I have kept a diary in one form or another since about the age of twelve. Sometimes a few words, sometimes pages and pages, but either way every day of my life is documented – apart, that is, from the ages of fifteen and sixteen. As with all teenagers, these were my angst years, and in a grand gesture of would-be liberation, symbolic of new beginnings, I burned them. The intended significance of the event was somewhat eclipsed when I set off my parents’ fire alarm and I realized that burning paper in my bedroom was exceedingly stupid. But anyway, those years are now lost, except for what I can recollect or others can tell me (which is pretty much the case for most people other than those who keep a daily diary).

Recently I spent some time looking through my old diaries. Sometimes forgotten memories were brought to the surface, but at other times I was reading about things I believe actually happened, because I wrote them, but of which I nevertheless had no recollection. At times, these recorded events reshaped my current self-perception: on some occasions they made me consider myself to be a better person than I currently thought, and at other times a worse one.

What constitutes our personal identity over time has long been the subject of debate, but what I am concerned with here is how much influence we can have over our own identity and self-perception. If memory is integral to identity, and memory can be manipulated, then it follows that we can manipulate the ‘self’.

The ‘memory criterion’ of identity is usually attributed to English philosopher John Locke (1632–1704) and can be found in the chapter ‘Identity and Diversity’ in Essay concerning Human Understanding. This interpretation of Locke is not without its critics, but nevertheless it is the most popular interpretation, and the one that will be adopted here. Locke distinguishes a ‘person’ from a ‘man’. The ‘man’ means the organism, an animal like any other, whose identity over time consists in continuity of biological life. This means that although parts can be gained and lost (we grow and shed skin cells, for example) there must be continuity within this change for us to be talking about the same man.

Concerning the identity of the person themselves – the thinking being – perhaps surprisingly for his time and culture, Locke claims that personal identity is not tied up with the soul. This is because he thinks that the same soul could in fact play host to different consciousnesses. It is your consciousness which makes you the same person over time; specifically it is the continuity of your memories.

The continuation of personal identity through memory is crucial for justice. For instance, in order to properly see the consequences of our actions and maintain our full responsibility for them, we must be able to contemplate our future selves as connected to the person about to carry out an action now and we also must remember an action for it to qualify as really being ‘me’ who did it. (However in practical terms, since we cannot check the accuracy of someone's memory, the courts often end up punishing a man who was demonstrably physically connected to a crime, regardless of his memory claims.)

The implication of Locke's memory criterion for identity seems to be that my identity changes over time as my memories fade, or perhaps reappear after a period of absence. One famous objection to Locke's view along these lines was from Thomas Reid (1710–96). I'll give an adapted version. Suppose that as a ten-year-old I am given a bike for Christmas. When I am thirty, I am given an iPhone for Christmas, and I can remember being given the bike. When I am eighty, I can recall being given the iPhone, but have no recollection of being given the bike. The argument is that according to Locke's memory criterion, the eighty-year-old ‘me’ is the same person as the thirty-year-old ‘me’ but is no longer the same person as the ten-year-old who received that bike. The thirty-year-old is the same person as the ten-year-old as they can remember the bike. However this cannot be true according to the rules of logic. The eighty-year-old (A) is identical to the thirty-year-old (B) and the thirty-year-old (B) is the same as the ten-year-old (C), but the eighty-year-old (A) is not identical to the ten-year-old (C). But the laws of logic state that if A = B and B = C then A = C. So Locke must be wrong.

A standard response to this objection is to refer to chains of interconnected or overlapping memories. The eighty-year-old does not have to be directly connected to the ten-year-old through memory, provided there is a chain of memories intertwined, so to speak. Admittedly I sometimes struggle to remember what I did a couple of weeks ago, but hopefully the day after I could remember what I did, and the day after that I can remember the day before and so on.

A different problem with the memory criterion is that of false memories. It might seem that the very term is a contradiction: either we remember truly, or we don't remember at all. But it is certainly possible to have a first-person experience of remembering being present at an event when one was not present and this could be indistinguishable from a ‘real’ memory. If I woke up with vivid apparent memories of being Lady Gaga and performing at Wembley, wouldn't this make me Lady Gaga the person (if not the physical woman) according to Locke's criterion of identity?

Again, we can reply to this with a qualification. Perhaps the state of consciousness we are experiencing as a memory needs to have an appropriate causal relationship with the event being experienced for it to be called a genuine memory. So unless my ‘memory’ of singing at Wembley is caused by my actually singing at the concert, then it's not a memory at all, and can't be included in Locke's theory.

We could also revise Locke's theory to incorporate an appeal to a causal dependence on broader psychological factors going beyond memory – to include beliefs and character, for example (see Sydney Shoemaker, Personal Identity (1984), with Richard Swinburne). This revised theory would say that a future being will be me if this being's future psychological states are dependent in a typically human way on my current ones. In other words, this future being can be said to be psychologically continuous with my present self, provided there is a chain of suitably connected psychological states.

The above is old ground, and most people would agree that memory is an important part of our identity, whether or not they would go so far as to say it is either necessary or sufficient for it. But my question now is: how much influence can we have in shaping our identity and self-perception through memory?

As we go through life we forget some things and remember others. Often this seems to be an involuntary process. Indeed there are many things we may wish we remembered, and many we wish we could forget. Paul Ricoeur (1913–2005) in Memory, History, Forgetting addresses the issue of what we forget. This is often seen as weakness, forgetting is the enemy of memory. But we regard with horror the possibility of a memory that never forgot anything. Ricoeur refers to the fable of Funes the Memorious by Jorge Luis Borges by way of illustration of what a curse this would be. Could there be an appropriate measure of memory and forgetting, negotiated to find some sort of balance reminiscent of a golden mean? We should remember and forget, but neither in excess.

Ricoeur refers to two ‘figures of profound forgetting’. The first is forgetting through the erasing of traces. The second is ‘backup’ forgetting, which is reversible. The memories are somehow kept in reserve. Traces can be written (historical documents which can be public or private diaries), psychical (impressions/affections left after an event has left a ‘mark’) or cerebral (the domain of neuroscience). All of these traces are open to willing or unwilling abuse or alteration. I've been focusing here on the traces left by documentation, which in the past has been largely associated with the collective, rather than individual memory via personal diaries etc. A diary, of course, in the traditional sense of being something private, is very different from memoirs. The latter are intended to form part of the collective memory of a person and focus more on the identity of that person through the eyes of others than on self-identity.

In social terms, anthropologists and sociologists widely believe that the collective memory of a group is a social construction. There are those in society with the power to construct the collective memory: we are reminded of what is deemed to be important to group identity through media, religion, ritual, and taught history. Of course there are extremes of this, where the history books have literally been rewritten. But the collective memory of the World Wars, and other significant historical events, are maintained and reinforced in folk culture, so that even those of us who were not present at those events have mental images of them, and they are important in shaping social or national identity. For more on this, see the work of Maurice Halbwachs (1877–1945), particularly Collective Memory.

Could the same element of active construction be true of individual memory and self-perception? If one keeps a diary, there exist potential triggers for memory which otherwise would have been lost, or at least inaccessible. Say I sit and read the diary entries for my eighteen-year-old self. There may be events I had forgotten that I now recall. Let's say there's an entry for some hilarious student prank, such as stealing a shopping trolley and riding it down the street. I read this entry and I think, ‘Oh yes, I'd forgotten about that night.’ One interesting implication is that, according to a straightforward understanding of Locke's criterion – that is, should we take a direct memory criterion rather than an interconnected memory view or a broader psychological view – the student who stole the trolley ceased to be me, and then, when I read about the incident, became me once again. This rescued memory also affects my immediate self-perception. I may think to myself ‘wow, I'm actually an idiot’. Reading about past events in my life may also help me to understand my current beliefs, actions and character. Of course there is the question of whether I do indeed remember the events (with a tone of ‘oh yes!’) or whether the event does not exist as a latent memory at all, but comes as new information (with a tone of ‘oh really?’). But either one would affect my self-perception, provided the event was believed to have happened.

So now the element of choice becomes even more important. The ill-advised burning of diaries was an active decision not to remember, a cutting of ties with a previous self that one wants no connection with, so that they can have no impact on one's future self. On a lesser scale, when I sit down each night to write the day's events, I have a decision to make. What I write will affect my future identity, my self-perception, when I read it back. This is the case either for a memory criterion or for a broader psychological interpretation. If one takes a strict interpretation of Locke, then at a future date when I use these words as a memory trigger, my present self will become identical with a future self whose memory is triggered. What I choose not to include runs the risk of being forgotten, and so of ceasing to be ‘me’ at a future date. (If we take the interpretation of overlapping or interconnected memories as being sufficient for personal identity through time, then forgotten memories are less important.)

Ricoeur says the following regarding the manipulation of collective memory:

The principles that apply to a historical textbook intended for public reading could also apply to the diary kept for the eyes of the individual only. What do I include or not include about the events of my life? As a prime example of Sartre's Bad Faith, I can retell the events of my life narrative with the knowledge that I am manipulating my future identity. What I chose to forget is as important as what I chose to remember. Ricoeur quite rightly points out that that forgetting is a semi-passive, semi-active process, and there is much ‘forgetting in avoidance’. Perhaps I argued with someone today. I don't want that to be ‘me’. So I don't include it in my diary. My eighty-year-old self will not remember it. The implication is that I can actively shape my future self by modifying my future memory, based on what I decide to record or not record as future memory triggers.

If we take a broader psychological interpretation of the nature of identity, what I later read about my current endeavours will also affect my sense of identity in terms of my self-perception. If I read the pages of a diary and then think better or worse of myself, or even change my beliefs, would this qualify as the psychological connectedness we mentioned, in which the psychological state of my future self is caused at least in part by the psychological state of my present self? Here we could even remove memory completely, and concentrate solely on self-perception rather than continuity of identity. For instance, I could, at the age of eighteen, record a completely fictitious day, specifically intended to make my eighty-year-old self think in a certain way about their identity (‘oh really?’ rather than ‘oh yes’).

But we no longer live in such a simple age of pen and paper. The discussion of memory and identity is going to need to be revisited in the age of social media. Here we carve out a public record of our identity on sites such as Facebook or Instagram in a digital space. What we expose to public scrutiny is often carefully selected and photographs of a night rather forgotten are hastily untagged in order that we may erase them from our digital identity. The construction element is much more obvious in this arena, and acknowledged in the faux wish ‘may your life someday be as awesome as you pretend it is on Facebook’. No need to revisit an old journal here, as Facebook is handily bringing up ‘your memories on Facebook’. Here, a photograph or post from the same day a number of years ago is brought up on a users’ feed. They are often shared by users, who presumably are happy for this to remain part of their self-perception and identity in the eyes of others. Unfortunately, participation in social media does put the construction and maintenance of individual memory, identity, and self-perception at least partly in the hands of other people. Ricoeur's ‘forgetting by avoidance’ is going to be much harder in the digital age.

What we remember of our past is a huge factor in identity, whether as a strict link ensuring a connection between past and present selves or as part of a broader psychological approach where my present psychological states can in some way affect my future psychological states. What I find most interesting is that when we consider a diary, whether the old-fashioned written journal or a digital version, the link can be lost and regained, and also manipulated to actively shape our future selves. So when I sit and write a diary tonight, I must ask myself … who exactly do I want to be?