5. Memory as an epistemically generative source

This week, I’m writing a series of posts on my new book Memory: A Self-Referential Account (Oxford University Press, 2019). The post today concerns the topic of chapter seven, the epistemically generative character of memory.

There is a debate on whether memory generates justification for beliefs about the past, or merely preserves it. The issue is whether the following claim is true:

When a subject remembers some fact, the subject is only justified in believing that the fact took place because, in the past, the subject was justified in believing that the fact was taking place through some other source.

Preservatists about memory endorse this claim, whereas generativists about memory reject it. I side with the generativists, but I disagree with some of the ways in which generativism has been defended in the literature.

Here I will only mention two of those defences. One of them is by Jennifer Lackey. Lackey offers a case in which a subject perceives a fact, but their attention was focused elsewhere. Later, when they are prompted to remember, they can episodically remember that the fact in question was the case. Thus, Lackey concludes, the subject justifiedly believes, now, something on the basis of memory which they did not justifiedly believe in the past. And I agree. But notice that all the work, in this case, is being done by the fact that the subject did not believe, in the past, that the fact they were perceiving was the case. (They were perceiving it, but they were not attending to it.) In the past, the subject did not justifiedly believe the relevant fact because they did not form a belief about that fact in the first place. Thus, this case does not show that memory is capable of generating epistemic justification. It only shows that memory is capable of generating beliefs that we did not have in the past.

A different defence of generativism is by Kourken Michaelian. Michaelian appeals to ‘boundary extension’ cases in which, by remembering a scene, a subject visualises parts of the scene which were just outside of the subject’s visual field when they originally saw the scene. In some boundary extension cases, the structure of the objects in the scene in the past is very simple, and the spatial distribution of the objects in the scene is very simple as well. (Think of a row of rubbish bins against a picket fence, for instance.) For that reason, when the subject remembers the scene, and visualises a detail which they have inferred from what they saw in the past even though they did not themselves perceive it, it is likely that the memory that the subject has constructed is correct. Assuming reliabilism, Michaelian claims, this means that the subject is justified in believing that the detail of the remembered scene that they have built into their memory was in the scene. And yet, they never perceived it in the past, so they were not justified in believing, through perception, that the detail was there.

My response is that if, when the subject remembers the detail in the scene, they are justified in believing that the detail was there for reliabilist reasons, then the subject was justified in believing that the detail was there when they perceived the scene in the past, for the same reasons. My defence of generativism appeals, once again, to the view according to which memories represent themselves as coming from perceptions of certain facts. If this is correct, then there is something that our memories represent, and yet the perceptual experiences in which those memories originate did not represent. There is something, that is, which is present in the content of our memories and absent from the content of the perceptual experiences in which those memories originate. This is the causal origin of those memories; the fact that they are causally connected with the relevant perceptual experiences.

If the view about the content of memories defended in chapter three of the book is correct, then this is something that our memories put us in a position to believe, but no other faculty could have put us in a position to believe in the past. Thus, when we remember a fact, we are justified in believing that our memory comes from a perception of that fact even though, at the time that we had the perception of the relevant fact, we were not justified in believing, through any other source, that we would have a memory which would come from that perception. Generativism is thus vindicated.