In this earlier post, I tried to capture a broadly Lockean thought about the role of memory in personal identity in this principle:
(ID2) x is the same person as y iff x and y are persons [perhaps better: person stages]; and x can extend their consciousness back by having a memory of the form ‘I performed action or thought token a’, and y did in fact perform action or thought token a.
In the previous discussion, I began to ward off a threat of triviality, which I diagnosed as due mostly to the pernicious understanding of the phrase ‘criterion of identity’ as flagging a method of discovery of identity. (ID2) is pretty useless if one is trying to figure out whether x really had a memory when thinking ‘I performed action or thought a‘, since the evaluation of the personal indexical in that sentence is precisely what the issue of identity turns on.
This has been thought to make (ID2) trivial:
it seems to belong to the very idea of remembering that you can remember only your own experiences. To remember paying a fine (or the experience of paying) is to remember yourself paying. That makes it trivial and uninformative to say that you are the person whose experiences you can remember—that is, that memory continuity is sufficient for personal identity. It is uninformative because you cannot know whether someone genuinely remembers a past experience without already knowing whether he is the one who had it. (Olson, ‘Personal Identity’, §4)
It is true that one can’t become ‘informed’ as to whether x is y by applying (ID2). But it is not clear that makes it ‘trivial’, unless of course that is just an objection to the claim that (ID2) is a necessary truth connecting personal identity and memory. (Every correct philosophical theory will be trivial in that sense.) But (ID2) is far from trivial. We can see this by looking at Olson’s claim that we can’t know whether x remembers something without knowing whether x is the one who experienced it. This claim is false, since we do know that kind of thing all the time. I can come to know about some past event on the basis of eyewitness testimony, which seems to involve knowing that the eyewitness remembers that past event. If I’m concerned that it may be a mere pseudo-memory rather than a genuine one, I may rule out the common sources of such pseudo-memories—is the eyewitness deluded or otherwise mentally ill? Are they prone to exaggeration, expanding their own role in the event? Are they over-eager to be helpful and thus telling me what they think I want to hear? Is the causal connection between the event of their having this apparent memory and the original event deviant (mediated by some deceitful demon), or ordinary? If none of these circumstances apply, we have good reason to think this is a genuine memory, and we can know on the basis of that x, who has the memory, and y, who was present at the event, are the same person. That is, we can use all sorts of facts which are reliably correlated with the presence of genuine memories as evidence that x is the same person as y. That means (ID2) can be used in a forensic context too, since the focus on memory does give us useful additional constraints on which kinds of facts we should be looking at to decide whether x is y.
If all we could say about memory was this:
(TMC) x remembers e only if x experienced e,
then it would indeed be trivial to connect memory and personal identity, since memory would suffice for personal identity with the e-experiencer, and that is the only thing we could say about it. But the point of the memory condition on personal identity isn’t to affirm this Trivial Memory Condition, but rather to note that all sorts of other facts bear on whether x genuinely remembers something that they seem to remember—including, crucially, facts about the causal dependence of the candidate memory on the event which is its content. Since these more interesting connections between memory and causation exist, and are strongly correlated with a cluster of other connections collectively termed ‘psychological continuity’, the memory condition can be a useful sufficient condition.
Moreover, there is a direct argument against the triviality of the memory condition. Compare it to animalism, the view that we are identical to the thinking human animal currently coincident with us. If the human animal can survive the loss of its cerebrum—as it apparently can, given the existence of human beings in a vegetative state with no discernible higher brain function—there can be a possible case where the cerebrum is transplanted to another body. In that case, the apparent memories had by the other body can have the right causal connections to the prior events to ensure that the apparent memory is in fact genuine. So this is a genuine memory; and hence even the trivial condition on memory entails that the new body person is the one who experienced the event, even though the old body was present at the event. Since (ID2) is thus inconsistent with animalism, it can hardly be trivial.
The main worry that people have had about (ID2) and the like is not triviality, but transitivity. Since this post is getting a bit long, I will deal with transitivity in another post.