Much of the perplexity that motivates modern discussion of the nature of mind derives indirectly from the striking success of physical explanation. Not only has physics itself advanced at a remarkable pace in the last four centuries; every hope has been held out that, in principle, all science can be understood and ultimately studied in terms of mechanisms proper to physics. Seeing all natural phenomena as explicable in terms appropriate to physics, however, makes the mental seem to be a singularity in nature. Chemistry and biology may well be reducible to physics, but the same seems hardly possible for the mental.
The gulf etween mind and physics seems too great to bridge, and the success of physics guarantees its standing. The place of mind in nature is thereby rendered problematic. This line of reasoning has tempted thinkers since Descartes to see the mind as not only independent of other natural phenomena, but as even somehow lying outside the natural order itself. A variety of particular problems about how the mental fits with the rest of nature have been widely discussed in recent years. Less often noticed, however, is that similar problems appear to affect our understanding of the concept of the self in relation to the natural order.
For something to be, or have, a self, two conditions seem intuitively necessary. There must be some sort of unity in the mental life of that being. In addition, to have or be a self, one must be distinguishable from other beings, and in particular from other beings of the same or of a relevantly similar sort. Intuitively, nothing can be a self unless it functions in some suitably unified way and unless there is a reasonably clear contrast between it, considered as a self, and other things distinct from it.
It is perhaps plausible to see the conceptual resources of modern physics as adequate to explain the physical nity and individuality of the macroscopic objects around us, including the biological integrity of living organisms such as ourselves. But it may well seem that, however well physics can accomplish those tasks, its conceptual resources are simply insuf- ficient to explain the special functional unity and individuality involved in something’s being a self.
The difficulty emerges especially vividly when we contrast the way a common sense view of things represents the place of the self in nature with the way this might be represented by the view of the world suggested by physics. On the common sense view, the orld consists of macroscopic objects of various kinds, which behave in the ways they do because of their intrinsic properties. This sort of view leaves it open to regard those organisms which have mental abilities as simply special sorts of organisms.
There may still be a problem in saying exactly what mental abilities are and what properties and capacities are special to such organisms. But the problem is not that of explaining how the mental can arise and exist in nature, but the problem of articulating just what the mental is and trying to determine what is special to those organ- isms that have mental states. On a common sense view of this sort, the special unity and individuality that characterize selves also poses no special problem. Organisms generally exhibit a remarkable degree of functional unity because of their biological makeup.
This functional unity results in a clear contrast between individ- ual organisms and their surrounding environment, including other individuals. Organisms that have more elaborate abilities and properties will function in ways that are more distinctively unified and more effectively resist impingement from the environ- ment. Such organisms will therefore be selves to a greater degree han less elaborate organisns. Various psychological and social abilities will presumably be the culmination of such functioning.
By contrast, modern physics conceives of natural phenomena as due to the behavior and interaction of nonmacroscopic entities governed by mathematically formulable laws. Such laws must gener- ate and explain the relative autonomy of biological organisms by appeal to the nomological behavior of interacting parts, ideally at the biochemical and biophysical levels of analysis. The emergence of selves among the furniture of physical reality needs special xplanation, therefore, as does the existence of mind itself. Many inanimate objects exhibit functional unity and individuality, and yet are not selves.
The common sense view accounts for this by appeal to the elaborateness and kind of unity of function. But from the viewpoint of physics, distinctions of elaborateness and kind of functional unity will at best seem artificial and ad hoc. The attempt to explain the functional unity and individuality of selves on the basis of the intrinsic properties of macroscopic objects may strike one as too easy; the concept of the self seems lmost to be built into the common sense view of things. But by the same token the attempt to explain such unity and individuality in terms proper to physics may well seem simply hopeless.
It is natural to conclude that the self cannot be understood at all on the basis of physical consideration, and must instead be accounted for in terms of the mental. If neither mind nor self can be located within the natural order, the temptation will be great to understand each in terms of the other. Moreover, only the mental seems available as an explanation of the self once the physical has been excluded. The most dramatic example of this line of reasoning is Descartes’ identification of the self with a being whose essence is simply that it thinks.
But an account of the self in terms of the mental need not invoke mental substances; it can rely instead on mental states. Though avoiding commitment to mental substances has evident advantages, it is far from clear that an account of the self cast solely in terms of mental states can succeed. In section I, I consider the three most common approaches to such an account, and argue that each fails. These approaches rely, respectively, on ropositional mental states, sensory mental states, and desires.
In section II, I turn to an examination of the intuitive connection that holds between being a self and being a center of conscious ness. I argue that this Cartesian idea also cannot do justice to our conception of the self. Finally, I argue in section III that a better account of the self can be constructed that is based not on propositional, sensory, or desiderative mental states, but on the emotions. And I urge that the close connection that holds between the emotions and our concept of the self indicates both why a urely mental account of the self is likely to fail and what sort of account holds promise for success.
I. Among attempts to account for the self on the basis of mental states, that which conceives of the self in terms of propositional mental states seems to many to be the most intuitively compelling. Propositional mental states–\”thoughts,\” for short–can be charac- terized in two distinct respects: their propositional content and the mental attitude held toward that propositional content. Nothing about the propositional content of thoughts seems to give us any special grip on the concept of the self.
Even if your ental states are distinctively yours and mine are distinctively mine, the propositional content of our thoughts is that aspect of them which can be shared. If I say that p and you understand me, we have thoughts that have the same propositional content: I, if sincere, believe that p and you understand me to say that p. And if you believe me and I was sincere, we end up both believing that p; what we believe, that is, the propositional content of our beliefs, is the same, even though we each have our own occurrent and dispositional beliefs.
Propositional content does not help us distinguish among selves; it elps us bridge the gap between them. The mental attitudes one holds toward the propositional content of thoughts, by contrast, does intuitively reveal a connec- tion between mental states and the self. Typical of such attitudes are believing, doubting, suspecting, wondering, guessing, disbe- lieving, anticipating, and assuming.
These and other attitudes provide a kind of mental frame for the propositional content of one’s thoughts, analogous to the illocutionary force of speech acts. Just as the illocutionary force of an utterance determines what sort of speech act it is, so the mental attitude of a thought etermines the kind of mental act it is. Propositional content is what can be shared among distinct mental acts; my mental acts at different times, my mental act of surmising now and my mental act of being confident later, your mental act of doubting and mine of affirming, all these can have their propositional content in common.
The distinct types of mental attitude, and the distinct occasions of their being held serve to differentiate these distinct mental acts. Without some suitable mental attitude, however loose or noncommittal, no mental act is possible; propositional content lso can supply no more than the object of that attitude. These considerations suggest a particular conception of the self. On this view, a self somehow unites thoughts, by holding a variety of mental attitudes towards various propositional contents.
The self is what holds the mental attitudes and, because it holds various attitudes, they are unified in a single functioning mind. Without such a self, mental acts would simply exist on their own, unrelated to one another except to the extent of sometimes sharing their propositional content. In this situation, no two thoughts would be more closely united than any others. Accordingly, we would be unable to distinguish between somebody’s having two contradictory thoughts and two beings having thoughts that contra- dicted one another.
Our ordinary conception of mental acts rules out their existing in this disconnected way. But to make sense of that ordinary notion we must presuppose that selves exist and that they are able to unite mental acts by underlying the relevant mental attitudes. The foregoing argument may convince us that an acceptable account of mental acts requires some notion of a unifying self. It is less clear that the argument helps us develop a conception of uch a self, or understand what a self is. Rather, the needed conception of the self must be supplied from some other source.
The inability of the argument to help here becomes obvious when one asks what would be left out of an account of mental acts that made no reference to the self. It is difficult to see how to avoid giving the question-begging answer that what is left out is the unity of mental acts with respect to distinct selves. In addition, the foregoing argument sheds little if any light on what sort of thing a self is. Presumably, considerations of the sort just advanced convinced Descartes that the self is a hinking thing.
But, as Hobbes, in effect, pointed out, overt actions presuppose unification in a self no less than mental acts. 2 An act of walking belongs to a self no less than a mental act of doubting; it may or may not be the same being that walks and doubts, or walks and eats, just as it may or may not be the same self that doubts and suspects. Argument would therefore be neces- sary to establish the privileged position of mental acts in regard to the nature of the self. It is not easy to see what sort of argument could accomplish this other than one that establishes a distinction between mental and physical substances.
For simply regarding mental states as nonphysical does not warrant seeing the self as more centrally bound up with mental activities than with overt actions such as walking. 3 An account of the self based on the need to unify mental acts also fails to be informative about how a self could do such a thing, and what sort of thing a unifying self would have to be. This difficulty does not affect a conception of the self based, at least in part, on the need to unify overt actions. An act of walking belongs to the same agent as an act of eating, say, because both involve a single body.
Short of a commitment to unextended, hinking substances, no parallel move is possible with mental act. Moreover, it is unclear that even the postulation of unextended substances would help. Such substances are definitionally designed to be substrata of mental states and, in particular, for thoughts. But simply saying that a particular entity discharges some function does not itself make it intelligible how that function is dis- charged. Eating and walking can involve a single body by involving movements of that body.
The idea that unextended substances can unify distinct mental acts is not modeled on such unity of walking nd eating, but on the unity that a physical substratum is thought to confer on distinct properties of physical objects, such as shape and color. But such unification by substrata does not explain any unification that is otherwise in need of explanation. Such considerations presumably are the sort that led Kant to conclude that the unity of our mental acts does not involve any empirical tie among them, but is rather a transcendental unity.
The difficulty in understanding how a self could succeed in unifying mental acts, and what sort of entity such a self would have to be, causes problems for the individuating of selves, so onceived. If it is unclear, for example, how my mental acts are bound together, then it will remain at least as unclear why yours and mine are bound into two distinct unities. The unifying and individuating functions of the self are thus correlative. Without some clear account of why your mental states are not unified with mine, I cannot understand why we are distinct selves.
For a conception of the self based on mental states allows no other method of individuating. I return to this difficulty below. The effort to understand the self as that which binds mental acts together has a distinctively a priori flavor. We do ot, on this conception, encounter the self in anything we experience, but must instead presuppose it to have an adequate grasp of our thinking. Largely because of dissatisfaction with this kind of a priori model, many post-Cartesian thinkers have sought to isolate the mental determinants of the self not in our propositional mental states, but rather in our perceptual and bodily sensory states.
Sensory states not only differ from propositional states in lacking conceptual content; they also differ in having, unlike thoughts, some qualitative or phenomenal character, such as the redness of a visual sensation or the painful haracter of pains. Both thoughts and sensations are described in ways that parallel our descriptions of particular sorts of nonmental states and events. Both mental acts and speech are described by means of propositional clauses governed by verbs of mental attitude or illocutionary force.
Our descriptions of sensory states, on the other hand, parallel the nonmental terminology we use to talk about the perceptible properties of physical objects and processes, and perceptible events that affect our bodies. Bodily sensations, such as pains, can be throbbing, stabbing, or dull, just as wounds to our bodies can be. Perceptual sensations can be red, triangular, loud, or salty, just as the perceptible physical objects and processes are that, in standard circumstances, cause corresponding sensations. For present purposes, the ordinary distorting effects of perspective, distance, and intervening medium can be disregarded.
Likewise, difficult problems about the precise status of properties such as color, sound, smell, and taste, when these are conceived of as properties of physical objects, can here be set aside. Though we can correctly characterize both physical objects and sensations in sensory terms, for example, as being red, the edness of a sensation is not the same property as the redness of a physical object, but a mental analogue. A table can be red, at least in common sense terms, and its redness is a first-order property of the table. But a sensation is not an entity, on a par with a table or a person; it is rather a state of a sentient being.
So if a visual sensation is red, the redness is a second-order property; the sensory state itself is the first-order property of the sentient being. Moreover, if somebody has a sensation of red, there need be nothing relevant in the sentient being that has the physical color of red. The relations between physical and mental qualities are complicated and notoriously difficult to articulate. But even if some reductionist effort to define one kind of property in terms of the other can succeed, it is important that, short of some such reduction, the two sorts of property are distinct.
Like thoughts, sensations are intuitively tied together, and it is tempting, therefore, to suppose that the self is somehow responsible for this tie. But the unity of sensations seems also to be less abstruse and more readily accessible to analysis than the unity of mental acts. Visual sensations, for example, are nified by virtue of their belonging to a single visual field, and similarly with other sensory modalities. Moreover, if a red and triangular sensation belongs to a particular visual field, intuitively this is because that sensation bears discernible sensory relations to other sensations in that visual field.
For example, the sensation may be visually adjacent and to the left of some other visual sensation. These considerations suggest that sensations are unified in the sensory fields that contain them because of factors connected with the sensory content of those sensations. Similarly, an auditory sensation and a visual sensation can eem both to occupy the same, or a similar, location in their respective sensory fields. This is not because some overarching or neutral phenomenal field exists, which subsumes or envelopes the special fields of the various sensory modalities.
Rather, we correlate locations in our auditory field with those in our visual field because we associate the sensations that occur in various locations in one field with sensations that occur in various loca- tions in the other field. Such association of auditory with visual sensations is natural, since we ordinarily take sensations of the two modalities to have common, nonmental causes. Correlations of phenomenal places in the field of one sensory modality with places in the field of another result from associations of the various sensory contents that occur at the relevant places.
In any case, since sensations of distinct modalities can seem to be located in corresponding phenomenal places, we can join together the relevant sensory fields into a single experiential field, including percep- tual sensations of all modalities. Again, this merging of fields is intimately bound up with the qualitative content of the particu- lar sensations involved; for it is one aspect of the qualitative ontent of a sensation that it occurs at a particular position in the relevant field. The unity of sensory mental states into an experiential field suggests a second way of accounting for the self in terms of mental states.
One can regard the self as a matter of the unity of this experiential field; this view would echo Hume’s claim that the self is \”nothing but a bundle or collection of different percep- tions, which succeed each other with an inconceivable rapidity, and are in a perpetual flux and movement. \”5 Such an account would have an advantage, intuitively, over an account that relies on the unity of mental acts. For, if the foregoing considerations are correct, we presuppose the unity of mental acts and the consequent need for a self to unify them because of a priori arguments about the nature of mental acts.
The unity of sensations, however, seems instead to be based on empirically discernible characteristics of these mental states. It is not surprising that seventeenth- and eighteenth-century rationalists should have generally adopted the first approach toward understanding the self, whereas empiricists of the same period should have tended toward the second. Once we turn from the unifying to the individuating role of he self, however, an account in terms of the unity of sensory mental states is far less promising. The unity of such states does seem to be partly a function of their sensory qualities.
But the kind of unity that results from considering sensory qualities is inadequate to explain why sensory mental states belong to distinct selves. If two sentient organisms are similarly located and have similar sense organs and cognitive apparatus, their sensations will presumably be strikingly similar. The sensations of one sentient being would, in these circumstances, be qualitatively as close to hose of another as they are to any others that belong to distinct selves than the unity of propositional mental states was.
The operation of the self must therefore be presupposed in the unifica- tion of sensory mental states, just as it is in the unity of mental acts. Indeed, the apparent advantage of relying on the unity of sensory mental states proves to make the problem of individuating all the more pressing. The selves that are supposed to unify propositional mental acts are conceived on the model of substrata, which operate in an unknown way and whose existence is established by a priori considerations. Such reasoning may fail to explain how distinct selves are individuated.
But such as model does no better in explaining how selves unify mental acts; on this model, one simply stipulates that the self can unify mental acts into distinct, individual groups. But, if one conceives of the self as depending, instead, on the occurrence of many sensations in a single perceptual field, the unifying and individuating functions of the self will be on different footings. Mental states will then be unified by the operation of the perceptual field, rather than by any presupposed substratum.
But, in the absence of such a substra- um, the individuating of distinct selves will be all the more inexplicable. The ability of the model based on sensations to avoid substrata in unifying mental states makes the subsequent individuating of selves seem impossible, unless some other model is invoked to supplement the perceptual field. Perhaps reasoning of this sort is responsible for the pronounced tendency towards solipsism that results from the adoption of a model of the self based solely on the unity of the perceptual field.
Even invoking some sort of individuating substrata will not fully dispel this tendency toward solipsism. For, as noted above, unless the operation of an underlying self is understood in unifying mental states, it will remain mysterious how they can be unified into different groups corresponding to distinct selves. It may seem that desires might help in giving an acceptable account of the self. The way in which mental acts such as believing, doubting, and suspecting are unified is obscure partly because of the abstract and intellectual character of these mental acts.
At best, beliefs and doubts have an oblique connection with nonmental reality, a connection mediated by the occurrence of other ental states that involve perception and action. Those which involve perception seem not to help. If their sensory field uni- fies them, this unity is powerless to explain how distinct selves can be individuated if something other than perceptual fields is responsible for such unity, an account of the self based on these mental states has no advantage over an account based on thoughts.
Desires resemble thoughts in suggesting a need for an underlying self needed to unite them; like thoughts, desires involve both a mental frame and some propositional content. Unity is presumably o be achieved by a single self holding the relevant desiderative attitudes. But in this case, the unifying function may seem more accessible than in the case of suspicions, doubts, and beliefs. For that which underlies the unity of desires involves the notion of agency; desires belong to a single self because a single agent holds these desires.
The concept of agency cannot help, however, in the attempt to give an account of the self in strictly mental terms. If the concept of agency involves overt actions, then agency would be sufficient to individuate distinct selves. For the physical ehavior involved in action would be a matter of movements of distinguishable bodies, and we could then individuate selves by reference to these bodies. But, if agency is understood in strictly mental terms, no such method of individuation will be available.
For, if overt action is ruled out, there is nothing to the concept of agency beyond the unity of a variety of desiderative and aversive mental acts, together with other interacting proposi- tional attitudes and perceptions. And by itself, the unity of desiderative mental states is no more comprehensible than that of other propositional attitudes. The situation is parallel to that of perceptual mental states. If perceptions are regarded as achievements, so that seeing and hearing things imply the presence of nonmental objects that are seen and heard, then there is no problem about individuating selves by reference to perception.
Different selves can be thought of, then, as having distinct causal relations to various nonmental objects, which in turn can be independently individuated. But, if perceptions are understood in strictly mental terms, so that nonmental objects of perception are excluded from consideration, such mental states cannot help ndividuate selves. II. It might be objected that the discussion has thus far not touched on the central insight that motivates a purely mental account of the self, an insight that might well be invoked to deal with the difficulties raised above.
Intuitively, the unity of mental states consists in their belonging to, or being immediately accessible to, single centers of consciousness. The idea of centers of conscious- ness explains much of the appeal of the Cartesian identification of the self with the mind. For centers of consciousness presumably connect all the mental states that exist in or for a particular elf; the unity of the self, on this view, consists in the presence of a variety of mental states to a single center of consciousness.
If this line of reasoning is correct, the failures noted above in the accounts based on thoughts, sensations, and desires should come as no surprise. Without the operation of centers of consciousness, such mental states cannot be unified in any comprehensible or effective way, nor will it be intelligible how they can belong to distinct selves. The notion of centers of consciousness is doubtless similar to the idea of individual points of view, which Thomas Nagel has orcefully and eloquently argued is necessary for an adequate account of subjectivity.
But it is far from obvious what subjective points of view or centers of consciousness might amount to. In particular, if we are to understand the unifying and individuating functions of the self in terms of such centers of consciousness, we shall need some grasp of how centers of con- sciousness operate in performing these functions. Intuitively, mental states are unified by being present to a single center of consciousness, and distinct selves are individuated because mental states are present to distinct centers of consciousness.
But without an account of what it is for a mental state to be present to a particular center of consciousness, these explanations will be no less idle than explanations that appealed just to the mental states themselves. Two models of the consciousness of mental states are available that might help give some account of what is involved in a mental state’s being present to a center of consciousness. On the first, all mental states are automatically conscious states, and such consciousness consists in every mental state’s somehow being transparent to itself.
This model is implicit in much that as been written in the Cartesian tradition. All mental states know themselves, on this account, and all are therefore, in Ryle’s useful phrases, \”self-intimating\” or \”self-luminous. \”8 Substantial difficulties face this model of consciousness, difficulties that seem decisive against it. 9 But in any case this model is unable to help with the present problem. For even if mental states know themselves or are otherwise self-transparent, this does not explain how two mental states can be present to one center of conscious- ness, or two others can be present to two distinct centers of onsciousness.
Each individual mental state may be self-luminous without thereby belonging to one self, rather than another. This model of the consciousness of mental states may initially appear to help because it echoes the Cartesian doctrine that all the mental states of a Cartesian mind are transparent to that mind. But that doctrine, once again, presupposes rather than explains the way in which mental states are present to distinct, individual selves.
And by itself, the self-luminosity of individual mental states cannot supply the unity necessary for the concept of the self. The second model of the consciousness of mental states relies on the idea that being conscious of something means being aware of it, and being aware of something amounts to having a thought about it. If I have a thought about a tree, I am aware and therefore conscious of the tree; if I have thought about a mental state, I am correspondingly aware and therefore conscious of that mental state.
This model seeks to capture the immediacy that char- acterizes our consciousness of our own mental states by stipulating that the higher-order thought one has about the conscious mental state be noninferentially arrived at. But we have noninferential thoughts about the mental states of others, as well as about our own. So this model specifies that the higher-order thought must involve reference to oneself; it must be a thought that one is, oneself, in the mental state in question.
It is not problematic for this account that we are generally unaware of the occurrence of such higher-order thoughts. If the consciousness of mental states consists in having a higher-order thought about them, those higher- order thoughts will themselves not be conscious thoughts unless one also has a still higher-order thought about those thoughts. Suitable care in developing this kind of account allows one to do intuitive justice to the phenomena of conscious mental states at least as well as an account based on the Cartesian notion of intrinsic self-transparency.
One might expect, however, that such a view of consciousness could do little if anything to capture or explain the unity of centers of consciousness. For, on the current account, the con- sciousness of mental states is an extrinsic property they have, and extrinsic properties may seem too accidental to hold out promise of an adequate account of such unity. But the intrinsic character of he consciousness of mental states, on the Cartesian model just considered, provided no help in understanding the unity of the self.
And the conception that relies on higher-order thoughts actually fares better on this score. For the higher-order thought that confers consciousness on other mental states represents itself as belonging to the same self as the mental state that it is about. Higher-order thoughts thereby succeed in yolking themselves toge- ther with the mental states that they make conscious. Moreover, higher-order thoughts are possible that, by being about two or more mental states, would bring unity to them.