Cartesian Composites

Paul Hoffman. Journal of the History of Philosophy. Volume 37, Issue 2. April 1999.

Towards the end of a paper in which I argued that Descartes thinks a human being is a genuine unity, I invited other commentators to come to Descartes’s defense by accounting for his apparently contradictory claims that a human being is an ens per se and that it is an ens per accidens. These claims seem to be contradictory, because in saying that a human being is an ens per se, Descartes appears to be asserting that a human being is genuinely one, and in saying that a human being is an ens per accidens, he appears to assert that a human being is not genuinely one, but instead is a mere heap or aggregate. In the ensuing eleven years no one has taken up my invitation, except to argue that I was mistaken in claiming that Descartes thinks a human being has per se unity in any robust sense.

In this paper I will take up the challenge myself, having noticed a similarity between Descartes’s account of the unity of composite substances and his account of composite figures having true and immutable natures. After showing how he can consistently maintain that a human being is both an ens per se and ens per accidens in roughly the same way he can consistently maintain that composite figures both do and do not have true and immutable natures, I will try to respond to criticisms of my claim that he thinks of a human being as a substance or an ens per se.

Let me begin with an analysis and comparison of two passages from the Objections and Replies. In the first, from the Replies to the First Objections, Descartes discusses composite figures. In the second, from the Replies to the Fourth Objections, Descartes discusses composite substances. Next, to remove the other part of the difficulty, it should be noted that those ideas which do not contain true and immutable nature’s, but contain only ones which are fictitious and composed by the intellect, can be divided by the same intellect, not only through abstraction, but through a clear and distinct operation, so that any ideas which the intellect cannot so divide were undoubtedly not composed by it. As, for example, when I think of a winged horse, or an actually existing lion, or a triangle inscribed in a square, I easily understand that I can also in opposition think of a horse without wings, a non-existing lion, a triangle without a square, and so on; from which it follows that these do not have true and immutable natures. But if I think of a triangle or a square (I do not speak here of a lion or horse because their natures are not clearly evident to us), then certainly whatever I apprehend as contained in the idea of` a triangle, as that its three angles are equal to two right angles, etc., I will truly affirm of the triangle, and of the square whatever is found in the idea of the square. For even if I can understand a triangle, abstracting from the fact that its angles should equal two right angles, I cannot nevertheless deny that of it by a clear and distinct operation, that is, correctly understanding this is what I say. Furthermore, if I consider a triangle inscribed in a square, not in order to attribute to the square those things which pertain only to the triangle or to the triangle those things which pertain to the square, but in order to examine only those things which arise from the conjunction of the two, then its nature will be no less true and immutable than the square or triangle alone, and it will be permitted to affirm that the square is not less than twice the triangle inscribed in it and similar things which pertain to the nature of this composite figure. (AT VII 17-8; CSM II 83-4)

I am not unaware that some substances are commonly called `incomplete.’ But if they are said to be incomplete because they cannot exist per se alone, I confess that it seems contradictory to me that they should be substances, that is, things subsisting per se and at the same time incomplete, that is, unable to subsist per se. But in another sense they can be said to be incomplete substances, namely such that insofar as they are substances, they have nothing incomplete, but only insofar as they are referred to some other substance, with which they compose something one per se.

Thus a hand is an incomplete substance when it is referred to the whole body of which it is a part; but it is a complete substance when it is considered alone. And in just the same way mind and body are incomplete substances when they are referred to the man which they compose; but, considered alone, they are complete. (AT VII 222; CSM II 156-7)

In the first passage Descartes might readily seem to contradict himself.4 In the first half of the passage he explains his test for distinguishing ideas containing true and immutable natures from those that contain natures that are merely fictitious and composed by the intellect. If an idea can be divided by a clear and distinct operation of the intellect, it does not contain a true and immutable nature. If an idea cannot be divided by a clear and distinct operation, but merely by abstraction, it does contain a true and immutable nature. Descartes says that triangles pass this test, but triangles inscribed in squares do not. In the second half of the passage he turns around and asserts that the nature of a triangle inscribed in a square is no less true and immutable than that of the square alone or the triangle alone.

So in the same paragraph Descartes asserts both that the nature of a triangle inscribed in a square is true and immutable and that it is not. He avoids contradiction in the only way possible, which is to relativize the notion of having a true and immutable nature. He says that if we consider a triangle inscribed in a square in order to examine those things which arise from the conjunction of the two, and not to attribute to one of the figures what belongs to the other, then its nature will be true and immutable. This language suggests two candidates for that to which a composite’s having a true and immutable nature is being relativized, but for our purposes we need not worry about the difference between them. First, he might be relativizing its having a true and immutable nature to different purposes we have in thinking about the composite: it has a true and immutable nature relative to our examining it with the aim of determining which properties arise from the conjunction of its constituents and it does not have a true and immutable nature relative to our examining it with the aim of attributing properties to any of its constituents alone. Second, he might be relativizing a composite’s having a true and immutable nature to different sets of properties: it has a true and immutable nature relative to those properties that arise only from the conjunction of its constituents and it does not have a true and immutable nature relative to those properties that arise from any of the constituents alone.

That Descartes would appeal to such a relativization strikes me as a plausible response to an obvious objection one might have about his criterion in the Fifth Meditation for distinguishing things having true and immutable natures. He claims there that the mark of things having true and immutable natures is that we can demonstrate properties of them that we did not previously recognize and that once recognized we cannot deny (AT VII 64; CSM II 45). The obvious objection to this criterion is that the composition of things typically generates new properties that are not immediately recognizable. So, for example, we might not immediately recognize all the properties that arise from the composition of a horse and wings or even from the composition of a lion and existence. But this would imply, given the Fifth Meditation criterion, that a winged horse and an existent lion have true and immutable natures, contrary to what appears to be Descartes’s intention in the Fifth Meditation.

Rather than abandon his Fifth Meditation criterion for having a true and immutable nature in light of this conflict, Descartes’s strategy in the Replies to the First Objections is to relativize the notion of having a true and immutable nature. So Descartes can allow that a winged horse and an existent lion do have true and immutable natures relative to the consideration of new properties generated by the composition of a horse and wings and by the composition of a lion and existence. One might object that this strategy undermines his principal aim in introducing the notion of true and immutable natures, which is to ground his ontological argument. If we grant that an existent lion has a true and immutable nature, we should be able to infer with certainty that there is an existent lion in the same way we can infer that God exists.

I think Descartes can block this inference. According to the relativized account, an existent lion does not have a true and immutable nature with respect to the properties of being a lion and existence because those properties are separable. It has a true and immutable nature only with respect to any new properties F generated by their composition that we cannot deny of an existent lion once they are recognized. So we can infer with certainty that an existent lion is F once we recognize that we cannot separate the idea of F from the idea of existent lion, but that does not entitle us to infer that a lion exists.

It is my thesis that just as Descartes can assert without contradicting himself that a figure which is a composite of two figures each of which has a true and immutable nature both does and does not have a true and immutable nature, so he can assert without contradicting himself that a composite of two substances each of which is complete is both an ens per se (or has per se unity) and an ens per accidens. To do so will require him once again to relativize the relevant notions, but, as we will see, this relativization has a different structure.

One important piece of evidence in support of my claim that there are important similarities in Descartes’s accounts of composite substances and composite figures is that he uses almost the same separability test to determine which ideas contain natures that are not true and immutable that he uses to determine whether substances are really distinct. In the case of natures, if the idea of AB can be divided into the idea of A and the idea of B not merely by abstraction but by a clear and distinct operation, then AB does not have a true and immutable nature. In the case of substances, if the idea of A can be clearly and distinctly conceived apart from B, not merely by abstraction, but by exclusion, and vice versa, then A and B are really distinct substances (AT IV 120; CSMK 236). Descartes’s account of the distinction between abstraction and exclusion is as follows. To conceive of A without B by abstraction is to think of A without thinking of B; when we think of A and B together we see a connection between them (AT III 420-1; CSMK i88). To conceive of A without B by exclusion is to be able to deny A of B while thinking of both A and B (AT III 474-6; CSMK 201-2).

Let me turn now to an examination of the second passage quoted above, the one that concerns substances. Descartes’s agenda as he describes it in the first paragraph is to explain how there can be such a thing as an incomplete substance, given that such a notion appears to be contradictory. He grants that the notion of an incomplete substance is contradictory if what we mean by calling something “incomplete” is that it cannot exist per se alone, because a substance is that which can subsist per se. An incomplete substance would then be something that both can and cannot subsist per se. (To subsist per se, as understood by both Descartes and Aquinas, is to exist without a subject.) But if by calling something “incomplete” we mean that it is referred to another substance with which it constitutes something that is one per se, then there is no contradiction in calling something an incomplete substance.

In the second paragraph he relativizes the notion of incompleteness. He says that something can be a complete substance when considered alone and an incomplete substance when it is referred to a whole of which it is a part. The whole, according to the first paragraph, is one per se. So Descartes’s idea here is that something which can subsist per se, and hence is a complete substance when considered alone, can still be an incomplete substance if it is part of a whole that is one per se.

Descartes’s first example to illustrate this claim is a hand. He says a hand is a complete substance when considered alone and an incomplete substance when referred to the whole body of which it is a part. This implies, given his accounts of what it is to be a complete substance and what it is to be an incomplete substance, that he thinks that a hand can subsist per se and also that the human body is one per se.

His second example is the one of special concern to us. He says that the mind and body are incomplete substances when referred to the human being which they compose, but considered alone they are complete. These two paragraphs constitute the most important passage in the entire Cartesian corpus in support of my claim that Descartes thinks the human being is an ens per se. While I referred to them in my earlier paper, I did not give them appropriate emphasis.

There are three reasons why these two paragraphs deserve great attention. First, it is the only place in his published works where Descartes commits himself to the view that a human being is one per se. Having said in the first paragraph that a substance is incomplete when it is referred to something else with which it composes something one per se, his assertion in the second paragraph that the mind and body are incomplete when referred to the human being entails that a human being is one per se.

Let me explain this point more formally. I read Descartes as asserting in the first paragraph that any substances A and B composing C are incomplete when referred to C only if C is one per se. His subsequent assertion in the second paragraph that mind and body are incomplete when referred to the man which they compose thus entails that a man is one per se.

Second, these paragraphs were published just a few months prior to Descartes’s two letters to Regius in which he asserts that a human being is an ens per se and also an ens per accidens. Other commentators have tended to discount the letters to Regius on the grounds that Descartes was simply trying to avoid controversy with the authorities at Utrecht. But since Descartes had already committed himself in his reply to Arnauld to almost exactly the same view he urges upon Regius, his remarks to Regius must be treated with the same respect and seriousness that are typically granted to his reply to Arnauld.

The third significant feature of these paragraphs is the comparison of the mind to the hand. It needs to be emphasized that Descartes’s mind is a complete substance in the same sense that a hand is. To say that the mind can subsist per se should not be taken, as it almost always is, to mean anything stronger than what it means to say that a hand can subsist per se. The vitally important implication for understanding Descartes is that his notion of what it is to be a complete substance is very weak, much weaker than is commonly attributed to him. When Descartes asserts that the mind is a complete substance he is read as using the term “substance” in the same sense that Aristotle did when Aristotle said that a human being is a substance. But Aristotle did not think a hand is a substance in the same sense that a human being is, and this passage makes it clear that it is a serious mistake to think that Descartes thought a mind is a substance in that strong sense. An Aristotelian human being is not incomplete in relation to anything else, but a Cartesian mind is.

Let me turn now to Descartes’s two letters to Regius. In the first letter he admonishes Regius for asserting in his controversy with Voetius that a human being is an ens per accidens and he proposes a remedy:

In your theses you say that a human being is an ens per accidens. You could scarcely have said anything more objectionable and provocative. The best way I can see to remedy this is for you to say that in your ninth thesis you considered the whole human being in relation to the parts of which he is composed, and in your tenth thesis you considered the parts in relation to the whole. Say too that in your ninth you said that a human being comes into being per accidens out of body and soul in order to indicate that it can be said in a sense to be accidental for the body to be joined to the soul, and for the soul to be joined to the body, since the body can exist without the soul and the soul can exist without the body. For the term ‘accident’ means anything which can be present or absent without its possessor ceasing to exist-though perhaps some accidents, considered in themselves, may be substances, as clothing is an accident with respect to a human being. Tell them that in spite of this you did not say that a human being is an ens per accidens, and you showed sufficiently, in your tenth thesis, that you understood it to be an ens per se. For there you said that the body and the soul, by their very nature, are incomplete substances; and it follows from their being incomplete that what they constitute is an ens per se. That something which is an ens per se may yet come into being per accidens is shown by the fact that mice are generated or come into being, per accidens from dirt, and yet they are entia per se. (AT III 460; CSMK 200)

Descartes’s proposed remedy has two elements. The first element is to relativize. He advises Regius to say that in the ninth thesis when he referred to the human being as an ens per accidens he was considering the whole human being in relation to its parts and in the tenth thesis he was considering the parts in relation to the whole. The second element is to distinguish between being an ens per accidens and being made per accidens. He advises Regius to deny that the human being is an ens per accidens and to say instead that it is made per accidens. He claims that something that is made per accidens can still be an ens per se. In the second letter to Regius, Descartes says that he understands that Regius meant something innocuous when he said that a human being is an ens per accidens, but that he was using it in a different sense from that used by the Scholastics so that the dispute between Regius and the Scholastics is merely verbal:

when you said that a human being is an ens per accidens I know that you meant only what everyone else admits, that a human being is made up of two things which are really distinct. But the expression ens per accidens is not used in that sense by the scholastics. Therefore, if you cannot use the explanation which I suggested in a previous letter—and I see that in your latest paper you have departed from it in some degree, and not altogether avoided the hazards-then it is much better to admit openly that you misunderstood this scholastic expression than to try unsuccessfully to cover the matter up. You should say that fundamentally you agree with the others and that your disagreement with them was merely verbal. And whenever the occasion arises, in public and in private, you should give out that you believe that a human being is a true ens per se, and not an ens per accidens, and that the mind is united in a real and substantial manner to the body. You must say that they are united not by position or disposition as you assert in your last paper-for this too is open to objection and, in my opinion, quite untrue– but by a true mode of union, as everyone agrees, though nobody explains what this amounts to, and so you need not do so either. (AT III 492-3; CSMK 206) Nevertheless, in the rest of the letter Descartes drafts a response for Regius, where he makes it clear that he thinks it is appropriate to refer to a human being as an ens per accidens, provided we understand that expression in the appropriately relativized sense:

We affirm that human beings are made up of body and soul, not by the mere presence or proximity of one to the other, but by a true substantial union. (For this there is indeed a natural requirement, on the bodily side, of an appropriate positioning and arrangement of the various parts; but nevertheless the union is different from mere position and shape and the other purely corporeal modes, since it relates notjust to the body but also to the soul, which is incorporeal.) The idiom which we used is perhaps unusual, but we think it is sufficiently apt to express what we meant. When we said that a human being is an ens per accidens, we meant this only in relation to its parts, the soul and the body; we meant that for each of these parts it is in a manner accidental for it to be joined to the other, because each can subsist apart, and what can be present or absent without the subject ceasing to exist is called an accident. But if a human being is considered in himself as a whole, we say of course that he is a single ens per se, and not per accidens; because the union which joins a human body and soul to each other is not accidental to a human being, but essential, since a human being without it is not a human being. (AT III 508; CSMK 209)

The picture that emerges from these two letters is this. The relativization of the notions of being an ens per accidens in the letters to Regius is the complement of the earlier relativization of the notion of completeness in the reply to Arnauld. When the parts of a composite are complete (when considered) in themselves because they can subsist by themselves, the composite is an ens per accidens (when considered) in relation to those parts. In both the reply to Arnauld and the letters to Regius, it appears that the notion of the parts being complete in themselves is the fundamental notion and the notion of the whole being an ens per accidens is defined by means of it.

Similarly the relativization of the notion of being an ens per se in the letters to Regius is the complement of the relativization of the notion of incompleteness in the reply to Arnauld. If the composite whole is an ens per se (considered) in itself, then its parts are incomplete (considered) in relation to the whole. However, there are some subtle but revealing variations in the reply to Arnauld and the letters to Regius which have to do with the order of explanation. That is, there are different accounts of which notion is fundamental and which is derivative-that of the parts being incomplete or that of the whole being an ens per se. In the first letter to Regius, Descartes suggests that the fundamental notion is that of the parts being incomplete in their very nature, from which it follows that the whole which they compose is an ens per se. In the reply to Arnauld the explanation seems to go the other way: since what body and mind constitute is one per se, it follows that they are incomplete in relation to it. In the second letter to Regius it appears that the fundamental notion is that of the union of the parts being essential to the composite whole, from which it is inferred that the whole is an ens per se.

At stake here is what Descartes thinks justifies the claim that a human being is an ens per se. Is it just something Descartes thinks we should take for granted as I argued in the earlier paper and as is suggested by the reply to Arnauld? Do we start instead with the idea that the union of mind and body is essential to the whole, as is suggested by the second letter to Regius? Or do we start with the idea that mind and body do not have natures that are complete in themselves, as is suggested by the first letter to Regius?

I will not try to settle just yet the issue of whether Descartes thinks one of these notions is more fundamental than the others, though my view will emerge. However, it is worth noting now that if the notion of the parts being incomplete in their nature is fundamental and the notion of being an ens per se is defined in terms of it, then it follows that the incompleteness of the part is not defined by its relation to the whole. Instead Descartes is using two distinct notions of incompleteness that are determined solely by consideration of the part. In one sense of incompleteness, a part is incomplete if and only if it cannot subsist per se (considered) alone. In the other sense of incompleteness, a part is incomplete if and only if considered by itself alone its nature is not complete.

That Descartes would say that mind and body are incomplete in their nature, although they are complete in the sense that they can subsist by themselves, resonates with his comparison in the reply to Arnauld of the mind with the hand and reinforces the claim that he has a very weak notion of what it is to be a complete substance at work in the real distinction argument. A substance that is complete in the sense that it can subsist per se, such as a hand, need not be complete in its nature. And so it does not follow from the fact that mind has a nature that determines a complete substance, that is, something that can subsist per se, that the mind is complete in its nature. That the notion of completeness at stake in the real distinction argument is a very weak notion is brought out nicely in another passage later in the reply to Arnauld:

Now someone who says that a man’s arm is a substance that is really distinct from the rest of his body does not thereby deny that the arm belongs to the nature of the whole man. And saying that the arm belongs to the nature of the whole man does not give rise to the suspicion that it cannot subsist per se. (AT VII 228; CSM II 160)

Descartes’s point here is that a thing’s being able to subsist per se does not entail that it is not part of the nature of a whole. If something is part of the nature of a whole, and we take this to mean that it is a fact about the nature of the part that the part is part of the nature of the whole (as seems true of an arm or a hand), then it seems perfectly plausible to say that its nature is incomplete (even though it is complete in the sense of being able to subsist per se). This point will be discussed more fully below.

How does the relativization that enables Descartes to assert without contradiction that a composite substance is both an ens per se and an ens per accidens compare to the relativization that allows him to assert without contradiction that a composite figure both contains and does not contain a true and immutable nature? Let me mention two differences. First, the relativization pertaining to figures is in regard to different sets of properties (or to our consideration of different sets of properties), but the relativization pertaining to composite substances is not in regard to different sets of properties. It is interesting, however, that later, in the Principles Descartes does draw a distinction that would enable him to relativize the notion of being an ens per se in exactly the same way he relativizes the notion of containing a true and immutable nature. He distinguishes those characteristics that arise from the mind alone and from the body alone from those that arise from the union of mind and body, namely, appetites like hunger and thirst, passions that do not consist in thought alone, and sensations (AT VIIIA 23; CSM I 209). Thus he might have said that a human being is an ens per se in relation to those characteristics (or when considered for the purposes of examining those characteristics) that arise from the union of mind and body.

Second, Descartes does not introduce a notion applicable to the parts of a composite figure having a true and immutable nature that corresponds to the notion of incompleteness that pertains to the parts of a composite substance having per se unity. That is, once Descartes points out that a triangle inscribed in a square has a true and immutable nature, he does not also assert that there is a sense in which the natures of a triangle and a square are incomplete. But there seems to be no obstacle to such a notion. It seems perfectly plausible to say that the nature of a triangle is incomplete in relation to the nature of a triangle inscribed in a square.

What is the significance of these differences? My intuition is this. It seems to me that given Descartes’s purpose in introducing the notion of true and immutable natures-to provide an ontological grounding for the certainty of our inferences about which characteristics we can attribute to things-he can allow that every composite has a true and immutable nature. But given his purpose in introducing the notion of having per se unity-to provide an account of which composites are truly one and which are mere aggregates-he might not want to allow that every composite of things complete in the sense that it can subsist per se has per se unity. If he were to relativize the notion of an ens per se to (consideration of) the properties that arise from the union of the parts, then it would seem to me to follow that every composite of substances would be an ens per se. It is worth noting that Descartes’s account of the per se unity of a human being in the second letter to Regius provides some evidence that Descartes was willing to allow that every composite whose parts can subsist per se is an ens per se. It seems true of every composite whole that the union of the parts is essential to it. We will return to this issue below. We are now in a position to respond to objections to my claim that Descartes believes a human being has per se unity. Marleen Rozemond argues that even though Descartes says that a human being is an ens per se because mind and body are incomplete, he does not mean what his Scholastic predecessors meant in saying that soul and body are incomplete. She offers three related reasons in support of this assertion. First, in contrast to the Scholastics, Descartes thinks that mind and body are complete when considered on their own. Second, the Scholastics thought that soul and body are incomplete by their very nature, that it is part of their essence to belong to a composite, whereas Descartes does not. Third, unlike the Scholastics, Descartes does not think that the mind has an aptitude for union with the body in the sense that the Scholastics thought it did. Rozemond concludes from the fact that Descartes does not think that mind and body are incomplete in the appropriate Scholastic sense, he does not really think that a human being has per se unity in any robust sense.

One of the key passages she cites in support of her interpretation is the continuation of Descartes’s first letter to Regius: It may be objected that it is not accidental to the human body that it should be joined to the soul, but its very nature; since, when a body has all the dispositions required to receive the soul, without which it is not a human body, it cannot be without a miracle that the soul is not united to it. And also that it is not accidental to the soul to be joined to the body. All these things should not be denied in every respect, so that the theologians are not again offended. But it must be responded nevertheless that those things can be said to be accidental because, when we consider the body alone we simply perceive nothing in it because of which it desires to be united to the soul; just as we perceive nothing in the soul because of which it must be united with the body; and for this reason I said a little earlier that. it is in some sense accidental, but not absolutely accidental. (AT III 460- i; CSMK 200) She places special emphasis on the fact that Descartes says, “When we consider the body alone we simply perceive nothing in it because of which it desires to be united to the soul; just as we perceive nothing in the soul because of which it must be united with the body.”

The first point I want to make in response is that it is false that all the leading Scholastics denied that the soul is complete when considered on its own. Aquinas himself asserts in his Questions on the Soul that the soul subsists per se because it has a per se operation, and moreover, in response to the objection that the soul is united to the body accidentally, he concedes that the soul has complete being. Aquinas also says that the soul separated from the body has perfect being. The notions of being complete and perfect were apparently used interchangeably by Aquinas, as they were by Regius. Thus Descartes is really endorsing fundamentally Thomistic views in claiming that the mind is a complete thing in the sense of being capable of subsisting per se. Descartes, like Aquinas, is what we might call a subsistence dualist.

The second point is that while Rozemond is right to emphasize Descartes’s assertion in the reply to Arnauld that it is not properly of the essence of mind that it be united to the body (AT VII 219; CSM II 155), he might not mean by this anything inconsistent with his endorsement of Regius’s assertion that the mind is an incomplete substance by its very nature. Let me elaborate by building on some helpful terminology that Rozemond has used in correspondence. We can distinguish between i) the nature or essence of A demanding union with B and ii) the nature or essence of A calling for union with B. If the nature of A demands union with B, then it is impossible for A to exist without being united to B. If the nature of A only calls for union with B but does not demand it, then it is natural or proper for A to be united to B but A can still exist without being united to B. Rozemond maintains that in asserting that it is not properly of the essence of the mind that it be united to the body Descartes’s point is that the nature of the mind does not call for union with the body, which she thinks is inconsistent with the mind being incomplete in its very nature on account of its relation to the body.

By drawing a further distinction between a weaker and a stronger sense of calling for union I think Descartes can be defended against the charge of inconsistency or insincerity in asserting that it is not properly of the essence of the mind that it be united to the body while endorsing Regius’s view that the mind is incomplete by its very nature. The distinction between the strong and weak senses is as follows. The nature of A calls for union with B in the strong sense if A cannot be defined without reference to B. The nature of A calls for union with B in the weak sense if A can be defined without B. I cannot think of any non-controversial examples to illustrate this distinction, so let me instead rely on different hypothetical views someone might take who thinks that a hand does not demand union with the whole body. One might maintain that even though a hand does not demand union with the whole body, it still cannot be defined independently of a human body. Alternatively, one might maintain that even though it is natural and proper for a hand to be united to a human body, not only does it not demand union with a human body but it can be defined independently of the human body.

When Descartes asserts that it is not properly of the essence of the mind that it be united to the body, he is surely implying at the very least that the mind does not demand union with the body. It also seems likely that he means to imply that the mind can be defined independently of the human body. So it seems right that Descartes thinks the mind does not call for union with the body in the strong sense. But there is still room for Descartes to maintain that the mind does call for union with the body in the weak sense, in other words, he can still maintain that it is natural or proper for the mind to be united to the body. And if that is what Descartes means when he endorses Regius’s view that the mind is incomplete by its very nature, then his remarks are not inconsistent. (In the passage Rozemond quotes from the first letter to Regius, Descartes grants for the sake of avoiding controversy that the soul by its very nature is united to the body and that it can exist apart from the body only by a miracle. This suggests a possible further distinction between two sorts of demand. The nature of A might demand, union with B if it is impossible for A to exist without being united to B or it might demand, union with B if it can exist without being united to B only by a miracle. Descartes apparently is prepared to concede for the sake of avoiding controversy that the mind demands, union with the body, even though he himself thinks that the mind does not demand, union with the body. To say that the mind demands2 union with the body amounts to saying that it calls for union with the body.)

Here again I think Descartes’s views are similar to those of Aquinas. Since Aquinas also thinks that the fact that the soul can exist apart from the body gives it complete being, he can’t think the soul demands, union with the body. Nevertheless, like Descartes he still thinks that the soul remains incomplete in another sense. It is not an individual in the genus of substance because it does not have through itself the complete nature of a species. To be sure, Descartes does not talk of complete species in the way that Aquinas does. However, Aquinas, like Descartes, also compares the soul to the hand to illustrate the sense in which the soul is incomplete. Aquinas maintains that the hand is a substance because it subsists per se in the weak sense of not inhering in anything in the way that an accident or material form does. But neither the soul nor the hand subsists per se in the strong sense because they are parts and do not have a complete species through themselves. So the situation is that Aquinas and Descartes both compare the soul to the hand to illustrate how the soul subsists per se in one sense of the term, but Aquinas provides a theoretical structure missing in Descartes-the notion of having a complete species-to explain the sense in which the soul and the hand are incomplete. That is, the reason why the nature of the soul calls for union with the body is that the soul does not have through itself the complete nature of some species. What should we make of this difference?

One response would be to infer that Descartes must be making implicit use of the notion of complete species to justify his agreement with Aquinas. That is, Descartes must also think that a human being has a complete species in a way that the mind does not to back up his claim that the mind is incomplete in its nature. This would imply that Descartes thinks a human being is a more robust substance than a mind, because it is complete in a way that a mind is not. It would also have the further implication that when he says that a human being is an ens per se he would mean something stronger than when he says that a mind is an ens per se, for a mind is an ens per se only in the weak sense that it can exist apart from a subject.

Even on this first response there would remain some important differences between the views of Descartes and Aquinas. In asserting that the soul does not have the complete nature of a species Aquinas seems to be asserting that the soul calls for union with the body in the strong sense, that the soul cannot be defined independently of the body. But Descartes at most would seem to be committed to the view that the soul calls for union with the body in the weak sense. In other words, Descartes’s view would be that a human being has a complete species in the way a mind fails to because it is natural and proper for a mind to be united to something else, whereas it is not natural and proper for a human being to be united to something else.

A second response would be the opposite-to deny that Descartes is making use of a notion of complete species. Certainly there is reason to think that Descartes’s physics involves the rejection of the notion of complete species as it pertains to bodies. One can easily read him as supposing that bodies can be compounded without limit into larger and larger bodies that still count as individual substances (AT VIIIA 53-4; CSM I 233). So one might try to argue that Descartes thinks that something can be incomplete in relation to a whole that is an ens per se, even if that whole does not have a nature that is complete in its species. This would imply that the whole in relation to which the parts are incomplete need not be more robust in its being than the parts themselves. And in this respect there would be a significant point of dissimilarity between Descartes and Aquinas—for Aquinas does think that a human being is more robust in its being than a soul precisely because a human being does have a nature that is complete in its species and a soul does not. But so long as Descartes is read as holding that a human being is no less a substance than a soul, then I would argue that their views are more similar than dissimilar.

If this is Descartes’s view, then it would seem that he cannot be using the notion of the parts being incomplete in their nature as fundamental. Nothing in his physics suggests that any body calls for union with any other body. Instead he must be starting the other way round. That is, instead of defining the completeness of the whole in terms of the incompleteness of the parts, he would have to be defining the incompleteness of the parts in relation to the whole. So we identify certain bodies as complete, for example, planets, and infer from this that their parts are incomplete in relation to them. Furthermore, this second response suggests that Descartes does not use the notion of an ens per se to mean anything stronger than the weak notion of being able to exist apart from a subject. Thus, any aggregate of substances would be an ens per se.

Both of these responses have merit even though they pull in opposite directions. They reflect what I consider to be a deep tension in Descartes’s philosophy that arises, on one side, from his attempt to preserve the commonsense Aristotelian conception of a human being, and, on the other side, from the implications of his new account of body. If we emphasize the first line of thought, then a human being is a more robust substance than a mind because it is not incomplete with respect to anything else, and not every aggregate of substances is a composite ens pe se. If we emphasize the second line of thought, then a human being, while still a substance, need be no more robust a substance than a mind, and all aggregates of substances would be entia per se.

The third point in response to Rozemond is that she is not justified in discounting the passage (that had not before caught my attention) in which Descartes in defending Regius asserts that he did not deny that the mind and body have a natural aptitude to substantial union:

They [the proponents of the thesis that the union of mind and body arises per accidens] did not deny the substantial union by which mind and body are conjoined nor a natural aptitude of each part to that union, as was clear from the fact that immediately afterwards they added: “those substances are called incomplete by reason of the composite which arises from their union.” (AT VII 585)

She discounts this passage on the ground that since Descartes supports his view that they did not deny the natural aptitude for union by pointing to the claim that mind and body are incomplete, whether he really agrees with the Scholastics depends on whether they have the same account of incompleteness. She maintains that they do not, because she argues that Descartes thinks mind and body are incomplete only in the sense that both are needed to constitute a human being. She bases her interpretation of Descartes’s account of incompleteness on the second letter to Regius, where, as we have already seen, Descartes asserts that a man considered in himself has per se unity because the union of mind and body is essential to the man.

I agree that the account of per se unity in the second letter to Regius does not provide evidence in favor of the more robust account of incompleteness I have attributed to Descartes. It does not follow from the claim that the union of mind and body is essential to the man that mind and body are incomplete in their nature. However, this passage does not conflict with the passage from the first letter to Regius, which Rozemond does not take into account, in which Descartes does assert that mind and body by their very nature are incomplete substances. If Descartes believes mind and body are incomplete substances by their nature, he might well believe that they have a natural aptitude to be united, even if their natures do not demand, that they be united. Again this sounds very much like Aquinas’s account of the soul.

Finally, the fact that Descartes does not think the body has a desire to be united to the soul should not be construed as evidence that he thinks it has no aptitude to be united to the soul. Desires are one thing, natural aptitudes another.

Vere Chappell argues that a human being cannot be an ens per se, because Descartes says that mind and body are one and the same thing only by a unity of composition and not by a unity of nature.

Notice that if we have different ideas of two things, there are two ways in which they can be taken to be one and the same thing: either in virtue of the unity, or identity of their nature, or else merely in respect of unity of composition. For example, the ideas which we have of shape and of motion are not the same, nor are our ideas of understanding and volition, nor are those of bones and flesh, nor are those of thought and of an extended thing. But nevertheless we clearly perceive that the same substance which is such that it is capable of taking on a shape is also such that it is capable of being moved, and hence that which has shape and that which is mobile are one and the same in virtue of a unity of nature. Similarly, the thing that understands and the thing that wills are one and the same in virtue of a unity of nature. But our perception is different in the case of the thing that we consider under the form of bone and that which we consider under the form of flesh; and hence we cannot take them as one and the same thing in virtue of a unity of nature but can regard them as the same only in respect of unity of composition-i.e. in so far as it is one and the same animal which has bones and flesh. But now the question is whether we perceive that a thinking thing and an extended thing are one and the same by a unity of nature. That is to say, do we find between thought and extension the same kind of affinity or connection that we find between shape and motion, or understanding and volition? Alternatively, when they are said to be `one and the same’ is this not rather in respect of unity of composition, in so far as they are found in the same man, just as bones and flesh are found in the same animal? The latter view is the one I maintain, since I observe a distinction or difference in every respect between the nature of an extended thing and that of a thinking thing, which is no less than that to be found between bones and flesh. (AT VII 423-4; CSM II 285-6)

My response to this objection is that the conclusion follows only on a controversial understanding of what Descartes means by a unity of nature. If in denying that mind and body are one by a unity of nature, Descartes were denying that the natures of mind and body can be united or joined, then the conclusion would follow. It would follow because the condition for the per se unity of a composite C constituted by A and B is that A and B are incomplete in their very nature, which suggests that the natures of A and B can be united or joined. But I don’t think this is what Descartes means in denying that mind and body are one by a unity of nature. I think his point is instead that mind and body do not have the same nature. And there is no contradiction in asserting both that A and B do not have the same nature and that A and B are incomplete in their natures. On this understanding Descartes can consistently assert both that mind and body are not one and the same by a unity of nature and that a human being is an ens per se.

Once again Descartes’s examples are helpful. The examples of things that have a unity of nature are all modes of the same attribute. This suggests that to have a unity of nature is to have the same nature. His example of a unity of composition is an animal composed of flesh and bones. By inviting us to suppose that an animal is a unity of composition, Descartes is thereby inviting us to suppose that flesh and bones do not have the same nature. Of course, this is not something he believes, since he thinks all extended things have the same nature. But since flesh and bones are incomplete in the same sense that a hand or arm is incomplete, the example serves to illustrate the point that things we suppose to have different natures can still be incomplete and hence can constitute something that is an ens per se.

Chappell also objects that a human being cannot be an individual substance because Descartes is committed to the principle that every finite substance has exactly one principal attribute or essence. This principle also plays a fundamental role in Rozemond’s interpretation of Descartes, for she maintains that it is the crucial premise in Descartes’s argument for the real distinction between mind and body. Once we see that thought is a principal attribute and that extension is a principal attribute, it follows that mind and body are distinct substances.

Descartes does appear to endorse such a principle in a prominent passage in the Principles, Part I:

53. To each substance there belongs one principal attribute; in the case of mind, this is thought, and in the case of body it is extension.

A substance may indeed be known through any attribute at all; but each substance has one principal property which constitutes its nature and essence, and to which all its other properties are referred. Thus extension in length, breadth and depth constitutes the nature of corporeal substance; and thought constitutes the nature of thinking substance. (AT VIIIA 25; CSM I 210)

Nevertheless, it is my conviction that it is one of the most serious mistakes of contemporary Cartesian scholarship that this passage is taken to endorse the principle that no substance can have more than one principal attribute. That interpretation is called into question by a fuller explanation of attributes in the Comments on a Certain Broadsheet:

As for the attributes which 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. (AT VIIIB 349-50; CSM I 298)

Here Descartes makes it clear that he believes even a simple subject can have two or more attributes of the sort that constitute the natures of things. This can happen when the concept of one is contained in the concept of the other. So it can’t be a basic metaphysical principle that no simple subject can have more than one attribute of the sort that constitute the natures of things. The true metaphysical principle is that no simple subject can have more than one attribute of the sort that constitute the natures of things when their concepts are independent. Moreover, this passage makes it clear that Descartes thinks it is true only of simple subjects that they can have only one principal attribute conceived independently of other attributes. Composite subjects can have more than one such principal attribute, and I see no good reason for thinking that such a composite subject is not a substance, provided that it has per se unity of the sort analyzed in this paper.

In claiming that the Principles passage is superseded by the Comments passage, I am not claiming that the two passages cannot be reconciled. I think that the Principles passage reflects Descartes’s view that as a matter of fact the simple substances created by God have one principal attribute. The mistake is to read it as implying that it is some sort of metaphysical truth that there can be no simple substance that has more than one principal attribute. Indeed I am tempted to ascribe to Descartes the view that God is a simple substance with more than one principal attribute, none of which can be conceived independently of the others.

In conclusion, I have tried to explain and defend Descartes’s strategy of relativization to justify seemingly contradictory assertions: that a triangle inscribed in a square both does and does not contain a true and immutable nature, that the mind is both complete and incomplete, and that a human being is both an ens per accidens and an ens per se. The fact that Descartes applies this strategy in all three of these cases requires us to take it seriously. The most significant result is that Descartes’s account of the sense in which soul is complete and the sense in which it is incomplete is very similar to that of Aquinas. A soul is complete in the same sense that a hand is complete and it is incomplete in the same sense that a hand is incomplete. And the solution to my worry of several years ago is that in saying that a human being is an ens per accidens, Descartes is not denying that it is an ens per se, but rather asserting the familiar Scholastic view that it can have constituents that can subsist per se.