The Meditations is a fertile source of self-conceptions. Over the course of this work, the meditator7 considers herself: (i) as a ‘human being [homo]’ (M1, AT VII 19/CSM II 13; see also M2, AT VII 25–26/CSM II 17), (ii) as a disembodied being without ‘hands or eyes, or flesh, or blood or senses’ (M1, AT VII 23/CSM II 15), (iii) as having a body with ‘a face, hands, arms, and the whole mechanical structure of limbs which can be seen in a corpse’ (M2, AT VII 26/CSM II 17), (iv) as ‘a thinking thing’ (M2, AT VII 27/CSM II 18), and, more specifically, (v) as ‘a thing that doubts, understands, affirms, denies, is willing, is unwilling, and also imagines and has sensory perceptions’ (M2, AT VII 28/CSM II 19), (vi) as ‘a thing that thinks and is not extended’ (M3, AT VII 44/CSM II 30; and M6, AT VII 78/CSM II 54), (vii) as ‘something intermediate between God and nothingness, or between supreme being and non-being’ (M4, AT VII 54/CSM II 38), (viii) as ‘not merely present in my body as a sailor is present in a ship,’ but, rather, ‘very closely joined, and, as it were, intermingled with the body’ (M6, AT VII 81/CSM II 56), and, finally, (ix) as ‘composed of body and mind’ (M6, AT VII 81/CSM II 56). As Schechtman (2014: 499) observes, some of these conceptions are artifacts of the meditator’s progress through the Meditations and do not reflect Descartes’s considered position.8 The meditator’s conception of the self as lacking ‘hands or eyes, or flesh, or blood or senses’ is at least incomplete, whereas the conception of the self as a mere body is incorrect. Nevertheless, I will argue that, for Descartes, the self of the Meditations may be correctly considered as either a mind or a human being, and that the self has different properties relative to these ways of considering it.

Let’s now work through the Meditations more slowly, following the evolution in the meditator’s understanding of herself. In Meditation 1, the meditator initially conceives of herself as a human being. This conception is inchoate at first, but is centered on the belief that she has a human body of a certain kind. This conception gradually comes into focus as a target of the meditator’s various skeptical arguments. These arguments are not just about raising doubts about the world external to the meditator. They are as much about dislodging the meditator’s previous conception of herself, which is resilient in the face of all but the most radical skeptical scenarios.

After observing that her senses sometimes deceive her, the meditator suggests that the mere possibility of sensory error does not yet provide her with any reason to doubt that ‘these hands or this whole body are mine,’ from which we may infer that her previous self-conception involved having a body (M1, AT VII 19/CSM II 13). The meditator wonders whether the madness hypothesis might dislodge this belief, as madmen are often confused about what they are, for example, when they maintain ‘that their heads are made of earthenware, or that they are pumpkins, or made of glass’ (M1, AT VII 19/CSM II 13). Although the meditator sets this hypothesis aside, her description of it suggests that she previously considered herself as having a human body with a specific constitution, made of flesh and blood, rather than earthenware, pumpkin, or glass.

The dream argument is more successful at raising doubts about the meditator’s previous belief that she had ‘such hands or such a body,’ but it does not raise any genuine doubt about the more generic belief that the meditator has some kind of body or other (M1, AT VII 19/CSM II 13). Someone might dream that she is seven feet tall, thereby raising doubts about the dimensions of her body, without raising any doubts about the more fundamental claim that she is embodied. Indeed, the meditator introduces dreaming as something that is characteristically done by human beings, and so cannot be used to doubt the meditator’s belief that she is a human being, and, hence, has a body:

Because the dream argument presupposes that the meditator is a human being lying in bed, it cannot undermine the meditator’s conception of herself as a human being. With the evil deceiver hypothesis, finally, the meditator lands on a way of calling this conception into doubt: by raising the more radical possibility that there are no bodies at all, and, hence, no bodies to be had. In the culminating paragraph of Meditation 1, the meditator writes:

We might hear an implicit ‘therefore’ between these two sentences, in which case the meditator would be doubting that she has a body—‘hands or eyes, or flesh, or blood, or senses’—in virtue of doubting the more general belief that there is an external world.

Thus, in Meditation 1, the meditator initially considers herself as having a body, and as a human being. Indeed, she regards these ways of considering herself as more or less interchangeable. But she is vague about the relevant sense of ‘having’ and what exactly a human being is. We cannot assume that Descartes regards this conception of the self as correct, since it is part and parcel of the pre-theoretical, sense-based picture of the world that the Meditations aims to revise. But we also cannot assume that this conception is wholly incorrect either, given that the meditator rehabilitates many sense-based beliefs in Meditation 6.

In Meditation 2, the meditator briefly reviews the skeptical results of the previous Meditation, and in the process gives more clues about how she formerly considered herself. The meditator is now supposing that she has ‘no senses and no body,’ from which we may infer that she formerly considered herself as having them. The meditator wonders whether she ‘is so bound up with a body and with senses that I cannot exist without them,’ which, again, suggests that she formerly considered herself as essentially having a body and senses (M2, AT VII 25/CSM II 16). All these beliefs are doubtful, however, and she sets them aside.

Once she strips away her old ways of considering herself, the meditator discovers that, nevertheless, there is one thing she cannot doubt: that she exists.9 ‘After considering everything very thoroughly, I must finally conclude,’ explains the meditator, ‘that this proposition, I am, I exist, is necessarily true whenever it is put forward by me or conceived in my mind’ (M2, AT VII 25/CSM II 17). What should the meditator make of this discovery? The meditator is certain that she exists, but not what she is. ‘I do not yet have a sufficient understanding of what this “I” is,’ says the meditator, ‘that now necessarily [i.e., certainly] exists’ (M2, AT VII 25/CSM II 17). To answer this question, the meditator’s strategy is to revisit what she formerly believed herself to be and to subtract any features that are so much as conceptually separable from the ‘I’ whose existence she has just discovered, in order to build a new and improved conception of herself.

After the brief flash of clarity that occurs during the cogito reasoning, the meditator dives back into her old conception of herself as a human being:

Let me highlight three points. First, the meditator’s former self-conception as a human being was not especially theoretical: it was not something she needed to be taught in school.10 Instead, her former self-conception arose ‘spontaneously and quite naturally,’ which suggests a sensory origin (M2, AT VII 25–26/CSM II 17; see also M6, AT VII 74/CSM 51–52).11 Second, the meditator’s previous self-conception included body and soul. But this conception emphasized the body more than the soul, and was clearer about the body’s nature. Third, the meditator’s conception of herself as a human being included various actions that she took herself to perform, namely, nourishment, self-motion, sense perception, and thinking, which she ‘attributed to,’ or, took to be explained by, the soul (M2, AT VII 25–26/CSM II 17).

Next, the meditator considers which elements of her previous self-conception, if any, are conceptually inseparable from herself. If an element is conceptually separable—that is, if she can conceive of herself as lacking a property—she sets it aside. If an element is conceptually inseparable—that is, if she cannot conceive of herself without it—then she builds it into the new conception of herself that she is in the process of constructing. Nutrition, movement, and sense perception—the meditator sets these aside. The only property that the meditator cannot separate from herself is thinking, which leads to the conclusion that she is a thinking thing: ‘Thinking? At last I have discovered it—thought; this alone is inseparable from me…. I am, then, in the strict sense only [praecise tantum] a thinking thing; that is, I am a mind, or intelligence, or intellect, or reason’ (M2, AT VII 27/CSM II 18). The meditator subsequently fills in this new conception of the self by clarifying what thinking is. ‘But what then am I? A thing that thinks. What is that? A thing,’ the meditator explains, ‘that doubts, understands, affirms, denies, is willing, is unwilling, and also imagines and has sensory perceptions’ (M2, AT VII 28/CSM II 19).

Thus, in Meditation 2, the meditator both spells out her previous conception of herself as a human being, and then executes a transition to the new self-conception as a thinking thing or mind. This new conception is clearly supposed to be correct, even if the meditator cannot yet say whether it is complete. Despite some of the rhetoric in Meditation 2, the meditator has not shown that her old conception of herself as a human being is false or incorrect. She has shown that it is dubitable, and so she has set it aside for the moment. The point of the qualification ‘in the strict sense only’ is to flag that the only thing she knows about herself at this point is that she is a thinking thing (M2, AT VII 27/CSM II 18).12 It does not follow that the meditator is only a thinking thing. A final decision on her previous self-conception must wait until Meditation 6.

In Meditations 3–5, the meditator widens her focus to consider not merely her own nature, but also her relation to God and the essence of material things. Throughout this process, the meditator elaborates her conception of herself as ‘a thing that thinks and is not extended’ (M3, AT VII 30/CSM II 44). In the causal argument for God’s existence in Meditation 3, for example, the meditator argues that when she is considered as a mind or thinking thing, she was not created by her parents:

The meditator goes out of her way to qualify this claim about her causal history: she is just talking about herself considered as a mind.

In Meditation 4, the meditator examines the implications of the fact that God is her creator. The meditator holds that she is ‘something intermediate between God and nothingness, or between supreme being and non-being,’ and that her mixed nature corresponds to two different ways of considering herself:

These conceptions of the self—as created by God, and as participating in nothingness—do not obviously map onto the meditator’s conceptions of herself as a human being and a mind. Nevertheless, this passage illustrates that two very different ways of considering oneself might be simultaneously correct, and both do explanatory work. In so far as God created her, the meditator’s cognitive faculties tend towards the true and the good; in so far as she participates in nothingness, the meditator is fallible. Hence, we should remain open to the possibility that the conceptions of the self as a human being and mind might be simultaneously correct as well.

Finally, in Meditation 6, the meditator examines whether her previous conception of herself as a human being is correct in light of her new and improved conception of herself as a mind. This occurs as part of the meditator’s more general reevaluation of her sense-based beliefs, which she introduces as follows:

Since the meditator’s previous conception of herself as a human being was one of her most important, and arguably among the most resilient, sense-based beliefs, she unsurprisingly reevaluates this conception.14

The meditator starts by going back over her old conception of herself as a human being: ‘First of all then, I perceived by my senses that I had a head, hands, feet, and other limbs making up the body which I regarded as part of and perhaps even as the whole of myself’ (M6, AT VII 74/CSM II 51–52). In this passage, the meditator highlights the sensory origin of her previous self-conception. She considered herself to be human because she had a sensory experience of herself as having a body, and, more specifically, because she experienced herself as inseparable from this body, and felt all her sensations ‘in’ and ‘on account of’ this body (M6, AT VII 76/CSM II 53).15 Someone might object that the meditator isn’t referring to her previous conception of herself as a human being (homo) in these passages, but merely her previous belief that she had a body. But, as we saw above, the meditator treated these conceptions more or less interchangeably in Meditations 1 and 2. Second, the meditator rehearses her reasons for doubting her previous conception of herself as a human being, mentioning cases of phantom limb pain to illustrate the fallibility of its basis in the internal senses or bodily awareness (M6, AT VII 76–77/CSM II 53). Again, the only skeptical argument that strikes at the heart of the meditator’s previous self-conception is the deceiver hypothesis.

Third, the meditator considers what she should now believe about her previous conception of herself as a human being. During the meditator’s review of her sense-based beliefs she begins (‘First of all then …’) with the conception of herself as human, and the meditator’s reevaluation begins in the same place, with the real distinction argument. On the basis of her clear and distinct idea of herself as ‘a thinking, non-extended thing’ and her clear and distinct idea of body ‘as a non-thinking, extended thing,’ the meditator argues that ‘it is certain that I am really distinct from my body and can exist without it’ (M6, AT VII 78/CSM II 54). So it looks like the meditator is telling us what she should now believe about herself: that she is a thinking, non-extended thing, really distinct from the body, and, a fortiori, not a human being.

If the Meditations ended at this point, that would be a reasonable conclusion. We might justifiably conclude that the meditator’s new and improved conception of the self as a mind is correct, her old conception of herself as a human being is wrong, and that the argument I am trying to make—that both conceptions are correct—is doomed to failure. But the Meditations doesn’t end there. In the second half of Meditation 6—which is sometimes referred to as Meditation 6.5—the meditator rehabilitates many of her former sensory beliefs, including her previous conception of herself as a human being. Thus, the meditator’s final position—and, hence, presumably Descartes’s—is that the self is correctly considered as a thinking thing or mind, but also as a human being.

The meditator argues that because God created her sensory faculties, her sense-based beliefs must contain some truth.16 Her previous conception of herself as a human being arose ‘spontaneously and naturally’ from her senses.17 So the divine guarantee applies. In the famous ‘teachings of nature’ passage, the meditator explains what can be salvaged from her previous self-conception: namely, that (i) she has or possesses a body, (ii) her relationship to this body is not like a pilot in a ship, but is closer and more intimate, and (iii) her body is surrounded by other bodies, such ‘that some of these are to be sought out and others avoided’ (M6, AT VII 81/CSM II 56). To this point, everything the meditator has said is (arguably) compatible with her new identity as a mind or thinking thing. But then the meditator throws a curve ball. She argues that agreeable and disagreeable sensations also ‘make it quite certain that my body, or rather my whole self, in so far as I am composed of body and mind [meum corpus, sive potius me totum, quatenus ex corpore & mente sum compositus], can be affected by the various beneficial or harmful bodies which surround it’ (M6, AT VII 81/CSM II 56, emphasis added).18 This is a striking claim: the meditator is saying that there are some things that are true about her—namely, that she can be affected by bodies in her vicinity—in so far as she is composed (compositus) of body and mind.

The language of composition (‘compositus’) shows that the meditator is not merely adding to the conception of the self as a mind, but is considering herself in a radically different way, namely, as a human being. As the meditator explains a few pages later, when she considers herself as a mind, that is, ‘in so far as I am merely a thinking thing,’ she discovers that she is simple and wholly lacking parts (M6, AT VII 86/CSM II 59). Thus, the claim that the self is composed of mind and body is incompatible with conceiving of the self as a mind. In this pivotal passage, then, the meditator rehabilitates a conception of the self as a human being (M6, AT VII 81/CSM II 56). Crucially, this conception is true or correct. The meditator introduces the teachings of nature precisely to identify the truth in the senses.

The human being the meditator rehabilitates differs from the one described in Meditations 1 and 2. A point of comparison may be helpful here. When the meditator proves the existence of material things in Meditation 6, she replaces ‘the Aristotelian sense-based conception of body with [her] purified intellectual conception of body,’ as Simmons (2014: 262) puts it. Similarly, when the meditator argues that she is correctly considered as a human being, she replaces the confused sense-based conception of a human being with a corrected Cartesian one. The transformation in the case of human beings is not as radical, however, as in the case of bodies. The sense-based conception of bodies as colorful, smelly, tasty, etc. bears little resemblance to the intellectual conception of bodies as geometrical objects made actual.19 In contrast, the meditator retains much of her pre-theoretical conception of what human beings are like. Considered as a human being, she is composed of soul and body, which is what she always believed. Considered as a human being, she engages in many of the same activities as before. She is nourished, moves about, senses, and thinks. But the significance of these claims has changed. Her understanding of the human being’s constituents—soul and body—has been transformed, as has the division of labor between them. Whereas the meditator previously attributed all her activities to the soul (nourishment, self-motion, sense perception, and thinking), now she attributes many of her vital functions to the human body on its own, which she considers ‘as a kind of machine’ (M6, AT VII 84/CSM II 58).20 The mind or soul is no longer the principle of life.21

Let me add one last piece of evidence from Meditation 6 for the claim that the self may be correctly considered as both a mind and a human being: the meditator’s description of her ‘own nature in particular.’ The meditator explains that by ‘my own nature in particular [naturam meam in particulari],’ she means ‘nothing other than the complex [complexionem] of things bestowed on me by God’ (M6, AT VII 80/CSM II 56). Her own nature includes ‘many things that belong to the mind alone—for example my perception that what is done cannot be undone, and all other things that are known by the natural light,’ ‘much that relates to the body alone, like the tendency to move downward, and so on,’ but also ‘what God has bestowed on me as a composite of mind and body’ (M6, AT VII 82/CSM II 57, emphasis added; see also M6, AT VII 88/CSM II 61 and AT VII 90/CSM II 62). This suggests that (i) the meditator can be correctly considered as a mind, a human being, and perhaps even as a human body, and that (ii) her properties vary accordingly. For example, the meditator writes that it is specifically her nature ‘as a combination of mind and body’ that teaches her ‘to avoid what induces feelings of pain and to seek out what induces feelings of pleasure’ (M6, AT VII 82/CSM II 57; see also M6, AT VII 85/CSM II 59). In contrast, ‘knowledge of the truth about such things [viz. things located outside us] seems to belong to the mind alone, not to the combination of mind and body’ (M6, AT VII 82–83/CSM II 57).

In the final sentence of the Meditations, the meditator invokes the fallibility of the human aspect of her nature: ‘it must be admitted that in this human life [humanam vitam] we are often liable to make mistakes about particular things, and we must acknowledge the weakness of our nature [naturae nostrae]’ (M6, AT VII 90/CSM II 62). The meditator has come full circle … almost. She is back to considering herself as a human being, but now with a deeper understanding of what it means to be human. The conception of the self as a mind is still available to her, whenever she meditates. But, crucially, the account of the self the meditator develops over the course of the Meditations incorporates both perspectives.

Thus, the meditator’s final position—and, hence, presumably, Descartes’s—is that we may correctly consider the self as either a mind or a human being. The question that I am ultimately interested in is how the self can be correctly considered in both these ways. In the remainder of this section, however, I will provide further textual evidence that the ‘I’ of the Meditations admits of this conceptual variability, by looking at texts where Descartes is pressed to clarify the views presented in this work: namely, the Objections and Replies (1641) and in his correspondence with Elisabeth (1643). In these texts, Descartes speaks in his own voice, and so they are useful for confirming which of the views expressed by the meditator are in fact Descartes’s.

Whereas the meditator refines the conceptions of the self as a mind and a human being over the course of the Meditations, in the Objections and Replies Descartes takes these hard-won conceptions for granted. Together, these conceptions provide Descartes with a powerful dialectical tool, as they allow for flexibility in the claims that he can make about the self. In the First Replies, for example, Descartes revisits the Meditation 3 claim that the meditator’s parents did not create her. He is now even more explicit that he was only talking about the self qua mind, not qua human being:

This passage suggests that one could legitimately ask about what caused the self qua consisting of mind and body (i.e., qua human being). Indeed, a few sentences later Descartes allows that there’s a sense in which ‘my father begot me’ (O/R 1, AT VII 107/CSM II 78).22 But, in fact, that was not the way Descartes was considering the self in Meditation 3. Thus, Descartes’s different conceptions of self allow him to simultaneously make counter-intuitive philosophical claims about the self (‘my parents did not make me’), while also speaking with the vulgar (‘yes, they did’), by indexing these claims to the different conceptions of self qua mind and qua human being.

Descartes generalizes this point in the Second Replies when he argues that the mind and the human being each have their own kind of truth:

Although the distinction between the two kinds of truth is murky, this passage illustrates the philosophical work that considering the self in different ways can do. Descartes can accommodate both an ordinary, everyday kind of truth and the philosophical, naked truth, by indexing them to the self qua human being and the self qua mind, respectively.

Finally, Descartes’s dual perspectives on the self play a pivotal role in his exchange with Arnauld. In the Fourth Objections, Arnauld launches a two-pronged attack on the Meditations’ account of the self. First, Arnauld criticizes Descartes’s real distinction argument, which is supposed to show that the self can exist apart from the body. Arnauld objects that Descartes cannot infer ‘from the fact that he is aware of nothing else [besides thought] belonging to his essence, that nothing else does in fact belong to it’ (O/R 4, AT VII 199/CSM II 140). Thus, Descartes cannot rule out the possibility that his essence includes a secret connection to the body. Second, Arnauld objects that even if Descartes could establish the real distinction, this would be a bad result, leading to an unacceptably Platonic view of the self:

Arnauld has set a formidable dialectical trap, as any response to the first objection will set Descartes up for the second one.

In the face of Arnauld’s criticism, Descartes does not abandon the claim that the self is really distinct from the body. Instead, Descartes appeals to the conception of the self qua mind in response to the first objection, while he wheels out the self qua human being in response to the second. This strategy presupposes that both conceptions are correct. In a bit more detail, Descartes responds to Arnauld’s first objection as follows:

Descartes argues that each of us is aware of ourself as having a property—thinking—that is sufficient for the possibility of our independent existence. We might see this as the conjunction of two claims: (a) thinking is sufficient for the self’s existence, and (b) thinking is self-sufficient, that is, it is possible for God to create an instance of thinking without creating any other attributes. Descartes clearly uses the conception of the self qua mind in this argument, since the conception of the self qua mind just is a conception of the self qua thinking, non-extended thing.

In contrast, Descartes deploys the conception of the self qua human being in order to deflect Arnauld’s criticism that he (Descartes) has proved ‘too much’ (O/R 4, AT VII 203/CSM II 143). Making sense of Descartes’s response to this objection is tricky, however, because Arnauld’s use of the Latin term ‘homo’ differs from Descartes’s typical usage. To mark this difference, I have translated ‘homo’ as ‘man’ in Arnauld’s objection, whereas I have been translating ‘homo’ as ‘human being’ in Descartes.

In his objection, Arnauld uses the term ‘homo’ to refer to the kinds of beings we are, whatever that might turn out to be. His willingness to switch back and forth between questions about ‘our essence’ and questions about ‘the definition of man’ indicates this usage. Thus, when Arnauld asks ‘what is man?’ that is more or less equivalent to asking, ‘what are we?’ or ‘what am I?’23 In contrast, Descartes typically uses the term ‘homo’ to refer to a specific kind of entity: namely, the mind-body composite that results when a mind is substantially united to a human body. When Descartes asks ‘what is a human being?’ that is not equivalent to asking ‘what are we?’ or ‘what am I?’ For one thing, ‘a human being’ is supposed to be an informative answer to the ‘what am I?’ question, as we saw in Meditation 2 (M2, AT VII 25–26/CSM II 17). Rather, in Descartes’s hands, the question ‘what is a human being?’ is a request for more information about a certain kind of composite, for example, is the human composite a substance? What is its principle of unity?24 That being said, Descartes adopts Arnauld’s usage of the term ‘homo’ in responding to his second objection. This is a perfectly reasonable thing to do. Using the language of one’s interlocutors helps avoid talking past one another. But it makes our job more difficult as interpreters. When Descartes uses ‘homo’ à la Arnauld, I will also translate it as ‘man’ to help keep things straight.

With these terminological preliminaries out of the way, let’s take a look at what Descartes says:

Descartes claims to have avoided the implication that man or the self is ‘only [solum] “a soul which makes use of a body,”’ on the grounds that he proved both the real distinction and the substantial union of mind and body in Meditation 6. Recall what happened at these two moments. In the real distinction argument, the meditator argues that her conception of herself as a thinking, non-extended thing—that is, as a mind—is correct, and, hence, that there is a sense in which she can exist apart from the body. But the meditator isn’t only (‘solum’) a mind, because her sense-based conception of herself as composed of mind and body—that is, as a human being—is also correct, at least when suitably corrected. Thus, Descartes resists Arnauld’s charge of taking us back to the Platonic view of man, by pointing out that he (Descartes) recognizes two views of the self: the admittedly Platonic self qua mind, but also the embodied self qua human being.25 This response presupposes the correctness of both ways of considering the self.

In the next paragraph, Descartes drives home the difference between these two ways of considering the self by clarifying their respective essences. The essence of the mind consists only in thought, whereas the essence of the human being includes both mind and body (O/R 4, AT VII 228/CSM II 160; see also O/R 4, AT VII 219/CSM II 155, Letter to Hyperaspistes, August 1641, AT III 422/CSMK III 189). From this essential difference it follows that the self qua mind is really distinct from the body and can exist apart from it, whereas the self qua human being is not really distinct from the body and cannot exist without both a mind and a body. Descartes explicitly draws out this latter implication in a letter to Regius: ‘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’ (Letter to Regius, January 1642, AT III 508/CSMK III 209).26 There can be disembodied minds, but there cannot be disembodied nor mindless human beings.

To sum up: Arnauld objects to Descartes’s account of the self (or I) on two fronts. First, Arnauld objects that Descartes has failed to establish that the self can exist apart from the body, and, second, that Descartes is going down the wrong path by even trying to establish that conclusion. Descartes responds to the first objection by arguing that the self considered as a thinking, non-extended thing has everything it requires for separate existence, while he responds to the second objection by reminding Arnauld that although the self can be correctly considered as a mind or thinking thing, the self can also be correctly considered as a human being. Hence, it would be a grave error to infer that the self is ‘only’ (solum) a mind or thinking thing. Moreover, Descartes proved the real distinction and the substantial union ‘at the same time’ (simul etiam), which suggests that he takes the corresponding conceptions of the self qua mind and qua human being to fit together somehow (O/R 4, AT VII 227–28/CSM II 160).

In the Objections and Replies, we see the fruit of conceiving the self as both a mind and a human being. Although these dual perspectives introduce complexity into Descartes’s account of the self, this complexity earns its keep by allowing him to defend the account against objections.

In 1643, so two years after the publication of the Meditations and the Objections and Replies, Princess Elisabeth of Bohemia sends Descartes a letter asking him to explain himself: if, as his Meditations claims, ‘the soul of a human being’ is ‘only a thinking thing,’ how can it ‘determine the bodily spirits, in order to bring about voluntary actions’ (AT III 661)?27 Elisabeth frames her question as being about the soul and its power to move the body, rather than in the first-personal terms of the self (or I) of the Meditations. In his response Descartes refers to the soul as well, much as he picks up Arnauld’s usage of the term ‘man’ in the Fourth Replies. Nevertheless, I would like to suggest that Descartes hears Elisabeth as objecting to his account of the self, since the idea of the mind as ‘only a thinking thing’ is introduced as a conception of the self in the Meditations. We might then reformulate Elisabeth’s objection like this: if the self is only a thinking thing, how can it possibly set the body in motion, when, for example, someone wills to raise their arm?

Descartes responds to Elisabeth’s question by claiming that the causal interaction between mind and body depends on their union, admitting that he has not said much about this topic up until this point:

Descartes’s response to Elisabeth resembles his strategy in the Fourth Replies. When Arnauld accused Descartes of taking us back to the Platonic view of man, Descartes responded by saying that, actually, he recognizes two views of the self: the self considered as a mind or thinking thing, and the self considered as a human being. Similarly, when Elisabeth presses Descartes to explain how the soul can act on the body, Descartes responds that there are two ways of looking at the soul or self: as a thinking thing, or as a mind-body union.

In his letter to Elisabeth, Descartes undertakes to give a more explicit account of how to conceive the union. He argues that we have a ‘primitive notion of union’ alongside the clear and distinct ideas of thought and extension:

The primitive notion of union seems to be a relabeled version of the sense-based conception of the human being we encountered in Meditation 6. Admittedly, this point is not wholly obvious in this passage. But, in Descartes’s next letter to Elisabeth, from June 28, 1643, Descartes argues that the idea of the union is primarily a sensory idea of oneself as composed of mind and body. Descartes tells Elisabeth that the notion of union is something:

Experiencing the union just is a matter of sensorily experiencing oneself as composed of mind and body. The idea of the union can be misapplied to other things, when, for example, Scholastics use this idea to form a confused idea of heaviness (Letter to Elisabeth, 21 May 1643, AT III 667–68/CSMK III 219). But this idea’s proper object is the self.

Reading Descartes’s exchange with Elisabeth as concerning the conception of the self qua human being, we may draw three main conclusions. First, the exchange with Elisabeth confirms that the conception of the self qua mind is an intellectual idea, whereas the conception of the self qua human being is sensory. As Descartes writes, ‘the soul is conceived only by the pure intellect,’ whereas ‘what belongs to the union of the soul and body is known only obscurely by the intellect alone or even by the intellect aided by the imagination, but is known very clearly by the senses [elles se connoissent tres-clairement par les sens]’ (Letter to Elisabeth, 28 June 1643, AT III 691–92/CSMK III 227). Second, Descartes argues that the primitive notion is ‘known very clearly by the senses [elles se connoissent tres-clairement par les sens]’ (Letter to Elisabeth, 28 June 1643, AT III 691–92/CSMK III 227, emphasis added). This indicates that the conception of the self qua human being is true or correct, since knowledge is factive. Third, Descartes’s claim that the union is known best through the senses helps explain why the conception of the human being that the meditator rehabilitates in Meditation 6 is not that different from her initial conception. The senses provide the best access to the nature of human beings, and so the meditator’s initial sense-based conception of her human nature was not that far off the mark. As Descartes writes, ‘it is the ordinary course of life and conversation, and abstention from meditation and from the study of the things which exercise the imagination, that teaches us how to conceive the union of the soul and the body’ (Letter to Elisabeth, 28 June 1643, AT III 692/CSMK III 227).

Descartes holds that the self (or I) of the Meditations can be correctly considered as a mind, but also as a human being. The problem is that the mind and human being have different, and, indeed, incompatible properties. This might seem to imply that a single subject of predication—the self—has incompatible properties. For example, we might worry that if the self qua mind is simple, while the self qua human being is composite, then the self is both simple and composite, which is a contradiction. As Descartes writes to Hyperaspistes, ‘a human being, being a composite entity, is naturally corruptible, while the mind is incorruptible and immortal’ (Letter to Hysperaspistes, August 1641, AT III 422/CSMK III 189). Similarly, if the self qua mind can exist apart from the body, while the self qua human being cannot, then it might seem to follow that the self both can and cannot exist apart from the body. More generally, we might worry that Descartes’s qualified predications imply corresponding unqualified predications.28 In response, Descartes might argue that the whole point of the qualifications ‘qua mind’ and ‘qua human’ is to block these sorts of implications. But that seems like a placeholder rather than a proper account. How does the addition of a ‘qua’ or a ‘considered as’ prevent the self’s seemingly incompatible properties from crashing into one another? To answer this objection, we need an account of the significance of considering the self in different ways.

Consider an expression of the form ‘S is P qua Q.’ This claim attributes a property P to a subject of predication S with the qualification Q. On the explicative analysis, both the predicate P and the qualification Q are attributed to the subject S. The ‘qua’ articulates an explanatory connection between the property P and the qualification Q. On this analysis, S is P qua Q if and only if (i) S is P, (ii) S is Q, and (iii) the fact that S is Q explains the fact that S is P. For example, the self is simple qua mind if and only if (i) the self is a mind, (ii) the self is simple, and (iii) the fact that the self is a mind explains the fact that the self is simple, perhaps because simplicity is built into the nature of the mind. On this option, expressions of the form ‘S is P qua Q’ specify explanatory relations between a thing’s properties.

Although the explicative analysis of qualified statements was historically important, this analysis cannot explain how Descartes avoids contradiction when he relativizes incompatible properties to different ways of considering the self.30 Consider, for example, the following pair of claims (which I will keep coming back to):

On the explicative analysis, (1) says that the self is a mind, that the self is simple, and that the self is simple because it is a mind. (2) says that the self is a human being, that the self is composite, and that the self is a composite because it is a human being. Thus, Descartes’s endorsement of (1) and (2) would imply that the self is simple and composite, which is precisely the kind of contradiction that we are trying to avoid.

Another option is that expressions of the form ‘S is P qua Q’ spell out the content of a certain representation of S. On this analysis, the claim that S is P qua Q does not directly describe S, but, rather, describes a way that S is represented. More specifically, S is P qua Q if and only if a representation of S as Q includes or entails a representation of S as P. On this analysis, for example, the self is simple qua mind if and only if representing the self as a mind implies representing the self as simple, in roughly the same way that representing a shape as a triangle implies representing the shape as having three sides. Whereas the explicative analysis takes ‘qua’ expressions to articulate the relations between a thing’s properties, the representational analysis understands ‘qua’ expressions as articulating the relations between representations of a thing’s properties.

The representational analysis does a better job of removing the contradictions. Consider, again, the following pair of claims:

On the representational analysis, (1) says that a representation of the self as a mind includes simplicity. Similarly, (2) says that a representation of the self as a human being includes compositeness. Descartes’s endorsement of (1) and (2) would then imply that the self is represented as both simple and composite. Does this imply a contradiction? No. Although someone cannot have incompatible properties, there is nothing contradictory about someone being represented in incompatible ways, so long as at least one of these representations is erroneous. This last qualification is important, since being truly represented as simple and composite implies being simple and composite, in which case we would be back where we started.

According to this option, the apparent variation in the self’s properties plays out at the level of representations, rather than at the level of the object. Commentators who defend this kind of reading arguably include Williams (1978) and Curley and Koivuniemi (2015).31 Williams argues that Descartes’s descriptions of the embodied self are mere reports of the way the self is represented by the senses: ‘the entire content of Descartes’s denial that he is a pilot in a ship is phenomenological—it is exclusively about what the experience of being embodied is like’ (1978: 267, emphasis added). Curley and Koivuniemi (2015) argue that Descartes’s puzzling characterizations of the human being or union—for example, his claim that the mind is the ‘form’ or the ‘substantial form’ of the human body, or that the human being is an ‘ens per se’—describe the way someone sensorily represents her connection to her body, which is otherwise merely causal.32 These commentators suggest, in effect, that Descartes’s claims about the self qua human being are really just claims about the way the self is (erroneously) represented by the senses. In a similar vein, we might add that claims about the self qua mind are claims about the way the self is (truly) represented by the intellect.

This strategy avoids contradiction by saying that, for every pair of representations attributing incompatible properties to the self, at least one of these representations is erroneous. Hence, for every pair of incompatible properties, this reading forces Descartes to choose one as truly applicable to the self, while rejecting the other as an illusion or misrepresentation. Simple or composite, really distinct or substantially united, created by its parents or not—which is the self really? Should we trust the intellect or senses here? But Descartes appeals to conception-relativity precisely in order to ‘have it both ways,’ as Baier (1981: 169) puts it, and he doesn’t seem to think that he needs to choose.33 He insists that the self is both a mind and a human being, both simple and composite, both really distinct from the body and substantially united to it, both created by its parents and not. As Descartes reminds Arnauld in the Fourth Replies, ‘in the Sixth Meditation, where I deal with the distinction between the mind and the body, I also proved at the same time [simul etiam] that the mind is substantially united with the body’ (O/R 4, AT VII 227–28/CSM II 160). The representational analysis does not explain how both these perspectives on the self can be simultaneously correct, but instead denies the phenomenon to be explained. Hence, the representational analysis fails as a reading of Descartes’s various ways of considering the self.

The next option for analyzing statements of the form ‘S is P qua Q’ suggests that the qualification Q is built into the predicate P. On this proposal, we attribute the qualified property P-qua-Q to the subject S. For example, John is good qua cobbler if and only if the subject John has the qualified property good-qua-cobbler, which is just to say that John is a good cobbler.34 The claim that John is a good cobbler does not imply that John is good. In general, on the predicate analysis, the claim that S is P qua Q does not imply that S is P. Consider, then, our pair of seemingly conflicting claims:

On the predicate analysis, (1) does not entail that the self is simple, nor does (2) entail that the self is composite. Instead, (1) implies that the self has the qualified property simple-qua-mind, and (2) implies that the self has the qualified property composite-qua-human-being. And perhaps something can be both simple-qua-mind and composite-qua-human being without contradiction. Hence, the predicative analysis can avoid the looming contradiction, at least at a formal level.

It is unclear, however, how to understand the qualified properties in question: namely, the properties simple-qua-mind and composite-qua-human-being. Although she does not spell out the details, Baier suggests one option: namely, that the self’s seemingly incompatible properties consist in relations to the theoretical and practical standpoints:

On this Baier-inspired reading, the self is simple-qua-mind if and only if the self is simple relative to the theoretical standpoint, whereas the self is composite-qua-human-being if and only if the self is composite relative to the practical standpoint.35 Descartes might then argue that the qualified properties—simple-qua-mind and composite-qua-human-being—are compatible, since these properties are relational, and something can stand in different relations to different things. And similarly for all of the Cartesian self’s other incompatible properties.

The problem with this Baier-inspired reading is that the self’s incompatible properties do not lend themselves to reinterpretation as relations. Simplicity and compositeness do not seem to consist in relations to different standpoints (whatever those might be). As Van Cleve writes, it is unclear how ‘the mereological structure of an object can be one thing in a certain relation and another thing apart from that relation’ (1999: 149). Or consider the real distinction. Descartes holds that the self qua mind is really distinct from the body, while the self qua human being is not. The self’s real distinction from the body does not seem to consist in a relation to the theoretical standpoint, nor does its substantial union to the body seem to consist in a relation to the practical standpoint.36 So the predicate analysis fails as an account of Descartes’s qualified claims about the self as well.

The final and most promising analysis for Descartes’s statements of the form ‘S is P qua Q’ interprets the qualification Q as attaching to the subject S. The property P is attributed not to S simpliciter, but to the qualified subject S-qua-Q. On this analysis, S is P qua Q if and only if S-qua-Q is P. For example, John is red qua beard if and only if John-qua-beard is red, or, equivalently, John’s beard is red. The claim that John’s beard is red does not imply that John is red, however. In general, on the subject analysis, the claim that S is P qua Q does not imply that S is P. Consider, one last time, our pair of examples:

On this proposal, claim (1) implies that the self-qua-mind is simple, while claim (2) implies that the self-qua-human being is composite. Hence, claims (1) and (2) attribute properties—namely, simplicity and compositeness—to logically distinct subjects of predication. Since we cannot, in general, infer that S is P from the claim that S-qua-Q is P (recall John’s beard), we cannot infer that the self is simple from the claim that the self-qua-mind is simple, nor that the self is composite from the claim that the self-qua-human-being is composite, at least not without some further argument. Hence, (1) and (2) would not be formally contradictory.37 Whereas the predicate analysis of ‘qua’ allows more precision about the properties we attribute to the self, the subject analysis enables more precision about the subjects of predication to which these properties are attributed.

Still, this strategy requires that we give an account of these special qualified subjects of predication, viz. the self-qua-mind and the self-qua-human-being. This is analogous to the task in the preceding section of giving an account of the special qualified properties simple-qua-mind and composite-qua-human-being. The qualified subjects of predication are not mysterious, however. The self-qua-mind just is the mind. The self-qua-human-being just is the human being. The main textual evidence for this claim is Descartes’s willingness to treat claims about the self qua mind interchangeably with claims about the mind. In Meditation 3, for example, Descartes treats the expression ‘I in so far as I am a thinking thing [me quatenus sum res cogitans]’ interchangeably with ‘the mind [mentem]’ (M3, AT VII 50/CSM II 35; see also M6, AT VII 86/CSM II 59). Passages like this suggest that the expressions ‘I in so far as I am a mind or thinking thing’ or ‘I considered as a mind’ are just roundabout ways of referring to a certain mind, namely, mine. And similarly, the expression ‘I in so far as I am a human being’ refers to a certain human being, namely, the one I normally take myself to be.

One presupposition of this strategy is that the mind and human being are distinct subjects of predication. If the claim ‘the self is simple qua mind’ implies that the mind is simple, and if ‘the self is composite qua human being’ implies that the human being is composite, then Descartes avoids contradiction only if the mind and human being are distinct subjects of predication. Let me ward off a possible misunderstanding here. By saying that the mind and human being are distinct subjects of predication, I am not assuming that the mind and human being are really distinct substances. As I use the term, a subject of predication is just something with properties. Subjects of predication come cheap. Anything with properties counts. A sensation of red, for example, is a subject in the relevant sense, since it has properties that distinguish it from a sensation of green, or an intellectual idea of a triangle. A sensation of red is not a substance, however, because it lacks the independence Descartes requires of substances (Principles I.51, AT VIIIA 24/CSM I 210). In short: a subject of predication is something with properties; a substance is a suitably independent subject of predication.38 We might then wonder how to characterize the sense in which the mind and human being are distinct, if not in terms of a real distinction between substances.39 Without begging substantive metaphysical questions at this point, we cannot say much more than that they are numerically distinct subjects of predication.40 Different accounts of the mind and human being will then suggest more precise characterizations of the distinction between them. A trialist like Hoffman (1986) or Skirry (2005), for example, will give a very different account of the mind’s distinction from the human being, than, say, Rozemond (1998) or Chappell (1994).

Although the subject analysis provides the best reading of Descartes’s various ways of considering the self, it raises difficulties of its own. I will mention two. First, if the expressions ‘I in so far as I am a mind’ and ‘I in so far as I am a human being’ refer to the mind and human being, why refer to them in such a clunky, roundabout way? One reason is that these more elaborate expressions allow Descartes to specify whose mind and which human being is at stake: namely, my own mind, and the human being that I typically take myself to be. Reflecting on ourselves, rather than on an arbitrary mind or human being, is important for Descartes, because each of us stands in a unique or privileged epistemic relation to ourselves. Descartes holds that I know my own mind—through consciousness or reflection—in a way that differs from the way I know yours. This first-person perspective grounds many of Descartes’s signature claims about the self or mind, for example, the certainty that I exist, and that I am a thinking thing. Expressions like ‘I in so far as I am a mind or thinking thing’ are one of Descartes’s linguistic strategies for getting his readers to adopt this perspective. Similar considerations explain why Descartes uses qualified expressions like ‘I in so far as I consist of mind and body.’ Descartes also holds that each of us has a first-person perspective on our humanity. I know my own human nature—through proprioceptive or bodily awareness—in a way that differs from the way I know yours. The first-person perspective afforded by bodily awareness also grounds many of Descartes’s claims about human nature, for example, the claim that mind and body are closely joined and intermingled (M6, AT VII 81/CSM II 56), and that dropsy—or, more generally, sickness—is ‘a true error of nature’ with respect to the composite (M6, AT VII 85/CSM II 59). By using the expression ‘I in so far as I am a human being,’ Descartes encourages his readers to adopt this embodied first-person perspective as well.

The second objection to the subject analysis is more serious. On this analysis, the expression ‘the self qua mind’ refers to the mind, while the expression ‘the self qua human being’ refers to the human being. We can then explain how Descartes avoids contradiction when he attributes incompatible properties to the self, because he attributes one property to the mind, the other to the human being. In a sense, this procedure explains how Descartes can ‘have it both ways’ (Baier 1981: 169). But we have not yet explained how Descartes can have it both ways with respect to the self. So the mind is simple, and the human being is composite. What does any of that have to do with the self? Without a story about how the self fits into the picture, and, more specifically, about how the self relates to the mind and human being, it is mysterious what bearing, if any, the mind’s and human being’s properties have for the self. Thus, we have removed the threat of contradiction, but at the expense of saying anything informative about what the self is like.

An analogy may be helpful here. Consider the claim that John is red qua beard. On the subject analysis, John is red qua beard if and only if John-qua-beard is red, or, equivalently, John’s beard is red. Although it may be tempting to assume that if John-qua-beard is red, then John is partly red, this conclusion would be somewhat hasty. On its own, the claim that John’s beard is red does not imply any conclusions about what color John might be: it all depends on how John relates to his beard. To make this point vivid, let’s consider some options for how John might relate to his beard. If John’s beard covers part of his face, then his beard’s redness would entail that John is partly red. In contrast, if John cut off his beard long ago and keeps his old beard in a box in the attic, then his beard’s redness would not entail that John is even partly red. Finally, if John were extremely hairy, so that his beard covered every inch of his body, then his beard’s redness would imply that John is red all over. My point is this: on the subject analysis, the claim that John is red qua beard tells us virtually nothing about what John is like. The claim that John is red qua beard only becomes informative about John when it is supplemented with an account of how John relates to his beard. Similarly, the claim that the self is F qua mind (or human being) is only informative about the self when this claim is supplemented with an account of how the self relates to the mind (or human being).

In a way, the solution to this problem is obvious: we need to say what the self is, and how it relates to the mind and human being. We need an account of the unqualified self, or the self simpliciter, so that we can make sense of how Descartes’s qualified claims bear on what the self is like. Such an account would make clear how claims about the mind and human being are informative about the self: it would allow us to move in some cases—but not all—from premises about what the self-qua-mind is like, or what the self-qua-human being is like, to conclusions about what the self is like. But, as we shall see in the next section, giving an adequate account of the self simpliciter is not so easy.

Someone might object that there is no such thing as the self simpliciter in Descartes, and that my call for an account is misplaced. This would imply that there are just two first-person pronouns. One refers to the mind, while the other refers to the human being. Their referents overlap, since the mind is part of the human being. But neither has priority over the other. This would mean that the first-person is ambiguous unless suitably qualified. Any unqualified claim of the form ‘I am F’ would be ambiguous between ‘I-qua-mind am F’ and ‘I-qua-human-being am F’ (Brown 2014: 245). This ambiguity would infect many of the questions Descartes cares most about: What is my essence? Can I exist apart from my body? What is my relationship with matter? Can I survive death and the destruction of my body? Despite Descartes’s willingness to qualify the first-person pronoun when pushed into a corner, he would not accept that these questions are fundamentally indeterminate. Consider that last question again. Can I survive death and the destruction of my body? If someone were to respond, ‘Well, it all depends, you-qua-mind can survive the destruction of your body, you-qua-human-being cannot,’ we might reasonably be unsatisfied with this answer.41 And Descartes would be too. Hence, Descartes needs to specify a referent for the unqualified first-person pronoun. He needs an account of the self simpliciter.42

On the traditional reading, the Cartesian self just is the mind.43 At face value, this reading fits nicely with the subject analysis of Descartes’s various ways of considering the self. Here’s the idea. The self is the mind. In fact, the self just is the mind: they are identical. So the self has all the same properties as the mind. In this life, the self/mind is united to a human body. The compound resulting from this union is the human being, which is composed of the self/mind and a human body. The self/mind is related to the human being as part to whole. The self/mind and the human being are distinct subjects of predication, and they have many different properties. The essence of the self/mind consists only in thought, whereas the essence of the human being includes thought and extension, mind and body. The self/mind is simple, whereas the human being is composite. The self/mind is really distinct from the body, whereas the human being is not. And so forth.

On the traditional reading, Descartes’s qualified claims about the self bear on the self simpliciter as follows. When Descartes claims that the self is simple qua mind, this implies that the self simpliciter is simple. In contrast, when Descartes claims that the self is composite qua human being, this implies that the human being of which the self is a part is composite. When Descartes talks about the self qua mind, he is really just talking about the self. When Descartes talks about the self qua human being, he is talking about the human being that envelops the self.

Although the self is intimately related to the human being, as essential part to whole, this account presupposes that the self is numerically distinct from the human being. This is built into the claim that the self/mind and human being are distinct subjects of predication, and is required to avoid contradiction.44 Hence, the self does not really have any of the properties proper to the human being. The self isn’t really composed of mind and body. Rather, the self is closely related to something that is composed of mind and body. Attributing the human being’s properties to the self is a mere façon de parler, in roughly the same way that a husband might say ‘we are having a baby’ when his wife is in labor, even though he is clearly not the one having the baby. This result should give us pause.

In Meditation 6, when Descartes claims that the senses teach him that he is ‘composed of body and body,’ and, hence, capable of being affected by surrounding bodies, this does not sound like a mere façon de parler (M6, AT VII 81/CSM II 56). In the Fourth Replies, when Descartes claims that his arguments against the Platonic view of man—that is, the view that the self just is the mind—are ‘as strong as any I can remember ever having read,’ this does not sound like loose speech (O/R 4, AT VII 228/CSM II 160). Similarly, in his correspondence with Elisabeth, I take Descartes literally when he says that the ‘union is known very clearly through the senses,’ and that one of the things someone can know in this distinctively sensory, first-personal way is that ‘he is a single person with both body and thought so related by nature that the thought can move the body and feel the things which happen to it’ (Letter to Elisabeth, 28 June 1643, AT III 692–94/CSMK III 227–28).

We can put the objection like this. Descartes holds that the self—the self simpliciter and not just the self-qua-human being—has a characteristically human property, namely, the property of being composed of mind and body.45 That’s the force of saying that the self may be correctly considered as a human being. But the mind is not composed of anything, let alone a mind and body. The mind, even a mind united and intermingled with a body, ‘is something quite single and complete’ and lacks parts of any kind (M6, AT VII 86/CSM II 59, emphasis added). So the self simpliciter has a property—namely, being composed of mind and body—that the mind lacks. Assuming the indiscernibility of identicals, it then follows that the self is not identical to the mind. And that would mean that the traditional reading of Descartes is wrong.

Let me add one last piece of textual evidence against the traditional reading. After Gassendi suggests that Descartes regards himself ‘not as a whole human being but as an inner or hidden component—the kind of component you previously considered the soul to be’ (O/R 5, AT VII 260/CSM II 181), Descartes vehemently disagrees:

Descartes recoils from the view that the self is ‘an inner or hidden component’ inside the body, as the traditional view would seem to imply. Descartes is adamant that he should not be addressed as a mind. But what else might he be?

Dissatisfaction with the traditional reading of Descartes has led some commentators to argue that, in fact, the Cartesian self just is a human being.46 Call this the re-humanized reading of the Cartesian self.47 For our purposes, we can assume that the mind and the human being stand in all the same relations as before. The only difference is that we are now taking the self to be identical with the human being, rather than the mind. So the self and human being will have exactly the same properties. Let’s see how that looks. In this life, the mind is united to a human being. The compound resulting from the union just is the self/human being. The self/human being is composed of mind and body, so that the self/human being relates to the mind as whole to part. As before, the self/human being and the mind are distinct subjects of predication, and have different properties. The essence of the self/human being includes thought and extension, mind and body, whereas the essence of the mind consists only in thought. The self/human being is composite, while one of its parts—the mind—is simple … Well, you get the idea.

The re-humanized reading can also be used to explain how the self simpliciter relates to Descartes’s various ways of considering the self. When Descartes claims that the self is composite qua human being, this implies that the self simpliciter is composite. In contrast, when Descartes claims that the self is simple qua mind, this implies that a part of the self—viz. the mental part—is simple. More generally, when Descartes talks about the self qua human being, he is talking about the self. When he talks about the self qua mind, he is talking about a part of the self. One advantage of this reading is that it goes further than the traditional one in vindicating Descartes’s view that the self has incompatible properties. The self is wholly identical to the human being, and partially identical to the mind, and so we can make more sense of the idea that the self participates in the properties of both. The claim that the self is simple, for example, is not a mere façon de parler. This claim is about the innermost part of the self, although the self as a whole is composite.48

Nevertheless, there are problems with the re-humanized reading as well. This reading implies that the self simpliciter cannot exist apart from the body. This follows more or less immediately from the claim that the essence of the self/human being includes mind and body. As Descartes writes to Regius, for example, ‘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’ (Letter to Regius, January 1642, AT III 508/CSMK III 209; see also Letter to Hyperaspistes, August 1641, AT III 422/CSMK III 189).49 The problem for the re-humanized reading, then, is that Descartes holds that the self—the self simpliciter—can in fact exist apart from the body. Rozemond puts this point beautifully: ‘In the Meditations (and, in fact, often elsewhere) [Descartes] does not state dualism by saying that mind is really distinct from body, but that he is really distinct from body’ (1998: 159).50 As the meditator writes in Meditation 6, ‘it is certain that I am really distinct from my body, and can exist without it [certum est me a corpore meo revera esse distinctum, & absque illo posse existere]’ (M6, AT VII 78/CSM II 54, emphasis added).51 In the next paragraph, the meditator argues further that she can clearly and distinctly understand herself as a whole (totum me) in an even more attenuated sense, namely, as lacking not merely a body, but also the faculties for imagination and sensation (M6, AT VII 78/CSM II 54; see also M6, AT VII 86/CSM II 59). This clear and distinct perception suggests that the self can exist not merely in a disembodied state, but as a pure intellect.

We can then put the objection like this. Descartes holds that the self—the self simpliciter, and not just the self-qua-mind—can exist apart from the body. But the human being cannot exist apart from the body. So the self simpliciter has a property—a way in which it can exist—the human being lacks. Assuming, again, the indiscernibility of identicals, it then follows that the self is not identical to the human being. So the re-humanized reading is wrong too. Notice that this problem has exactly the same structure as the one faced by the traditional reading.

The proponent of the re-humanized reading cannot downplay the conclusion of the real distinction argument in Meditation 6, as they might for the meditator’s claim in Meditation 2 that she is ‘only a thinking thing’ (M2, AT VII 27/CSM II 18). By the time the meditator reaches this point in Meditation 6, she has emerged from the skeptical doubts of the previous Meditations and is speaking in Descartes’s voice. This is confirmed by the fact that Descartes claims that he is really distinct from the body in other texts, such as the Fourth Replies, which are free of the interpretative complexities raised by the Meditations’ narrative and developmental structure (O/R 4, AT VII 219/CSM II 154–55). A defender of the re-humanized reading might argue that Descartes doesn’t really mean it when he suggests that the self simpliciter can exist apart from the body. Perhaps Descartes is implicitly considering the self as a mind.52 But that seems like a stretch to me. I think Descartes means it when he says that you and I can exist without our bodies.

We need an account of the self and how it relates to the mind and human being. Otherwise Descartes’s qualified claims about the self won’t tell us anything about the self simpliciter. I have argued against the traditional reading, on which the self is (identical to) a mind, as well as the re-humanized reading, on which the self is (identical to) a human being. We can summarize this negative argument as follows:

Therefore,

Both the traditional and the re-humanized reading latch onto an aspect of the self—the mind or the human being—and then treat this aspect as if it were the whole truth about the self. But Descartes holds that both the sensory and intellectual perspectives reveal important aspects of the self. The intellect tells us that we are thinking beings capable of existing apart from our bodies, while the senses tell us that we are embodied, composite creatures, tightly bound to our bodies. The traditional and re-humanized reading each respects one of these perspectives, but at the expense of the other. The inadequacy of these readings might suggest that Descartes lacks a coherent account of the self.53 The way to avoid this pessimistic conclusion is to sketch a new account of the Cartesian self that integrates them. We need to explain how the self can be composed of mind and body, and yet capable of existing in disembodied form.

My account presupposes that sensory and imaginative thoughts differ in kind from purely intellectual thoughts, and that the human being and mind produce these different kinds of thoughts. There is ample textual evidence that Descartes accepts both these claims.60 In Meditation 6, for example, the meditator distinguishes ‘certain special modes of thinking, namely imagination and sensory perception’ from pure understanding (M6, AT VII 78/CSM II 54). Imagination and sensory perception are imagistic forms of thinking, whereas the pure understanding is non-imagistic. The meditator illustrates this difference by contrasting the experience of imagining a geometrical figure and understanding one. When I imagine a triangle, I form a mental image of this shape, such that I ‘see the three lines with my mind’s eye as if they were present before me’ (AT VII 72/CSM II 50). I can understand a triangle, in contrast, without forming any kind of image: I simply grasp its nature as a three-sided figure.61

The meditator argues that the special modes of thinking and pure understanding are activities of different things. Sensory and imaginative thinking are activities performed by the mind and body together. In Meditation 6, the meditator suggests that imagining depends on corporeal images in the brain (M6, AT VII 73/CSM II 51), whereas she claims that sensory perceptions, as we saw above, ‘arise [exorti] from the union, and, as it were, intermingling of mind and body’ (M6, AT VII 81/CSM II 56). In both the imaginative and sensory cases, the mind joins forces with the human body to bring about these distinctively embodied forms of thought. In contrast, the mind performs its acts of pure understanding without any assistance from the body. ‘When the mind understands,’ the meditator explains, ‘it in some way turns towards itself and inspects one of the ideas which are within it’ (M6, AT VII 73/CSM II 51).62

When Descartes revisits this material in the Fifth Replies, he argues that ‘the powers of understanding and imagining do not differ merely in degree but are two quite different kinds of mental operation’ (O/R 5, AT VII 385/CSM II 264). The imagination is bound up with the body in a way that the pure understanding is not: ‘For in understanding the mind employs only itself, while in imagination it contemplates a corporeal form’ (O/R 5, AT VII 385/CSM II 264). In this passage, Descartes focuses on the imagination’s dependence on the body, in contrast to the pure understanding. But there are many places where Descartes argues that sensory perception is similarly dependent. In Principles I.48, for example, Descartes argues that sensations, natural appetites, and passions ‘must not be referred either to the mind alone or to the body alone,’ but that ‘these proceed [proficiscuntur] from the close and intimate union of our mind with the body’ (AT VIIIA 23/CSM I 209). And in Principles II.2, he writes:

Sensory perceptions do not ‘come from’ the mind alone. Rather, sensory perceptions are produced by the joint agency of mind and body, or, in other words, by the compound or human being.

Descartes also reiterates that the mind can think independently of the body. In the Comments on a Certain Program (1647), Descartes responds to Regius’s contention that the human body ‘so long as it is in the body, is organic [organica] in all its actions. Thus, as the disposition of the body varies, so the mind has different thoughts’ (Comments, AT VIIIB 355/CSM I 295). Descartes interprets Regius as saying that the mind ‘does not act on its own [per se nihil agat], but is something of which the body makes use, just as it makes use of the arrangement of its limbs and other corporeal modes’ (Comments, AT VIIIB 355–56/CSM I 302). Descartes signals his disagreement with Regius in no uncertain terms. Descartes argues that if the mind were just the body’s tool, as Regius suggests, then the mind would be ‘nothing but a mode of body’ (Comments, AT VIIIB 356/CSM I 302). But this consequence flies in the face of the mind’s real distinction from the body, which Regius and Descartes both accept. This exchange provides further evidence that Descartes holds that the mind can act on its own, namely, to produce intellectual thoughts of God and common notions (Comments, AT VIIIB 359/CSM I 304).

In arguing for this division of cognitive labor, Descartes further specifies his new account of the human being. Consider, again, the meditator’s list of the human being’s activities in Meditation 2: the human being is nourished, moves, and engages in sense perception and thinking (M2, AT VII 25–26/CSM II 17). The meditator’s pre-theoretical conception of a human being attributed all these operations to the soul; the correct Cartesian account is more fine-grained. By the end of the Meditations, as we saw above, the meditator attributes the human being’s vital functions—such as nourishment and breathing—to the human body reconceived as a machine. This section shows that Descartes attributes sense perception and imagination to the composite of soul and body, while intellectual thinking (and perhaps willing) is the only human action reserved for the soul on its own.

Now, we might wonder about the relevant sense of production in these passages, especially since I am claiming that the self is constituted by the entities jointly producing its mental life. Someone might worry, for example, that many entities outside me play a role in producing my mental life—surrounding bodies, my community, God, and so forth. Reflecting on the influences on our mental lives might lead to the conclusion that the universe as a whole always plays a role in the production of our mental lives, and, hence, that the universe as a whole constitutes each of us.64 To answer this objection, we need to identify a sense of production in which the mind-body union or human being uniquely produces sensory and imaginative thoughts, and in which the mind alone uniquely produces intellectual thoughts.

My strategy is ostensive. Descartes clearly holds that there is an important sense of production—an important sense of cause or activity—in which it can be truly said that the union of mind and body is the producer of sensations and imaginative thoughts, and in which the mind alone is the producer of intellectual thoughts. This commitment emerges in the passages we have looked at so far, and provides a fixed interpretive point for any reading of Descartes. When Descartes claims, for example, that sensations ‘arise from’ or ‘proceed’ from the mind-body union, rather than, say, one’s complete causal history, that’s the sense of production that I am appealing to when I say that the Cartesian self is constituted by the entities that jointly produce the totality of its mental life. Similarly, when Descartes claims that ‘in understanding the mind employs only itself,’ that’s the sense of production I’m appealing to (O/R 5, AT VII 385/CSM II 264; see also Comments, AT VIIIB 356–61/CSM I 302–305).65

Commentators disagree about how to understand the relevant kind of production in these passages, and especially about how to understand the way mind and body jointly produce sensory and imaginative thoughts.66 Some commentators—like Wilson (1978: 211) and Chappell (1994)—argue that sensations ‘arise’ from the union of mind and body through a purely causal process. The body causes the mind to have sensations and imaginative thoughts, and that’s all there is to it. Other commentators—like Rozemond (1998: ch. 6)—argue that the cooperation of mind and body is not merely causal, but includes a metaphysical dimension. As Rozemond writes, ‘the union of mind and body is such that it affects the mind so that it becomes susceptible to a new type of mode of the mind [viz. sensation], a subspecies of thought, which the mind in separation does not have’ (1998: 188). Adjudicating between these options would take us beyond the scope of my current project. All I need is that there is some such sense of production, according to which the mind-body union is singled out as the producer of sensory and imaginative thoughts, while the mind is singled out as the producer of intellectual thoughts. Whatever the best interpretation of this kind of production turns out to be, we can substitute that into my account of the self.

Whereas the self’s variable constitution is explained by the variety of thoughts a single mental life can incorporate, the self’s identity—both at a time and over time—is explained by the fact that we are dealing with a single mental life. Suppose that self₁ is having a visual experience of a goldfinch, while self₂ is having a purely intellectual thought about a triangle. Self₁ and self₂ are identical just in case their respective thoughts are integrated into a single mental life, whether at a time or across time. The unity of a mental life is what justifies saying that we are dealing with one self. To be clear, my proposal is not that every thought corresponds to its own little self, in which case a multiplicity of thoughts would imply a multiplicity of little selves. Rather, a unified collection of thoughts—a unified mental life—grounds a self. We count selves by counting unified mental lives.

Despite his keen appreciation for the complexity of our mental lives, Descartes takes the unity of one’s mental life more or less for granted. When each of us reflects, we discover that all our various thoughts belong to one mental life—a single consciousness or perspective.67 As Descartes writes in Meditation 2:

Descartes suggests that the synchronic unity of a mental life—that is, its being a single mental life—is given in reflection. And presumably he also takes the diachronic unity of a mental life—that is, its being a single mental life over time—as similarly unproblematic. The crucial claim for my reading is that this unity of a mental life, consciousness, or perspective grounds the self’s identity.

My reading is compatible with various substantive accounts of what the unity of a mental life consists in, just as it is compatible with various accounts of the way the union ‘produces’ sensations. All I need is the claim, which Descartes clearly accepts, that mental lives are strongly unified. For the purposes of this paper, then, I will remain neutral on what makes it the case that a given mental life is unified.

Still, what might some of the options be? We might argue that since thoughts are just ways in which a mind exists, the question of what unifies a collection of thoughts into a single mental life is really just the question of what unifies the multiplicity of ways in which a mind exists into a single mind. One possibility—defended by Nelson (2005)—is that Descartes takes the unity of a mental life, or of the mind, as primitive. After all, the meditator says that ‘the fact that it is I who am doubting, and understanding and willing is so evident that I see no way of making it any clearer’ (M2, AT VII 29/CSM II 19, emphasis added). Another possibility—defended by Shapiro (2013)—is that the unity of a mental life is brought about through the psychological relations between its constituent thoughts, most vividly in the case of memory. As Shapiro writes, ‘memory unifies the meditator’s thoughts by effecting a continuity of the awareness of those thoughts. And in providing this unity, memory further helps make the meditator the thinking thing—that is, the self—she is’ (2013: 236).68 Adjudicating between these options would, again, take us too far afield. Whatever the best interpretation of the unity of a mental life turns out to be, we can substitute that into my account of the self as well.

My account of the Cartesian self incorporates the insights of both the re-humanized and traditional readings. The re-humanized reading recognizes that, for Descartes, the self is composed of mind and body—not merely in a figurative or loose sense, but literally. On my reading, the self is constituted by a human being in this life, and, hence, is composed of mind and body. The traditional reading, in contrast, highlights that the self can exist apart from the body. On my view, the possibility of the self’s disembodied existence is grounded in virtue of the facts that (i) purely intellectual thoughts are operations of the mind’s alone, and (ii) it is possible for one’s mental life to consist solely in intellectual thoughts. Thus, my account explains how the self simpliciter is composed of mind and body and yet can exist apart from the body. The ‘self qua human being’ refers to the way the self actually exists, whereas the ‘self qua mind’ refers to the way the self can exist (as well as to one of its actual constituents). By splitting the difference between the re-humanized and traditional readings, my reading does a better job of allowing Descartes to have it both ways.

Before considering an objection to my account, let me clarify one last point about the self’s capacity for disembodied existence. This possibility depends not merely on the distinction between the special modes of thinking and pure understanding, but also on the possibility of purifying one’s mental life of any of the special modes.

Suppose, for example, that someone is currently enjoying a visual experience of a goldfinch, imagining other birds she has seen, and intellectually understands what it is to be a goldfinch. She will then be constituted by those entities that produce the totality of the thoughts she is currently having. These various entities will be woven into a single self because they all contribute to producing a single mental life. In this case, there is overlap. Her visual experience and imaginative act are produced by the mind and body together (i.e., the human being as a whole), while her intellectual understanding is produced by the mind alone (i.e., a part of the human being). So she is composed of mind and body. Thus, if someone has even a single sensory or imaginative thought, then she will be composed of mind and body in that moment. In contrast, if someone could purge her mental life of every sensory and imaginative trace, so that all of her thoughts were purely intellectual, then in that moment all of her thoughts would be produced by the mind alone, and that is what she would be.

Can our mental lives be thus purified?69 Descartes suggests that it won’t be easy. Our mental lives are typically suffused with sensory and imaginative thoughts. We clothe even abstract topics in images. ‘Your imagination insistently mixes itself up with your thoughts,’ Descartes writes, ‘and lessens the clarity of this knowledge by trying to clothe it with shapes’ (Letter to [Silhon], March or April 1648, AT V 138/CSMK III 331). The method of the Meditations can help to some degree. It can help cure ‘the lifelong habit of confusing things related to the intellect with corporeal things, and replace it with the opposite habit of distinguishing the two’ (O/R 2, AT VII 130–31/CSM II 94).70 But this method does not wholly extinguish the sensory and imaginative thoughts in question. It allows us to hold these thoughts at arm’s length, but it does not make them vanish. Even when we are deep in meditation, our stomachs still grumble, we feel the temperature of the room, and the pressure of the floor beneath our feet. We always feel our own bodies. As the meditator writes, ‘I will now shut my eyes, stop my ears, and withdraw all my senses. I will eliminate from my thoughts and images of bodily things, or rather, since this is hardly possible, I will regard all such images as vacuous, false and worthless’ (M3, AT VII 34/CSM II 24, emphasis added; see also M6, AT VII 76/CSM II 54; and July l 1648, Letter to Arnauld, AT V 219/CSMK 356). Hence, the skeptical exercises of the Meditations do not allow the self to exist as, or be constituted by, a mind alone.71

While the method of the Meditations approximates a purely intellectual state, death sets us free. When the body breaks down, the mind’s union with the body snaps (Passions I.6, AT XI 330/CSM I 329). The mind loses its ability to sense and imagine. The imagistic layer of our mental lives melts away. As Descartes writes in a letter to More, ‘the human mind separated from the body does not have sense-perception strictly so called’ (Letter to More, August 1649, AT V 402/CSMK 380; see also Letter to [Silhon], March or April 1648, AT V 138/CSMK III 331). Pure understanding is all that’s left. In death, and perhaps only in death, the self is constituted by the mind alone.72

On my reading, the Cartesian self is what Marshall calls ‘an effect-relative composite’ (2010: 14). An effect-relative composite exists when ‘a set of entities … compose[s] a single individual on the basis of their jointly making a distinctive contribution to a single effect’ (2010: 14). In the case of the Cartesian self, the relevant effect is a unified mental life. Someone might object that effect-relative composites have no place in Descartes’s substance-mode ontology.73 Are effect-relative composites substances? No. Are effect-relative composites modes? No. Therefore, we might conclude, there cannot be any effect-relative composites in Descartes’s world. And that would a bad result for my analysis of the Cartesian self, for if there’s anything we know about the self by Descartes’s lights, it’s that the self exists.

In response to this objection, we might point out that the Cartesian self is not the only effect-relative composite Descartes recognizes. Drawing on Des Chene’s (2001) influential work, I would like to suggest that living things, like plants and animals, are plausibly effect-relative composites as well. The place of living things is not entirely clear in the world of res extensa. In order to identify some configuration of matter as an animal, for example, Descartes needs to have some way of distinguishing the animal from the swirling matter surrounding it. That is, he needs to specify an appropriate principle of unity that tells us which bits of matter are parts of the animal and which bits are not, much as I specified a principle of unity for the Cartesian self. Des Chene puts the point like this:

Common motion and rest—which is the principle of unity Descartes offers for the bodies of his physics—is not a good candidate to explain the unity of living things (Principles II.55, AT VIIIA 71/CSM I 246; see also Principles II.25, AT VIIIA 53–54/CSM I 233). A dog wagging its tail is one living thing, even though the tail and the rest of its body have different motions. Conversely, a dog with a thorn stuck in its paw might move as one with the thorn, but presumably the thorn does not thereby become part of the dog. A better strategy, it seems, is to analyze living things in terms of their ability to produce certain characteristic effects. And that is tantamount to analyzing living things as effect-relative composites.

This is Des Chene’s strategy. Des Chene (2001: 132) argues that Descartes wavers between dispositional and functional analyses of living things. On the dispositional analysis, a collection of material parts constitutes a dog just in case these parts are arranged in such a way that they produce a certain kind of motion, for example, the pattern of motion characteristic of dogs (which includes tail wagging, chasing sticks, and so forth). The functional analysis includes a teleological twist. On the functional analysis, a collection of material parts constitutes a dog just in case the arrangement of these parts is for the sake of producing a certain kind of motion, for example, the pattern of motion characteristic of dogs. Crucially, the dispositional and functional analyses both treat animals as effect-relative composites. Assuming that the broad brushstrokes of Des Chene’s reading are correct, this would mean that Descartes appeals to a notion of effect-relative composites in his biology, and, hence, plausibly does so in other contexts as well. Although other commentators disagree with Des Chene about the details of his reading, they seem to agree that animals are effect-relative composites of some kind: the debate is primarily about how to specify the relevant effect. Brown (2012: 88), for example, argues that Descartes’s animals are constituted by collections of matter that are configured in such a way that they preserve their own configurations. This suggests that the concept of an effect-relative composite has legitimate application in Descartes’s framework.

If Descartes can appeal to effect-relative composites in his biology, then there cannot be any general prohibition on effect-relative composites in his framework. Somehow they have got to fit into the substance-mode ontology, or that ontology needs to be relaxed. This ‘partners-in-guilt’ response does not tell us which of these options we should pursue. I am inclined to follow Normore (2011) in relaxing the framework, so as to allow for composite beings that are somehow grounded in substances and modes. Defending Normore’s approach would take us beyond the scope of the current paper, however. In any case, the comparison to living things suggests that the concept of an effect-relative composite is not foreign to Descartes’s thought, and, hence, that we may employ this concept when giving an account of the self.