Personal Identity and Memory, part I: some reflections on Locke

(This is the first part of a two three-part post on personal identity and memory. This part concentrates on Locke; the next part will look at the role of memory as a sufficient condition for personal identity; the final part will look at the infamous transitivity worries about memory-based theories of personal identity.)

Modern discussion of diachronic identity originates with Locke, who introduces many of the standard moves in the literature—for example:

  • connecting identity with same-F-as, for any sortal predicate F;
  • distinguishing the question of what it is to be an F at a time from the question of when an F picked out at one time is indeed the same F as another F picked out at another time;
  • and noting that general bans on coincident entities must themselves be relativised to sortals: ‘whatever exists any where at any time, excludes all of the same kind, and is there it self alone. (Essay, 2.27.1—all future unattributed references are to the Essay).

This last point is important in some controversial issues. Lowe (1995: 106) for example argues that identity is ‘absolute’, and thus one object cannot be an instance of two distinct sortals with two different criteria of identity, so that—contrary to appearances—Locke is not actually allowing that there might be a man who is also a person. But if Lowe is right, then Locke needn’t have included the italicised qualification in the quote above; the fact that he does is some evidence that the natural treatment of the text is correct.

Aside: Lowe’s own view seems to be that ‘criteria of identity’ are exactly that: the world is full of things, each falling under one specific sort, and what identity consists in varies from thing to thing, being one relation in the case of persons, and another in the case of animals, etc. (I suppose the natural way to take this is as endorsing a kind of contextualism about `is identical with’, with context fixing the appropriate identity relation.) The alternative view, one which I suppose I favour, is that an individual can be fall under many sortals, and stand in the same-F-as relation to some things, and the same-G-as relation to other things. The identity relation proper—that relation expressed by the ‘=’ of logic—is a univocal relation, one especially closely connected with counting, and perhaps to be identified with the same-thing-as relation, the sortal identity relation for the most general sortal ‘thing’. Since some things are temporally extended, we can have genuine identity between an F at one time and an F at another time (both being parts of the same thing), even while the F at the first time is also a G, and identical to a G at another time which is not an F (in virtue, again, of those two Gs being part of the same thing). The fact that an F and a G can overlap in this way doesn’t seem to provide much reason for thinking that the F and the G must be identical. Nor is this account ‘incoherent’ as Lowe suggests.

Enough of that; Locke’s own text is pretty equivocal on the relationship between identity in the strict and logical sense, and same-F-as for sortal F. It’s Locke’s account of identity for the sortal ‘person’ that I’m primarily interested in here. He begins by defining the synchronic conditions for an entity to be a person—for that entity to be ‘a thinking intelligent being, that has reason and reflection, and can consider it self as it self, the same thinking thing in different times and places’ (where that last clause does not presuppose diachronic identity, but rather means ‘the person right now can consider itself at different times and places’, thus having memories about the past, expectations about the future, and the capacity for plans and goals that is properly characteristic of persons and in part is responsible for giving persons the moral status they have).

Having stated what it takes for something to be a person at a time, he goes on:

[S]ince consciousness always accompanies thinking, and ’tis that, that makes every one to be, what he calls self; and thereby distinguishes himself from all other thinking things, in this alone consists personal Identity, i.e. the sameness of a rational being: And as far as this consciousness can be extended backwards to any past Action or Thought, so far reaches the Identity of that Person; it is the same self now as it was then; and ’tis by the same self with this present one that now reflects on it, that that Action was done. (2.27.9)

Or again, a little later on:

For as far as any intelligent Being can repeat the Idea of any past Action with the same consciousness it had of it at first, and with the same consciousness it has of any present Action; so far it is the same personal self. For it is by the consciousness it has of its present Thoughts and Actions, that it is it self to it self now, and so will be the same self as far as the same consciousness can extend to Actions past or to come… (2.27.10)

So for Locke, unity of consciousness in the referents is what permits two terms for persons used at different times to refer to the same person. If consciousness is taken as Locke intends—as ‘inseparable from thinking, and…essential to it’ (2.27.9)—we might well attribute to Locke some kind of psychological view of personal identity. That may extend to character traits as well as the more conscious cognitive processes of thinking and reflection. But it’s pretty clear that it’s not at all implausible to view Locke as the forerunner of this way of thinking about personal identity, and as offering an account of a type that is generally plausible, notwithstanding the existence of viable rivals. Note also that Locke, immediately after introducing the term ‘personal identity’ in the first passage, glosses it as ‘sameness of a rational Being’, which looks very much like he’s adopting the sensible kind of attitude to understanding contentious claims of identity simpliciter as sameness claims for a given sortal.

People objected pretty quickly to this proposal. And fair enough too: Locke is pretty cryptic, and somewhat misleading. In the first quote, for example, he certainly speaks as if consciousness is distinct from thinking, and that it is a kind of second-order attitude to one’s own thinking for one to be conscious of it. Locke thinks this awareness of one’s own thought is inseparable from the thought itself—’It being impossible for any one to perceive, without perceiving, that he does perceive’ (2.27.9). And we are nowadays, many of us, pretty suspicious of the requirement of luminosity of our own cognitive states that this view of consciousness apparently involves (Williamson 2000). Locke apparently wants this because he wants memory to be just a special case of conscious awareness—it is conscious awareness of a past thought or action. If this is right, for Locke synchronic and diachronic identity are fundamentally the same. Many (most?) commentators assume that the relevant state for Locke is that of a present person being occurrently conscious of their present state, remembering that P; I’m suggesting instead that Locke thinks a present person remembers that P just when they are occurrently conscious of the past state, experiencing P (of course they must be conscious of this past state as past, but requiring that kind of internalist introspective access is not foreign to Locke).

Be that as it may, we probably don’t want to go in for this kind of conception of conscious awareness as crucial, and—tellingly—most of Locke’s own discussion focusses not on the role of self-consciousness, but on the role of memory in grounding identity. So suppose we offer, as an updated version of Locke, the following:

(ID) x is the same person as y iff x and y are persons; and x‘s memory ‘can be extended back’ to y‘s actions or thoughts, as being x‘s own.

The last clause attempts to exclude remembering what someone else did as counting as remembering being them; we might wish here to invoke remembering ‘from the first-personal perspective’. This can sound very mysterious, and objectionable. Lowe does indeed object at this point:

[W]e may have some doubts as to whether [this definition] can serve to explain, non-circularly, what personal identity ‘consists in’, since the very description of first-person memories seems to rely upon an antecedently given notion of personal identity. My first-person memories are those in which I recall some past episode in the life of a person as being one involving myself, as opposed to any other person, and this seems to presuppose that I already need a grasp of what constitutes the difference between myself and another person in order to enjoy, and recognise, distinctly first-person memories. (Lowe, 1995: 109–10)

But this objection seems confused. Firstly, Locke only says that for x to be y, x‘s consciousness needs to be able to be extended back to y‘s consciousness, not that—in addition—x needs to recognise this extension of consciousness as an extension of consciousness. That seems to involve conscious awareness that one is consciously aware of remembering that P, and nowhere does Locke explicitly require this (it may fall out a general internalism, but we externalists can keep the good in Locke without buying into those doctrines). Secondly, the claim seems to be that for x to remember one of y‘s experiences first-personally is for x to have a memory with the following content: y performed action or thought a; and x is the same person as y. That would indeed be a thought with the problematic circular content Lowe suggests. But why think that a first-personal memory has to have that content? Ever since Perry (1979), we’ve known that indexical content isn’t straightforwardly equivalent to purely third-personal content like the boldfaced content above. So if we’re serious that identity-grounding memories are first-personal states, they better have the proper kind of first-personal indexical content. That gives us something like this:

(ID2) x is the same person as y iff x and y are persons; and x can extend their consciousness back by having a memory of the form ‘I performed action or thought a‘, and y did in fact perform action or thought a.

Some kind of triviality worry remains in the air. Since ‘remembers that P‘ is a factive operator, if x really does have a memory of performing a, then it must be that the proposition expressed by x‘s token utterance or thought of ‘I performed a‘ is true. But then ‘I’ must (in fact) refer to the performer of a, who is y. But the suspicion about this is that all the work is being done by the indexical ‘I’, and hardly anything by the memory criterion of identity.

This is to mistake the nature of what is being offered. We have necessary and sufficient conditions for a person picked out at one time, x, and by one description, the first-personal rememberer of a past action a, to be the same as a person y picked out at another time and by another description, the performer of a. It turns out, because of the factivity of memory, that it is a necessary condition on x having such a first-personal memory that x did indeed perform action a, and that x and y are therefore able to be correctly picked out by the same definite description and hence are identical. Moreover, it seems sufficient for identity that x be in a position to have such a memory—for it to be possible that x‘s consciousness ‘can be extended back’ to y‘s actions or thoughts in the appropriate way.

(ID2) may seem trivial because of its epistemic uselessness. One can’t use (ID2) to figure out whether x is y, because unless you already know the temporally extended individual that x‘s use of ‘I’ picks out, you won’t be able to tell whether x‘s thought is a genuine memory or a pseudomemory (i.e., like a memory in internal phenomenology but false), and if you can’t tell whether x remembers or not, you can’t make use of (ID2). This is one of the highly misleading things about the phrase ‘criterion of identity’, suggesting as it does that the proper criterion of identity might give us something like a checklist, using which we can establish to our satisfaction whether x is y. Really, of course, (ID2) purports  to give us a relation that is necessarily coextensive with identity, and whether the description of that relation also gives a recipe for establishing when it holds is another issue.

This fact doesn’t make (ID2) trivial. For note that we could define another relation:

(ID3) x is the same person as y iff x and y are persons; and x can extend their consciousness back by having a memory of the form ‘I saw someone perform action or thought a‘, and y did in fact perform action or thought a.

(ID3) is surely much worse than (ID2), but note that it too has the feature of involving the personal indexical. So it must be not just the involvement of the personal indexical in (ID2), but rather the way in which it is used—namely, that Locke requires there to be a first-personal memory of performing some action or having some thought, not merely third personal awareness by me of some action or thought. While of course I am identical with whatever I was in the past, it is not so trivial to recognise that I am identical with a thing whose past actions I’m first-personally conscious of, as opposed to the many other things I have other kinds of conscious awareness and memory of.


  • Lowe, E. J. (1995), Locke on Human Understanding, Routledge: London.
  • Perry, John (1979). ‘The Problem of the Essential Indexical’, Noûs 13: 3–21.
  • Williamson, T. (2000), Knowledge and its Limits, OUP: Oxford.

One thought on “Personal Identity and Memory, part I: some reflections on Locke

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s