Parfit’s Retreat: “We Are Not Human Beings”

In a new paper, “We Are Not Human Beings,” Derek Parfit argues that persons are identically their conscious, thinking parts, which he identifies as their cerebrums.  This is a significant departure from the position he defended in Reasons and Persons, that personal identity consists in non-branching psychological continuity and connectedness with any cause:

Our identity over time just involves (a) Relation R—psychological connectedness and/or psychological continuity, either with the normal cause or with any cause, provided (b) that there is no different person who is R-related to us as we once were. [Parfit, 1984, p 216]

I call Parfit’s new view a “retreat” because it is a move away from the radical insights about what we are which illuminated Reasons and Persons, to a ‘conservative’ account of persons as physical substances. I find the move puzzling, because I can’t see that Parfit is compelled to make it, and disappointing, because it raises once again the fog of mysteries about persons that looked well on their way to being dispelled.

Parfit’s claim that persons are their cerebrums has as a direct consequence that persons cannot survive information-based teleportation. If  I plan to be teleported to Mars, I should accept that my replica on Mars will not be me, because my replica’s cerebrum is numerically different from my cerebrum. The cerebrum is a body part, which, like any other ordinary material object, ceases to exist when it is destroyed. Its replica on Mars is a different cerebrum—hence, if Parfit is right, a different person.

Parfit has long thought that survival—a person’s continued existence—is different from what matters in survival. His new view on what persons are could perhaps coexist with his earlier position that information-based teleportation preserves everything that matters in survival. But such coexistence, I will argue, is an uneasy truce between fundamentally warring ideas. An alternative account of what persons are—informational entities—is a better fit to Parfit’s intuition (which I endorse) that nothing important need be lost in teleportation of persons.

“Why We Are Not Human Beings” is Parfit’s response to animalism—the view put forward by Eric Olson and others that persons are identical to animals, or biological organisms. (In the animalist literature, “human being” is used as a synonym for “human animal.”) In this review of Parfit’s paper, I raise the following points:

  1. The arguments Parfit brings to bear against animalism rely on an intuition that has equal force against the paper’s conclusion that persons are their cerebrums.
  2. The claim that persons are their brains was strongly rejected in Reasons and Persons for reasons that supported a central argument of that book. If we were identical to our cerebrums, Parfit’s main argument against the Self-Interest Theory would be undercut in the same way that it would be if we were identical to our (whole) brains.
  3. The claim that we are our cerebrums weakens Parfit’s argument in Reasons and Persons that “ordinary survival is about as bad as being destroyed and having a Replica.” If I am my cerebrum, it is hard to believe that destruction of my cerebrum is not especially bad for me, even if a replica of my cerebrum is manufactured in its stead.

The Cerebrum and the Brain

The cerebrum consists of the two cerebral hemispheres and the neural tissue connecting them. It is distinguished from the whole brain by excluding the brain stem. Parfit prefers the cerebrum to the whole brain as the bearer of personhood on grounds that the cerebrum is “the conscious, thinking part” of a human being.

That is scientifically doubtful. Neuroscientist Antonio Damasio reports clinical experience with hydraencephalic children who, although born with little more than a brain-stem, show fear, emotional bonding, and musical appreciation, leading him to conclude, “The condition gives the lie to the claim that sentience, feelings, and emotions arise only out of the cerebral cortex.” [Damasio, 2010, p. 82]  Parfit’s argument would not, however, have been importantly different had he claimed that persons are identical to their whole brains instead of to their cerebrums. Animalists claim that persons are identically their entire biological organisms. It is open to Parfit to oppose them by championing the view that a person is identical to whatever physical organ constitutes her “conscious, thinking part.”   If his argument is sound, its force should not be diminished if science discovers that some organ other than the cerebrum fits that description.

Parfit’s reasons for denying that we are animals are also reasons for denying that we are our cerebrums

Parfit’s central arguments against animalism turn on thought-experiments in which his cerebrum is removed from his body and kept alive. In one case, his entire brain remains within his head, which is transplanted to Bernard Williams’s body; in another, Parfit’s head with his cerebrum, but without the brain-stem, is transplanted; in a third, his head is attached to an artificial life support system; and in the fourth, his detached cerebrum is put on life support, where it remains conscious and is given means of communicating with visitors and dictating the remainder of Parfit’s unfinished book. Parfit argues convincingly that in all such cases, most people would be strongly inclined to regard the survivor of the operation as Derek Parfit. About the head-transplant case, he says:

Suppose that you knew both Williams and me, and you visit the resulting person in the post operative recovery room. You see my head on the pillow, and have a long conversation with someone whom you assume to be me. If some nurse then lifted the blankets on the bed, and you saw the rest of what you knew to be Williams’s body, you wouldn’t conclude that you weren’t, as you assumed, talking to me. You would believe that the person with my head would be me. As many Animalists concede, this widely held belief, which some call the Transplant Intuition, provides a strong objection to their view. [Parfit, 2012, p 9]

A parallel argument supports the claim that information-based teleportation preserves personal identity. Suppose we are living in the not-too-distant future, around 2080, when teleportation is no longer a novelty, but a widely-used replacement for air travel. One Monday morning, my colleague Bob shows up late for work, looking tired.  I ask him about some problems with his project which were identified in last week’s review. Between yawns, he explains how he solved them, including details which only Bob would know. Reassured as to the project status, I enquire about his weekend. He tells me he spent it in Melbourne, Australia, which he just left this morning (Vancouver time) at the end of a long and active day (Melbourne time) spent surfing with his brother’s family. That, I now realize, explains why he’s so tired. It also informs me that he is a replica of the Bob who spent a weekend in Australia, who was a replica of the Bob who worked with me in the Vancouver office last week. I am not disturbed by this; after all, I too have been replicated many times.

Learning this fact about Bob, I would not conclude that he is not, as I assumed, the person who shared my office last week. I would continue to believe that this Bob, who spoke with such evident, intimate knowledge of his project, is the same man who lent me $50, of which he now reminds me, and which I gratefully repay. In a society in which teleportation is commonplace, the belief that Bob would have survived his trip will be widely held. A philosopher in that social milieu (not so far in the future that we in 2012 cannot imagine it clearly) could argue that the teleportation-survivor is Bob, with the same force and the same reason that Parfit today can argue that the survivor of his imagined head-transplant operation would be himself.

Parfit’s argument depends entirely on the intuitions of his hospital visitors—their stubborn belief that the survivor of the head transplant is Parfit, not Williams. He does not offer a reason for giving the beliefs of his imagined visitors any more weight than the beliefs of my imagined users of teleportation technology. If anyone objects that the intuitions of people in a teleporting society cannot be assessed because teleportation is not yet technically feasible, I will reply that neither are head transplants technically feasible, so how can we assess the intuitions of people in an imagined society that transplants heads? Arguably, such a society would be stranger to us than a society of teleportation-users.

I have argued elsewhere for two predictions of fact: that information-based teleportation will be one day be commonplace, and that its users will think of it as a transformation which they survive. Here, I will comment briefly on the second. Users of teleportation and other non-branching applications of human replication technology (for medical treatments, or life insurance) will certainly invest their special concern for the future (which, following Ray Martin, I call “self-concern”) in the surviving replica; if they did not, they would not use this technology. Excepting a handful of philosophers, everyone’s investment of self-concern stands or falls with his or her first-person judgements of personal identity. There is a strong psychological association between believing that someone is oneself and having the distinctive emotions of self-concern for that person. The strength of that association explains the widespread resistance to the claim that personal identity is not what matters in survival. That is a psychological fact about people, which I predict will not change as a result of adopting human replication technology, provided that branching does not become commonplace. Since human branching would cause many social and legal problems, I also predict that it will not become commonplace because it will be legally proscribed, on pain of severe penalties for the survivors.

If anyone finds Parfit’s head-transplant case more convincing than my teleportation case, I suspect that is because he is influenced by the idea that persons must be some sort of physical object—either their living bodies, or some part of their bodies. This assumption that persons, whatever they are, must be physical substances might be thought to be required by scientific materialism. I have argued that it is not—that a account of persons as informational entities, which are immaterial in themselves although instantiated in material substances—is compatible with a scientific view of the world. If my argument succeeds, I don’t see any remaining reason to favour the intuitions of Parfit’s hospital visitors over the intuitions of my teleportation users.

If persons are identically their cerebrums, Parfit’s argument against the Self-Interest Theory fails

In Appendix D of Reasons and Persons, “Nagel’s Brain”, Parfit argues against Nagel’s view that we are our brains.

On my view, one of the two relations which matter, psychological connectedness, holds over time to reduced degrees. This is an essential premise of my argument, in Chapter 14, against the Self-Interest Theory. This argument would be undermined if Nagel’s view is true. The continued existence of the same brain, in our actual lives, is not a matter of degree. [Parfit, 1984, p 479. Emphasis added]

Like that of a brain, the continued existence of a cerebrum is all-or-nothing, not a matter of degree. Parfit’s argument against the Self-Interest Theory is similarly undermined if we are our cerebrums.

The Self-Interest Theory, which Parfit set out to defeat:

…claims that, for each person, there is one supremely rational ultimate aim: that things go as well as possible for himself. A rational agent should both have, and be ultimately governed by, a temporally neutral bias in his own favour. It is irrational for anyone to do what he believes will be worse for himself. [Parfit, 1984, p 307]

Parfit insists it is essential to the Self-Interest Theory that the bias be “temporally neutral.”

Central to this theory is: The Requirement of Equal Concern: A rational person should be equally  concerned about all the parts of his future. [Parfit, 1984, p 313]

In support of this claim he invokes Sidgwick:

My feelings a year hence should be just as important to me as my feelings next minute, if only I could make an equally sure forecast of them. Indeed this equal and impartial concern for all parts of one’s conscious life is perhaps the most prominent element in the common notion of the rational…[Sidgwick, 1907, p. 124]

Parfit adds that according to the Self-Interest Theory “it is clearly irrational to postpone an ordeal if one knows that this would make this ordeal worse.”

This temporal neutrality of the Self-Interest Theory is a weak point, which Parfit proceeds to attack.  He invokes his previously-defended claim that:

What fundamentally matters are psychological connectedness and continuity.

Offering examples of other relations that can plausibly be described as less important when they hold in reduced degrees—friendship, indebtedness, and kinship—Parfit suggests that psychological connectedness shares this characteristic. He imagines being able to  choose between a day of pain either tomorrow or in forty years, and concludes that “it cannot be irrational for me to care less” about the pain in the distant future “when there will be much less connectedness.” On these grounds, Parfit rejects temporal neutrality as a rational requirement of self-interested concern for the future.

He then considers the possibility of revising the Self-Interest Theory so as to exclude that requirement. On the revised theory, it may be rationally permitted to prefer a smaller benefit in the near future to a larger (and equally certain) benefit in the distant future. But that, Parfit points out, directly contradicts the Self-Interest Theory’s core tenet that each person is rationally required to prefer that his own life go as well as possible. The revision of the Self-Interest Theory amounts to its denial.

Rejection of the Self-Interest Theory means that acts of great imprudence—such as smoking in one’s youth, heedless of the risk to the health and longevity of one’s mature self—cannot be adequately criticized on grounds of irrationality. Parfit proposes that they can be criticized on moral grounds instead. Our obligations towards ourselves in the distant future, when we may have changed greatly, are more like moral obligations towards other people than rational obligations towards ourselves. I endorse this conclusion, which goes some way towards levelling the playing field in the contest between the claims of morality and the claims of self-interest. I think there are stronger arguments for it than Parfit presents here. (Elsewhere I have argued for the Extreme Claim that prudential concern is never rationally required.)  But Parfit’s argument does have some force, if paraphrased along these lines:

If there is any rational requirement for prudential concern for oneself in the future, it must be grounded in real relations that hold between one’s present self and one’s future self. Given that there is no “deep, further fact” of personal identity, the real relations that can plausibly be thought to have rational significance are psychological continuity and connectedness. Psychological connectedness holds to a reduced degree over time, so it is plausible that its rational significance for prudential concern also diminishes over time.

But if we allow (with Nagel) that persons are identically their brains, or (with the later Parfit) that they are their cerebrums, then there is another real relation—that of being the same brain (or cerebrum)—that may plausibly be thought to ground the rational requirement for self-concern. And since the identity of the brain (or cerebrum) does not diminish over time, such a relation would support the Self-Interest Theory with its temporal neutrality condition intact. That would effectively neutralize the moral significance of most of the contemporary discussion of personal identity.

The claim that we are our cerebrums undermines Parfit’s view that “ordinary survival is about as bad as being destroyed and having a Replica”

The My Division argument in Reasons and Persons, which Parfit used to show that the relation of personal identity does not necessarily coincide with the relation that matters in personal survival, is compatible with his position in this paper. A cerebrum consists of two cerebral hemispheres. If Parfit’s hemispheres were separated, and both were transplanted into the living bodies of his two brothers, his cerebrum would thereby cease to exist, although most of what matters in Parfit’s survival might be preserved in the living hemispheres that would then become the “conscious, thinking parts” of the two post-operative persons. There are good grounds for saying that, in this case in which the hemispheres survive and continue to function, destruction of the cerebrum is not nearly as bad as dying.  Persons have, after all, survived the surgical removal of a single hemisphere.

Parfit’s claim that “ordinary survival is about as bad as being destroyed and having a Replica” rests on the plausibility of the view that what mainly matters in personal survival is psychological continuity and connectedness with any cause. ‘Any cause’ includes the cause envisaged in the teleportation example, in which I am scanned and destroyed at one location while my information is transmitted to another location where it is used to construct a living replica of me.

Much of what matters to most people about their survival consists in continuing to have experiences in the future. If I believed I was my cerebrum, and I expected to undergo the branching operation Parfit describes in My Division, I could anticipate having the experiences of both survivors of the operation, even knowing they would not be identical to each other. My grounds for anticipating having both streams of experience would lie in my knowledge that both my cerebral hemispheres would survive the operation, and that a great deal of parallel function and redundant memory storage makes my two hemispheres fairly similar. The loss of my cerebrum’s identity doesn’t matter greatly because its two parts will continue to function in much the same way as before the split. The extensive parallelism and relative independence of the hemispheres makes this possible. Further subdivision—of each of my hemispheres into two hemi-hemispheres—would amount to a death sentence.

No parallel argument supports my hopes of having the experiences of my replica on Mars. If I believe I am my cerebrum, then since I know my cerebrum will be atomized in the scanning terminal on Earth, I am bereft of one strong reason to anticipate having the experiences of my replica. But if I believe my identity coincides with non-branching Relation R, I can anticipate having the experiences of my replica on Mars, because I believe he is myself.

I can anticipate having those experiences even if I believe, with Parfit, that identity is not what matters in survival. Parfit used branching cases to argue that personal identity is not necessary for what matters in survival. My anticipation is based on the belief that identity is sufficient for what matters in survival. That belief is justified because to continue to exist, alive, is to survive, and survival itself must preserve what matters in survival.

Or must it? It’s an odd position for a Lockean like Parfit to take. There are actual cases in which the cerebrum survives, but the ‘Lockean person,’ identified by its psychological attributes such as memories and personality, arguably does not. Severe, rapid-onset dementia seems to fit this description. Perhaps Parfit would reply that all actual cases of dementia are gradual in onset, and thereby preserve sufficient day-to-day continuity to maintain what matters in survival. If so, and if I am diagnosed with galloping dementia, I can rationally fear experiencing the confusion and anxiety of the survivor. To counter this reply, it is easy to concoct a thought-experiment in which continuity is decidedly broken. A high dose of an amnesia-inducing drug, something like Midazolam but with permanent effect, could be used to ravage a person’s psychology while preserving the life and consciousness of his cerebrum. As a Lockean, Parfit would surely hold that such a devastating drug treatment is equivalent to  death—and therefore, the person facing it would have no reason to fear any pain that might be inflicted on the survivor. In identifying the person with the cerebrum that survives the drug treatment, Parfit makes it difficult to block that fear.

Parfit argued in Reasons and Persons that Relation R (psychological continuity and connectedness, with any cause) preserves what matters in survival, and that Relation R, when further constrained so as to exclude branching cases, coincides with personal identity. He also argued that personal identity is lost in branching cases. Branching examples are central to his argument for the claim that personal identity does not matter in personal survival, a claim he admits is counterintuitive. His argument that Relation R preserves what matters in survival supports the claim that personal identity coincides with the relation that matters in survival, further constrained so as to exclude branching cases.

In his new paper, Parfit endorses a different view of personal identity—that we are our “conscious, thinking parts,” our cerebrums. On this view, as on his earlier view, branching examples can effectively be invoked to support the claim that personal identity is not the relation that matters in personal survival. The argument is even stronger under Parfit’s new view, because the identity criteria for bodily organs are less controversial than the identity criteria for persons. If it were established that persons are identically their cerebrums, it would be uncontroversial that when a cerebrum ceases to exist because its hemispheres are separated, the person ceases to exist.

On Parfit’s new view, his argument that Relation R preserves what matters in survival is weakened. If personal identity coincides with the relation that matters in survival, further constrained so as to exclude branching cases, then non-branching teleportation, which destroys the cerebrum and thus fails to preserve identity, also fails to preserve what matters in survival. If Parfit wishes to maintain both that persons are identically their cerebrums, and that Relation R preserves what matters in survival, he must now say something like personal identity coincides with the relation that matters in survival, further constrained so as to exclude branching cases and also constrained to include only cases where the relation that matters in survival is supported by a continuously existing material substance. The kinds of material substances that support the relation that matters in survival (relation R) are in fact limited to whole brains, or whole cerebrums, or, at a minimum, single cerebral hemispheres. So the effect of this modified position is to distinguish personal identity from what matters in survival by not one, but two, additional constraints. That is perhaps a tenable position, but to me it seems strained. In fact, people justify their special concern for the future not in terms of psychological continuity and connectedness, but in terms of identity. (“I care about old-age security because my interests are on the line.”) That fact is hard to explain if the relationship between identity and what matters in survival is indeed so tangential.

If I am my cerebrum, the difference between my own life and the lives of others is greater than if I am anyone who stands in non-branching Relation R to me. If I believed I were my cerebrum, I would not say:

My death will break the more direct relations between my present experiences and future experiences, but it will not break various other relations. This is all there is to the fact that there will be no one living who will be me. [Parfit, 1984, p 281]

An Alternative View of What Persons Are

The case for animalism rests on a fundamental materialist assumption bolstered by a family of arguments against the proliferation of entities, of which the Two Lives Argument is an example.

The materialist assumption behind animalism is that whatever a person is, it must be some sort of material thing: either a living organism or part of a living organism. Its intuitive appeal may depend in part on the implausibility of Dualism. Any theory which claims the existence of mental or spiritual substances is rightly suspected of being unscientific.

In claiming that persons are their cerebrums, Parfit has aligned his views with this materialist assumption. He need not have done so, because there is a better theory of what persons are—bodies of information, or informational entities. Although informational entities are not material things, they have good standing in a scientific ontology. Informational entities constitute a broad class of replicable entities which includes both genes and memes. That genes exist is uncontroversial. That memes belong in a scientific ontology is sometimes disputed, but I maintain, with Dennett, that they do.[Dennett, 2006, pp 341-343]  Words (word types, not word tokens) are memes. Words and other elements of language are entities with multiple physical instances, that, as Darwin noted, evolve in accordance with laws of natural selection. A theory that excludes word types from its ontology will be hard-pressed to explain the relationship between word tokens inhering in media as different as sound-waves, ink on paper, computer memory chips, and the human brain, that enables a recipe for ginger snaps to be successfully transmitted from one such medium to the others.

The information which constitutes a person is the same information that must be transmitted from Earth to Mars for that person to be teleported successfully. Obviously, if a substantial portion of that information were lost in transmission, neither he, nor what matters in his survival, would survive teleportation.

A book—the intellectual work, not the bound volume—is another example of an informational entity. Parfit’s intellectual work, Reasons and Persons, came into existence when he began to write it in the early 1980’s, and will continue to exist until its last copy is destroyed. The bound volume on my desk, also entitled Reasons and Persons, was printed in 1984, and will continue to exist until it is destroyed—but no longer, regardless of the fate of other copies. Both are called ‘books,’ but they are two different entities, of very different kinds, and with different properties. Only one has weight. Only one is found in libraries around the world. On analogy with teleportation, I can email the intellectual work, but not the bound volume, to Australia (not yet to Mars).

The printed volume and the intellectual work have important attributes in common: for instance, they contain the same sequence of sentences. This commonality, between an informational entity and its physical instance, affords an answer to the Two Lives objection about persons, variants of which are central to the discussion in Parfit’s new paper.

A central motivation for accepting animalism stems from its promise to avoid a confusing, and intellectually distasteful, multiplication of entities.  Human beings are animals that think, have experiences, and are self-conscious.  Persons also think, have experiences, and are self-conscious.  Eric Olson argued that it is implausible that our bodies house two kinds of conscious entities: animals and persons.

I argued in an earlier piece entitled “What We Are Not” that the Two Lives Objection is answered by the theory of persons as informational entities.

In “We Are Not Human Beings,” Parfit answers this family of objections by developing the Embodied Person View. He argues that the first-person pronoun is ambiguous; it can refer either to ‘inner-me’,  the thinking part, or to ‘outer-me’, the animal. Inner-me is the ‘Lockean person’, a part of the animal, which is embodied in the animal. The person thinks ‘directly,’ the animal ‘indirectedly’ by virtue of having the person as a part. Parfit’s development of this view succeeds, I think, in sapping the force of the variants of the Two Lives objection considered in the paper. It offers a consistent way of talking about persons as both animals and as the thinking parts of their brains.

What it doesn’t offer is any further insight into the mysteries surrounding our self-interested concern about the future—what matters in survival. The divorce between personal identity and reasons for self-interested action, which Parfit introduced in Reasons and Persons, becomes more entrenched if a person is identical to his brain or cerebrum—a bodily organ which clearly does not survive transformations like information-based teleportation, which, according to Parfit (and me) preserve everything that matters in survival.

Parfit appears content with this divorce, although he acknowledges that others are not.

On the true view, I claimed, though we have reasons for special concern about our future, these reasons are not given, as we assume, by the fact that this will be our future. Nor will our death be as significant as most of us believe. In my somewhat misleading slogan, personal identity is not what matters.

In defending these claims, I appealed in part to the imagined case in which two future people would be psychologically continuous with me as I am now, because each person would have one half of my cerebrum. But this is only one example. And I have found it hard to convince some people that, in other cases, personal identity is not what matters. I cannot persuade these people, for example, that if they were about to be destroyed and replicated, it would not matter that their future Replica would not be them, so that they would never wake up again.

If I am right that the primary criterion that most people actually use for first-person judgements of personal identity is the presence or absence of an emotional response—an attitude  consisting of dispositions to have the distinctive emotions of self-concern directed towards the target of that response—it is not at all surprising that Parfit has had, and will continue to have, difficulty convincing his colleagues (let alone the general public) that personal identity is not what matters in survival.

The idea that persons are identical to informational entities is something like the idea that they are types. I have avoided calling them “types” because that word suggests implausible views that Parfit attacks in Reasons and Persons.[Parfit, 1984, pp 293-297] If persons were types, he said, they might be thought to be immortal, like abstract types such as the number five; but persons are not immortal. And if a person were a type in the way that The English Rose is a type, a person would not be a suitable object for love, since only individuals, not types, can be loved in the way we love persons.

Persons are not types in either of those senses. Is Shakespeare’s Hamlet a type? Perhaps, under some definition; but it is not immortal (since it only began to exist when Shakespeare wrote it, and it may someday cease to exist if all copies are destroyed), and it is not a type of play like Elizabethan Tragedy. I love Hamlet, not as a representative of the type of Elizabethan Tragedy (to which I am indifferent), but as an individual literary work; I love it for its unique language, for the characters of Hamlet, Ophelia, and Polonius, for its many jokes.

The idea that persons are informational entities has recently entered the philosophical literature. Mark Walker’s “Personal Identity and Uploading,”  argues that person-types should be distinguished from person-tokens, and he uses literary works by analogy to clarify what he means by person-types. Walker writes:

I will say a little about the type/token (TT) account of personal identity and then see whether the no-branching argument has any traction against it. In the case of literature, the tokens of Hamlet are individuated according to the physical implementation: my Hamlet is in a different spatial location from your Hamlet. The Hamlet type is an abstract entity, which particular tokens of Hamlet embody. Similarly, the TT solution to personal identity says that tokens of a person type are individuated in terms of physical implementation: each replica will have a different spatial location. The person type is the abstract entity, which the various tokens are all embodiments of. [Walker, 2011]

Walker concludes his paper with an argument that the survival of person-types is more important than the survival of person-tokens.  His usage of “type” is close enough to that of my “informational entity” that I would agree.

Unlike books, which become fairly static once completed, persons are dynamic bodies of information, constantly changing throughout their lives. If relations of continuity and connectedness hold, changes in informational content do not threaten the identity of an informational entity. The informational content of databases, or of websites like Wikipedia, is in flux throughout their existence. During the time when a book is being written, it too changes substantially from day to day; what makes it the same book are, once again, relations of continuity and strong connectedness between successive versions.


Damasio, Antonio (2010) Self Comes to Mind, Pantheon Books, New York

Dennett, Daniel (2006) Breaking the Spell, Penguin Books, London

Olson, Eric (2003) “An Argument for Animalism”, in Martin and Barressi, ed. (2003) Personal Identity, Blackwell, Oxford.

Parfit, Derek (1984) Reasons and Persons, Oxford University Press.

Parfit, Derek (2008) “Persons, Bodies and Human Beings”, in Dean Zimmerman, Theodore Sider & John Hawthorne (eds.) Contemporary debates in metaphysics, Blackwell, Oxford

Parfit, Derek (2012) “We Are Not Human Beings,” Philosophy, vol. 87, issue 1.

Sidgwick, H (1904) The Methods of Ethics, MacMillan, London

Walker, Mark (2011) “Personal Identity and Uploading,” Journal of Evolution and Technology, vol. 22, issue 1.


Return to the Phantom Self home page.

Tags: , , , , , ,

6 Responses to “Parfit’s Retreat: “We Are Not Human Beings””

  1. Karin S. Fester says:

    I found your article interesting but lacking some important details, one of which concerns a citation. You have excluded a significant source when referring to the cases of children born with hydranencephay who showed awareness, emotions, emotional bonding, love for music etc. It was NOT Antonio Damasio who first reported these unusual mental phenomena, but D.A. Shewmon’s research in 1989 (see the paper authored by Shewmon and G.L. Holmes in 1999). Perhaps you need to thoroughly check the origin of the sources cited in Damasio’s work. Karin S. Fester

  2. Thanks, Karin, for pointing that out. Damasio was my source, but you’re right, the 1999 Shewman, Holmes and Byrne paper, “Consciousness in Congenitally Decorticate Children” was Damasio’s source. I did not mean to suggest that Damasio did the original research.

  3. thelonious27 says:

    You claim that Parfit has retreated from the radical position of Reasons and Persons to a materialist view that persons are physical substances. But you appear to have confused identity with what matters to survival, a distinction Parfit has always been at pains to emphasize. We are identical to our brains, or enough of our brains to be conscious etc. But having the same brain/enough etc. is not necessary for survival in Parfit’s view, old and new. The idea that Information (psychological continuity/connectedness etc.) matters to survival and is what we care about appears to be completely consistent with Parfit’s case against animalism.

  4. admin says:

    I’m amazed, thelonious27, that you accuse me of confusing identity with what matters in survival. I tried to maintain Parfit’s distinction carefully throughout this piece. I admit to doubting that Parfit can succeed in showing that the concepts of personal identity and what matters in personal survival can be shown not to coincide. Here, again, is my view of the matter in a nutshell:

    “If I am right that the primary criterion that most people actually use for first-person judgements of personal identity is the presence or absence of an emotional response—an attitude consisting of dispositions to have the distinctive emotions of self-concern directed towards the target of that response—it is not at all surprising that Parfit has had, and will continue to have, difficulty convincing his colleagues (let alone the general public) that personal identity is not what matters in survival.”

  5. thelonious27 says:

    Your view about emotional response etc. is completely irrelevant to the INTERNAL criticism that you start out your article making, and that I took issue with. You say that Parfit has retreated from his position in Reasons and Persons from a psychological approach when he states in his newest work that we are identical with our cerebrums. These two positions are NOT inconsistent, and in fact do not present any change at all in his position about identity or what matters to survival — he holds, early and late, that personal identity consists in identity with our cerebrums (more or less), but that identity does not matter to survival — psychological continuity etc. is what matters. Again — once one recognizes, as you fail to in this particular criticism of Parfit, that identity is not what matters to survival, one can hold both that we are identical to our cerebrums and that psychological continuity with ANY cause is what matters to survival.

  6. Thelonius, I didn’t claim that Parfit’s new view is inconsistent with the central claims of Reasons and Persons. I took the much weaker position that Parfit’s earlier and later claims represent “an uneasy truce between fundamentally warring ideas.”

    If you revisit Appendix D of Reasons and Persons, you’ll find that Parfit there denied that persons are identical to their brains (and by extension, their cerebrums). I believe my opening quote represents Parfit’s view, of that time, on personal identity fairly. “Our identity over time just involves (a) Relation R—psychological connectedness and/or psychological continuity, either with the normal cause or with any cause, provided (b) that there is no different person who is R-related to us as we once were. [Parfit, 1984, p 216]” (NB: This is a definition of personal IDENTITY, not of what matters in survival.)

Leave a Reply