The Functional Character of Memory

Today's post is by Jordi Fernández. He is an associate professor of philosophy at the University of Adelaide. Jordi's research interests are mainly in philosophy of mind, epistemology and metaphysics. He is particularly interested in self-knowledge and memory. 

He is the author of Transparent Minds (2013), a monograph on self-knowledge, and he is currently working on a monograph on memory. He is also interested in cognitive science and continental philosophy. Jordi's post is the second of a series on chapters from the New Directions in the Philosophy of Memory collection. (See here for the first in the series.) He discusses his chapter "The functional character of memory".

Consider the question of what is to remember something, as opposed to imagining it. This is a question that I have tackled in a recent article. I try not to appeal to the phenomenology of memories, or the content of memories, or the kind of knowledge that memories provide. The reason is that, once I determine what kinds of mental states qualify as memories, I intend to shed some light on what the typical phenomenological features of memories are, what their characteristic content is, and what knowledge they usually provide. So it would be circular to appeal to those aspects of memory to answer the more basic question of what mental states qualify as memories.

One answer, popular within philosophy, is the causal theory of memory: A mental state representing some event is a memory just in case it causally originates in the subject's past experience of the event. But this condition rules out cases in which the memory is 'reconstructed'. And it rules in cases in which, as far as the person themselves are concerned, the event being represented never happened to them. Neither of those outcomes, I argue, is desirable.

Another answer, popular within psychology, is the narrative theory of memory: A mental state representing some event is a memory just in case the subject can use that representation to tell a story of their own lives. But this condition rules out cases in which we are confident that we witnessed the represented event, but we cannot relate it to any other event in our lives. And it rules in cases in which the subject confabulates and the represented event never took place in the past. Neither of those outcomes is desirable either.

What we learn from those theories is that one needs to walk a fine line between, on the one hand, allowing for some error in a mental state while it still qualifies as a memory and, on the other hand, not allowing for just any mental representation that we can cook up in our minds to qualify as a memory. I draw on the literature on functionalism to try to achieve the required balance. My suggestion is that a mental state is a memory if it is typically caused by an experience of the event, and it typically causes, in the subject, both the belief that the event happened and the belief that they experienced it.

This allows for a memory of an event to be 'reconstructed' as long as, in normal circumstances, the reconstructed memory is the mental representation of the event that the subject would have had, had they witnessed the event. But it does not allow for a confabulatory experience to count as a memory. Typically, if the confabulatory patient had experienced the event that they claim to remember, they wouldn't be representing it mentally now. This last claim, which is crucial for my discussion of memory, hinges on confabulation involving amnesia.