Hostname: page-component-74656f9884-8r8nv Total loading time: 0 Render date: 2026-11-09T23:54:51.090Z Has data issue: false hasContentIssue false
  • English
  • Français

EXPERIENCE, AGENCY, AND PERSONAL IDENTITY

Published online by Cambridge University Press:  15 June 2005

Marya Schechtman
Marya Schechtman
Affiliation:
Philosophy, University of Illinois at Chicago
Rights & Permissions [Opens in a new window]

Abstract

Psychologically based accounts of personal identity over time start from a view of persons as experiencing subjects. Derek Parfit argues that if such an account is to justify the importance we attach to identity it will need to provide a deep unity of consciousness throughout the life of a person, and no such unity is possible. In response, many philosophers have switched to a view of persons as essentially agents, arguing that the importance of identity depends upon agential unity rather than unity of consciousness. While this shift contributes significantly to the discussion, it does not offer a fully satisfying alternative. Unity of consciousness still seems required if identity is to be as important as we think it is. Views of identity based on agential unity do, however, point to a new understanding of unity of consciousness which meets Parfit's challenge, yielding an integrated view of identity which sees persons as both subjects and agents.

Information

Type
Research Article
Information
Social Philosophy and Policy , Volume 22 , Issue 2 , July 2005 , pp. 1 - 24
Copyright
© 2005 Social Philosophy and Policy Foundation

I. Introduction

Questions of personal identity are raised in many different philosophical contexts. In metaphysics the question at issue is that of personal identity over time, or of what relation a person at one time must bear to a person at another time in order for them to be, literally, the same person. There are two standard responses to this question in the current literature. One considers a person as essentially a biological entity and defines the identity of the person over time in terms of the continued existence of a single organism. The other follows John Locke in accepting a distinction between persons and human beings and defines the identity of a person in terms of the continued flow of psychological life.1

There are, of course, other positions that are and have been defended. Historically it has been popular to define personal identity in terms of the persistence of an immaterial soul, and this view still has its defenders. Others see identity as irreducible and unanalyzable. Continuity of organism and continuity of psychological life are, however, generally acknowledged as the basic candidates for an account of personal identity in the current discussion, and the vast majority of authors in the area defend one or the other of them.

This second response—the “psychological continuity theory”—has a great deal of appeal and enjoys a great many supporters. Both positions are still very much alive in the current discussion, suggesting that at the very least each response expresses some important aspect of our thought about what we are and how we continue. In what follows I will leave aside the dispute between these two accounts of identity and will focus only on the psychological side. Understanding what fuels the idea that personal identity should be defined in psychological terms, and what a viable psychological account of identity would look like, is an important goal in its own right and will put us in a better position to understand the relation between psychological and biological accounts of identity.

A central element in the motivation and defense of the psychological continuity theory of personal identity comes from the intimate connection between the concepts of identity and survival. One obvious way of understanding what it means for a person to survive is for there to be someone in the future who is she. Since surviving is, to put it mildly, a matter of the utmost practical importance, it seems that whatever relation defines personal identity over time should also bear the importance we attach to survival. Psychological continuity theorists argue that when we ask ourselves which kind of continuation—physical or psychological—seems to bear this importance, the latter emerges as the better candidate. This view trades on the intuition that a situation where a person's body continues to function but his psychological life comes to an end (as in, e.g., a case of irreversible coma or total possession) seems like a failure to survive, while a situation in which psychological life continues despite the death of a person's body (as in, e.g., some of the traditional religious depictions of an afterlife) seems like a case of personal continuation.

It remains, however, for psychological continuity theorists to describe in more detail the kind of psychological continuation that constitutes survival. The initial and most natural idea is that persons should be construed as experiencing subjects, and that personal identity should be defined in terms of a deep unity of consciousness throughout an entire life, but this unity turns out to be difficult to define coherently. For this and other reasons, some identity theorists who support a psychologically based view have shifted from a picture of the person as a subject to a picture of the person as an agent, and from a definition of personal identity in terms of a deep phenomenological unity to a definition in terms of the unity of agency. This shift leaves us with the question of whether unity of agency is a strong enough relation to provide an account of what it is for a person, literally, to survive.

Some resources for answering this question can be found in other areas of philosophy—in particular, in discussions of free will, practical reason, and the theory of action, which I will refer to collectively with the term “practical philosophy.” A prominent strand of thought there suggests that having deeply held projects and commitments is essential to being a person, and describes how a person can literally disintegrate from a lack of such projects and commitments. While this discussion does make a strong case that unity of agency must be part of any acceptable psychological account of the persistence of the person, the original intuition that some sort of deep unity of consciousness is required for personal survival runs deep, and a view couched entirely in terms of unity of agency is unlikely to be fully satisfying. Fortunately, the insights gleaned by following out the shift toward a view of the person as agent can help point to a way that something like phenomenological unity of consciousness might be introduced without so much difficult metaphysical baggage.

My main goal here is to follow out the way in which a picture of persons as essentially agents has found its way into the metaphysical discussion of personal identity, and to assess where this leaves us in our attempts to understand the conditions of personal persistence. I begin by showing how the internal pressures of the metaphysical discussion lead to the introduction of an understanding of persons as agents, and how work in practical philosophy can help support the plausibility of this understanding as a response to the metaphysical question of personal persistence. After this I discuss why a response to this question wholly in terms of agency is likely to remain unsatisfying, and I point very briefly to a possible strategy for reintroducing an understanding of the person as subject as part of a more integrated account.

II. What Matters in Survival and Peter Unger's Distinction

As I have already mentioned, the central considerations supporting the psychological continuity theory are based on the idea that it is psychological rather than physical continuation that captures what matters to us in survival. As discussion of this view has unfolded, many psychological continuity theorists have found reason to put aside the question of personal identity per se and ask directly about the conditions of survival and what we find important in it. The main impetus for this switch is a difficulty with using psychological continuity as a criterion of personal identity. Psychological continuity—at least as usually defined by these theories—does not have the logical form of an identity relation. The continuity relation, unlike identity itself, can be one-many, intransitive, and a matter of degree. Derek Parfit famously argues that this does not undermine the psychological continuity theory but rather reveals that what really underlies our interest in the question of personal identity over time is the issue of survival.2

In fact, Parfit shifts his use of the term “survival” over the course of his work. In Derek Parfit, “Personal Identity,” Philosophical Review 80 (1971): 3–27, he suggests that survival can come apart from identity, and suggests that we are interested in questions of survival. In his book Reasons and Persons (Oxford: Clarendon Press, 1984), Parfit seems to imply that, strictly speaking, survival requires identity, and that we are interested in “what matters in survival.” I will be using the terminology in the former sense. For a full discussion of the evolution in Parfit's use of this term, see Marvin Belzer, “Self-Conception and Personal Identity,” elsewhere in this volume, Section IIC.

In ordinary circumstances, survival will involve identity; the person as whom one survives will be oneself. However, in certain imaginable science fiction scenarios, Parfit argues, it may be possible to survive as someone else. This might happen, for instance, if a person could split amoeba-like into two qualitatively identical people. Such a division hardly seems like death, he argues, yet the two resulting people, since they are clearly not identical with one another, cannot both be identical to the original person on pain of intransitivity. This “fission” scenario is thus a case in which a person enjoys what we take to be significant in survival despite the fact that there is no one in the future to whom she is identical.3

Parfit, Reasons and Persons, chap. 12.

Parfit holds that our interest in questions of identity is an artifact of the fact that in ordinary circumstances survival requires identity with some future person. Once we see that we can have what matters in survival without identity, we can see that our main interest is with survival rather than identity itself. The question of whether or not we are going to survive undoubtedly bears a special significance for us. As John Perry puts it: “You learn that someone will be run over by a truck tomorrow; you are saddened, feel pity, and think reflectively about the frailty of life; one bit of information is added, that someone is you, and a whole different set of emotions rise in your breast.”4

John Perry, “The Importance of Being Identical,” in Amélie Oksenberg Rorty, ed., The Identities of Persons (Berkeley: University of California Press, 1976), 67.

What we are after with questions of diachronic personal identity (that is, questions about the conditions for the persistence of a single person over time), says Parfit, is the question of what relation to the future carries this importance—“what matters” in survival. Many theorists of personal identity who would define it in psychological terms have thus shifted their emphasis from identity to what matters in survival.

It might seem as if the shift of focus to “what matters” automatically turns the discussion to a more metaphorical or psychological notion than that with which we started, since we are now asking what makes our continuation a good rather than what makes us, literally, still alive. Most psychological continuity theorists would deny this, however. The survival they are trying to define is, they insist, quite literal, and the sense of “what matters” at issue should be read that way. This point is made most explicitly by Peter Unger, who says that there are different things we could mean by “what matters in survival,” and that it is important for us to keep clear on which of them is under consideration. He distinguishes between three different uses of this term. One, to which I will return later, concerns the question of which features of survival constitute it as survival. For present purposes, the most important distinction is between the other two senses, which he calls the “desirability” and the “prudential” senses of “what matters.” According to the desirability sense,

‘what matters in survival’ will mean much the same as this: what it is that one gets out of survival that makes continued survival a desirable thing for one, a better thing, at least, than is utter cessation. On this desirability use, if one has what matters in survival, then, from a self-interested perspective, one has reason to continue rather than opt for sudden painless termination.5

Peter Unger, Identity, Consciousness, and Value (New York: Oxford University Press, 1990), 93.

This is contrasted with the prudential use, which determines,

from the perspective of a person's concern for herself, or from a slight rational extension of that perspective, what future being there is or, possibly, which future beings there are, for whom the person rationally should be “intrinsically” concerned. Saying that this rational concern is “intrinsic” means, roughly, that, even apart from questions of whether or not he might advance the present person's projects, there is this rational concern for the welfare of the future being.6

Ibid., 94.

Unger suggests that there is a question of whether a future will be mine to experience—whether I will undergo what happens in it—which at least seems quite independent of my values, goals, ideals, or character traits.

If we are looking to answer literal, metaphysical questions about our survival, Unger argues, we should be looking for the relation that underlies the more fundamental, prudential connection between different moments of a person's life. The desirability sense of what matters may be an important thing to define in its own right, but it is not, he says, “highly relevant to questions of our survival.”7

Ibid., 93.

To test for the prudential sense of what matters, Unger suggests that we use the “avoidance of future pain test.” We are to imagine ourselves connected to a future person in some way, and then imagine that person being subjected to horrible torture or excruciating pain. The question we are to ask is whether our horror of the future pain is sympathetic in nature or rather the sort of concern we would feel if we were anticipating our own future pain. The latter response suggests that we have what matters in survival in the fundamental sense at issue in metaphysical discussions of personal survival.

Unger himself believes that attention to this question will lead us to a more biologically based account of personal survival, albeit one that requires the continuity of minimal psychological capacities (what he calls “core” psychology). Traditional psychological continuity theorists have rejected Unger's solution, but they seem to accept his basic insight. They suggest that the avoidance of future pain test reveals that what matters fundamentally in survival is the legitimate anticipation of future experiences. To survive we must bear a connection to the future that makes it rational for us to expect to have experiences then. These theorists do not accept Unger's solution because they deny, for reasons given by Locke, that the continuity of the body can itself be what legitimates anticipation of future experiences. (Locke famously argues through the use of hypothetical cases that continuity of body—or soul—with no continuity of consciousness cannot provide what we seek in survival.) For such anticipation we need a deep phenomenological connection to the future—a unity of consciousness over time. Continuity of the body might turn out to be one source of such a phenomenological connection; it might even turn out to be the only source. This would, however, be a contingent fact, and even if it were true it would not be the continuity of the body per se, but rather the unity of consciousness it produced, that would make our intrinsic concern for some future person rational. On this view, then, a person is most fundamentally an experiencing subject, and survival must involve the unity of an experiencing subject over time. This is, in many ways, a very natural and appealing picture of what is involved in our survival as persons, but it is difficult to develop a coherent theory based on this picture, as the next section will explain.

III. Derek Parfit's Argument against the Person as Experiencing Subject

A compelling initial picture of what constitutes literal survival for those who take a psychological approach to this question is, as we have seen, a unity of consciousness over time.8

As mentioned in note 2, for terminological simplicity in what follows I will refer to having what matters in this fundamental sense simply as “survival,” meaning that in my use survival does not automatically imply identity.

Derek Parfit has argued quite forcefully, however, that this natural first idea of what defines survival cannot be defended—or even defined—in the end. He provides several arguments explicitly directed at the idea that personal identity cannot be defensibly defined in terms of the unity of a subject of experiences, but most of the fundamental insights behind these arguments can be found in the Teletransportation and Branch-Line cases with which he opens his discussion of personal identity in Reasons and Persons.9

Parfit, Reasons and Persons, 199–201.

This pair of cases uncovers both the hold this conception of personal continuation has on us, and its problematic nature. In Teletransportation, a subject enters a booth where his body is scanned and dematerialized. A molecule-for-molecule duplicate of the original body is built on the destination planet, a duplicate who by hypothesis replicates the intrinsic character of the original person's psychological states exactly. The question is whether this is an efficient means of travel to distant planets, or death and replacement by an imposter. Parfit acknowledges that we are quite likely to think of it as the latter. Despite the exact similarity of the newly formed individual to the original person, we may well fear that the replica will lack a deep attachment to that person present in ordinary survival. We fear that while in ordinary survival one will feel one's future experiences, in Teletransportation one will feel nothing in the future—one's point of view, which would continue in the ordinary case, is snuffed out.

The presupposition underlying this fear is that there is a deep unity of consciousness within a given life that may be absent in our connection to a replica. It is not immediately obvious exactly what this deep unity of consciousness is, but the basic idea can be understood on the model of co-consciousness at a time. At a given time there is a brute difference between a set of experiences being present in a single consciousness and their being parts of distinct consciousnesses. This is why, for example, a decision about whether to tolerate a small pain in my shoulder now to avoid an excruciating pain in my foot now is different in kind from a decision about whether to tolerate a small pain in my shoulder now to avoid your experiencing an excruciating pain in your foot now. The fear of Teletransportation reveals the supposition that we have a relation to our future experiences which is relevantly like the relation we have to our own present ones. This is why the decision about whether to suffer a small pain now to avoid an excruciating pain later is also different from the decision about whether to suffer a small pain now to keep someone else from suffering an excruciating pain later.10

Of course, in either case we may decide that suffering the small pain is the better thing to do. We can care as much or more about the pains of others as we do about our own. The point is just that the considerations that go into the decision making are different in each case.

This difference is, of course, what the avoidance of future pain test proposed by Unger is testing for. When we fear that we will not survive Teletransportation, we are expressing the possibility that this connection to the future may not hold in cases of replication. In more concrete terms, we suspect that my attitude toward the replica's potential pains should be more like my attitude toward the potential pains of others than toward my own.

Since a teletransported replica is, by stipulation, completely identical to the original person in terms of the contents of her consciousness, whatever difference there is between ordinary survival and having a replica cannot be defined in terms of these contents. The previous discussion makes it clear, however, that it is supposed to be a difference in the quality of experience. To make the case that there could be such a difference, we thus need to be able to understand a phenomenological distinction between really continuing and being replaced by a replica which is not defined in terms of the contents of consciousness. Parfit's claim is that there is no defensible way to define such a distinction. The difference we are seeking cannot be explained by the fact that in one case we retain our original body, since the psychological account is based on the intuition that we could in theory have what counts as survival without continuation of the body. If the sameness of body plays a crucial role, it will have to be because continuing in a single body is experientially different from being replaced by a replica, and we still need to say something about the nature of this phenomenological difference.11

Of course, at this point we might revert to a biological criterion, saying that our special concern about our own futures is not based on a deep unity of consciousness, but only on the fact that the organism currently having experience will be present, sentient, and in pain in the future. There are many considerations in favor of this position, and Parfit also offers many arguments against it. For the present, I am only interested in following out the intuitions behind psychological accounts of identity. I will, therefore, only consider what would be required to make this approach work, and see where this leads us.

Parfit thinks that we might have been able to make sense of this mysterious unity of consciousness over time if we accepted the view that we are simple, immaterial souls (although the Lockean arguments make this questionable in any event), but argues that there is no reason to believe such a thing, and many reasons not to.12

Parfit, Reasons and Persons, 227–28.

Of course, the inability to define the deep unity of consciousness that we assume in our own lives—and fear might be absent in replication—may not be enough to convince us that it does not exist. It seems as if we actively experience this continuity and connection in our own lives. Moreover, it seems all too possible to imagine the flow of our own psychological life terminated or diverted despite the existence of a replica with qualitatively identical contents of consciousness. Kathleen Wilkes makes this point quite effectively in a footnote in Real People: Personal Identity without Thought Experiments. Discussing the transporter depicted on the science fiction series Star Trek, she says:

Captain Kirk, so the story goes, disintegrates at place p and reassembles at place p*. But perhaps, instead, he dies at p and a doppelganger emerges at p*. What is the difference? One way of illustrating the difference is to suppose there is an afterlife: a heaven, or hell, increasingly supplemented by yet more Captain Kirks all cursing the day they ever stepped into the molecular disintegrator.13

Kathleen Wilkes, Real People: Personal Identity without Thought Experiments (Oxford: Clarendon Press, 1994), 46n.

Once the viewpoint of the original Captain Kirk is reintroduced, it seems natural to assume there is a unified flow of consciousness between the original Kirk and the disembodied Kirk—a flow of consciousness that does not exist between the original Kirk and his doppelganger. If we now simply take out the assumption of an afterlife, we can see that the connection between the original and doppelganger remains phenomenologically deficient.

Parfit himself is very sensitive to the pull of this idea, and this is why he introduces the variation on his Teletransportation case that he calls the “Branch-Line case.” In the Branch-Line case, the traveler is scanned and a blueprint made, but the original body is not dematerialized. Instead, it is damaged in such a way that it can live only a few more days. At the same time, a healthy replica is built on the distant planet (here Mars) as in simple Teletransportation. There are now two people who are psychologically continuous (in Parfit's sense) with the original person: the replica on Mars and the person with the original body on Earth. It seems clear, Parfit acknowledges, that the connection between the person on Earth (now traveling on a branch line of psychological continuation) and the replica does not contain what matters most fundamentally in survival. Parfit imagines himself on the branch line talking to his replica on Mars:

Since my Replica knows that I am about to die, he tries to console me with the same thoughts with which I recently tried to console a dying friend. It is sad to learn, on the receiving end, how unconsoling these thoughts are. My Replica then assures me that he will take up my life where I leave off. He loves my wife, and together they will care for my children. And he will finish the book that I am writing…. All these facts console me a little. Dying when I know that I shall have a Replica is not quite as bad as, simply, dying. Even so, I shall soon lose consciousness, forever.14

Parfit, Reasons and Persons, 201.

The replica is as like the branch-line person in terms of contents of consciousness as a person is like herself in the near future, but there is all the difference in the world for this person between her psychological life continuing several more decades and its being terminated while the replica continues for those decades—or so it seems. This is why, Parfit acknowledges, “it is natural to assume that my prospect, on the Branch Line, is almost as bad as ordinary death.” Nonetheless, he goes on to say, “[a]s I shall argue later, I ought to regard having a Replica as being about as good as ordinary survival.”15

Ibid.

He offers a variety of considerations in favor of this claim. These include a collection of arguments for a reductionist view of persons, arguments against the possibility of directly experiencing a deep unity of consciousness over time in our own lives, and thought experiments aimed at making it more plausible that the person on the branch line could survive as the replica. The case most directly addressing this issue is, importantly, not a science fiction case, but one that is part of a real life. Parfit describes a sleeping pill he has taken that takes a short time to work. After taking the pill, the person remains fully conscious for a while, but when he awakens he has no memory of the time between taking the pill and falling asleep. Parfit reports that having once taken such a pill, he apparently solved a practical problem before falling asleep, for he found a note the next morning under his razor advising him of the solution, despite having no memory of solving the problem or writing the note. He argues that the person who has taken such a pill can know, like the person on the branch line, that his stream of consciousness will end before morning, but he is unlikely to face this prospect as if he is facing death. Yet, Parfit argues, in terms of intrinsic relations of consciousness or phenomenology the relation of the person who takes this sleeping pill to the person in the morning is no different from the relation of the branch-line person to his replica on Mars. There are, indeed, a host of differences between the two cases, but they do not seem to be intrinsic differences, and so if we want to say that these make the difference between surviving and not surviving we will have given up on the idea that the importance of survival is linked to a deep diachronic unity of consciousness.

While there is much to understand—and to take exception to—in Parfit's arguments, the underlying theme is, I think, fairly simple and quite persuasive. What his arguments do at their most effective is direct us to commonplace examples of survival and undermine our conviction that there is any deep unity of consciousness there. The sleeping-pill case, for instance, tells us less about a heart-damaged person with a replica on Mars than about what happens to us all the time. We do not need to take sleeping pills to have gaps in our consciousness, or to know sometimes that what we are thinking and experiencing in the present will be entirely lost to future consciousness—this is why people make lists and keep journals. This challenge is reinforced by the sort of empirical work used by Daniel Dennett and Andy Clark, among others, to reveal that our unity of consciousness—even at a time—is not all that casual introspection might lead us to believe.16

See, for instance, Andy Clark, Being There (Cambridge, MA: MIT Press, 1998); and Daniel Dennett, Consciousness Explained (Boston: Little, Brown, and Co., 1991).

Whatever the upshot of detailed engagement with Parfit's many arguments, he raises a powerful and serious challenge to one deep-seated understanding of what is involved in our survival, and its importance. A tempting and natural way to think about our continuation is in terms of a flow of consciousness that connects current experience to future experience in something like the way our current experiences are connected to one another. It is this flow of consciousness that seems to allow us to really have experiences in the future—to be there in the sense that is most important. Parfit's repeated challenge to this picture, emblemized in the Teletransportation and Branch-Line cases, is essentially David Hume's—that if we take the time to really look, we can find no such unity in our own lives. This is why Parfit finally says of Teletransportation: “I want the person on Mars to be me in a specially intimate way in which no future person will ever be me. My continued existence never involves this deep further fact. What I fear will be missing is always missing.”17

Parfit, Reasons and Persons, 208.

He thus draws the following conclusion: it is not, as he claimed earlier, that Teletransportation is about as good as ordinary survival, but rather that “ordinary survival is about as bad as, or little better than, Teletransportation. Ordinary survival is about as bad as being destroyed and having a Replica.”18

Ibid.

If Parfit succeeds in his challenge—which many think he does—the only psychological connection between one's present and one's future is a relation of similarity or association between the contents of consciousness at different times. Since our sense of the importance of survival seems to be based—at least largely—on the assumption of a deep, phenomenological connection to the future, that importance needs to be reassessed in light of Parfit's challenge. The most extreme possibility is that the psychological relations we really do bear to the future have no intrinsic importance, and hence that survival itself is not actually very important. As counterintuitive as this conclusion is, Parfit suggests that it is defensible, and in some ways liberating.19

Ibid., 281–82.

This conclusion is not forced on us, however. To avoid it, one need only show that some relation that we really do bear to the future justifies, at least partially, the importance we attribute to survival. This might be done by revisiting the original arguments that favor a psychological account of identity and turning instead toward a biologically based view—a view that places the significance of survival in biological connections.20

This approach has become increasingly popular lately. A prominent example is Eric Olson's book The Human Animal: Personal Identity without Psychology (New York: Oxford University Press, 1997).

But the importance of survival might also be salvaged within the framework of a psychologically based view if it could be shown that the psychological relation that we do have to our futures can still bear the importance we attribute to survival despite the fact that it is not the deep phenomenological connection we thought we had. This strategy has been fairly popular, and it is here that we see a turn to a view of persons as agents in the metaphysical literature on personal identity.

IV. The Turn toward Agency

A. John Perry's distinction between the importance and constitution of identity

One good example of the switch to an agency-based approach to survival can be found in John Perry's work. Like Unger, Perry distinguishes between different senses of “what matters in survival.” His distinction, however, is not quite the same as Unger's. He says that the phrase “what matters in survival” can mean either “what is of importance” in survival or “what makes a case of survival a case of survival.”21

Perry, “The Importance of Being Identical,” 85–86.

Unger's prudential and desirability senses would thus both fall under Perry's category of importance. Although Perry is a psychological continuity theorist, he holds a version of this view in which identity-constituting psychological continuity is intimately interlinked with continuity of body. He therefore claims that not only is a person's replica not identical to her, but, strictly speaking, she does not survive as her replica either (he thus uses “survival” in a different sense than I have been using it). The puzzles which Parfit and others find concerning identity and survival, Perry argues, frequently involve a conflation of the two senses of “what matters” as he has defined them. The confusion stems from the expectation that the relation that defines survival will be intrinsically important, and when it is found not to be, this is taken as a challenge to the psychological continuity theory. What we need to understand, says Perry, is that the importance of survival—in both Unger's prudential sense and his desirability sense—is derivative rather than immediate.

To make this point, and to show from what the importance of survival derives, Perry uses a variant on Unger's avoidance of future pain test. He imagines a scenario in which a person, by pushing a button, can prevent someone from experiencing terrible pain the next day, and asks whether a person has different sorts of reasons to press the button if the pain she prevents will be her own—reasons that go beyond the simple fact that anyone's pain is bad and should, all things equal, be prevented. Perry argues that a person does, at least usually, have special reasons to push the button to prevent her own pain, but also that they are not the sorts of reasons we typically take her to have. On Perry's view, the legitimacy of our concerns about our future pains is not based on a deep phenomenological connection to the future, but instead derives from our concern for our projects. Each of us has many projects, and we have reasons to see those projects carried out. If we are in excruciating pain, we are less likely to be able to see our projects through, and so we have reasons to avoid excruciating pain. Given that we are the people most likely to carry out our own projects, we thus have special reasons to avoid our own pains. Perry is aware, of course, that in some instances there may be others who can carry out our projects better than we can. He adds, however, that it is extremely improbable that there is anyone else who could and would carry out all of our projects better than we would, and therefore our concern for our projects gives us reason to be specially concerned for ourselves.

Perry also acknowledges that the asymmetry between the likelihood that a person will carry out her projects and the likelihood that someone else will is a contingent rather than a logically necessary one. In the types of thought experiments used in the personal identity literature, this truism may not hold. Teletransportation, for instance, replicates a person's entire psychological life. In Perry's terminology, the people in these cases will not “survive” as their replicas, but the replicas will, by definition, be just as likely to carry out their projects as they would, and in just the same way. What is important in survival in the most fundamental sense will thus be captured in these cases, even though the original person does not, in Perry's sense, survive.22

That is, the person does not survive as Perry uses the term—as I am using the term “survival,” however, the person does survive.

Perry agrees with Parfit that the importance of identity and survival is derivative and does not depend upon a deep, phenomenological connection between the different temporal parts of a person's life. It depends instead, he says, on our interest in carrying out our projects—that is, our interests as agents.

B. Carol Rovane and the reconnection of constitution and importance

The shift to a view of the person as agent is also made by Carol Rovane in The Bounds of Agency: An Essay in Revisionary Metaphysics. Rovane does not enter the debate in direct response to the challenge of explaining the importance of survival or capturing phenomenological unity, and so she does not explicitly address these issues. She does, however, undertake to offer a psychological account of personal identity that will capture its practical importance (and, although she does not herself use the distinction between identity and survival, the practical importance of survival as well). More specifically, Rovane sets herself the goal of providing an account of identity that explains the normative injunction that rational unity should be achieved within the point of view of a single person, but not beyond.23

This normative injunction says that “when I deliberate I ought to resolve all of the contradictions and conflicts within my own point of view, but I need not resolve all of my disagreements with you. Likewise, I ought to rank all of my preferences, but this ranking need not reflect your preferences. And more generally, when I arrive at all-things-considered judgments I should take into account all of my beliefs, desires, and so forth, but not yours.” Carol Rovane, The Bounds of Agency (Princeton, NJ: Princeton University Press, 1998), 24.

To achieve her goal Rovane offers a normative analysis of personal identity. “According to the normative analysis, the condition of personal identity is the condition that gives rise to a certain normative commitment, namely, the commitment to achieve overall rational unity.”24

Ibid., 23.

Although her view is quite complex, the features relevant for our purposes are well-articulated in an analysis she offers of where her view stands relative to Locke's. She agrees with Locke, she says, that personal identity should be defined in terms of the unity of a first-person point of view, but adds that her “analysis will reject Locke's specific account of what a first person point of view is. He construed it as the phenomenological point of view of a unified consciousness. In contrast, the normative analysis takes it to be the rational point of view of an agent.”25

Ibid., 19.

Central to Rovane's view is the claim that a rational and phenomenological point of view need not—indeed, likely will not—coincide.

Rovane thus argues that the view that personal survival should be defined in terms of a phenomenological unity does not explain the normative injunction to take into account all of one's own states, but not any one else's, in deliberations about what to do. Instead she offers a view in which identity is defined in terms of unity of agency. This unity requires as a background condition many of the types of psychological connections that are present in the psychological continuity theory, but “in addition, the set must include certain substantive practical commitments that serve as unifying projects.”26

Ibid., 31.

There is, of course, a great deal more to her view than I have described here, but it is the fundamental insight that is important for our purposes. Translated into the terminology we have been using, Rovane says that the bounds of self-interested concern will be set not by phenomenological unity, but by agential unity, and thus that both what constitutes survival and what matters in it will be our continuation as rational agents, not as experiencing subjects.

C. Resources in practical philosophy—Christine Korsgaard and Harry Frankfurt

The idea that persons should be viewed primarily as agents, and the unity of persons as the unity of agency, is not, of course, limited to discussions in the metaphysics of personal identity. It is a standard position in practical philosophy as well. For the most part, the discussions of identity in metaphysics and practical philosophy have proceeded independently. This is unfortunate, as insights found in practical philosophy provide useful resources for those interested in developing an agent-based view of personal identity. However, one important point of contact between these two discussions can be found in Christine Korsgaard's essay “Personal Identity and the Unity of Agency: A Kantian Response to Parfit.”27

Christine Korsgaard, “Personal Identity and the Unity of Agency: A Kantian Response to Parfit,” Philosophy and Public Affairs 18 (Spring 1989).

Here Korsgaard explicitly takes on Parfit's arguments against a deep unity in the life of a person. She urges the Kantian position that “subject” and “agent” represent two perspectives we can take on persons, but says that if we wish to find what unifies a person, we must understand persons under the latter description rather than the former. The unity of a person's life, she says, is a unity of agency, not of experience.

Korsgaard accepts Parfit's argument that there is no unity of consciousness of the sort we tend to assume. She very elegantly describes the naive view of phenomenological unity as follows:

The sphere of consciousness presents itself as something like a room, a place, a lit-up area, within which we do our thinking, imagining, remembering and planning, and from which we observe the world, the passing scene. It is envisioned as a tunnel or a stream, because we think that one moment of consciousness is somehow directly continuous with others, even when interrupted by deep sleep or anesthesia. We are inclined to think that memory is a deeper thing than it is, that it is direct access to an earlier stage of a continuing self, and not merely one way of knowing what happened. And so we may think of amnesia, not merely as the loss of knowledge, but as a door that blocks an existing place.28

Ibid., 116.

Like Parfit, Korsgaard denies the accuracy of this picture of consciousness and acknowledges that if we are looking for a significant unity within the life of a person we will not find it by looking for a unified subject of experiences. If we think about persons as agents, however, a meaningful, nonarbitrary unity is easy to find. Korsgaard writes: “[S]uppose Parfit has established that there is no deep sense in which I am identical to the subject of experiences who will occupy my body in the future…. I will argue that I nevertheless have reasons for regarding myself as the same rational agent as the one who will occupy my body in the future.” She goes on to clarify that “these reasons are not metaphysical, but practical.”29

Ibid., 109.

When we start out thinking of persons as agents, we will immediately come upon the obvious fact that as agents, we must act, and that means we must decide what to do—or not do, as the case may be. Once we concentrate on this aspect of human existence, Korsgaard says, it will be easy to see that in order to act we must conceive of ourselves as unified agents. To show this, she begins by asking us to consider what underlies our sense that we are unified at a time, and says that there are two elements of this unity. First, the need for unity is forced upon us by the fact that, at least in our world, we control one and only one body. If we are divided about what to do, we need to overcome our division to be able to act coherently and effectively. “You are a unified person at any given time because you must act, and you have only one body with which to act.”30

Ibid., 111.

The second element is the “unity implicit in the standpoint from which you deliberate and choose.” She says that “it may be that what actually happens when you make a choice is that the strongest of your conflicting desires wins, but that is not the way you think of it when you deliberate. When you deliberate, it is as if there were something over and above all your desires, something that is you and chooses which one to act on.”31

Ibid.

According to Korsgaard, this sense of ourselves as choosing necessitates that we have reasons for choosing one course of action over the others, and “it is these reasons, rather than the desires themselves, which are expressive of your will.”32

Ibid., 113.

You must have a sense of yourself as a unified chooser with reasons in order to act.

Korsgaard goes on to argue that this unity, unlike unity of consciousness, can easily be seen as stretching over time as well as holding at a time. A person needs to coordinate action and motives over time as well as at a time, since almost anything we want to do will take at least a bit of time to unfold. Imagine, she says, that my body really is occupied by a series of subjects, changing from moment to moment. It is clear that they had better learn to cooperate “if together we are going to have any kind of a life.”33

Korsgaard therefore holds that personal identity will, in our world, be tied to the identity of the human being. She stresses that this is a contingent fact, however, in the sense that in a world where people had their bodies replaced with replicas on a regular basis, and these replicas continued their lives in something like the ordinary way, the unity of the agent could involve more than one body. See ibid., 113.

But more than this, the need to think of oneself as a unified agent acting from reasons means that we must identify ourselves with those reasons, and those reasons may, and probably will, “automatically carry us into the future.”34

Ibid.

If you understand yourself as an agent “implementing something like a particular plan of life, you need to identify with your future in order to be what you are even now.”35

Ibid., 113–14.

The basis for our unity over time is thus the necessity of acting once we are thrust on the scene, and not some metaphysical fact. As Korsgaard sums it up: “You normally think you lead one continuing life because you are one person, but according to this argument the truth is the reverse. You are one continuing person because you have one life to lead.”36

Ibid., 113.

Korsgaard realizes that to many it will seem that she has ignored the central issue. There may be some good practical reasons to act as if we will persist into the future, and to view ourselves as unified, but this does not really unify us. She rejects this description of the situation, however, for two reasons. First, she agrees with Parfit that our naive conception of the unity of a person rests on the assumption of a deep unity of consciousness over time, and that there is no such unity to be found in our actual lives, let alone in bizarre science fiction cases. Second, and most important, while our sense of our unity as agents may rest on practical rather than metaphysical facts, these facts are by no means trivial or arbitrary. It is not just an accident or convention that we take the body as the most relevant unit of agency; things would go badly for us if we didn't—indeed, after a while there might well not be any “us” for whom things would go badly. She allows, as does Rovane, that there can be units of agency bigger (and presumably also smaller) than the body—families, tribes, bowling leagues, or our teenaged selves might qualify. There is no denying, however, the necessity of viewing ourselves as unified agents (whatever the size of the unit), nor of taking a human life as a particularly basic such unit. Such unification is necessary, even if the necessity is not metaphysical.

The basic picture behind this view is developed further in recent work in practical reasoning that starts from the picture of the person as an agent and stresses the fundamental importance of our unity as agents, suggesting that our very survival depends upon it. Much of this work is Korsgaard's own. This picture is laid out in great detail in her book The Sources of Normativity and in her Locke Lectures. In the latter she tells us:

Because human beings are self-conscious, we are conscious of threats to our psychic unity or integrity. Sometimes these threats spring from our own desires and impulses. The element of truth in the image of the miserable sinner who must repress his unruly desires in order to be good rests in the fact that we deliberate in the face of threats to our integrity, and act against them. What is false about the picture is the idea that we must repress these threats in order to be good. Rather, we must repress them in order to be one, to be unified, to be whole. We must repress them in order to maintain our personal or practical identity.37

Christine Korsgaard, Locke Lectures, Lecture I, 18. Available online at http://www.people.fas.harvard.edu/∼korsgaar/Korsgaard.LL1.pdf.

If we do not unify ourselves as agents, we are in danger of real disintegration.

This general picture is shared by Harry Frankfurt. Although his view of personal identity differs from Korsgaard's in many important respects, he shares with her the idea that we must maintain unity of will to have any sort of life at all. Frankfurt tells us: “[T]here is, I believe, a quite primitive human need to establish and to maintain volitional unity. Any threat to that unity—that is, any threat to the cohesion of the self—tends to alarm a person, and to mobilize him for an attempt at ‘self-preservation.’ ”38

Harry Frankfurt, “Autonomy, Necessity, and Love,” in Frankfurt, Necessity, Volition, and Love (Cambridge: Cambridge University Press, 1999), 139.

Although the term appears in scare quotes, elsewhere Frankfurt links these issues to “self-preservation” “only in an unfamiliarly literal sense—in the sense of sustaining not the life of the organism but the persistence of the self.”39

Harry Frankfurt, “On the Usefulness of Final Ends,” in Frankfurt, Necessity, Volition, and Love, 89.

By giving us a way to think about unity of agency as central to our real persistence as persons, the work of Korsgaard and Frankfurt in practical philosophy adds a great deal to the plausibility of the approach suggested by Perry and Rovane. It provides a deep, nonarbitrary unity over the course of our lives that is not a unity of consciousness.

V. Unity of Consciousness Again

The switch to an agency-based view of persons and personal identity, especially supplemented by work in practical philosophy, represents an important development in our thought about these issues. There is a compelling case for the view that unity of agency is a real and important feature of personal identity. The question that remains, however, is whether this sort of unity is really strong enough to bear the kind of importance we attribute to survival. It may be that there is no unity of consciousness of the sort we typically assume there to be throughout the life of a person, and also that unity of agency is a more important unity than we generally take it to be, but that does not mean that it has the kind of visceral significance our naive-but-problematic notion of unity of consciousness did. To use Parfit's idiom, it is not clear whether an agency-based approach to personal identity is an alternative above and beyond those he conceived, or whether it is a version of the “moderate claim,” giving survival some importance, but not as much as we originally believed.

When we look at a view like Perry's or Rovane's, it certainly seems the latter. Of course we care about our projects, but if all there is to survival is the continuity of projects, the difference between surviving and being survived by like-minded folk does seem trivial. Korsgaard and Frankfurt offer resources for seeing an agency-based view as more of a real alternative to the Parfitian approach. The brute importance of unity of agency, not as a nicety or refinement of our lives, but as a condition of them, suggests that this sort of unity may well be considered the condition of survival as we ordinarily conceive it. On this view, unity of agency is not a consolation for not having unity of consciousness, but the unity that carries the significance we usually attribute to survival. This perspective, however, works better when we look at the short term, and it becomes more problematic when we look on the time scale of a human life. Korsgaard makes an excellent case for the need for unity of agency in the short term as a precondition for taking any action, and thus as a precondition for anything we might recognize as a person. The need for unity of agency over the long term, however, is a more tenuous thing. I do need a unified will at a time to eat, drink, or get out of bed, but I don't necessarily need a unified will over time to do those things.

Korsgaard, of course, argues that in a real sense I do need a unified will over time to take even minimal actions now. The move here depends rather heavily on the need to have reasons for action. To unify my will now, on her view, I need to see myself as having reasons for my choices, and it is these reasons that carry me into the future by taking the form of general laws. This is why the necessity of unification of agency over time is based on the view of oneself as “implementing a life plan” or having “any kind of a life.” So here the necessity of unification looks less like it comes from the desire to have a life and more like it comes from the desire to have a meaningful life. Frankfurt's position is similar in this respect. Although he does not think we need reasons to act, he does think we need to have fairly stable commitments to what we care about. His rationale for this claim is the same as Korsgaard's; such stability is necessary to have a meaningful life. In their later work, both philosophers thus stress the connection between unity of agency and a life that is “worth living.” Korsgaard does this, for instance, in her discussion of our identification with a variety of “practical identities”—identities as mother, friend, patriot, union member, etc.—which provide us with reasons for action. She defines these practical identities as “a description under which you value yourself, a description under which you find your life to be worth living and your actions to be worth undertaking.”40

Christine Korsgaard, The Sources of Normativity (Cambridge: Cambridge University Press, 1996), 101.

And Frankfurt describes a life of ambivalence as follows: “For someone who is unlikely to have any stable preferences or goals,” he says,

the benefits of freedom are, at the very least, severely diminished. The opportunity to act in accordance with his own inclinations is a doubtful asset for an individual whose will is so divided that he is moved both to decide for a certain alternative and to decide against it. Neither of the alternatives can satisfy him, since each entails the frustration of the other. The fact that he is free to choose between them is likely only to make his anguish more poignant and more intense.41

Harry Frankfurt, “The Faintest Passion,” in Frankfurt, Necessity, Volition, and Love, 102.

This provides us with a powerful motivation for seeking volitional unity over time, to be sure, but it does not make an overwhelming case for the fact that such unity is what defines our persistence. The distinction between a life integrated enough to be a life and one integrated enough to be a meaningful life thus becomes blurred.

This is problematic for agency-based accounts of personal identity because it seems all too possible for a person to live a life that has no meaning or direction, or to lose meaning only to find it again, and then perhaps lose it again. These are all crises within the life of a person and not crises that signify the end of a life. This idea is well-expressed in the opening question of Camus's Myth of Sisyphus—“why not suicide?” Camus denies the existence of a transcendent meaning that could unify a life, and asks whether and why we should go to the trouble of living if life has no meaning. While I do not mean to imply that Camus has a completely convincing conception of the nature of human existence, I do think he represents well one strain of thought that persists in our thinking about ourselves and our lives. The relation of this strain of thought to a position like Korsgaard's is complex, and worth investigating.

To begin, it might seem as if the very coherence of Camus's question challenges Korsgaard's position as I have interpreted it. If we are worried about having to live in the future despite the fact that life has no overarching significance, it would seem that no such significance is required to continue living a life. But this is too quick. Even if we accept (as I do) that Camus is talking about suicide not only as the act of killing an organism, but as the act of killing a psychological self, his question in many ways supports Korsgaard's position far more than it challenges it. Camus is saying that without meaning we wonder whether it is worth the effort of living. This means we do feel as if we have to decide what to do about the future—whether to perpetuate ourselves into it in a way that guarantees that we will have to keep acting and so take on the responsibility of deciding what to do, or whether to spare ourselves that effort. Considering such a drastic step, and for this reason, certainly suggests that we are identifying ourselves as the very agents who will be called upon to act in the future. This is why we need to make decisions now about what to do then. Suicide is a decision about what to do in one's future; it is the decision to do nothing.

The challenge to Korsgaard raised by Camus, such as it is, comes rather in the way that he argues that suicide is not an appropriate or rational reaction to the recognition that life has no meaning. Life can be lived all the better, he says, because it has no meaning. We can find a reason to live in the perpetual, self-expressive struggle against the arationality of the world. It is important to note that in some respects Camus and Korsgaard offer the same solution to the problem of meaninglessness. Neither relies on a transcendent purpose, outside of our natures, that provides our reason to live. Reasons for living and value must be provided by us, and are provided by us via our expression of ourselves as value-conferring or reason-giving beings. In fact, toward the end of The Sources of Normativity, Korsgaard gives an analysis of what is wrong with suicide that is in many ways very like Camus's.42

Korsgaard, The Sources of Normativity, 160–64.

There is, however, an important difference. Korsgaard's picture involves our taking the reasons we give ourselves seriously, which means recognizing that they give meaning to our lives insofar as we are true to them, and threaten to make our lives disintegrate into meaninglessness if we are not. Camus, on the other hand, suggests that we can only have a life worth living if we refuse to take these human meanings seriously, if we reject stability and find our value instead in the raw fact of our consciousness. Now, sometimes by “consciousness” he means consciousness of our condition—the continued awareness that we assert ourselves by struggling hopelessly against the absurd. But he also often means just the capacity to have experience—any experience. He tells us that “what counts is not the best living but the most living,”43

Albert Camus, The Myth of Sisyphus and Other Essays, trans. Justin O'Brien (New York: Vintage International, 1991), 63.

that “no depth, no emotion, no passion, and no sacrifice could render equal in the eyes of the absurd man (even if he wished it so) a conscious life of forty years and a lucidity spread over sixty years,”44

Ibid.

and that “the present and the succession of presents before a constantly conscious soul is the ideal of the absurd man.”45

Ibid., 63–64.

I do not mean to urge Camus's view as more accurate than Korsgaard's, but in its extremity it points to an important element that is not given sufficient emphasis in her account—the value that we put on the sheer fact of being conscious, independent of any practical implications. Of course, Korsgaard does not deny that we care about consciousness, but I think Camus's picture shows more than just that; it shows how much we care about being perpetuated into the future as the same experiencing subject. The subject who is conscious now wants to act, and wants meaning, but she also just plain wants to be there in the future, with or without purpose. Two more months, even if there are no loose ends to tie up or projects to complete, are two more months. This does not mean that pain and depression cannot make the negative value of continuation so great that it is no longer deemed worthwhile—this is when we see suicide as an understandable response. But the significance of being there, just to be there, just to keep having experiences, just to have a life and not necessarily a meaningful life is, I think, an important part of “what matters in survival.”

So, we seem to be back to the idea that the agency-based account of personal identity is some form of Parfit's moderate claim. The kind of survival that seems so important to us—the survival that psychological accounts of identity are trying to capture—requires a deep unity of consciousness over time, something that we now have reason to believe we do not, and cannot, have. This would mean that Parfit was right in claiming we are merely under an illusion that this unity exists, and therefore take identity and survival to have an importance that they do not actually have. But this is too quick. Our discussion of agency-based accounts of identity has hinted that they may be able to give us something more like our naive conception of psychological continuation than we might have thought. These hints are the suggestions that our conception of ourselves as unified agents can have a profound impact on the quality of our conscious experience, and on our experience of ourselves as persisting subjects as well as agents. Korsgaard talks about the way in which our reasons for action “carry us into the future” as agents, but they may also do so as subjects. In conceiving of ourselves as continuing as agents, we project ourselves into the future as subjects as well—imagining what will bring us satisfaction, when we will be filled with regret, what is likely to tempt us away from our purpose, and so on. To experience ourselves as unified agents, we may well also have to experience ourselves as unified subjects, and vice versa.

Another aspect of this connection is mentioned by Frankfurt in his discussion of why we need final ends. He has told us already that we need to have stable commitments that constrain our will (with our approval) because otherwise life will be meaningless, but he puts a particularly interesting gloss on this claim when he tells us that if we fail to find something we care about in this way we will get bored, and, he says,

the avoidance of boredom is a profound human need. Our aversion to being bored is not a matter simply of distaste for a rather unpleasant state of consciousness. The aversion expresses our sensitivity to a much more basic threat. It is of the essence of boredom that we don't care about what is going on. We therefore experience an attenuation of psychic vitality or liveliness. In its most familiar manifestations, being bored involves a reduction in the sharpness and focus of attention. The general level of mental energy diminishes. Our responsiveness to conscious stimuli flattens out and shrinks. Distinctions are not noticed and not made, so that our conscious field becomes increasingly homogeneous. As boredom progresses, it entails an increasing diminution of significant differentiation within consciousness.46

Frankfurt, “On the Usefulness of Final Ends,” 89.

He goes on to explain:

At the limit, when consciousness is totally undifferentiated, this homogenization is tantamount to the cessation of conscious experience altogether. When we are bored, in other words, we tend to fall asleep. Any substantial increase in the extent to which we are bored undermines, then, the very continuation of conscious mental life. That is, it threatens the extinction of the active self. What is expressed by our interest in avoiding boredom is therefore not simply a resistance to discomfort, but a quite primitive urge for psychic survival.47

Ibid.

It is not just activity that is threatened if we do not have long-term commitments; it is our very consciousness. The danger, of course, is not just literally falling asleep, but rather failing to project one's consciousness into the future in the ways we usually do.

Interestingly enough, Camus makes just the opposite claim. He says that we need to reject long-term commitments and the meaning derived from them if we are to retain our consciousness. Commitment breeds complacency, and a rote, mechanistic implementation of a life-plan dulls our consciousness. If we are to be really alive, and remain really conscious, he says, we must think of ourselves as faced with the world afresh at each moment. Undoubtedly there is some truth to both of these perspectives. A failure to have passions may well dull consciousness, and so may falling into a routine. Moreover, there is probably less of a distinction between Frankfurt and Korsgaard on the one hand and Camus on the other than it at first appears. While Frankfurt and Korsgaard do require stability, they also emphasize that as agents we are forced to keep our commitments in place, essentially by perpetually rechoosing them. These issues are too complex to sort out here. For the present, I am interested only in making the more general point that the quality of our consciousness is going to be affected in a variety of ways by our recognition that we are agents and by our conception of ourselves as unified agents. The agency view thus need not be seen simply as a substitution of a picture of persons as agents for a picture of persons as subjects, but rather as a new way of thinking about persons as subjects, and the way in which subjects might be unified over time.

This discussion suggests that while we do, indeed, feel that a continuation of consciousness is a crucial part of the survival of the person, it is not the bare continuation of sentience isolated by Unger's prudential sense of what matters, nor the formal connections of similarity and association between the contents of consciousness offered by psychological continuity theorists. It is, rather, a continuing sense of ourselves as subjects moving through the world and experiencing it. This experience of ourselves as unified subjects may well require at some point the structure and organization that comes from conceiving of ourselves as unified agents, and may arise only under the practical exigencies of living in the world. This does not mean that this experience is not of independent value, nor that it cannot outlast the sorts of identifications and reasons that unify an agent over the long term. There may thus be a meaningful and significant question of whether I can anticipate having experience in the future that cannot simply be answered by looking at how I unify myself as an agent, even if the answer is not totally independent of that fact. The relations to which I am pointing are, admittedly, very vague at this point. If, however, there is reason to believe that they really exist, there is reason to think that we can come up with a more satisfying account of personal identity by uncovering and understanding them.

VI. Conclusion

The psychological approach to problems of personal identity seems to capture one very important aspect of our conception of what we are and the conditions of our persistence. It is not easy, however, to articulate the insight it provides. Originally it seems that the insight is that we are fundamentally subjects of experience, and that our survival requires a persistence of experiencing subjects. The naive understanding of this persistence is in terms of a deep, metaphysical unity of consciousness over time. Parfit and others have argued effectively that it is hard even to make this naive understanding coherent, let alone give reasons for believing that we have it in ordinary survival. This argument seems to imply that ordinary survival cannot have the importance we usually attribute to it, perhaps having no importance at all.

An alternative approach has developed suggesting that persons should be conceived as most fundamentally agents rather than subjects of experience, and thus that the unity of a person over time should be defined in terms of unity of agency rather than unity of subject. This approach is seen in both metaphysics and practical philosophy. Perry and Rovane offer their alternatives within the context of the metaphysical debate, while Korsgaard and Frankfurt develop theirs mostly within the context of discussions of practical reason. I have suggested that as revealing as these views are, they still seem to leave out the importance our conception of ourselves as experiencing subjects has for us, an importance that is described in Camus's Myth of Sisyphus. Using hints found in the development of agency-based views of personal identity, however, I have pointed to a possible reintroduction of the picture of the person as subject into the agency-based views we have considered.

Coming from a Kantian perspective, Korsgaard says that persons are both subjects and agents, and we can take one perspective or the other depending on our purposes. I have suggested that for these purposes we must think of subjectivity and agency as inherently intertwined, and thus of unity of agency and unity of consciousness as separate but interdependent unities. In this way we can hope to regain a meaningful unity of consciousness as part of personal survival. Like the unity of agency described by Korsgaard, it will not have a metaphysical basis, but it will be nonarbitrary and deeply important nonetheless.

Footnotes

I am indebted to many friends and colleagues for their input in the course of writing this essay. I would like especially to thank David DeGrazia, Anthony Laden, Ray Martin, Marc Slors, and the editors of Social Philosophy and Policy.