligible relation between cause and effect, has untoward and unappreciated skeptical consequences. This is so both in general, and in the particular case of individual souls and bodies (‘why cannot a soul move wherever it will any body … why can Paul’s soul move his own body, but not that of another, such as Peter’). But if these skeptical consequences are false, and if Cartesian Interactionism really does have them as consequences, then we can see the disagreement between Descartes and Spinoza differently than Schechtman would have us see it. On her reading, their disagreement boils down to whether causation is an ontological dependence relation, or perhaps whether a genetic view of natures is correct, with each side occupying defensible, and indeed actually well defended, territory. Given, however, that I have just argued both that Descartes did think that minds and bodies were, by their nature, dependent on God, and given Descartes’s obvious concern to avoid any sort of skeptical conclusions, it seems that things are far less rosy for Descartes than Schechtman suggests. But it takes Spinoza’s unnoticed argument from the Short Treatise, the one I’ve tried to bring to light here, to help demonstrate why.

I turn finally to another recent argument along lines similar to those I develop here. In recent work, Hübner offers an extremely insightful treatment of the mind-body relation in Spinoza, one that emphasizes the importance of Barrier to Spinoza’s thought and tries to resolve outstanding puzzles while preserving it.⁶⁹ There, Hübner cites EVPref to argue that the unintelligibility of causal relations between minds and bodies, due to the lack of a common measure between them, motivates Barrier.⁷⁰

Now, Hübner is undoubtedly correct that Spinoza finds mind-body interactions unintelligible. This was precisely the point I just emphasized in arguing that Schechtman appears to raise and concede the central issue of this paper. However, as noted in section 1, Barrier requires justification that precedes Spinoza’s other commitments, since it is used as a premise in arguing for them. But this passage occurs far too late in the Ethics, in the Preface to Part V, and outside its overall propositional structure, to play that role.⁷¹ Moreover, since Descartes accepts the conceptual independence of the attributes, Spinoza once again requires an independent argument for Correlation to get Barrier.⁷² Otherwise, we cannot infer the absence of a common (causal) measure from the absence of a common (conceptual) one. Newlands (2018) does argue for Correlation on Spinoza’s behalf, as does Schechtman, but unfortunately what Newlands says is no use to Spinoza here, since it clearly appeals to Barrier at crucial points, and can thus not justify it without circularity.⁷³

Finally, as Deviance shows, Spinoza develops his skeptical problem only after stipulating the intelligibility of Interactionism, in [A], for the sake of argument. That is, Spinoza initially allows the prima facie intelligibility of the Cartesian position, subject to a careful examination of its consequences. Of course, having developed his skeptical objection and having concluded that rejecting Interactionism for Barrier contributes to its resolution, as he does in the Ethics, Spinoza is then free to label the Cartesian view unintelligible in EVPref. But he earns this view with the rule following paradox analogue in the Short Treatise; he cannot simply help himself to it. So the central disagreement between my interpretation and Hübner’s concerns what exactly the argument for the unintelligibility of Interactionism is, and whether it results from sustained meditation on the intractable skeptical problems raised in the Short Treatise, or whether by contrast it is posed late in the Ethics, outside its justificatory structure, and without explicit argument.

Interestingly, echoes of the Deviance passage from the Short Treatise do show up in the parts of Part V Preface that Hübner doesn’t cite. Descartes’s views on the connection between the soul and pineal gland come in for sustained mockery there, and Spinoza mixes in a number of different criticisms. Spinoza argues, for instance, that Descartes got the anatomy wrong: ‘this gland is not found to be so placed in the middle of the brain that it can be driven about so easily’ (EVPref, II/280, 17–19). He also repeats Hübner’s charge of unintelligibility in different ways. But in the midst of all that, some of his rhetorical questions take on new meaning in light of the Short Treatise. Most clearly, Spinoza says: ‘I should like very much like to know how many degrees of motion that Mind can give to that pineal gland. … For I do not know whether this gland is driven about more slowly by the Mind than by the animal spirits, or more quickly’ (EVPref, II/280, 4–7, emphasis added). How many degrees of motion indeed? Why can the Mind ‘not move all other bodies where it will,’ as Deviance asks? Spinoza’s focus on the extent of the mind’s control over the pineal gland, and its competition with animal spirits for its manipulation, seems to refer, however obliquely, to the sorts of considerations the Short Treatise develops and responds to at greater length.

5

In this paper, I have argued that Barrier is an ineliminable plank in the arguments for all Spinoza’s major conclusions in the Ethics. An unnoticed passage in the Short Treatise states an instance of a skeptical problem that parallels Leibnizian and Wittgensteinian rule following paradoxes in various respects, from which I constructed an argument for Barrier. I then argued that reflection on the issues raised by the Short Treatise both better grounds Barrier and more clearly explains Spinoza’s fundamental motivations for believing it than those considerations offered by other recent interpreters. I close with some outstanding questions. Can Spinoza really show that the rule following paradox is strictly worse for Interactionism, and that it does not arise for Spinoza’s own view?⁷⁴ The dialectic is tricky here, but I do think it is insufficient simply to claim that Interactionism, in the Cartesian sense, cannot be what gives rise to the skeptical paradox, simply because other views do so also. For it may be that even non-Cartesian or ostensibly non-interactionist views, for example property dualist views, have problems that mirror those Spinoza identifies with Interactionism in key respects. That would make these shared central features, whether present in full-blown substance dualism or in nearby views, responsible for generating skeptical problems. And what follows from the fact that Spinoza’s necessitarianism seems to cause problems for normativity in general, and thus for epistemic normativity in particular.⁷⁵ If Spinoza’s only answer to the justificatory questions about the truth of our causal claims concerning the limits of the mind’s activities is that what is the case is necessarily so, does that really provide sufficient justification? If not, how should that affect our interpretation of the argument for Barrier? These are difficult issues, to be sure. But I hope to have shown that Spinoza’s barrier between the attributes, as given in EIp10, can be defended surprisingly well using the resources of Spinoza’s texts, to an extent far greater than even many of his most sympathetic interpreters have supposed.

Competing Interests

The author has no competing interests to declare.