(HF1) Persons persist through time.

(HF2) Persons can survive (sometimes quite drastic) physical and psychological changes.

(HF3) "I know who I am (i.e., that I am the same person, that I have survived) without needing to check with my five senses."

In order to be adequate, a CPI must give a wholly general, informative, necessarily true "if and only if" statement specifying the changes that a person can (and the changes that a person cannot) continue to exist through (survive).

(SCPI) person A at time t = person B at time t+ if and only if B at t+ has the same soul as A at t.

Cartesian Dualism provides a clear characterization of soul.

The No Interaction Argument and the Population Explosion Argument independently raise serious questions about the acceptability of SCPI. Like the Problem of Other Minds, the Population Explosion Argument relies on the fact that souls in others' bodies are undetectable. This means that it is consistent with any evidence that you could have about an individual human body that a continuous stream of psychologically very similar souls flows through the body.

Population Explosion Argument against SCPI
(1) If SCPI is true, then for all anyone knows, each of us a different person from moment to moment
(2) We do know that not each of us is a different person from moment to moment
(3) Therefore, SCPI is not true

Psychological criteria promise to preserve the insights that underlie SCPI while avoiding its problems. These criteria are often stated in terms of periods of time in, or "stages" of a person's existence. For present purposes, almost any period of time - milliseconds to a year or more - can be thought of as a stage, as long as one chooses some one interval of time and sticks with it during the discussion of the CPI.

Using "stage" terminology, a CPI sorts sets of stages by saying what relationship links those stages together to make up exactly one person.

(PCPI1) person stage A at time t is a stage of the same person as person stage B at time t+ if and only if B at t+ has all of the same psychological characteristics as does A at t.

PCPI1 is too strong since it conflicts with HF2: people can and do change their psychological characteristics over time.

(PCPI2) person stage A at time t is a stage of the same person as person stage B at time t+ if and only if B at t+ has some of the same psychological characteristics as does A at t.

PCPI2 is too weak, since almost any two individuals are sure to share some psychological characteristics. For example, Carolina Smythe and Wyoming Jownes might both be happy at the same time, without Smythe's being the same person as Jownes.

Since PCPI1 is too strong, and PCPI2 is too weak, we need to find some reasonable way of steering between the two extremes they represent.

Memory Criteria are intended to steer between these two extremes.

(MCPI1) person stage A at time t is a stage of the same person as person stage B at time t+ if and only if B at t+ contains memories of experiences had by A at t.

Amnesia victims seems to constitute counterexamples to MCPI1, although there may be a good reply to such counterexamples.

More serious: MCPI1 does not distinguish between real and apparent memories. It does not allow us to distinguish between a person and a deluded pretender.

The most obvious way to make this key distinction is:

(MCPI2) person stage A at time t is a stage of the same person as person stage B at time t+ if and only if B at t+ contains real memories of experiences had by A at t.

where:

B at t+ contains real memories of experiences had by A at t =df. B at t+ contains apparent memories of experience had by A at t; and B at t+ is a stage of the same person as is A at t.

But this way of making the distinction makes the criterion circular and therefore uninformative.

One popular proposal about how to distinguish non-circularly between real and apparent memory is that real memories are caused in a particular way: so as to preserve the information contained in the initial experience. Whence:

(MCPI3) person stage A at time t is a stage of the same person as person stage B at time t+ if and only if there is an information-preserving causal chain leading from experiences had by A at t, to memories (of those experiences) had by B at t+.

Duplication cases, in which the chain forks, pose the most serious problem for MCPI3. Suppose that the Star Trek transporter malfunctions and instead of transporting you to the surface of planet Zorgon, it creates a precise, atom-for-atom duplicate there, without moving you anywhere. MCPI3 says that since there is an information-preserving causal chain leading from you now to both you then and your duplicate, both of them are you! Yet the duplicate on the planet could be devoured by the dreaded Sharp-Fanged Phlabingo without your coming to any harm (or even noticing).

MCPI4 (below) is intended as a response to the duplication problem, but it takes a radical approach. Since the duplicate should not be counted as you, and this causal chain approach must treat the information-indiscernible branches of an information-preserving causal chain on a par, MCPI4 says that neither continued branch of the process is you:

(MCPI4) person stage A at time t is a stage of the same person as person stage B at time t+ if and only if there is an information-preserving causal chain leading from experiences had by A at t, to memories (of those experiences) had by B at t+; and B at t+ has no competitors (duplicates).

MCPI4 does NOT say that competitors (duplicates) are impossible. Rather it says that continued existence with competition is impossible.

Suppose that while you are sleeping, the Klingons secretly make an exact duplicate of you via remote scanning from their star cruiser. With the aid of their advanced technology, they have created a remote branch in your information-preserving causal chain. MCPI4 counts as you neither your continuation in your own body nor the duplicate in the new body on the Klingon star cruiser.

So, in such a case, MCPI4 has the wildly counterintuitive consequence that you could cease to exist while seeming to yourself to survive, without interruption, in your own body. This conflicts with HF3.

The inability of such memory criteria to count properly suggests that the body one actually has is after all an essential component of personal identity since it's your body and the duplicate's body which are distinct.

There are even some examples that suggest the continued existence of your body is all that matters. Suppose that you are about to undergo some dangerous surgery and for the duration of the procedure, as a precaution, all the information in your brain is off-loaded into a static memory bank, leaving your own brain temporarily blank. Lucky for you, the surgery goes well, and the information is restored to your brain afterwards. Who underwent the surgery? If it was you, then your body must be the identity-defining component since it, and not the information-preserving causal chain, was the only thing that persisted from before to after the surgery.

Even if you find the latter example unpersuasive, the problems for memory criteria are so severe that they alone make body criteria worth considering:

(BCPI) person A at time t = person B at time t+ if and only if B at t+ has the same body as A at t.

(RCPI) person A at time t = person B at time t+ if and only if B at t+ has the same brain as A at t.

The Problem of the Ship of Theseus (see below) raises serious questions about the coherence of the concept same physical object, and so raises serious questions about the coherence of the concepts same body and same brain

The problem arises because there seem to be two essential components to the concept same physical object:

(i) Complex physical objects can persist through changes (losses or replacements) in their parts:

Every time I wash my most valuable pair of white knee socks, a few sock molecules go down the drain. But they are still fairly new and largely intact. So I still have that pair of socks, thank goodness.

My car has gone through many repairs - new tires, brake jobs, air-filter replacements, a new starter motor, new spark-plugs, etc. - because I want my car to remain in good working order. I'd have been lying if I 'd said after any or all of the repairs, "I've got a new car!" It's the same car; I've just kept it in good repair.

(ii) A complex physical object is nothing more than its parts, arranged in a certain way.

Suppose, for example, that I am inordinately fond of the particular desk in my office - we've been through a lot together - and INSIST on keeping it when I change offices. Fearing for my sanity, my department head agrees. The desk is, however, too large to fit through the doors. So, it is disassembled, the smaller pieces moved to the new office and it is soon reassembled there, exactly as before. I have the same marvelous desk as I did in my old office. I am happy. My department head is relieved.

The example of the Ship of Theseus shows that these two components are in conflict with one another.

Theseus was a Greek ship owner who lived in ancient times, when ships were still made of wooden planks. Theseus began his career with one ship and was more fastidious than most ship owners. He kept careful track of each plank in his ship and would replace a plank with a precisely similar one if the original became "too filthy for words." His ship was made from ten thousand planks, and I'd like to show you a drawing of it, but I'm too lazy to draw all the planks, so I'll draw it as if it had only four planks. Here is Theseus's ship (side view):

When plank 1 got grungy, Theseus removed it and replaced it with plank A, yielding:

When plank 2 got grungy, Theseus removed it and replaced it with plank B, yielding:

When plank 3 got grungy, Theseus removed it and replaced it with plank C, yielding:

When plank 4 got grungy, Theseus removed it and replaced it with plank D, yielding:

These various replacements took twenty years, during which time Theseus was famous for having the cleanest one-ship fleet.

But he was not one to waste materials. He kept planks 1, 2, 3 and 4 in his garage and after cleaning them up very carefully, kept them from any further filth or damage. At the end of the twenty year period, Theseus used planks 1, 2, 3 and 4 to double the size of his fleet:

Now, which one of ABCD and 1234 is the original ship? (Remember that the two ships really have ten thousand planks each, and that the replacement process took twenty years.) Component (i) of the concept same physical object says ABCD is the original ship. But component (ii) says 1234 as reassembled is the original ship. It can't be both, since one thing can't be the very same thing as two different things (duh). On the other hand (or oar), it's got to be one of them, since Theseus was never shipless.

Of course, the assumption that planks were replaced is not essential to the point of the example. It could just as well have been stated as an example involving ship molecules or quarks, etc.

If you think that problem is readily solved, consider this slightly modified version. This time, Theseus begins with two ships:

Then he begins swapping planks, plank 1 for plank A, plank 2 for plank B, and so on. At no time does he have more than two ships. At no time does he have fewer than two ships. Throughout the process of swapping, the two ships reamin the same in look and function, yielding this final picture:

Neither 1234 nor ABCD ceased to exist at any point during the process. Or did they? And if so, at what point exactly, and why at that point and no other?

Perhaps the concept same physical object has less to it than is usually thought. Since brains and bodies are also complex physical objects that may undergo part replacement, naturally and artificially, their identities are also called into question in the same way.

Here's another example that suggests body criteria are wrong.

Suppose that your body is riddled with cancer and will very soon cease to function. There's no way to get rid of the cancer cells. But new, utterly reliable technology does allow all of the information stored in your nervous system to be transferred to a new body, at least as good as the old, healthy one you once had. (For a bit extra, you can opt for some improvements, but health insurance won't cover the additional cost. Sorry.)

You may stay in your old, rotting "home" and die (in about four hours), or you can move to a nice, new place and survive. If your response to this choice is something like, "GET ME OUT OF HERE! I WANT TO LIVE!," then your intuitions about personal identity are clearly inconsistent with the body criteria. And is there any other rational choice but yours? (This example does of course trade on the insights that motivate psychological criteria of personal identity. But those insights are worth gripping tightly.)

So, even if the Problem of the Ship of Theseus were solved, this Cancer Patient example would also cast independent doubt on the body criteria.

(BPCPI) person stage A at time t is a stage of the same person as person stage B at time t+ if and only if there is an information-preserving causal chain leading from experiences had by A at t, to memories (of those experiences) had by B at t+; and B at t+ has the same body (brain) as does A at t.

BPCPI inherits the problems of BCPI and RCPI, since it says that having the same body (brain) is necessary (though not sufficient) for continued existence. Not a good deal.

Issues concerning personal identity should be at the heart of debates about abortion and euthanasia. We will focus on the abortion debate, but similar remarks apply to euthanasia.

Anti-Abortion Argument #1
(1) This fetus is the same person as that baby
(2) It is always morally wrong to kill a person
(3) To abort that fetus is to kill it
(4) Therefore, it is morally wrong to abort this fetus

(1) is seldom even discussed in most public debates about abortion. At best, it is said that the fetus and baby have the same body (-and we have seen that same body is not the most secure concept). But why should that be taken as establishing (1)?

(2) is clearly false, though often asserted. Self-defense or defense of an innocent other constitute counterexamples. Even the more plausible:

(2*) It is always morally wrong to kill an innocent person

is refuted by the following trolley-car counterexample, and so (2*) is false, too. Finding a correct moral principle to put in its place is difficult, and is one reason that the abortion issue is difficult.

The Trolley-Car Counterexample to (2*): A fully-loaded San Francisco trolley-car, which is confined to the tracks, is on a very long, very steep hill when its brakes (and the back-ups) fail. There's no feasible way to stop it. It reaches over 50 mph at the bottom of the hill, where the tracks branch. To the left track, a madman has tied one innocent infant. To the right track, the madman has tied ten innocent infants. No one is able to free the infants in time.The trolley-car driver can either leave the steering bar in its current position, causing the car to go down the right track; or she can put it in its other position, causing the car to do down the left track. She knows all this. She knows also that there is nothing she can do to make the car leave the tracks. Whatever she does, at least one innocent infant will be killed, and she knows that as well.

What should the driver do? The best (least bad) of the only two awful alternatives she has is to push the steering bar into its other position, reducing the number of inevitable deaths from ten to one.

(You might find it useful to consider alternative examples in which one infant is on each track, or in which an infant is on one and its mother is on another.)

(3) is merely stipulative: again, "kill" means: terminate the existence of. In this sense of "kill", it is an open question whether killing is morally significant.

A Pro-Choice Argument
(1) This fetus is not the same person as that baby
(2) If (1) if true, then this fetus is not a person
(3) If this fetus is not a person, then it may be morally permissible to kill it
(4) To abort a fetus is to kill it
(5) Therefore, it may be morally permissible to kill this fetus

(1) is based on the observation that this fetus and that baby are psychologically discontinuous: before about month five and a half, a fetus is missing large portions of its central nervous system, and so is biologically incapable of having a human psychology. But why should lack of psychological continuity be taken to establish lack of personal identity?

(2) is based on the claim that: if this fetus is not that baby, then there is no other person for it to be, and so it is not a person at all. (For a reply to this premise, see Anti-Abortion Argument #2, below.)

(3) expresses the idea that if an organism is not a person, then it may be killed if some greater good (e.g., contributing enough to the well-being of some person) can be achieved by doing so.

(4) is true because we again stipulate that: "kill" means: terminate the existence of. In this sense of "kill", it is an open question whether killing is morally significant.

Notice that this argument does not support the conclusion that it is morally permissible to terminate a woman's pregnancy under all circumstances (e.g., at eight and a half months when there are no birth defects and no threats to the mental or physical health of the woman).

Anti-Abortion Argument #2
(1) This fetus is a person
(2) It is always morally wrong to kill a(n innocent) person
(3) To abort this fetus is to kill it
(4) Therefore, it is morally wrong to abort this fetus

(2) is false (see above). (3) is true by stipulation (see above).

(1) does not depend on any claims about same person. It is based on the more basic concept person. (CPI presuppose an understanding of person, and so do not try to provide an analysis it. CPI are in effect counting principles for persons, and of course one cannot count things of a kind unless one can identify things as being of that kind.) But what exactly is it that makes something a person? (Is it having a soul? We have seen that use of that notion in this context leads to serious problems. Is it having a human body? But why should that be essential?) It is no less difficult to answer this question than it is to analyze same person. One might try to avoid this question by defending (1) as follows:

(a) This fetus can become a person
(b) Whatever can become a person is already a person
(c) Therefore, this fetus is a(lready a) person [=(1)]

But why should (b) be accepted? Is (b) based on the general claim, "For any X, whatever can become an X, is already an X"? If so, then it is false, since for example, a pile of bricks and mortar can become a house, but is not one already. If not, then what reason is there for accepting (b)? What is 'special' about persons that would make (b) true?

Perhaps we can conclude that the truth about abortion (and euthanasia) is less obvious than some of the participants in the debates seem to believe.