Parfit's Reasons and Persons: Rationality, Identity and Population Ethics

Suppose you could prevent a future hardship only by changing which child will be born. Waiting would lead to a different conception and a different person. The child who actually arrives has a life worth living, although the alternative child would have had a better start. Something about choosing the worse future seems open to criticism, yet the familiar complaint that we made this particular child worse off becomes difficult to state.

Derek Parfit develops this difficulty in Reasons and Persons, first published in 1984. The book also examines self-interest, attitudes toward time and personal identity. These subjects are connected by questions about what we have reason to do, but they are not all solved by a single theory of the self. The population problems remain problems even after Parfit's account of personal identity is understood.

The value of reading the book lies partly in learning to keep these questions apart. A theory of what makes someone the same person does not automatically determine how much concern that person deserves. A judgment that a future is worse does not always identify someone for whom it is worse. Following those distinctions reveals why several ordinary ways of justifying conduct need more careful support.

Reasons and Persons has four parts. The first examines theories that can be self-defeating. The second concerns rationality and time. The third investigates personal identity. The fourth addresses future generations and population ethics. Beginning with the personal-identity thought experiments is possible, but it can conceal the earlier question about how theories of reasons should guide conduct.

A theory can tell us to pursue an aim while a particular way of following it frustrates that aim. Parfit distinguishes different forms of this problem. A theory's recommended dispositions or methods of deliberation can have poor results even when its criterion for assessing outcomes remains defensible. Consequently, identifying a self-defeating pattern does not always establish that the theory should be abandoned.

Consider a hypothetical friendship conducted through constant calculations of personal advantage. Someone may monitor every conversation for a better return and thereby undermine the trust they hoped would benefit them. This is an illustration of a possible tension between an aim and a way of pursuing it. It is not an empirical finding that every deliberate attempt to promote one's welfare damages friendship.

Parfit calls the Self-interest Theory S. As a theory of rationality, it gives each person the aim of making their own life go as well as possible. This differs from psychological egoism, the claim that people always act from self-interested motives. A normative theory can tell someone what to aim at while recognizing that they frequently want or do something else.

It also differs from a theory of well-being. To apply the self-interest aim, we need an account of what makes a life go well for the person living it. Parfit distinguishes hedonistic, desire-fulfillment and objective-list accounts. These locate well-being, respectively, in happiness or pleasure, fulfillment of relevant desires, and goods whose value need not depend on being desired. Rejecting one account of well-being would not by itself refute every form of self-interest reasoning.

The broader lesson is about levels of explanation. A criterion of rational action, an account of welfare and a recommended motive answer different questions. They may constrain one another, but we should not move between them unnoticed. An argument against constantly calculating personal advantage is not yet an argument that one's own welfare has no importance.

In the second part, Parfit examines attitudes toward different times. A familiar contrast concerns suffering that lies ahead and suffering already over. A person might strongly prefer that an unpleasant experience be in the past, even when the past experience would have lasted longer. The question is whether the preference can be justified, rather than merely whether it is recognizable.

His discussion separates temporal position from other relevant differences. Uncertainty about a distant outcome can matter to a present decision. A difference in what we can influence can matter too. Neither point establishes that a benefit deserves less concern simply because it occurs later. The reason for the discount needs to be identified rather than smuggled into the date.

Imagine choosing between helping your future self avoid discomfort next week and helping that same person avoid a much greater difficulty years from now. This invented example does not prescribe a particular financial or medical choice. It isolates the question of whether temporal distance alone supplies a reason to neglect the larger difficulty, assuming the other relevant conditions have been held fixed.

Parfit uses such questions to test the authority of self-interest over a whole life. Why should belonging to the same life make benefits at different times equally important to rational concern? Why should the distinction between different people have the significance assigned to it? These are questions about justification. The fact that we care strongly about our future does not answer them by itself.

Numerical identity means being one and the same individual. Qualitative similarity means sharing features. Two people can look alike without being the same person, while one person can change greatly and remain numerically identical. Parfit uses this distinction to prevent the phrase “the same person” from doing several jobs at once.

The distinction becomes useful in thought experiments involving replicas. Imagine a device that records a person's physical and psychological organization and creates a matching person elsewhere. We can stipulate very close resemblance while still asking whether the resulting individual is numerically identical to the original. Establishing that the replica would remember a childhood does not settle that question merely by repeating the word same.

These are philosophical cases, not descriptions of available technology. In Parfit's discussion, they separate bodily persistence from psychological continuation. A thought experiment asks what we would say if these came apart. Its conclusion depends on the case being coherent and on the relevance of the judgments it elicits.

Parfit's reductionism says that personal identity consists in more particular facts, rather than in the persistence of a separately existing subject over and above them. Physical and psychological accounts can both take a reductionist form. Reductionism therefore should not be defined simply as the claim that memories alone determine who someone is.

The view also does not say that people are imaginary. An institution can exist through facts about people and their activities without being a separate additional substance. That comparison helps explain the form of reductionism; it does not prove that persons are institutions. Parfit's substantive argument concerns which account fits the facts about persons and the possible cases.

Psychological connectedness concerns particular direct connections. Remembering an experience is one example; carrying out an earlier intention is another. Persistence of beliefs, desires or character can also contribute. The extent of these connections can vary. Someone may have many direct links with their recent past and fewer with a much earlier period.

Psychological continuity concerns overlapping chains of strong connectedness. The distinction matters when direct links fade. An adult need not remember every experience of early childhood for there to have been a chain of strongly connected stages through the intervening years. Continuity is not merely the presence of a current autobiographical memory.

Parfit calls psychological connectedness and/or continuity, with the right kind of cause, Relation R. The causal qualification prevents us from treating accidental resemblance as equivalent to the relationship at issue. He further argues for a broad view about admissible causes. The characterization of R and the defense of which causes count are distinguishable steps in his account.

These terms help make his central claim precise: identity is not what fundamentally matters in the prudential concern associated with survival. The relations captured by R can carry that importance. This does not mean that identity is never relevant to a practical question. It concerns the role identity plays in explaining why we should care about a person's future.

It would also be too quick to translate degrees of connectedness into a fixed schedule of concern. Nothing in the definition itself tells us exactly how much less to care when a direct memory is lost. An account of the psychological facts still needs an argument about their normative weight. Forgetting an episode does not by itself remove obligations or make injury unimportant.

In Parfit's division case, a person's brain is divided and its halves are successfully transplanted into two other bodies. The case stipulates that each resulting person preserves the relevant psychological life. The details are imagined to make both continuers equally good candidates; they are not a claim that such an operation is currently possible.

If the two resulting people are distinct, the original cannot be numerically identical to both. Identity is transitive: if each were identical to the original, they would be identical to one another. That conflicts with the stipulated existence of two different individuals. The argument concerns numerical identity, not whether both have similar memories.

Yet each branch can retain relationships that would ordinarily give the original person reason to anticipate a future with concern. Parfit asks why producing a second equally good continuer should turn the first one's otherwise valuable continuation into something equivalent to ordinary death. This is the pressure behind separating what matters in survival from identity itself.

The case does not force every opponent to accept reductionism. Someone can question its possibility, its implications or the account of prudential importance. The useful achievement is a more exact problem: what explains the significance we assign to numerical identity when psychological continuation can be stipulated without a unique continuer?

Nor does the case show that distinctions between people have no moral importance. Two people after division still have different experiences and may suffer different injuries. Their claims cannot simply be merged because of their common past. Parfit explores implications for morality, but those require arguments beyond the logical observation that one individual cannot be two distinct individuals.

The third part asks how a reductionist understanding should affect prudence and morality. If a separately existing ego is not what unifies a life, one proposed justification for an absolute boundary around self-interest loses support. That leaves room to reconsider the weight of psychological relationships and the significance of distinctions within and between lives.

This is a challenge to a justification of exclusive self-concern, rather than a proof that all special concern is irrational. Projects, attachments and commitments can supply reasons that do not depend on a further metaphysical fact. A person can continue to plan for old age while rejecting the claim that a simple persisting substance explains why those plans matter.

Parfit also describes changes in his own attitude toward death associated with reductionism. That is a philosophical and personal reflection, not evidence that adopting the view reliably reduces fear. A reader may find the view consoling, unsettling or unpersuasive. None of those reactions alone determines whether the account is true.

The distinction protects the inquiry from becoming advice it cannot support. An abstract identity argument does not tell someone to abandon relationships or disregard their future welfare. Its value is in making the grounds of concern available for examination. The ethical conclusions need the same care as the metaphysical claims on which they draw.

Part Four considers choices that affect which future people exist. If a decision changes when someone is conceived, it may result in a different person. In such a case, comparing two policies can mean comparing futures with different people, rather than better and worse lives for a single fixed group.

Parfit's example of a fourteen-year-old concerns a child given a bad start because the mother is so young. If she had waited, she would have had a different child with a better start. The actual child's life is stipulated to be worth living. The example should not be retold as though its defining premise were a known congenital condition; that changes the case.

If we criticize the earlier choice, we cannot straightforwardly explain the criticism by saying that waiting would have made this particular child better off. On the stipulation, that child would not have existed. This exposes a limitation in a simple comparative account of harm. It does not establish that the choice is permissible, or that every broader use of the word harm must fail.

The problem can also arise in choices about resources that alter the course of future lives. One policy may lead to worse living conditions but also to different people being conceived. Saying that those people would have had better lives under the alternative can then be false. We need a justification for preferring the better future that recognizes the change in identity.

The challenge is sharper than a general reminder to care about future generations. We may already agree that future people count. What remains unsettled is how to evaluate a choice that seems objectionable although it is worse for no particular person in the simple comparative sense. Reductionism about identity across one life does not dissolve that comparison between different possible people.

Population ethics introduces another dimension: alternatives may contain different numbers of people. A total view adds the well-being of everyone in a population. For a fixed finite population with very high well-being, a sufficiently larger population with small positive well-being per person can have a greater total. This comparison assumes additively comparable well-being and holds other morally relevant features equal; it does not establish that the total view is correct.

Parfit calls the resulting implication the Repugnant Conclusion. His formulation in chapter 17 starts with any possible population of at least ten billion people, all enjoying a very high quality of life. It says that some much larger imaginable population would be better, other things equal, even though its members' lives would be barely worth living. The troubling feature is the trade between the quality and quantity of lives, rather than a claim that lives below the worth-living threshold become good merely by multiplying them.

Replacing totals with averages creates different difficulties. A rule that attends only to average well-being can oppose adding people with good lives when their lives fall below the existing average, even if the original population is unaffected. This does not conclusively establish the total view. It shows why rejecting one counterintuitive implication does not finish the search for a satisfactory theory.

The Mere Addition discussion examines pressures created by comparisons among populations. Apparently acceptable steps can lead toward conclusions we find hard to accept. Each comparison needs its own justification. A diagram or a sequence of agreeable-sounding descriptions would not remove the need to examine the assumptions connecting the steps.

Parfit's search for a satisfactory Theory X is unfinished in this book. He does not present reductionism as a settled solution to the Repugnant Conclusion. The distinction matters because the book teaches us to recognize competing requirements whose reconciliation remains difficult. Understanding the problem is an achievement even when the desired comprehensive principle has not been supplied.

A useful reading exercise is to take one case and write down what changes between the alternatives. Does the same person exist with a different welfare level? Does one person have two psychological continuers? Do different people exist? Does the number of lives change? These variations produce different questions, and an answer to one should not be silently carried into another.

Next separate description from evaluation. List the stipulated physical or psychological relations, then identify the proposed reason for concern. If an argument moves from fading connectedness to weaker obligations, locate the premise that licenses that move. If it moves from no one being worse off to no wrong being done, examine that inference too.

Reasons and Persons makes these transitions difficult to ignore. It asks us to defend the significance of identity and time, and to examine futures whose quality cannot be described solely through gains to particular existing people. Its questions remain demanding because our judgments must answer to several considerations at once. Clearer distinctions give us a better way to test those judgments.

• Derek Parfit, Reasons and Persons. Read the four parts as connected inquiries rather than a single deduction from reductionism.
• Part Three, especially the discussions of the psychological criterion, division and what matters. These develop the distinction between identity and Relation R.
• Part Four, beginning with “The Non-Identity Problem,” then “The Repugnant Conclusion” and “The Mere Addition Paradox.” These examine the separate difficulties concerning future people and population size.
• Derek Parfit, On What Matters. The later work develops questions about reasons and the possible convergence of moral theories.