We are most fundamentally persons, i.e., thinking things;Footnote 1 but what else are persons? The widely held answer combines the following claims:Footnote 2
1. The concept of person is purely mental.
2. We—persons—have bodily properties.Footnote 3
The first claim expresses a basic Cartesian conception: persons with no physical properties are conceivable. The second claim is grounded in our ordinary practice: statements such as “I’m running as fast as you” are as natural, literal and serious as “I am thinking as fast as you.”
The aim of this paper is to argue from the premise that 2 is a truism to the conclusion that 1 is false: to ratify the Strawsonian claim that our concept of person is of a type of entity that has both mental and physical properties. The argument shows, in fact, that 1 → that 2 is a prima facie mystery, or a highly theoretical discovery, and that any theory that accepts 1 cannot defuse or mitigate that prima facie oddity.
Descartes’ move from 1 to the denial of 2 is considered to be nothing but a slide-over from conceptual into metaphysical separability, or from “separability into separateness”.Footnote 4 Opponents to the move from conceivability to possibility claim that 1 has no metaphysical implications whatsoever. We might just be brains or bodies, although our self-conception speaks against such identities. Opponents to the move from separability into separateness agree with Descartes that 1 implies that we are not identical with any physical entity, but deny that this amounts to substance dualism. The intuition that persons with no bodily properties are conceivable implies, at most, that persons are separable from their body, not that they are separate from it. I am a thing that thinks, certainly, but also a thing that runs. How can I be distinct from my body, and yet have bodily properties? In the same way that the statue is made of the block of marble: Descartes denies 2 on the basis of 1 only because he has not thought about constitution.
My argument avoids the alleged Cartesian fallacies, since I do not argue that 1 implies that 2 is false (as the constitution view argues), or necessarily so (as the identity view argues). My only argument is that 1 implies that 2 is a mystery. Furthermore, I do not base this conclusion on modal intuitions; my argument is not a variation on the “conceivability” and “explanatory gap” arguments against materialism. Utilizing an Anscombian scenario, in which reports on one’s own actions are made on the basis of observation, I argue that such “pilots in their ships,” as it were, are persons according to 1; and yet, for them, 2 is a mystery. The second and crucial stage of the thought experiment is that even if we change their predicament, so that all of a sudden they gain bodily sensations and become virtually like us, their reaction to 2 would remain unchanged. I conclude that the mere fact that we feel in our bodies unlike pilots in their ships, cannot generate the intuition that we are bodily: as long as 1 is true, the fact that we are bodily beings remains an inexplicable mystery.
The charge of mysterianism often elicits no more than a shrug. So what if we cannot grasp the identity between pain and certain brain states? As long as we are careful not to jump to metaphysical conclusions from this acknowledgment of the limits of our cognition, the veracity of the mental-physical identity thesis remains unchallenged. When it comes to persons and bodies, however, we cannot make do with this “so what?” response. “Pain = C-fiber stimulation” does indeed sound odd, or at least highly theoretical; whereas the claim that “I am running” is plain commonsense. If a philosophical doctrine renders such a truism a mystery, there is good reason to abandon it.Footnote 5
Given the heavy weight that is put on the claim that 2 is a truism, I begin by exploring the basis for the intuitiveness of 2. I then clarify the implications of premise 1 in order to show that it is not committed to the Cartesian fallacies (section 2). Section 3 presents the “pilots’” challenge that makes 2 a prima facie mystery. The challenge might provoke an objection from proponents of the constitution view of persons and body. After all, the very idea of identity between a mind and a physical entity is often considered a mystery in and of itself (“How can I be this small parcel of grey matter, wrinkled, moist, and soft?”). But the idea that I am constituted by my body is not the claim that I am this small parcel of grey matter; an appeal to constitution, so goes the objection, can make 1 compatible with the claim that 2 is a truism (“Constitution locates persons and bodies in the material world”Footnote 6). In section 4, I argue that the mysteriousness of 2 does not dissolve when we replace identity with constitution. In the last section, I move to the second stage of the argument that raises the objection that the intuitiveness, or naturalness, of our bodily self-attributions has nothing to do with the truth of identity or constitution; rather it is in virtue of the simple fact that unlike the pilots, we do have bodily sensations.Footnote 7 I argue that the intuitive basis of 2 requires nothing less than the “concept of ourselves and other people as beings who are both corporeal and conscious”.Footnote 8
1. What does 2 Claim, and Why is it a TruismFootnote 9
Let us take a closer look at what is meant by the claim that we have bodily properties. In particular, we must clarify what it is not; that is, we must first rule out one reading of 2.
Descartes agrees that we have bodily properties, and that our bodies have mental properties—in a certain sense. For example, I have the physical property of being connected to something, i.e., a body, whose muscles contract. And my body has the mental property of being connected to something that feels pains, viz., myself. No reasonable dualist would deny that. But Cartesian dualists believe that bodily properties are extrinsic properties, properties I have in virtue of standing in relation to something else. What Descartes denies, and 2 affirms, is the claim that I have intrinsic bodily properties.
For my limited purposes here, I do not have to commit myself to any particular definition of intrinsic property. I rely on an intuitive understanding of the distinction, according to which, roughly, an intrinsic property is one whose possession “speaks exclusively to what X is like in itself, without regard to what may be going on outside of X”.Footnote 10 All other properties are extrinsic, that is, properties that X has in virtue of standing in a relation to something else, outside of it, or wholly distinct from it. Redness, roundness, and having a mass are intrinsic, while being an uncle, and being the fattest in class are extrinsic. To say that we have only extrinsic bodily properties is to reject 2.
The claim that 2 is a truism does not imply that it is accepted by most philosophical accounts of the self, or even that it is accepted by non-dualist accounts. Shoemaker, for one, accepts that we are constituted by our bodies,Footnote 11 but denies along with Descartes, that we ascribe to ourselves intrinsic bodily properties. The claim that 2 is a truism means that it is an integral part of our non-reflective practice of bodily self-ascriptions. Therefore, the truistic status of 2 is explicated in our ordinary thinking and linguistic practice of ascriptions of bodily properties, mainly by three conspicuous facts: to whom we ascribe bodily predicates. How we ascribe them. And what kind of predicates they are. Let me take each fact in turn.
1. The subject of bodily predicates. The most obvious fact about the practice of self-ascriptions is that it does not differentiate between mental and physical predicates. We utter true sentences which ascribe bodily properties to persons, such as “I weigh 135 pounds,” as we do with mental properties. This implies that we ordinarily take 2 to be true—we ascribe bodily predicates to ourselves on a regular basis.
One might, of course, deny that bodily and mental self-ascriptions refer to a single entity, and argue that 2 only seems truistic. But in fact, the argument goes, 2 is vague,Footnote 12 since bodily self-ascriptions are ambiguous. Mackie argues that the reference of “I” is dual, referring to the body in bodily self-ascriptions, and to the mind in mental self-ascriptions. Shoemaker argues that physical predicates are ambiguous, so that “the rock weighs 135 pounds” is straightforward, whereas “I weigh 135 pounds,” is shorthand for a longer expression: “I stand in the embodiment relation to this body, and this body weighs 135 pounds.” Zimmerman suggests that generic exemplification has two species, “really having” and “having by being a soul tied to a body that really has.” These three alternatives—the expression “I” is ambiguous, the physical predicates are ambiguous, the concept of “having a property” is ambiguous—argue that in fact we attribute bodily properties to our bodies, and only extrinsically to ourselves. Appearance to the contrary notwithstanding, we do not really ascribe bodily properties to ourselves.Footnote 13
On the face of it, it is a serious revision of our practice to say: “You aren’t really bleeding; it’s just your body that is bleeding, and you stand in the embodiment relation to your body”.Footnote 14 When I weigh 135 pounds, am I less heavy than a rock of the same weight? But how can one argue for these claims, beyond emphasizing once again that we “really” have bodily properties, indeed, have them “in a strict and literal sense”Footnote 15—i.e., intrinsically?
One source of evidence for the fact that we do ascribe bodily predicates to ourselves in a strict and literal sense, is the flow of inferences such as the following:
I am F
I am G
Therefore, I am both, F and G
On the ambiguity thesis, when F is a mental predicate (“thinking of buying a new apartment”), and G is a bodily predicate (“fat”), the inference (“I am both, thinking of buying a new apartment and getting very fat”) is invalid. A proponent of the ambiguity thesis might convert the original argument to a more complicated one, such as:
I am F
My body is G
Whenever my body is G, it seems to me that my body is G
Whenever I am F, it seems to me that I am F
Therefore, it seems to me that I am F and my body is G
But this argument delivers a much weaker conclusion than the original one, and therefore, it is a form of revision, rather than an interpretation of the simple, original argument.Footnote 16
A second source of evidence is our linguistic intuitions. Stanley forcefully argues that we lack any model into which the ambiguity thesis could fit. The ambiguity in question cannot be understood by using the simple model of “bank,” which contains an ambiguity between “river bank” and “financial institution.” Because in utterances in which the two phrases are conjoined, such as
A. The bank was open for business and filled with swimmers, “bank” can mean only one thing: either the financial institution or the river bank. (A) cannot mean that the financial institution was open for business, and the river bank was filled with swimmers. But if “I” were ambiguous like “bank,” then sentences such as
B. I am both, thinking of buying a new apartment and getting too fat, should be unacceptable, because we could only understand the “I” as having two different senses, which, as (A) demonstrates, is disallowed in such contexts. (B), however, is perfectly acceptable. Therefore, the ambiguity thesis cannot be modeled on the two meanings of “bank”.Footnote 17
A natural model for the ambiguity of “I” could be utterances such as “I am out of gas,” in which the use of “I” seems to be an elliptical for “my car.” But Stanley has shown that in utterances such as “I am out of gas” the use of “I” cannot be an elliptical for “my car.” If it were, then the following two sentences should be on a par:
a. * Both Bill’s car and I are out of gas
b. Both Bill’s car and my car are out of gas
However, (a) is very odd, where (b) is perfectly acceptable, as Stanley argues.
The second version of the ambiguity thesis, which ascribes the ambiguity to physical predicates, might be analogous with utterances such as “I am out of gas.” Proponents of this kind of ambiguity can agree with Stanley that it is not the case that the “I” refers elliptically to my body or to my car, but that the predicates (“weigh 135 pounds,” and “out of gas”) are ascribed directly to my body or to my car, and to me only extrinsically, in virtue of my relation to my body and to my car respectively. However, even if “out of gas” in “I am out of gas” is shorthand for some longer predicate, such as “standing in the relation F to a vehicle which is out of gas,” the analogy with physical predicates fails, because
c. Both I and my car weigh 135 pounds,
is perfectly acceptable, but (a) is not. The fact that (a) is very odd indicates that the predicate “out of gas” has a different meaning than its usual meaning when applied to persons. On the other hand, the fact that (c) is perfectly acceptable means that the predicate “weigh 135 pounds” does not change its meaning when it is self-ascribed. Thus, it should not be ambiguous as is “out of gas”.Footnote 18
The third alternative, which locates the ambiguity in the generic exemplification, might claim that my car weighs 135 pounds in the way physical objects do, that is, by weighing 135 pounds; and I weigh 135 pounds in the way persons do, that is, by standing in the embodiment relation to a body which weighs 135 pounds. Advocates of this third alternative might rely on the following model:
d. Both the mayor and the committee believe that John is guilty.
(d) is acceptable, and it allows that the same predicate (“believe”) can apply in different ways: in the singular to the mayor, and in the plural to the committee. But as Stanley points out, if (d) is acceptable, it is only so because the predicate “believe” is ascribed to the committee: not in virtue of being ascribed to some wholly different objects, as it is in (a); but in virtue of being ascribed to objects which constitute it.Footnote 19 Dualists, who deny that we are constituted by our bodies, cannot provide a truth-preserving interpretation of our physical self-ascriptions.
These facts about our practice support the claim that we ordinarily take ourselves to have bodily properties. But it is not sufficient in itself for supporting the claim that 2 is a truism, since it might be that the basis for self-ascriptions of physical predicates differs from that of mental predicates.
2. The direct ascription of bodily predicates. The second fact about our practice is that we have unmediated grounds for bodily self-ascriptions. Just like self-ascriptions of mental predicates, self-ascription of bodily predicates does not involve an identification component. When I feel that I’m being pushed, I do not receive the information that someone is being pushed, and then identify the subject who is being pushed as me. Rather, I have a general capacity to perceive my body “from the inside.” The capacity of “bodily awareness” can be broken down into several distinguishable capacities—proprioceptive sense, sense of balance, of heat and cold and of pressure, which provide us with awareness of our bodies from the inside. This awareness grants what Evans and Shoemaker call the “immunity to error through misidentification (henceforth, IEM)” of bodily self-ascriptions. Shoemaker characterizes the notion of IEM in the following example:
I can be mistaken in saying “I see a canary,” or (in the case of hallucination) that there is anything at all that I see, but it cannot happen that I am mistaken in saying this because I have misidentified as myself the person I know to see a canary.Footnote 20
The judgment “I see a canary” is IEM-relative to the “I” because it makes no sense to ask, “someone sees a canary, but is it me?”Footnote 21 Bodily self-ascription is also IEM-relative to the “I”: it does not make sense to ask “someone is being pushed, but is it I who am being pushed?”
Shoemaker objects, contending that sometimes it does make sense to ask this question. We can imagine being wired up to someone else’s legs in such a way that I am conscious of the movements of those alien legs as immediately present. In such an abnormal situation there is room for the question: “someone’s legs are crossed, but are they my legs?” Evans replies that if I know about my poor situation, I can infer from my experience that someone’s legs are crossed. But our way of gaining knowledge from bodily awareness does not involve any kind of inference or identification. In any case, since my purpose is in articulating our actual practice, this dispute is irrelevant. The relevant point, with which Shoemaker agrees, is that when we receive information in the normal way, bodily self-ascriptions are IEM-relative to the “I.” This phenomenon indicates that we ascribe to ourselves physical predicates in the same way that we ascribe to ourselves mental predicates.
One might object that there is an alternative understanding of bodily awareness, whereby the phenomenon of the IEM is relative to “my body,” and is not relative to the first-person pronoun.Footnote 22 The Cartesian might argue that bodily awareness is a privileged source of information about this body, not about myself, so that it would not make sense to ask, “someone is falling, but is it my body?” But it does make sense to wonder whether this body is I. If this is the right account of IEM then the phenomenon does not imply that I am presented to myself as bodily, only that I am related in a special way to my body, namely: I am receiving information about this body and not about any other.
Now, if one proposes this reading of bodily awareness in order to show that we do not really ascribe bodily predicates to ourselves in a strict and literal sense, then one has to hold either the first or the third version of the ambiguity thesis. One need argue that the fact that bodily self-ascriptions are IEM only relative to my body, implies either that we do not have bodily properties (only our body); or that we have them, but only in the sense of being related to our body, which has them strictly and literally. But I have already argued that the ambiguity thesis falsifies the phenomenology of bodily self-ascriptions.
3. The nature of sub-group of bodily predicates. A third kind of evidence for the revisionist implications of the ambiguity thesis has not been discussed in the literature so far. The ambiguity thesis, in all of its versions, involves a partition of all predicates to purely mental, purely physical, and a combination thereof. However, an important group of our bodily predicates resists such an exhaustive division. Action-predicates—”hitting,” “running,” “painting,” “cooking,” etc.—are intentional and bodily. Since they are intentional, namely, they imply intentions, they cannot be ascribed to mere bodies; and since they involve bodily movements, they cannot be ascribed to purely mental beings. Therefore, actions form a special group of intentional bodily movements, which cannot be ascribed to either the subject or her body; rather, they would have to be decomposed into a mental component (intending, willing, trying), and a physical component (bodily movements). But this decomposition is itself a revision of our practice, in suggesting that in acting, all we ever do is will (or intend, or try), and “the rest is up to nature”.Footnote 23
That we ascribe action-predicates, together with the rest of bodily predicates, and we do so directly, shows that 2 is a common-sensical claim, not a theoretical discovery like the identity “water is H2O.”
2. What does 1 Claim?
Suppose I tell my daughter that I have a gift for her. She asks what it is. The reply, “something red,” or “something beautiful,” does not answer her question. Though it does say something about the object in question, it does not identify what this object most fundamentally is. If, however, I say “a ring,” then the question has been answered, because now she knows what sort of thing the present is, and in learning what sort of thing the present is, she has learned what it most fundamentally is. Let us call a predicate that gives the fundamental answer to the question—what is it to be that object (a statue, a piece of marble, a person, a ring)—a basic sortal.
“Person” is a basic sortal. The essential characteristic of 1 that makes it Cartesian is that the sortal “person” applies to purely mental beings, be it God, an angel or a human being. Let me elaborate on 1 by considering first its denial, and second, what does not follow from it.
The denial of 1 is Strawson’s claim that the concept of person applies to a type of being with mental and bodily properties. This allows a non-Cartesian understanding of constitution, like that of Wiggins’s, who claims that to be able to conceive of x as a person presupposes that x is conceived to be a human being. Thus, there is an interdependency of the concepts person and human being, which rules out the conceivability of non-human persons. Since all persons are essentially human, no wonder that they are constituted by bodies.
Wiggins’ position can be understood in light of the famous analogy of the statue and the clay. “Statue” is an aesthetic sortal, with different properties than that of “piece of clay”. However, in that case there is no question that David—as a statue—has colour or shape. “Statue” does not apply to things that could be conceived as existing with aesthetic properties alone, without matter. A statue is a sub-kind of material object, as are persons, according to Strawson and Wiggins. Statues have non-aesthetic properties: colour, shape etc., since they are material objects. That statues have weight, for example, simply follows from explicating what statues are.
But given 1, explicating what persons are does not reveal to us their bodily properties. Whatever the defining characteristics of “person” are—an ability to reason, the possession of self-consciousness, etc.—bodily properties are not on the list. The category of person is not a sub-kind of the category of material object, as it is for Strawson and Wiggins. What it is to be a person can be specified without saying that people must or even do have non-mental properties. 1 says that the concept of “person” with no bodily properties is coherent: we can conceive of mental properties without bodily properties; thus, non-human persons are conceivable.
Does 1 imply that non-human persons are possible? Baker answers positively:
To be a person—whether God, an angel, a human person, or a Martian person—one must have the capacity for a first-person perspective. Person is a genus, of which there may be several species: human, divine, bionic, Martian, and so on. It is in virtue of the capacity to have a first-person perspective that a person is a person.Footnote 24
First-person perspective is, as Baker argues, “underlying self-consciousness in all its forms”. The important thing for Baker is that since an angel, or God, could enjoy a first-person perspective, not all persons have bodies. This is not yet the Cartesian intuition that we (human persons) can exist without a body.Footnote 25 Perhaps we can. According to Baker, though, from the claim that “person” is a psychological sortal, it merely follows that there can be instances of this category that exist without bodies.
One may object that even this qualified modal claim does not follow from 1. Though the concept of person is entirely psychological, it does not follow that there could be a person who had only mental properties. Why must the concept of something specify all the properties that beings falling under that concept must have? Recently, Almog suggested a radical interpretation of Descartes, according to which the mind is necessarily united with a body, although the concept of mind is purely mental. Not “all properties necessary for being x are included in one’s idea of x”.Footnote 26
For the sake of argument, I concede that the claim that “person” is a psychological sortal (that is, 1), does not entail that there could be persons without bodies. It may follow only under a certain understanding of the role of sortal concepts, specifically, that they determine identity and persistence conditions. Alternatively, it may follow from a Kripkean move, according to which the conceivability of purely mental beings is a guide to possibility. But I do not assume the cogency of this move from conceivability to possibility. For the purpose of this paper, I do not take 1 to imply that there could be persons without bodies.
Rather, 1 only means that the conception of persons with no intrinsic bodily properties is coherent. Under this reading, 1 is compatible with the claim that persons are necessarily embodied. Moreover, that persons are necessarily embodied can be a conceptual truth, so that persons conceive themselves most fundamentally as necessarily having a body, i.e., they conceive themselves as mental subjects locked inside a body. This does not yet imply that they conceive themselves as bodily beings, i.e., as having bodily properties intrinsically. According to 1, it is not a conceptual truth that persons have intrinsic bodily properties. Even if persons are necessarily “in” a body, and even if they conceive themselves as necessarily “in” a body, they need not conceive themselves as bodily beings, that is, as having intrinsic bodily properties. The conception of a non-bodily subject—a subject who has no intrinsic bodily properties—is coherent, even if it does not reflect a genuine possibility.
3. The Challenge of the Pilots in their Ships
Anscombe imagines a society in which reports on one’s own actions are given from observation. Let us call the members of Anscombe’s society “the pilots in their ships.” The example is of a case in which the persons are completely numb with respect to their bodies; they feel estranged from their bodies because they ascribe bodily properties only on the basis of observation, and hence they can know of “their own” bodies only what anyone else, in principle, can know of them:Footnote 27
Reports of one’s own actions, which are given straight off from observation … are made not on the basis of observation alone, but also on that of inference and testimony or other information. B, for example, derives conclusions expressed by sentences with [“this body”] as subject, from other people’s statements using “B” as subject.Footnote 28
When one observes a fist being clenched, one can tell that the fist belongs to this body, understood as “the body at the front of whose head is the point from which the observed scene is seen”.Footnote 29 This judgment is based on B’s observation, and so can be mistaken insofar as the fist may actually belong to a body other than B’s.
By 1, mental life without bodily properties is conceivable, so let us conceive of such mental beings, by conceiving these pilots as ascribing only mental properties to themselves using the first-person pronoun. That is, let us imagine persons just like ourselves in their relation to their bodies, except for the mode of ascription of bodily predicates. One cannot argue that this scenario is inconceivable—if persons with no bodies at all are conceivable, why could we not imagine persons who only differ from us in the way they self-ascribe bodily properties? By 1, the pilots are persons, just like us,Footnote 30 and they use “I” to ascribe psychological predicates to themselves. Unlike us, they use “this body” to ascribe bodily properties to their bodies, on the basis of observation, testimony, and the like.
Nevertheless, metaphysically speaking, the pilots relate to their bodies just as we relate to our bodies—the import of imagining the pilots is that they differ from us only in the way they ascribe bodily predicates—the fact that they ascribe bodily predicates on the basis of observation alone. Surely this difference in the mode of ascription of bodily predicates cannot constitute a metaphysical difference, so let us stipulate that they are related to their bodies as we are to ours. Hence, according to any version of materialist theory—identity or constitution—they have bodily properties, and should ascribe bodily properties to themselves.
Unfortunately, they do not. It is not just that they think that 2 is false; what is more, the claim that these properties are theirs is a complete mystery to them, as is the suggestion that “being out of gas” is a property of mine, rather than of my car. This is not a stipulation; rather, I submit, the fact that they would claim not to understand what it is for them to have bodily properties is the intuition of this thought experiment, if there is one. This is because the body is a mere vehicle for the pilots, as our car is for us. It should be emphasized that the point is not about their preferred linguistic expression. The pilots could use the expression “this body,” or “a body,” or even may utter “I am running,” from time to time. Whatever the preferred linguistic expression, the intuitive prediction is that for them, it would be no more than our “I am out of gas.” They would claim not to understand what it is to self-ascribe the property “running” in any way other than extrinsically.
Some people argue that there are actual counterexamples to the supposed intuition that the pilots would find 2 to be completely mysterious. However, it is not so easy to actually make such a case. It is not enough to claim that people with all sorts of sensory deficits learn to navigate around the world using pilot-like observations. What is required is a case of someone with no bodily sensations at all, maybe one like the patient depicted in the movie “The Diving Bell and the Butterfly,” who is almost completely physically paralyzed but can ascribe physical properties to himself. The undoubted fact that the patient thinks of himself as a human person cannot, however, be used to show that the pilots might be able to do so as well. The main difference is that in this extremely rare condition, the patient has genuine memories of his past experience as a bodily being; the pilots have none. As far as I know, there is no known case of complete bodily blindness from birth.
Let me emphasize how modest my claim is at this stage. I do not claim that the pilots could not be aware of their metaphysical relation to their bodies, be it identity, constitution, or supervenience. Knowing this, one may wonder, why should the lack of feeling, and hence the reliance on observation for the ascription of bodily properties, make the fact that they have intrinsic bodily properties mysterious to them?Footnote 31 There is no reason why they should not know that they are identical to or constituted by their bodies, in the same way that they know that water is H2O, or that pains are C-fiber stimulation.
Well, I certainly agree. My point is that to the extent that the pilots would be able to understand the metaphysics of their situation, it remains for them a highly theoretical discovery, precisely as the supposed discovery that pains are C-fiber stimulation is for us. The thrust of my argument is that 1 is incompatible with the fact that we regularly and intuitively ascribe bodily predicates to ourselves. If all that can be established is that 2 is a theoretical discovery, which is not part of regular self-attributions, then this is a good reason to abandon 1.
An objection from the constitution theorist might suggest itself. The very idea of identity of mind and body implies that persons are material. That is why it cannot mitigate the mystery of 2 for the pilots; the identity itself is a mystery to them, or at best, a highly theoretical discovery. But constitution is not identity; once we retreat to the model of the statue, so goes the thought, the mystery disappears. All the pilots need is to understand that the mind and the body are “as tightly integrated as Goliath Statue and Goliath Atoms”.Footnote 32 Once this understanding becomes part of their unreflective practice, 2 becomes for them the intuitive claim it is for us.
4. The Constitution Reply
In order to explain how the appeal to constitution can mitigate bodily self-ascriptions for the pilots, constitution should be a genuine unity relation that sustains “mutual borrowing of properties”, to use Baker’s terms. Indeed, Baker claims that to say “I am a fast runner” is to say “I am constituted by a body that can run fast.”:
Strawson was right that states of consciousness are ascribed to the very same things as corporeal characteristics, and that thing is the person. But on my view, a person has corporeal features derivatively and first-person-perspective states non-derivatively.Footnote 33
Derivative properties are not extrinsic ones, because the body is not something “outside of me,” but a part of me. Understanding that they are constituted by bodies, the pilots might come to accept 2, and it might become second nature to them to derivatively ascribe bodily properties to themselves.
I doubt that the properties so derived from their bodies would become, for them, intrinsic properties; the very idea that they are constituted by their bodies cannot generate a new practice in their society. In order to understand why, we have to consider what constitution is.
It is one thing to claim that constitution is not identity, and another to spell out what it is. All the detailed accounts of constitution accept the following as necessary conditions:
i. Spatial coincidence—x constitutes y at t only if x and y occupy the same spatial region at t.
ii. Material coincidence—x constitutes y at t only if the underlying matter of x and y is the same at t.Footnote 34
Condition ii ensures that constitution is defined over individuals which are not composed of immaterial stuff (it cannot exist, say, between the body and the Cartesian soul, or ectoplasm). It implies that x and y are determined by the same underlying matter. But i-ii comes to mere coincidence, and a satisfactory answer must appeal to something weightier than “mere” coincidence, if it is to allow the pilots to have bodily properties. What else is needed, then, for a plausible account of constitution?
In the literature there are a number of attempts to make constitution a relation that can account for the fact that two distinct things, one of which is made of the other, share properties. All accounts emphasize the dependence of the constituted object on the constituting matter. (1) The parthood (or mereological) analysis requires that “at some level,” x, the constituting matter, and y, the constituted object, are made out of the same parts.”Footnote 35 (2) The destruction analysis requires x to be a potential “substratum” for y’s destruction. (3) The explanatory analysis requires that the fact that x is F at t explain the existence and persistence conditions of y at t.Footnote 36 (4) The modal analysis sets two requirements: x and y must belong to different basic sortals, and under special circumstances (defined relative to the different sortals in question) that exist at t, x necessitates y. For example, where x is a body and y is a person, it is necessary that for anything that has the property of being a human body as its basic sortal at t, and is in intrinsic and environmental conditions at t that are conductive to development and maintenance of first-person perspective, there is some spatially-coincident thing that has the property of being a person as its basic sortal at t.Footnote 37
Let us grant that the constitution relation—R(x,y)—satisfies certain spatial, mereological, parthood, modal or explanatory conditions. The appeal to constitution cannot make 2 for the pilots the truism that it is for us, due to the following dilemma. Either R(x,y) is understood to be a substantive relation, modeled on the relation between the statue and the piece of marble, or it is understood very weakly, as amounting to no more than a modal claim. Call the first alternative “the substantive account of constitution,” and the second, “the deflationary account of constitution.” If constitution is understood substantively, it is mysterious or theoretical for the pilots, just as the identity thesis. If it is understood to be deflationary, it cannot make 2 a truism for the pilots.
Let us begin with the substantive relation. The statue does not merely supervene on the piece of marble. It is composed of it, and the property of being composed of matter is essential to what a statue is, that is, a material thing (though it is not essential to the statue to be composed of this piece of matter). True, the statue is most fundamentally an aesthetic sort of thing, but it is also a sub-kind of material object. As I explained in section 2, statue does not apply to things that could be conceived as existing with aesthetic properties alone, without matter. Statues have non-aesthetic physical properties: weight, colour, shape etc., and these properties are shared with the matter which constitutes them. This account of R(x,y) implies that persons are material. But the pilots would regard the substantive claim that they are literally composed of material parts or of an aggregate of particles, in the same way that we would regard the absurdity that we are made of steering wheel and motor (or at least, as a piece of theory). Being so estranged from their bodies, as they certainly are, they cannot rid themselves of the intuition that they are not material. Thus, substantial constitution is mysterious for them. The intuition vis-à-vis this scenario is, I submit, that the pilots’ thought that they are material, and accordingly, any account which is substantial enough to generate this thought, is for the pilots what the thought “pains are C-fiber stimulation” is for us. Or worse, still, it is a mystery to them as the thought that “I am out of gas” would be to us, if we were asked to take it “literally and seriously.”
On the other hand, the pilots would be more than happy to accept constitution of the purely modal variety, which says merely that two particulars stand in some necessary correlation or supervenience, however strong. But the problem is that such an account cannot grant the truth of 2, let alone make it obvious to them. The pilots might admit that they supervene on their bodies, even by necessity, and yet insist that they have no intrinsic bodily properties.Footnote 38
The temptation to shift from identity to constitution in order to overcome the intelligibility problem rests on a conflation of two possible notions of constitution. If the constitution relation is substantive enough to imply that the pilots are material, then it is itself mysterious or theoretical for the pilots as the identity is. If the constitution relation is deflationary, than it cannot make 2 the truism it is for us.
Now, the pilots are persons by 1, and constituted (whatever that means) by their bodies; yet, the intuition is that they would find 2 to be wholly mysterious. Therefore, the truth of constitution cannot be the explanation of why we find the claim that we have bodily properties to be so intuitive. Indeed, as I argue in the first section, 2 is intuitive for us in virtue of facts related to our practice of bodily self-ascriptions.
But then, why should we care about the pilots at all? We would not—or need not—take the intuitiveness, or naturalness, of our self-attributions of bodily properties to have anything to do with the “truth” (if it is one), or the belief, that we are constituted by our bodies. Unlike the pilots, we do have bodily sensations and kinesthetic experiences that—as a matter of psychological fact—directly prompt us to believe that we have various bodily characteristics. And as argued in the first section, this seems to explain sufficiently why we find it intuitive, natural, and automatic to attribute bodily properties to ourselves. Why should it matter to us that those beings find some fact about themselves mysterious? Being in our bodies not as pilots in their ships, why should we not believe that 2 is obviously true?
5. What is the Difference between the Pilots and Us?
The answer is that given 1, bodily awareness cannot be an awareness of the self, and hence, cannot make 2 intuitive. The strategy of arguing for this claim is to fill in the pilots’ scenario with whatever was considered to be the grounds or basis for the truistic status of 2. My claim is that no matter how we enrich the pilots’ society, as long as 1 is true, 2 cannot become a truism.
The first evidence for the intuitiveness of 2 was the use of “I” in ascriptions of bodily predicates. Let us mandate such a use in the pilots’ society, switching “this body is running” to “I am running” regularly. Let us also imagine that the pilots are happy to cooperate, and correct each other so as to establish the new practice. I argue, however, that they would consider this change as merely terminological. For them, the ambiguity thesis would be the intuitive and natural understanding of their new practice, on the model of “bank.” For them,
B. * I am both, thinking of buying a new apartment and getting too fat, would be very odd, as would
A. * The bank was open for business and filled with swimmers, if it is understood to mean that the financial institution was open for business, and the river bank was filled with swimmers. Similarly, for them
c. * Both I and my car weigh 135 pounds, would sound as odd as
a. * Both Bill’s car and I are out of gas.
This is because, for them, the body is a mere object to which they are related, and this fact cannot be changed by stipulating a new linguistic convention. Provided that the first stage of the thought experiment established that 2 cannot be part of their practice, nothing improves at this stage, except the way they speak: they still have to observe in order to see that they are being pushed on a swing, and so, for them, “I am being pushed on a swing,” is what “I am out of gas” is to us.
Let us then enrich the pilots’ world in a way that brings it closer to our own. Let us therefore endow them with all of our sensory capacities—proprioceptive sense, sense of balance, of heat and cold and of pressure—the capacities that are part of the ground for the intuitiveness of 2, as explained in section 1. Those capacities make us feel unlike pilots in their ships. Now that the pilots become virtually like us (indeed, according to proponents of 1, they are identical to us), how would they react to their new condition?
I claim that their reaction would be as follows: “Admittedly, we now feel very comfortable in our bodies; it certainly feels not like being in a ship. But merely feeling that you are in your body not like a pilot in a ship, is not feeling that you have bodily properties. “If we felt a special hot feeling every time our car was out of gas, would we better understand the claim “I am out of gas” (that is, understand it as an ascription of an intrinsic property, rather than as a property we have in virtue of standing in a certain relation to our car)? What changed is only our evidence for bodily ascriptions, nothing else. How the fact that we can feel that this hand is raised, when before we only saw it, can make sense of the claim that now I am raising my hand, rather than my body?”
One might object that this reaction would fit, at most, the Cartesian philosopher, and it would not be the unreflective common-sensical view. After all, their bodily ascriptions are now IEM, and this, I argued in the first section, is an indication that 2 is part of our practice, so why wouldn’t they learn to self-ascribe bodily predicates in this natural way, once they do not have to observe the movements of their bodies?
I agree that bodily self-ascriptions would be IEM, but not relative to the “I”; rather, for the pilots, they would be relative to “this body” (it would not make sense for one of them in her new situation to ask: “someone’s body is falling, but is it this body?”). Correlatively, I argue that their new bodily self-ascriptions would not be strictly and literally self-ascriptions. In other words, the new variety of sensory capacities may be a special awareness of their bodies, but it would not amount to an awareness of their selves. To understand why, let me explain why the phenomenon of IEM of bodily self-ascriptions relative to the “I” is considered to be a form of self-awareness; that is, why it is taken to show that I am presented to myself as bodily, as I argued in section 1.
Evans’ point is that the phenomenon of IEM shows that we are presented to ourselves as bodily because the object we refer to in “I-judgments,” that is, the self, is determined in part as the one with which we are informationally linked, as the source of the information we get in these special ways. Since we receive bodily information in this special way, it follows that the self is the source of bodily information. To see why the object we refer to in such judgments is determined as the one with which we are informationally linked, compare a visual experience in which the judgment “This object is red, round, and on my left” is IEM-relative, to the use of the demonstrative. Why does it not make sense to ask: “something is red, round, and on my left, but is it this object?” Because the object I am referring to, namely, “this one” is partly determined as the source of the information I am acquiring (of being red, being round and being to my left). It is not the case that I have an independent identification of the object in question which does not include that it is red, round and to my left, so that there is an intelligible question whether the independently identified object is red, round and to my left. Similarly, it does not make sense to ask: “someone’s leg is crossed, but is it my leg?” because the object which is singled out—I—is determined in part as the source of the bodily information triggering such judgments. If the self is the source of the information in all judgments which are IEM, then it is a bodily self. As Evans argues,
There just does not appear to be a gap between the subject’s having information … that the property of being F is instantiated, and his having information … that he is F.Footnote 39
In bodily awareness, one is not merely aware of one’s body as a mere body (“this body,” or even as “my body”), but one is aware of it as oneself: as a bodily self. The source of bodily information is a genuinely mental subject. The fact that bodily self-ascriptions are IEM-relative to the “I” is thus “the most powerful antidote to a Cartesian conception of the self”.Footnote 40
The pilots, however, have their own self-conception from their pre-revolution times; that is, they have a complete conception of themselves as purely mental beings. The pilots’ selves are not determined, even in part, as the source of their bodily information. Their selves are completely identified independently of their bodily information. No matter how smoothly and directly the new bodily information is now being acquired, they cannot take its source to be their selves. The source of bodily information must be their body. For them, there should appear to be a gap between the subject’s having the information that F is instantiated (when F is a bodily predicate), and her having the information that she is F. Let me emphasize that their natural reading of the IEM is relative to “my body.” For the pilot, it would not make sense to ask “Someone is falling, but is it my body?” although it would still be a mystery for her, how she could be falling, rather than her body. The problem is that for the pilots, even in their new condition, there are two sources of information for “I-judgments.” The self is the source of mental information, and the body is the source of bodily information. The pilots cannot take the IEM of bodily self-ascriptions to be relative to the “I.” For the pilots, the IEM of bodily ascriptions does not show that they feel themselves as bodily subjects—they do not—only that they feel their bodies in a very special way. And if the conception of themselves as bodily subjects was a mystery (or highly theoretical) before the acquisition of the new sensory capacity, the IEM relative to “my body” cannot dissolve it; indeed, the question is: “My body is falling, but how could it be me who is falling?” Their practice would still be committed to one or another version of the ambiguity thesis.
The thought experiment of the pilots is only an effective means to understanding the implications of 1. According to advocates of 1, we are just like the new pilots. And this means that if 1 is true, that is, if we have a complete conception of ourselves as mental beings, the IEM of bodily self-ascriptions is only relative to “my body,” not relative to the first-person pronoun, because the self is not determined as the source of the bodily information. Given 1, the ambiguity thesis, and the understanding of IEM as relative to “my body” should be integral to the practice, and constitute the natural reading of it, not a form of revision. Although we have a special access to our bodies, for holders of 1, this cannot be a form of self-awareness. The conceivability of the pilots makes it vivid, because even when they are freed of their unfortunate situation, they cannot but retain their self-conception, and this self-conception implies that their “I-judgments” are based on two different sources: a person, the source of their mental properties, and a mere body, the source of their bodily information.
What differentiates us from the original pilots, according to advocates of 1, revolves around how we feel our bodies, and feelings alone cannot make a cognitive difference. As long as 1 is true, so that we, like the pilots, think about ourselves most fundamentally as purely mental beings, the fact that we feel in our bodies unlike pilots in their ships cannot generate the intuition that we are bodily. I conclude that given 1, the phenomenological difference by itself cannot dissolve the mystery. It can only serve to restate Descartes’ insight that we are not in our bodies as pilots in ships, but not to explain in virtue of what we ascribe to ourselves bodily properties ordinarily and regularly.Footnote 41
There is one last move for an advocate of 1: to admit that 1 implies that the IEM of bodily ascriptions is relative to “my body,” but maintain that this phenomenon, backed up with the constitution (or the identity) view, is compatible with the intuitiveness of 2. Baker, for one, might argue that we ascribe bodily properties to ourselves derivatively, by virtue of being constituted by their bodies. Yet, these bodily self-ascriptions are literally and seriously ascriptions to the self. The fact that bodily self-ascriptions are IEM-relative to “my body” just means that we have bodily properties derivatively, but we have them nonetheless. Recall Stanley’s point about the following sentence:
d. Both the mayor and the committee believe that John is guilty. Stanley argues that (d) is acceptable because the predicate “believe” is attributed to the committee not in virtue of being ascribed to some wholly different objects, but in virtue of being attributed to objects which constitute it. Baker might argue that indeed, our practice is just like that of the pilots-after-the sensory enrichment: we self-attribute bodily predicates derivatively, in virtue of the fact that we are constituted by our bodies, but we do so most naturally.
However, in conceding that bodily self-attributions are IEM only relative to “my body,” one revises ordinary bodily self-ascriptions. It might be helpful to go back to the pilots in their original state in order to understand why. A striking fact about the original pilots is that they did not ascribe to themselves action-predicates. Their conceptual repertoire lacked genuine action predicates. They ascribed to themselves intentions, desires, states of trying, but not any predication which involves doing something, “which indicate a characteristic pattern, or range of patterns, of bodily movement. … Such things as ‘going for a walk’, ‘coiling a rope’, ‘playing ball’, ‘writing a letter’….”.Footnote 42 The important point about action-predicates, as I explained in the first section, is that they are intentional and bodily, and as such cannot be ascribed either to the pilots, or to their bodies. The original pilots ascribed to themselves the intention to run, and to the body, the movements of the leg. But since the movements of the legs are known only observationally, not through the intention, these movements could not be conceived as “their actions.” All they ever did was will things to happen; the rest was up to their bodies.
How is the situation improved once they gained the new sensory capacities? All their bodily ascriptions are IEM-relative to “my body,” that is, they ascribe bodily predicates to themselves derivatively at best. How would they ascribe action-predicates, such as “running”? The pilot would feel that he “himself,” that is, the person, intends or decides to run, and now that he gained bodily awareness, he can also feel something else; namely, he can feel the movements of the body. But what he feels is movements of his body, not of himself. He feels his legs “doing” the movements. True, now that he has sensory capacities, he does not have to observe in order to discover that his body executes the order: he feels that immediately. But this is not to feel that he is running—he feels that he is deciding to run just now, and he feels his body starting to run at this very moment. As before, the pilots are bound to ascribe to themselves only the mental component, the willing or the intention. This decomposition serves as a constant reminder that they are not, in and of themselves, bodily, although they do not feel like pilots in their ships anymore.
The moral from the pilots’ tale is that actions require a genuine concept of a bodily subject, not a composition of self and body, which means the falsity of 1. Advocates of 1 must hold that we ascribe bodily predicates to ourselves derivatively at best, as do the pilots, and therefore persons cannot be the direct bearers of action-predicates. But the body cannot be the direct bearer of action-predicates either, since they imply intentions. As long as 1 is true, we would not be able to make sense of the third ground for the intuitiveness of 2: the self-ascription of action-predicates. 1 implies that we have two independent sources of information: the self, or the person, which is the direct bearer of mental predicates; and the body, which is the direct bearer of bodily predicates. Action-predicates must be derived from two different sources, and thus be decomposed into mental and physical components. If 1 is true, we lose our status as agents which do things like writing, running or jumping. We, like the pilots, only give the orders, and we feel our body following them.Footnote 43
The answer to the last objection is, then, that if persons are bodily derivatively at most, and their bodily self-ascriptions are IEM only relative to “my body,” then they are not genuinely agents; that is, they have no room for action-predicates which are bodily and intentional. Actions are intuitively intentional and bodily, that is why we must give them pride of place.Footnote 44 A genuine concept of person that is irreducible to subject and body, is possible only with a genuine kind of property that is irreducible to mental and physical properties. I think that this is what Strawson has in mind when he stresses that the concept of person is primitive: it comes with a distinct group of predicates, action-predicates, which can be meaningfully applied only to members of that category. The concept of person is the correlative of action-predicates.Footnote 45
