Derek Parfit’s Reasons and Persons began, “Many of us want to know what we have most reason to do.” He aimed to establish a foundation for ethics, a project which required addressing the conflicts between morality and self-interest. In asking what the claims of self-interest really are, Parfit came to grips with the question of whether or not there is a rational basis for the special concern that persons feel for themselves. Although Reasons and Persons did not answer that question, Parfit tried to show that the Self-Interest Theory—which claims that each person has the “supremely rational ultimate aim: that his life go, for him, as well as possible”—is false.
Parfit went on to write On What Matters, following an honoured tradition of searching for a kind of Unified Field Theory of ethics—one law, from the correct application of which all moral precepts can be derived. Reasons and Persons famously argued that personal identity is not what matters in survival; and it is surely not coincidental that the phrase “what matters” recurred in the title of Parfit’s new book. Normative ethics consists of giving reasons for action; and reasons for action matter. Because the survival of persons is very important, what matters in survival can be presumed to be a significant part of what matters.
On What Matters says little about personal identity. Parfit’s primary interest is in discovering reasons for action—a rational basis for decision-making. My primary interest is in gaining a better understanding of human nature—of what we are. Parfit’s work is driven by prescriptive aims, mine by descriptive ones. Parfit and I would probably agree that reasons for action are rooted in values. But where he asks what has value, I ask what people actually value, and—a more interesting question—why they value what they do.
Is Similarity What Matters in Survival?
According to the Information Theory of Persons, we are our attributes. That suggests a corresponding claim about the value of survival: that what matters in survival is to preserve as many attributes as possible. When contemplating a transformation—whether an exotic one like teleportation or a familiar one like aging—we should be concerned about how many of our attributes (how much of our informational content) will be preserved. Surviving a transformation, on this model, is a matter of degree; and what matters is how successfully we preserve ourselves unchanged.
This model nicely fits other informational entities, such as books. (As before, I mean the intellectual work, not the bound volume.) Books are entities that survive replication, and the normal goal of replicating books is to preserve their informational content. A replication error is a loss of some of the book’s original information, and can be viewed as ‘incomplete survival.’ A copy of Moby Dick with one page left blank because of a printing glitch is still Moby Dick, not a different book. Rules for reidentifying books are loose enough to allow for some variation. But a copy of just the first chapter of Moby Dick is too little to qualify as Melville’s classic novel.
These days, books are almost always reproduced by automated processes which leave little room for error. But when the only way to copy intellectual works was manually, with the constant risk of human error, their survival was a matter of concern at each iteration. Scholars use the language of partial survival to describe this process, as seen in this excerpt from The Manuscript Tradition of Polybius by John Michael Moore:
It is probable that even if the text of Polybius was not actually complete, virtually the whole of it survived at the period when the excerpts were selected; two lacunae were noted by compilers of excerpts.… These were presumably lacunae in the particular manuscripts of Polybius used by the excerptors, though it is possible that Book XIV was partially lost by then; similarly, it is possible that the Books noted above as not having been used for the surviving titles of the excerpts may have been already lost in the tenth century…. [Moore, 1965, p 128, emphasis added]
A good copy of a book preserves all its words and punctuation in the correct order. And as I have already argued, the survival of a book consists in the continued existence of at least one instance, or copy. A book begins to exist when its original manuscript is written, and ceases to exist when its last copy is destroyed.
But there is more to the survival of books than their preservation. Arguably, after a book is finished, all that matters in its continued survival is to maintain its informational content, either by preserving copies of the book or by making new copies. That’s because (most) finished books are (almost) static entities—ignoring later editions for now. But there is another, more interesting stage in every book’s existence when it is not at all static—when it is being written. A book in progress is a reidentifiable entity. We do not say that every chapter the author adds (or every word, or every character) destroys one intellectual work and begins another. The concept of a book covers works in progress.
Since persons are dynamic rather than static entities, we might expect the identity criteria of books in progress to be an illuminating analogue to the identity criteria of persons.
Works in Progress
A parallel exists between the rules for identifying literary works-in-progress and the rules for identifying biological species. Both literary works and biological species can branch. Darwin’s Tree of Life has a branching form. New species are created when populations separate geographically and genetically diverge, as natural selection favours different mutations that are adaptive for different environments. Daniel Dennett pointed out the impossibility of telling that one species have become two when it happens.The facts required to establish speciation are only available in hindsight. Only after the two populations have been separated long enough to diverge genetically to the point that they will no longer interbreed, can biologists know they are two species, not one. After establishing that, they backdate the start of the new species to the branching event, when the geographical separation occurred.
This is not a point about our epistemic limitations—as if we would be able to tell when speciation occurs if only we had better microscopes, or even if we could get in a time machine and go back in time to observe the appropriate moments. This is a point about the objective property of being a speciation event. It is not a property that an event has simply by virture of its spatio-temporally local properties. [Dennett, 1995, p 96]
Intellectual works can branch too, when more than one product emerges from what began as a single project. This often occurs in software, where individuation is driven by marketing considerations: that is why we have Photoshop and Photoshop Elements, two products with different user interfaces, functionality, and price points based on the same core code. Many other examples are on offer from different software vendors. Folk songs, altered and embellished by different singers and musicians over many years, are particularly prone to branching. Book examples are rarer, but not unknown. The Old Testament is the Torah with chapters added. The Qu’ran, composed a millennium later, also contains the stories of Adam and Eve, Cain and Abel, Noah’s ark, Abraham and Isaac, Sodom and Gomorrah, compelling evidence of a common origin. And many authors have kept material cut from one book to use later in another. But like speciation, the bifurcation of literary works in progress cannot reliably be detected until long after the split occurs. The fact that an author cuts a chapter from his manuscript counts for nothing until the possibility that he will put it back is no longer open. Thus hindsight allows us to correctly describe an author as starting two books in a certain year, although she viewed them at the time as a single project.
Although, in the early stages of a literary endeavour, there may be no fact of the matter as to how many books the author is working on, we can nonetheless sensibly reidentify her work-in-progress from one month to the next, distinguishing it from other projects the same author has on the go. What unifies a literary work during the period it is being written are relations of continuity and connectedness. A first sentence is written, and expanded into an opening paragraph. After adding much more, the author decides the most dramatically effective order of her novel’s events is not chronological, and moves the first scene of chapter 1 to chapter 4. Later, on the advice of her editor, she cuts the scene entirely. It is the same novel throughout, known by its working title (which may change several times during the writing, and change again for publication), clearly distinct from her other novel (still incomplete, though started earlier) which resides in a different folder on her hard disk, and which she revises from time to time when she needs a break from working on the new one.
What transformations could threaten the survival of a novel in progress?
One is that work on the manuscript stops, and all copies are destroyed. But even destruction of all electronic and paper copies does not guarantee that the novel ceases to exist. If it remains active in the author’s brain, it may yet be written. The brain is another medium in which literary works reside. In Homer’s day, it was the main medium. Even in the 21st century a few people memorize poetry.
But if the author (all authors, in the case of multiple authorship) dies, and all manuscripts outside the author’s brain are destroyed, the novel has truly met its end. It would be a mistake to call a book written later by someone else the same novel as the one whose completion was prevented by the death of its author, however similar they might be—unless, perhaps, that similarity was the result of copying.
Another kind of survival-threatening transformation occurs when a work in progress is not destroyed, but changes greatly in character. J.L. West III describes an early version of F. Scott Fitzgerald’s The Great Gatsby published, many years later, under the title Trimalchio, as “not the same story,” one “different enough from Gatsby to deserve publication on its own.” West cites “crucial differences,” including the entire contents of Chapters VI and VII, the character of Nick Carraway, and the “unfolding” of Gatsby’s character. My purpose is not to support or undermine West’s claim that the published Gatsby was a different novel from the manuscript now called Trimalchio. I only use it as an example of a case which people who take books seriously might prefer to describe as two books rather than one. West’s decision to speak of two novels was driven by his scholarly purposes, and is, in a sense, conventional (which is not to say arbitrary). Being the same novel is a matter of grey areas and fine points.
Despite the greyness, it is clear that the survival of a literary work in progress is not directly related to the degree of connectedness between its successive temporal stages. In describing the process which unites the first sentence written to the final galley proofs as the writing of a novel, we are not making a claim about the similarity of successive drafts. A novelist is allowed to tear up chapters, even the entire manuscript, and begin again, without losing the right to describe all of her activity, before, during, and after the destruction, as work on the same novel. She destroys material because she “didn’t get it right the first time.” What is this ‘it’? The novel as it should be, I suppose. I don’t suggest that ‘the novel’ exists before it is written, as a Platonic ideal which the novelist is striving to instantiate. I only mean that the novelist is guided by her own goals and artistic vision, which she is trying to express in the public medium of language. Not all such attempts succeed. If the writer is good, she can tell, after she has written it, whether a passage or a scene communicates what she wished to communicate. She may not be able to tell at once. Months may go by before she can re-read a chapter with the fresh eye that allows her to judge whether it succeeded or failed.
Lack of change in an incomplete manuscript may be a sign of death—that the author’s vision has become clouded, or she has lost interest. For literary works in progress, as for living creatures, change is vital. The simple thesis that what matters in the survival of informational entities like novels is similarity between each version and the next is neither illuminating nor true. The survival of a novel is more complicated than that.
Some other informational entities, such as commercial databases, have less interesting identity criteria. The identity of a database seems to be fully described by relations of continuity and connectedness, with emphasis on continuity. Nothing corresponds to the author’s goals or values; the content of changes is immaterial. A database can survive deletions of unlimited numbers of records, even all its records. It can even survive structural changes such as adding and deleting columns or entire tables. A database can branch in two interestingly different ways. It can branch by redundant replication, for purposes of data security. This may take the form of periodic static back-ups, or dynamic updates in parallel to maintain full redundancy at all times. In such cases, it is usual to speak of different instances of a single database. In other cases, a database may be allowed to branch for the purpose of disseminating information to different locations which are not served by a common network. If multiple copies of what was originally one database are allowed to change independently, with no intention of reintegrating them, they are normally described as different databases after the branching event.
The ways in which databases split raise questions of identity similar to the questions we face when persons branch. When a split occurs, does the original database cease to exist, being replaced by two new ones? Or were there two all along? Both descriptions make sense, and each one suits legitimate rhetorical purposes. If I ask how long a database has been corrupted, the reply, “Since it began,” meaning when it was created by a branching event, does not answer my question. I want to know how far back I must go to find an ancestral copy of the database that is not corrupted. Clarity in such discussions is greatly aided by distinguishing between the database and its versions.
What matters in the survival of a novel is very different from what matters in the survival of a database. That fact should caution us against expecting the interesting, complicated, identity criteria of literary works-in-progress to closely model the identity criteria of persons, even though, as I argue, both are informational entities. What matters about the survival of an informational entity depends on what kind of informational entity it is. One element—preservation of informational content—is common to all, but attempts to generalize much beyond that fail. Consideration of informational artifacts like books, songs, poems, databases, and the common programs running on our computers—none of which are conscious, or have internalized values—may not shed much further light on what matters in the survival of persons. Until artificial informational entities become much more sophisticated than they are today, the primary examples of what matters in the survival of persons must unfortunately all be human. That lack of variety makes our task harder than it otherwise would be.
Like literary works in progress, persons are dynamic, not static entities. Unlike literary works, persons never become entirely static while they live (although some are quite set in their ways).
Also unlike literary works, persons are to a large extent their own authors. Many people consciously and actively engage in self-transformation. They strive to change. Whether their goal is to become fluent in Mandarin, or to win political office, or to lose weight, they regard self-chosen change, not as threatening their survival, but as survival-enhancing.
This resembles the way an author regards her changes to a manuscript as progress. Both are processes of becoming. As the identity of a novel is determined primarily by the author’s vision and intent, it might seem that what matters in personal identity is each person’s vision and intent with respect to himself—preservation of his values and realization of his goals. But I advise caution against making too much of this similarity. However we may like to think of ourselves as self-written novels, many of us, much of the time, feel like databases, passively recording the slings and arrows of outrageous fortune.
Personal survival consists in preserving and replicating attributes. But not all a person’s attributes are equally valuable to him, and some are not valuable at all. I would willingly lose some of my attributes if I had the choice—my bald spot, and a tendency to shyness. What I want is to preserve those attributes that are important to me, and to change in ways that promote my current values.
Some of our attributes are precious to us, others we detest, and many are objects of our indifference. That is what is mainly wrong with the idea that similarity is what matters in survival. What matters is not our attributes as such, but how we value them.
This view of survival throws a new light on the relationship between the different stages of human life—in particular, between adults and the children that they were. Things that are important to a young child often become unimportant in adult life, and adults typically hold values to which they gave no thought as children, or which they may even have rejected.
Not only values and interests, but personality can change profoundly as we grow up. People “put away childish things;” they undergo religious conversions; they discover professional callings; their interest in the opposite sex often changes from mild disdain to fascination; they may develop the desire to have children, and, when they have children, adopt new values and goals related to their children, as well as radical personality changes related to their role as parents, with responsibility for the lives of others. By the time people are adults, even their memories of the first three or four years of childood are sparse and inaccurate, difficult to distinguish from autobiographical narratives generated later and often-told family legends. In their psychological attributes, many adults do not resemble the young children they were more closely than they resemble many other children, on any objective measure of similarity.
Although psychological connectedness may be almost entirely absent between a child of three and a man of thirty-three, psychological continuity—similarity from each day to the next—can still be invoked to explain the unity of a life. But this discussion is about what matters in survival; and, when I consider my own survival, I find the importance of continuity to be derivative. I do care about preserving many of my attributes from day to day—I need to maintain memories of recent events, and many short-term and medium-term intentions, in order to advance my goals effectively. I don’t much care about continuity for its own sake. If I face the prospect of some transformation—if, for example, I am diagnosed with a disease leading to catastrophic dementia—I am more concerned with the final result than with the pace of change. I would regret the loss of certain kinds of connectedness with my current self, but disruptions of continuity do not, in themselves, trouble me. And if the change is one I welcome, it does not become more welcome by coming gradually.
If I have psychological conflict of certain kinds, I may want to change radically. I may, like Hamlet, desire to change from a ditherer to a decisive man of action. I may desire a radical change in my character, which will change my behaviour and lead to other changes in beliefs, feelings, and lifestyle. I imagine surviving this change; indeed, the whole point of trying to change is predicated on me surviving the change.
Could someone desire to change so radically as to disrupt his psychological connectedness? Probably, yes. Two characters in the movie Eternal Sunshine of the Spotless Mind choose to have memories of their relationship—the most emotionally significant aspect of their lives—erased. If memory erasure were technically feasible, a likely candidate for it would be an ex-soldier suffering the hell of post-traumatic stress disorder. It’s not hard to imagine someone returning from the horror of war who chooses to erase, not only its memories, but personality changes that the war induced, the hair-trigger aggression, depression, guilt, shame, and panic attacks. Someone might understandably want to remove all traces of the war years from his brain, and restore his happy pre-war psychology. If he could do it in one quick operation, rather than gradually, so much the better. Although his comrades-in-arms might not recognize him after the procedure, we should not conclude that somebody died—unless we are prepared to say that the carefree young lad who signed up four years ago to serve his country also died, when he was transformed into a PTSD wreck. For the wreck to desire to change back, suddenly, to something like his pre-wrecked condition is not like desiring to die.
What makes the returned soldier’s desire to change back into someone like he was four years ago unlike the desire to commit suicide is the fact that he has self-concern for the survivor of the memory-erasing operation. That the change is of his own choosing may make it easier for him to extend self-concern across the transformation. Another ex-soldier might reject the offer of the mind-cleansing operation, on grounds that, even though he is greatly troubled by his war experiences, they are his experiences, which made him what he is today. He remembers his pre-war self as naïve, innocent of the grim knowledge of the world and human nature which he now possesses, and for which he paid dearly. Perhaps he intends to write a novel. If some benevolent veterans’ administration proposed to make him undergo the erasure for his own good, this soldier might regard it as close to a death sentence.
What we choose in the way of transformations does seem relevant to what we can expect to survive. But, as was shown in “Death, Revisited,” choosing a transformation is neither necessary nor sufficient for surviving it. We survive countless small changes we do not choose; and even suicide can be chosen.
I have defended what Derek Parfit called The Extreme Claim: that we have no special reason to care about our future experiences. You are not rationally required to have concern for your own future. No relation in nature provides rational justification for the strongly motivational attitude that is anticipation of future experience.
Parfit’s primary goal in Reasons and Persons was to defeat the Self-Interest Theory. He described its main claim as:
For each person, there is one supremely rational aim: that his life go, for him, as well as possible. [Parfit, 1984, p 4]
Parfit thought the Extreme Claim could not be proven, and chose another argument against the Self-Interest Theory. He acknowledged, however, that if the Extreme Claim is true, the Self-Interest Theory is false. We now see that naked self-interest has no rational force. A person’s reasons for action cannot be grounded in anticipation of future experience.
To say that self-concern is not rationally required is not to say we never have rational grounds for caring about ourselves in the future. Such grounds are provided by our present values and goals, excluding any values and goals which depend on the belief that anticipation of experience is rationally required. Those values and goals should be excluded because the belief they depend on is irrational.
Reasons and values are closely linked: to value an outcome is to have a reason to bring it about. The value I place on a habitable and pleasant natural environment provides me with a reason to act so as to protect the environment. Since I cannot do this alone, it also gives me a reason to care about the well-being of other people who are working to protect the environment, or who will do so in the future.
Promoting my present values includes working towards my goals, completing projects I have begun, and fulfilling my current intentions. It includes helping people I care for. It may even include having experiences I would like to have.
Our choices express our values. That explains how choices are relevant to survival. Transformations which a person chooses are likelier to promote his values than transformations he undergoes unwillingly. His values give him reasons to prefer those transformations. It explains why one soldier might regard undoing the psychological changes of his war years as “giving him his life back,” while another would think of the procedure as destructive of himself.
Our values can give us reasons for caring about ourselves in the future. If someone in the future is likely to promote my values and advance my goals, that fact gives me a reason for concern about that person’s well-being. Some of my goals and projects are more likely to be advanced by myself than by any other person. That fact (as John Perry argued) gives me a reason for special concern about my own well-being.
That reason would disappear if my values were unstable, changing radically from each day to the next. If I am inconstant in my enthusiasms, fighting for communism last week, unbridled capitalism this week, and as likely as not to support a hardline theocracy next week, the value I place on unregulated markets today does not give me a reason to stay alive to vote in next week’s election. I’m too likely to vote the wrong way.
Most people’s core values are fairly constant. That gives them a reason to care about their well-being in the future. It also gives them a reason to care about the well-being of anyone else who shares their values and is likely to promote them effectively.
Consider the principle:
Having shared values is a relation that provides a person with a reason for concern for someone in the future.
This relation can be called “value connectedness.” One difference between it and psychological connectedness is that value connectedness is more obviously a relation we can have to other people as well as to ourselves. We need not resort to future-tech examples of human replication or brain hemisphere transplants to appreciate that different people can hold values in common, just as a single person can hold the same values at different times.
My claim may seem obviously true, but too weak to be interesting. Yet this stronger version:
Having shared values is the only relation that provides a person with a reason for concern for someone in the future.
is false. Someone may value another person’s well being for its own sake. The love which many parents have for their children is not predicated on shared values—especially when the children are very young. Unconditional love for her child may provide a mother with a reason for protecting the child’s interests, which is independent of any values the child holds, or will ever hold.
Our actual values matter. Does anything else? If not, someone might defend a weakened form of the Self-Interest Theory by arguing that although it may not be true for everyone, it is true for anyone whose values are invested solely in himself. Someone whose values and goals are entirely selfish is an egoist. Is it true that an egoist’s supremely rational aim is that his own life go as well as possible?
Many ethicists would answer that each of us has reasons for action which are independent of the values he actually holds—reasons grounded in objective value, sanctioned by moral principle. Much can be said for that point of view, but it lies outside the descriptive realm which is the domain of science. Instead of weighing in on the question of whether values exist independently of what human beings actually value, I will comment on this weakened egoist position using principles already developed here.
The Narcissist’s Defence of Egoism
To cease to believe that anticipation is rationally required is to undergo a profound shift in one’s values. No longer are “Because it’s me,” and, “Because it’s mine,” fundamental reasons for action. If they are reasons at all, they are derivative ones, grounded in my values.
Could an egoist accept that, and remain an egoist?
Imagine someone who cares only about himself, not because he believes it is rationally required, but simply because he cares about nothing else. He is in love with himself, to the exclusion of everyone and everything else. He is a Narcissist. A Narcissist’s present values seem to provide him with a reason to care about his own future interests, which he does not have to care about the interests of others, except derivatively when his own interests depend on the well being of others.
Narcissus was a vain young man who spurned all admirers of both sexes. Robert Graves described how, after one committed suicide, the goddess Artemis punished Narcissus by showing him his reflection in a still pool.
As he cast himself down, exhausted, on the grassy verge to slake his thirst, he fell in love with his reflection. At first he tried to embrace and kiss the beautiful boy who confronted him, but presently recognized himself, and lay gazing enraptured into the pool, hour and hour. [Graves, 1955, p 287]
Unable to tear his gaze from his own image, Narcissus starved and died.
The legend of Narcissus’s unenviable life does not show that Narcissists are irrational, but that they are impoverished and unhappy. The argument against Narcissism appeals, ironically, to self-interest. Because there is no independent reason to prefer the values of self-interest to other values, the argument has force. To be concerned exclusively about oneself is a recipe for misery, for the Narcissist as well as for those around him. It may not be irrational, but it is a misfortune.
Losing One’s Soul
If rational grounds for concern for oneself in the future are only located in one’s present values and goals, then a transformation which compromises one’s values may threaten the only relation which in fact matters in one’s survival. The compromise itself is profoundly irrational, because it destroys what it aims to protect.
That is what is mainly wrong with decisions like the one made by the Bush-Cheney administration to use torture and indefinite arrest without charge in the so-called ‘War Against Terror.’ The use of torture and suspension of due process were contrary to fundamental values affirmed by the most important documents and traditions by which the United States has distinguished itself from its enemies—not only its Constitution and the Geneva Convention, but the lessons most of us, American or not, were taught in school: that what’s valuable in our civilization depends on the rule of law, and torture is a great evil. If reasons for action are rooted only in our values, a person can have no reason to act contrary to his most fundamental values.
The most important differences between persons are differences in their values and goals. If I adopt the values of my enemy in order to protect myself, I am left with little that is worth protecting. If I transform myself into someone indistinguishable from my enemy, my enemy wins.
Torture is correctly criticized as being an ineffective means of obtaining accurate information. But that should not be the main reason for rejecting it. Some other reprehensible practice might turn out to be an effective weapon against our enemies, yet we should not adopt it. The first reason for rejecting reprehensible practices should always be that they are reprehensible—contrary to our values, the ones that make our side worth defending.
To suppose that there is any reason to the contrary would be to fall victim to the illusion that one’s mere continued existence, regardless of how one changes, matters in the least.
Dennett, Daniel C. (1995), Darwin’s Dangerous Idea, Simon and Schuster
Graves, Robert (1955), The Greek Myths, Penguin Books
Moore, John Michael (1965), The Manuscript Tradition of Polybius+, Cambridge University Press
Parfit, Derek (1984), Reasons and Persons, Oxford University Press.
Parfit, Derek (2011), On What Matters , Oxford University Press.
Perry, John “The Importance of Being Identical,” in Rorty (ed.), The Identities of Persons (1976), University of California Press.
West, James L.W. (2000), “Almost a Masterpiece,” Humanities, January/Febuary 2000, Volume 21/Number 1