Our proposal #4 made use of the notion of “genuinely remembering” something — as opposed to merely seeming to remember it, or thinking that you remember it. And when discussing Reid, we wanted to see if Locke could make sense of a difference between evidence that you’re identical to someone, and the facts about whether you’re identical. We said we could make sense of such a difference, if people could sometimes seem to remember things, that they did not in fact genuinely remember.
But now a problem arises. When we try to make sense of genuinely remembering, it seems like we’ll end up saying that for you to genuinely remember doing X, X had to in fact have happened, and moreover that you have to be the person who did X. If you never pulled the cat’s tail, then you can’t remember pulling the cat’s tail. You can only seem to remember pulling it. It doesn’t help if someone else pulled the cat’s tail. You can only genuinely remember doing this if you are identical to the person who pulled the cat’s tail.
That’s the source of our problem. It looks like the notion of genuinely remembering an event presupposes the notion of personal identity. But then it shouldn’t be OK for us or Locke to define personal identity in terms of genuine memory (what Locke calls “consciousness of past states”). Locke’s critic Joseph Butler put the complaint like this:
[O]ne should really think it self-evident that consciousness of personal identity presupposes, and therefore cannot constitute, personal identity, any more than knowledge, in any other case, can constitute truth, which it presupposes.
That is, I can remember only my own experience; but this is because “being of my own experiences” is part of what memory is: it’s not my memory of the experience that makes it mine; I can remember it only because it’s already mine. Memory can only be how facts about my identity get known or revealed. Not what those facts consist in.
Another way to put it: arguably, we can’t tell whether your apparent memories of pulling the cat’s tail are genuine memories, unless we first know whether or not you are identical to the person who pulled the cat’s tail.
If this is right, then Locke’s theory and our Proposal #4 seem to be circular in a problematic way. They purport to tell us what personal identity consists in, but they do so by using notions like memory that already presuppose the notion of personal identity (see Perry Dialogue, pp. 27-30).
Butler gave the most famous and memorable presentation of this objection, but he wasn’t the first to raise it. That seems to have been John Sergeant, who was much less influential than Butler. Reid’s famous objection with the Brave Officer was also given earlier by another philosopher, Bishop Berkeley. In this case, Berkeley was a well-known and influential philosopher too. I have no idea whether Butler and/or Reid were aware of the earlier discussions; they may have thought up their objections on their own, even though they weren’t the first to publish them. At any rate, they are the ones whose presentations of their objections became best-known. This sort of thing happens all the time in philosophy.
How might we get around this circularity objection?
Let’s try to define a notion which is like memory, but which doesn’t presuppose the notion of personal identity.
Suppose I try to remember what my grandfather looked like. A certain memory image comes to my mind. Now this memory image might be distorted: my grandfather may never have looked exactly like the face in my memory image. What if there’s some guy, Joe Schmoe, whom I never met, who does look exactly like the face in my memory image? Would my memory image then be a memory of Joe Schmoe? Of course not. I never met Joe Schmoe, so I can’t be remembering him. It doesn’t matter that I have an apparent memory of someone looking a certain way, and Joe Schmoe in fact did look that way. This is not enough for me to be remembering Joe Schmoe.
Similarly, suppose I once stepped on the Queen’s toe, but I didn’t notice that I had done this; and no one ever told me that I did it. Some years later, a hypnotist implants an apparent memory in my mind of stepping on the Queen’s toe. Here I seem to remember stepping on the Queen’s toe, and I in fact did step on the Queen’s toe. But would we say that I genuinely remember stepping on the Queen’s toe? That doesn’t sound right. This is a case where I have an apparent memory of stepping on the Queen’s toe, and I in fact did step on the Queen’s toe. But that’s not enough for me to be remembering stepping on the Queen’s toe.
What’s missing in these cases? What more is required, for an apparent memory to be a genuine memory?
One suggestion is this: the apparent memory not only has to match something that actually happened, it furthermore has to be caused in the right way. If the experience of doing X leaves some physical traces in your brain, and as a result of those traces, you later seem to remember doing X, that would count as a genuine case of memory. If, however, your apparent memory matches the face of Joe Schmoe, whom you’ve had no causal contact with, then your apparent memory doesn’t count as a genuine case of memory. If a hypnotist causes you to seem to remember doing X, that would not count as a genuine case of memory, either. In the first case, your apparent memory is causally connected in the right way to an original experience that it’s a memory of. In the Joe Schmoe and the hypnotist cases, it is not.
Now, consider the case where you pull the cat’s tail, and the experience of doing this leaves some physical traces in your brain. Suppose it’s possible to carefully extract those physical traces from your brain, and transplant them into Maria’s brain. Maria would then seem to remember pulling the cat’s tail. However, Maria never did pull the cat’s tail. You are the one who pulled the cat’s tail. So we wouldn’t say that Maria remembers pulling the cat’s tail. She can’t remember pulling the cat’s tail because she never did pull the cat’s tail. But Maria is having an experience very much like memory. We can call it quasi-memory.
Maria quasi-remembers an event E iff: (i) she seems to remember E, (ii) E did in fact take place, and (iii) her apparent memory of E is caused in the right way by traces left by some person’s witnessing or performing E (however, that person need not be Maria herself).
Notice that quasi-remembering does not just mean “seeming to remember.” To quasi-remember an event, conditions (ii) and (iii) also have to be fulfilled. Notice also that every genuine memory is also a quasi-memory. One’s memories are simply quasi-memories of one’s own past life. Maria in addition has quasi-memories of (parts of) someone else’s life.
What is important for our purposes is that it’s possible to quasi-remember an event without being identical to the person who witnessed or engaged in that event. So the notion of quasi-memory does not presuppose the notion of personal identity. Neither does it exclude it. Some of what Maria quasi-remembers are things she did herself. She is identical to the original person.
Perhaps enough links of quasi-memory among a chain of person stages can be what unite those stages into a single person. That will be the idea behind Proposal #5.
Proposal #5 modifies Proposal #4 by replacing the notion of memory with the notion of quasi-memory:
Proposal #5: Stage A* and stage B* are parts of the same person iff they are parts of a chain, all of whose later stages are able to quasi-remember “from the inside” enough details of enough events experienced by the immediately preceding stages in the chain.
There is no circularity problem here. Quasi-remembering some event does not require Maria to be identical to the person who witnessed or engaged in that event. That’s why we have to add the qualifier about “enough”: you can have a few quasi-memories from someone else’s like without becoming identical to them. Where is the line between “a few” and “enough”? Hard to say — but for the time being let’s not fuss about that.
As before, we can state Proposal #5 more succinctly using a notion of psychological continuity. But now we say that a chain of person stages are psychologically continuous iff all of the later stages in the chain are able to quasi-remember “from the inside” (enough details of enough) events experienced by the immediately preceding stages. Then Proposal #5 can be expressed as:
Proposal #5: Stage A* and stage B* are parts of the same person iff they are parts of a chain of psychologically continuous person stages.
As before, we can also add more conditions for psychological continuity too. We might require that the stages undergo only gradual changes in their personality, opinions, interests, and values; and that their earlier intentions influence their later actions and choices.
To properly assess Proposal #5, we need to introduce some new ideas. That’s what we’ll do next.