Two Kinds of Memory

Matt McGrath sent me a note about memory that got me thinking a bit about it. I’ve never worked on memory, and haven’t checked the literature to see if this issue has been discussed, so this may already be out there. The bottom line point is that there seem to be at least two quite different kinds of memorial beliefs, and the distinction between the two threatens assumptions many, including me, have made about such beliefs.

Ask me how to get from the house I grew up in to the church we attended when I was young, and memory becomes active. I remember the street I lived on, I imagine travelling from that street, across the railroad tracks to Main Street, turning left, going 3 blocks and turning into the parking lot of the church. So now I have a belief that, even if it was present prior to this incident, is now based on the activity of memory just used in reconstructing the route.

There’s a different role for memory, however, that does not involve such activity. Take a belief originally formed on some basis–say, perception–and in the ordinary case, the belief will be retained across time. A week later, one will often still believe that there is a tree in one’s backyard, for example. Is this belief a purely perceptual one? It started out as one, but the retention of the belief seems quite clearly a function of something further–namely, the operation of memory. The belief began by the operation of perception, and continued to exist, we might say, in virtue of the operation of memory: after being formed, the belief is retained in memory, and thus it is a mistake, a week later, to call the retained belief a purely perceptual one.

So one kind of memorial belief is where the belief is based on the activity of memory itself, and another kind is where the belief is simply retained in memory but is not based on any activity of that faculty or power.

This raises the question of whether one kind of memorial belief can be explained in terms of the other, or if we are stuck needing two kinds of epistemic principles in order to account for the (doxastic) justification of memorial beliefs. More important, though, is the issue of the justification of belief retention in memory. If basing is causal, then the basis of such beliefs would seem to remain whatever prompted the belief in the first place (at least, there won’t be an evidential state that is memorial in character that sustains the belief in existence). If basing is explanatory, then memory might play a role here, but it won’t play a role in virtue of there being some state of recollection or recalling that matches, or is appropriately related to, the content of the belief state. So not only is there a question about what the evidence is for such a belief, there is also the question of what such a belief is based on.


Two Kinds of Memory — 18 Comments

  1. Hi Jon,

    Sorry to make trouble, but I wonder if there isn’t a third kind of memory, in addition to the two you mention.

    I open my dresser drawer and see lots of objects in it: pencils, coins, pieces of papers, paper-clips, etc. etc. I don’t notice each of the many things that I see. (I notice some of them, but not all of them.) Among the things I fail to notice is a Danish Kroner. I see it, but I don’t notice it, and I don’t believe that there is a Danish Kroner there. (If you were to ask me, right after I look into the drawer, “did you see any Danish Kroner in there?” I would have sincerely said “no”.) Now, time goes by, and several days later I’m packing for my upcoming trip to Denmark, and all of a sudden it occurs to me “I have a Danish Kroner in my dresser drawer!” I feel certain of it, and then I go look in the dresser drawer, and it is indeed there, right where I had previously seen it. In fact, the earlier seeing caused my present belief — only the causal chain from seeing to belief took several days.

    In this case, memory produces a belief that there is a Danish Kroner in my dresser drawer, but it does not do this by preserving any earlier belief that there is a Danish Kroner in my dresser drawer.

  2. Ram, man, you’re fast! And, of course, no apologies are needed for good points!

    One question about the example: when you say that the earlier seeing caused the present belief, do you mean the seeing that didn’t involve noticing? I’m not sure how I see that this has occurred. I don’t doubt it’s possible, but as you describe the example, it looks like the belief could be completely independent of the seeing but not noticing incident. Is it that the seeing but not noticing caused you to reflect on the various things you’ve put in that drawer, and that causes you to remember having put a Danish Kroner in it (so that transitivity of causation is part of the case)?

  3. Hi Jon,

    I should have been clearer. I meant to be describing a case in which there is a causal chain from:

    (a) seeing (but not noticing) the Kroner, to

    (b) believing (but only several days later) that there is a Kroner in my drawer.

    Of course, (a) alone isn’t sufficient to cause (b). (I think this is typical though: seeing, all by itself, seldom suffices for belief.) Other things need to happen for (b) to occur. Maybe as I’m packing for my Denmark trip I ask myself “do I have any Kroner in the house?”, and that event is the precipitating cause of (b). Still, (b) could not have occurred unless (a) had occured — or so I’m assuming for purposes of describing the case.

    Does that help to clarify?

    Notice that, in cases such as this, memory generates epistemic justification and knowledge, rather than preserving them.

  4. Ram, yes it helps, and there surely are cases where memory generates, rather than preserves justification. One way of reading my directions example works that way too, if one reads it as a case where belief didn’t exist until memory did its work.

    So, we can say there are at least 3 kinds of case: two where memory is active, one in preserving justification and another in generating it, and one where memory isn’t active in this way. But there may be other cases as well…

  5. Jon,
    Your original distinction initially struck me as also including something like a ‘memory how’ vs. ‘memory that’ distinction. I remember how to use a chainsaw and ride a horse, for instance. If someone asked me, I suppose I’d try to imagine myself as I did these things and report it.That’s different from remembering the color of my childhood home. But, although there is a difference between remembering how vs. that, (the former needn’t involve any overt beliefs), when I report on my memories how, it looks to be the same kind of memory. Why can’t we just say that what’s different isn’t the role of memory, but the object of the memory. In one case, its an activity, the other not?

  6. Ram and/or Jon,

    This is a minor question and not directly related to your issue here but I’m curious why you think that Ram’s case is a case in which memory generates knowledge rather than merely preserving a perceptually grounded justification or reason. Clearly, it can’t be memory preserving a justified belief as there is no belief to preserve, but why can’t it preserve the justification? Personally, I don’t want to say that the ordinary functioning of memory adds to whatever reasons you might have had for believing (then) that there was a Kroner in your drawer.

  7. Robert, I can see how you might interpret my example in terms of remembering how, since I do that to construct the path from house to church. But notice that the result of this is propositional knowledge (in the good case): I can inform you that the way to the church is to go here, turn there, go 3 blocks, etc. So the end product is not riding a bike or using a chainsaw, but belief that has propositional content which is justified in virtue of the activity of memory.

  8. Clayton, here’s what I’d say, at any rate. The memory experience that Ram is describing is one that makes it as rational to believe that there’s a coin in the drawer as it would be whether you actually saw one there or didn’t see one. It won’t count as knowledge, of course, unless you actually saw one there, since there’d be an external defeater.

    This result is an analogue of what I think we should say about deceiver worlds in general. Take a Russell world that is only 5 minutes old. The operation of memory will have to generate reasons to believe in such a world on pain of having to count most of our beliefs about our own personal history as unjustified.

  9. Jon,

    Thanks for that. I suspected that there was something like that as a rationale but wasn’t entirely sure. The claim that memory is an autonomous source of reasons strikes me as a reasonably controversial claim. I would have expected more resistance. While I suppose I share the intuition that in the 5 min. old world case, our Russellian manchild would be irrational to refrain from believing what he seems to remember, the Broomean in me wants to resist saying that it follows from that that memory generates good normative reasons for belief and the externalist in me most certainly wants to say that if I have no more reason to believe that I had cereal for breakfast than that guy does, I shouldn’t believe it. I wonder if there are less drastic measures that would allow one to deny that memory is an autonomous source of reasons.

  10. Jon,
    I’m not convinced yet that there is a distinct activity of memory, as opposed to a memory of an activity, when you construct from memory your route, or I how to use a chainsaw. There are just interlinking memories which, when taken together, constitute an accurate picture of how one did something. And that doesn’t look any different (yet, to me, the dim bulb) than remembering the color of my childhood home. Is it the stringing together of these memories that is the activity part? I’m not inclined yet to think of that as part of the remembering, but something one does with the things one remembers. I’m inclined to think the same thing about attending to parts of memories. Remembering is one thing, attending to it, stringing it together with other rememberings, and so on, another. Or at least that’s where I was stuck.

  11. Jon-

    I don’t doubt that there’s a phenomenological distinction between the two sorts of memory you identify, but I’m not sure there is more to it than that. As I understand the psychology of memory, it’s a three stage process of encoding, storage and retrieval. The retrieval stage is not a matter of passively looking up information but of reconstructing it on the basis of what was stored. Research on false memories (see especially what Henry Roedigger was doing in the late 1990s) highlights the reconstructive dimension even of cases like remembering there is a tree in the yard.

    Roedigger does experiments where people hear lists of 15 words, all of which are closely associated with a target word that is not on the list. Very, very often (often enough to make for fun classroom demonstrations), subjects remember hearing the target word, even though they didn’t. It’s not just that they think they heard it or suppose it must have been on the list. They have the specific memory of hearing the word, even though it was never read. I once did a demonstration in class, and a student refused to believe me when I disclosed that I hadn’t read the target word!

    Once we acknowledge that the retrieval process, even when it involves simple perceptual memories, is an active and constructive process, rather than a passive process of looking information up, I think the difference between the two sorts of memories you identify becomes much less marked. In the first case (remembering the route), there are constructive processes involved at the personal, voluntary level, and in the case of remembering the tree, there are constructive processes occurring at the subpersonal level as part of the normal functioning of human memory. I suspect it means that remembering there is a tree in the yard shouldn’t count as simply retaining a perceptual belief.


  12. Robert, let me try to summarize the “activity” point. There’s a piece of knowledge (let’s assume) that I’ve got now, after being asked the way from the house to the church. It is propositional knowledge, since I express it in the form of a sentence to the person who asked the question. If it is knowledge, the question is how I know that the sentence is true. And the answer appeals to the activity of memory that I just experienced. The point is that my memory did something in this case. In the other case, where a belief is preserved across time, it is not in virtue of some activity of memory that the retained belief counts as knowledge (at least there is no conscious activity of any sort).

  13. Chase, this sounds fun to do in the classroom. I think, though, that your tree case is different from the one I was describing. In my case, I see a tree and form the belief that there is a tree in my backyard. That belief displays inertia: it doesn’t disappear, but is retained. A week later, I’m at Rutgers, still believing that there is a tree in my backyard. My belief retention here could be caused by constructive processes occurring at the subpersonal level, but that sounds more like what would happen if the belief were pulled to the conscious level by being asked (at Rutgers) if there’s a tree in my backyard. But I’m thinking of the case where the belief is retained dispositionally. Now, it may be that the little homunculi never sleep and that’s what explains belief retention, but the more plausible account is one that appeals to the storage metaphor. When storage occurs, we presumably have a memorial belief (in some cases).

    There may be nothing of interest here epistemologically–I can’t quite tell yet–but if there is, it will be in the region of basing. One’s evidence for the belief is perceptual, but the belief, in some sense, is now a memorial belief. But memory is supplying no evidence. So what’s the belief based on? It’s supposed to be based on the evidence for it, on usual accounts of proper basing.

  14. Hi, I’m a little late; no fast connection on the weekend. It seems to me that we can understand Jon’s original cases better by distinguishing between a past and a present object for memory: remembering that an object was a certain way, versus remembering that an object is a certain way. The same distinction can be made for peception, too: seeing that there was a supernova, many years ago, versus seeing that there is a star in a particular place. Sometimes memory, or perception, produces beliefs about present objects, and when it does, it relies implicitly on some sort of “objects stay the same” principle. But sometimes it doesn’t, and only forms beliefs about past objects.

  15. Chris, I don’t think that’s essential to the difference. My belief “Student X isn’t in class,” formed on the basis of perception an hour ago, has smoothly transitioned into “Student X wasn’t in class,” preserved in memory. But it is preserved as in Jon’s second case rather than recovered as in Jon’s first case. (Well, it went non-occurent for a little while, but it didn’t require an effort of memory as in Jon’s first case.) You’ve got an interesting distinction, but I think it’s orthogonal to the one Jon draws.

  16. Matt,
    Maybe you’re right. I’m not sure that the transition into thinking of an object as past is so smooth, though; the reason may be just that I’m not sure exactly what is required for “reconstruction.” In the tree example, Jon does not require any change in how he thinks about the object: it is a present object in his yard. But in thinking about your class, there is a change from experiencing it as present to remembering it as past. That might count as reconstruction, though it’s something we do so frequently that it happens pretty seamlessly.

    One thing that is a little odd about the two examples is that in the “reconstruction” case, where the belief is about the past, the full propositional content of the belief stays the same, but in the “preservation” case, where the belief is about what is the case now, the full propositional content of the belief changes with the time; only the narrow content stays the same.

  17. That’s a thorny question, Chris! One of the reasons I’m interested in this is because of an interest in what it is to be able to have beliefs of the form “X is happening now”; I’m sympathetic to the view (expressed by Gareth Evans, if I remember correctly) that in order to have thoughts of the form “X is not in class now” we need to be able to keep track of time, so that later we can have thoughts of the form “X was not in class then.” That would militate in favor of the smooth transition.

    If I take your use of the terms correctly, note also that in my memorial case the narrow content changes, even if we agree that the full propositional content remains the same. (I don’t think I would agree that, but my reasons for thinking so are so eccentric as to not be worth mentioning here.) My inclination is to say that the semantic questions don’t settle the epistemological questions.

  18. In many cases where a belief is about the past, and therefore has a constant propositional content as we move through time, there will be a change in narrow content; the narrow content may be something like “P was true x length of time ago,” for constantly increasing x. On the other hand, we may have a fixed narrow content as well, something like “P was true in 1970.”

Leave a Reply

Your email address will not be published. Required fields are marked *