Brain Inspired
Brain Inspired
BI 157 Sarah Robins: Philosophy of Memory
Loading
/

Support the show to get full episodes and join the Discord community.

Check out my free video series about what’s missing in AI and Neuroscience

Sarah Robins is a philosopher at the University of Kansas, one a growing handful of philosophers specializing in memory. Much of her work focuses on memory traces, which is roughly the idea that somehow our memories leave a trace in our minds. We discuss memory traces themselves and how they relate to the engram (see BI 126 Randy Gallistel: Where Is the Engram?, and BI 127 Tomás Ryan: Memory, Instinct, and Forgetting).

Psychology has divided memories into many categories – the taxonomy of memory. Sarah and I discuss how memory traces may cross-cut those categories, suggesting we may need to re-think our current ontology and taxonomy of memory.

We discuss a couple challenges to the idea of a stable memory trace in the brain. Neural dynamics is the notion that all our molecules and synapses are constantly changing and being recycled. Memory consolidation refers to the process of transferring our memory traces from an early unstable version to a more stable long-term version in a different part of the brain. Sarah thinks neither challenge poses a real threat to the idea

We also discuss the impact of optogenetics on the philosophy and neuroscience and memory, the debate about whether memory and imagination are essentially the same thing, whether memory’s function is future oriented, and whether we want to build AI with our often faulty human-like memory or with perfect memory.

0:00 – Intro
4:18 – Philosophy of memory
5:10 – Making a move
6:55 – State of philosophy of memory
11:19 – Memory traces or the engram
20:44 – Taxonomy of memory
25:50 – Cognitive ontologies, neuroscience, and psychology
29:39 – Optogenetics
33:48 – Memory traces vs. neural dynamics and consolidation
40:32 – What is the boundary of a memory?
43:00 – Process philosophy and memory
45:07 – Memory vs. imagination
49:40 – Constructivist view of memory and imagination
54:05 – Is memory for the future?
58:00 – Memory errors and intelligence
1:00:42 – Memory and AI
1:06:20 – Creativity and memory errors

Transcript

Sarah    00:00:04    A lot of the work about memory taxonomies in particular, was really coming about at a time when there was great optimism about taking our psychological categories that, you know, the cognitive psychology had done a really good job of making precise, and imaging was starting up, and we were like, we were just trying to line everything up. I think of it as like really bound up in that optimism, but as I see it has come sort of crashing down. So the philosophy of imagination has been an area in philosophy for a lot longer than memory. Um, but I think our understanding of what imagination is, or the many different purposes to which it could be put has, is sort of all over the place. I mean, I think a lot of the early computer designs are very much based on how we thought memory worked. It was a sort of sensible way of thinking about memory, of having, you know, like taking everything, giving it an address, <laugh> storing it, and then calling it up by that address. And then here we go. Um, and that kind of view of memory in some respect or another is sort of in the background of a lot of philosophical and theoretical works that people wrote at those times.  

Speaker 2    00:01:13    This is brain inspired  

Paul    00:01:16    Greetings. I am Paul. My guest today is, uh, Sarah Robbins. Sarah is a philosopher at the University of Kansas, and she’s one of a handful of philosophers specializing in memory these days. Much of her work, uh, focuses on memory traces, which is roughly the idea that somehow our memories leave a trace in our minds, and we discuss memory traces themselves, how they relate to the Ingram, uh, which Sera calls a kind of ized subset of memory traces. One that is more focused on the physical implementation of a memory trace in our brains. Uh, we also discuss how memory traces may cross-cut the categories of memory in our current, uh, more or less accepted taxonomy of memory, which divides kinds of memory, like episodic memory, implicit memory, semantic memory, declarative and non declarative memory and so on. There are challenges to the idea of having a stable Ingram like, uh, memory trace in our brains, for example, knowing that all of our molecules and synapses are constantly changing and turning over and being recycled, uh, which is often called neurodynamics.  

Paul    00:02:31    So we discussed that. And another challenge, um, called memory consolidation, which refers to the process of transferring our memory traces from an early unstable version to a more stable, long-term, uh, version in a different part of the brain. If you’re into neuroscience, you’ve probably heard of the, um, recent technology called optogenetics, which allows neuroscientists to use light to precisely activate and record a defined population of neurons. And we talk about how that tool has affected how we think about memory. There’s an ongoing debate as well in neuroscience and philosophy and psychology, uh, about the distinction between memory and imagination. Are they essentially the same thing? Why or why not? So we discuss that. We also talk about whether memories function is future oriented or not, and whether we want to build AI with our kind of often faulty memory, uh, or if it should have perfect memory.  

Paul    00:03:31    So this episode is kind of a tour of many of the current issues in the philosophy of memory, um, many of which, uh, Sarah writes about. Uh, by the way, apologies if you’re watching her video, gets, uh, slightly grainy for a couple prolonged stretches. So I hope that’s not too distracting. I link to a handful of Sarah’s work. If you wanna dive deeper from what you hear, uh, on this episode, go to brain inspired.co/podcast/ 157. If you want full episodes of brain inspired and to join our Discord community, or just to show your appreciation, um, supporter on Patreon for super cheap thank you to my current Patreon supporters. You can go to brain inspired.co to learn more about that. All right, here’s Sarah. I didn’t know that the philosophy of memory was like a, a, a thing, you know, <laugh> and until I came across your chapter and then started diving in more. I mean, there’s a philosophy of everything these days, isn’t there? Yeah,  

Sarah    00:04:30    I think, I think philosophy of memory is sort of proof of concept of that idea. So it wasn’t a thing that long ago. Um, so I, a lot, in a lot of ways share your surprise and fascination, um, pleasantly at least. Um, yeah, it was a decade ago that I was writing, finishing my dissertation, and at that point it would’ve been presumptuous to say that there was a philosophy of memory. There were a few of us starting to do more of that sort of thing. And of course, philosophers have always talked about memory here and there, but in, in terms of it being an area of philosophy that people outside of doing that work would recognize as a topic, it just really wasn’t the case. So it’s been a very, um, a very active decade moving in that direction.  

Paul    00:05:11    Okay. So I, I want to, uh, ask you this, um, <laugh> before we move on. I just had the thought, you know, how philosophers talk about making moves, you know, and then I made that move and I’m making this move. Where, where did that come from? Like what, what is the deal with making a move?  

Sarah    00:05:26    <laugh>? That’s a great question. I think philosophers, really, maybe it’s because we don’t have labs and we aren’t doing, we, we crave action verbs. I think that especially, you know, physical action verbs that, um, people often talk about, you know, going in directions, making moves. Um, and I do think there has, maybe cuz of the history of the discipline, people lately have tried to move away from a lot of combative metaphors, <laugh>. So making it sound like you’re oh, um, starting doing war like, uh, activities when you’re, so maybe, maybe moving is less, um, aggressive and but still assertive and physical in a way that, yeah. Um, I don’t know. <laugh>  

Paul    00:06:06    You don’t know where it came from though? Like, how old is it? I I don’t even, I don’t even know, you know, how to, how, how I would look that up even. Yeah,  

Sarah    00:06:13    It’s a great question cuz I think it’s mostly, I don’t know that you see it so much in the writing, but you certainly see it in the way that people talk and engage with each other. Um, right. And yeah, it’s,  

Paul    00:06:23    So the cart wasn’t making moves. Was he  

Sarah    00:06:25    <laugh>? Well, he certainly was, but I don’t know that he called. I don’t know that he would’ve called <laugh> called it that. Um, yeah, that’s, I’m good. I actually, I’ll ask around. I think it’s, so that’s so in the, in the water of that we swim in that I don’t, I actually have never really thought about where that came from. So  

Paul    00:06:41    Yeah, I, I, I, I cuz I, I don’t know if I just came across the phrase the other day or, or what, but, um, it’s so vague and like I don’t actually really know what it means, you know, it’s not a very technical term. So  

Sarah    00:06:52    Anyway. Yeah, yeah. <laugh>.  

Paul    00:06:55    Yeah. So, so, um, you mentioned like the, the philosophy of memory is fairly young, right? Mm-hmm. <affirmative>, um, and you were, you know, 10 years ago were one of the, uh, first people really writing about it mm-hmm. <affirmative>, where, where is it now? Is it, is um, is it combative in the, in amongst philosophers? Is it, is everyone, um, nice to each other? What, what is the state of the philosophy of memory?  

Sarah    00:07:20    Um, I think as a community itself, it’s a great community of philosophers. So it’s been, maybe it’s because I think it’s brought together a lot of people. A a lot of people doing philosophy of memory have come from kind of philosophy of mind and science in ways that they want memory to be the topic they’re thinking about in that intersection. But there’s people that have also come from other areas like epistemology and the history of philosophy and other places. So I think maybe they’ve always felt like outsiders cuz they were talking about a topic. So it’s like we’ve created a community of people who everyone else was like, why do you care about this? And now we’re all hanging out together, <laugh> caring about the same stuff. So in those ways, it’s been a really, um, it’s a very supportive community and I think because yeah, it’s, it’s newer and those of us who have have been involved in starting it are just happy that it exists. Um, we’ve been, you know, not eager to kick anyone out who wants to genuinely contribute. So <laugh>, um, it’s still a bit of an issue, uh, convincing some people who are, uh, you know, more traditional about philosophy that, that, that it is an area that it’s kind of an inherently interesting set of topics. But, um, but philosophers have talked about memory for a long time. Those puzzles have, have, you know, been there, right? Um, it’s just kind of giving them a more proper focus.  

Paul    00:08:36    And how, how did you come on to be interested in memory and the philosophy thereof?  

Sarah    00:08:43    Yeah, I mean, for me, I mean, it was always interesting, but there wasn’t a literature. It was really actually my advisor, um, in grad school, uh, Carl Craver, who’s a philosopher of neuroscience mm-hmm. <affirmative>. Um, I was at that stage of dissertation writing where you’re very much spinning your wheels, um, at least in philosophy, that kind of how you’re trying to put together this big project <laugh>. And, um, I really wanted to talk about different levels of explanation between psychology and philosophy and neuroscience. I had these and it was just, it was, I can see now from the other side of it, it was just way too big of a project. Um, and it was just out of control. And he was like, well, you know, why don’t you just answer this question about like, let’s just focus it a little bit, like take a memory trace.  

Sarah    00:09:22    Is that a, a personal laris sub personal phenomenon? So that was the kind of way I was framing things in terms of that distinction between levels. And he was like, you know, go and answer that question. And he was leaving as I remembered he was leaving for like a semester in Germany or something. And, um, when he came back I was like, well, you know, I was supposed to be finishing an entire draft of the dissertation while he was gone. And I was like, well, good news, bad news. Like, I didn’t finish the whole thing. I’m only done with chapter one, the one about the memory trace thing, but now it’s like a hundred pages long <laugh>. He was like, I think actually you’ve just changed your dissert. I think you’re writing about that. Wow. Um, so that was sort of, um, yeah, kind of getting into the top, like realizing like, oh, there’s things to say here.  

Sarah    00:10:03    And, you know, there are, there’s a lot of memory psychology at, I was at Washington in St. Louis, um, and so Roddy Roeger and Kathleen McDermott and, um, a bunch of people working is, uh, on plenty of other memory <laugh> researchers as well, whose names I’m suddenly just completely blanking on. But, um, you know, there was a lot of memory in the air, so as soon as I kind of focused in on that and had taken classes with these people and, you know, been in conversations with them for several years, it was, you know, sort of clearer to see how you might come to focus on that as a, as a project in and of itself. So  

Paul    00:10:39    That, that, that seems like a really cool PhD program. Um, what is it, like the, what is it, what’s the name of it? But it seems very interdisciplinary and, uh, seems  

Sarah    00:10:47    Like a great place to be. It’s philosophy and, yeah, philosophy, neuroscience and psychology. So p n p, um, it’s a great, uh, yeah, there’s a range of things like it in philosophy. I, I mean, obviously have particular loyalties and affinities to that one, but it’s a really fantastic, um, way of doing that kind of the kind of work I’ve always wanted to do. So,  

Paul    00:11:09    Should I have Carl on the podcast?  

Sarah    00:11:11    You should. He is, uh, um, good luck getting him to sit still for an hour, but, um, yeah, he is, uh, he’s fantastic. Yeah.  

Paul    00:11:19    <laugh>. Yeah. Okay. So, um, you, you, you mentioned memory traces and I guess that was what, where your dissertation, um, took you, but you’ve since written about like tons of different, uh, topics in the philosophy of memory. So, and I could ask about any one, but I’m curious what you’re excited about right now. Is there a specific thing that you’re working on that, uh, you’re thinking about all the time that, um, maybe we could start with with that?  

Sarah    00:11:45    Yeah. Um, so I’m actually, I’m, I’m writing a book about memory traces <laugh>, um, at the moment. That’s what I’m sort of, uh, in the, the throes of working and writing about. And, and in some ways connections between that and some of the work that you’ve talked with people on here previously about, about the concept of the Ingram and, and what sort of contemporary research about that, sort of how those two things relate to one another. Um,  

Paul    00:12:09    You prefer saying memory traces to, to Ingram, right?  

Sarah    00:12:13    Uh, I do overall. Um, but I think some of the, I think they’re very related notions. I think they do get used interchangeably. Um, there’s reasons to keep them separate for commitments that might come on the Ingram side, that you could have a trace view without having that ingram, some of the commitments that an engram brings along with that.  

Paul    00:12:34    What, can you, can you spell that out for me a little bit? I thought they were equivalent.  

Sarah    00:12:38    Yeah. Yeah. So I guess, I mean, so certainly the memory trace idea is an older idea, right? So it’s, you know, you find versions of appeals to things, uh, to memory traces in Aristotle, you find rejections of it and other views of, um, of those kinds of times and appeals to it kind of throughout the history of, well, you know, philosophy, but when philosophy was, what were the distinction between philosophy and psychology and the kind of natural philosophy of the world wouldn’t have been distinguished. Hmm. Um, so it’s an older, it’s a, certainly, I think of it as something that people have long talked about when they’ve talked about memory, but it’s never been very precise, I think for, for an idea that’s had so much staying power. No one sat down and said, here’s kind of like, it’s the essence of what it is.  

Sarah    00:13:21    So it’s part of why I’m trying to write a book. So put something out there like, here’s what I think, you know, we should say it is. And I think without that, the engram is a kind of way of making that idea much more concrete. And so of course they’re very, you might think of it as a, as a kind of proper subset of that memory trace idea that commits you to the idea that it’s a physical brain mechanism. Oh, okay. Maybe it’s about certain carrying certain kinds of information. So Right. You could have been a philosopher of a different era and thought that minds and bodies were distinct in some kind of way. Maybe you would’ve thought about the memory trace in a way that wasn’t necessarily in the brain <laugh>, or wasn’t even necessarily a kind of physical thing.  

Paul    00:14:00    Okay. Um, so it’s more, the Ingram is more the implementation of the memory trace, like a physical  

Sarah    00:14:05    Yeah. Implementation.  

Paul    00:14:06    Okay.  

Sarah    00:14:07    Yeah. It’s kind of a, I sometimes, I mean, I’ve said sometimes and I think is still what I think that it’s a kind of ized version of a memory trace. You know, it’s kind of that concept <laugh> taken into that, you know, into the kind of contemporary space of neuroscience and biology.  

Paul    00:14:23    How, how’s the book coming along? When is it gonna be available?  

Sarah    00:14:26    <laugh>? Um, I mean, in some sense I’ve been working on it for a long time. Um, it’s, I’m committed to sharing drafts of chapters with, um, some colleagues for like a workshopy session starting in a couple months. So, um, that should, uh, so it’ll be shareable with a trustable small set in a couple months, and then hopefully, uh, more people, a broader audience not too long after that. So,  

Paul    00:14:51    So four or five years? No, I’m just <laugh>.  

Sarah    00:14:53    Exactly. <laugh>. I mean, guy. Yeah. I, i part of me wants to say like, of course not the given <laugh>, you know, the math of how these things work. Maybe that’s true. <laugh>.  

Paul    00:15:02    Yeah. Um, so, okay, well, let’s talk about memory traces because it’s se you know, like you sort of alluded to, a trace is kind of a vague <laugh> notion or term also, and I, I guess that’s why Ingram scient it, uh, as you say mm-hmm. <affirmative>. What are some alternatives to the idea of a memory trace? Like how memory could work?  

Sarah    00:15:24    Mm-hmm. <affirmative>. Yeah, I mean, so I think part of the trick in answering that question is first thinking a little bit more about what a memory trace is, cuz I think without, yeah. One of the things I find, uh, kind of frustrating in that kind of dialectic to begin with is that the kind of thing you hear people frequently say nowadays in philosophy, but also in psychology and neuroscience is like, well, we know that memory is not a perfect recording of, you know, of one’s experience. Or we, you know, they might start by saying like, we know that memories are not perfectly archived in a, you know, library in the mind or something like that. And so you get a kind of straw man as a philosopher, you know, kinda straw person articulation of what the idea is a very, like, caricature of the idea. Like, of course there’s not literal pictures stored in one’s head of past experiences.  

Sarah    00:16:13    So clearly that view is, doesn’t work, but then like what the alternative is, um, is not always well specified. So, um, so one first thing is to say like, there might be some ways of articulating what this commitment is that are somewhat less cartoonish than thinking that it’s an argument for a video recording or a perfect image in one’s head. Um, and so that’s sort of part of the goal is to give a more substantive idea of what that, that idea might be. Um, and also for philosophers who’ve written about these kinds of things, you either get the people who say like, you know, this sort of record archived recording idea is misleading. Um, or you get people who say like, well, sure, I still feel, you know, of course there are still traces, but they’re distributed, they’re not local. Or they say, of course there’s still traces, but they’re, um, they don’t have content to them.  

Sarah    00:17:00    Or of course there’s traces, but, um, they’re, uh, dis dispositional, not sort of, they’re not, they don’t, they’re not explicitly represented or something. And those are all interesting alternatives, but they’re really different from one another. And you might think like, if you are there some features that if you got rid of them, you just wouldn’t be a trace anymore. Like we have to, I think it’s interesting to sort of have an idea of what the kind of concept is such that we sort of know when you’ve moved beyond it or not. So, you know, I have a kind of collection of different things philosophers have said, and so it seems like you could have a trace view, but it could be radically, you know, very, very different from something that is also a trace view in terms of which things it supports or doesn’t. Um,  

Paul    00:17:41    So what’s, what’s your trace view? Yeah,  

Sarah    00:17:44    Yeah. So as I think about it, I think the most profitable way of thinking about it, it is that traces are the kinds of things, uh, that it, it’s an entity that’s the, the explanatory role that it plays. Like the, the function of them, why we think that they exist, why we go looking for them, um, is that we think is in order to explain how we retain specific events or particular past experiences, that the trace is really there to explain a particular kind of retention, right? The way when I remember, uh, you know, an some very specific occurrence, um, or something that’s only happened one time. Um, so it’s about a part, a way of retaining information that might exist alongside keeping track of general tendencies and trends and broader bits of information. But it’s, it’s about that kind of a commitment. So if that’s the right way to think about it, then non tracee views or views that try to explain the entirety of memory or the entirety of cognition without thinking you’re gonna have those discreet, you know, whether you wanna think that their representations or not, but discreet mental brain states that hold individual events,  

Paul    00:18:56    But, but a trace is not confined to episodic memory, right? I mean, it can be implicit memory non declarative or, or what is the reach of a trace in the ontology of Yes. So taxonomy of memory, <laugh>,  

Sarah    00:19:09    It’s a great question. I mean, so the text you had sent me that you might ask about taxonomies in membrane, I think that they’re kind of a mess in that way. And so I think that, I think of the idea of specific events or specific, you know, uh, individual experiences or, you know, pieces of information learning or something like that as being the kinds of things that could serve traces and that might crosscut classifications. So, you know, so you might have, um, I mean, depending on how richly you define episodic memory, there are lots of things that we do or encounter only one time that don’t meet that criteria because they don’t have the kind of phenomenological, the kind of mental time travel aspects to them. I mean, if you’re me, I have like, I think eight or nine memories that meet that <laugh> more elaborate requirement that toing wants.  

Sarah    00:20:00    Um, but a lot of memories are, you know, the information I only acquired, you know, it was just one time that that thing happened, um, or one time where I heard that and like, I’ve forgotten a lot of the details, but I, I remember that thing and that that’s what I, where I learned it or something like that. Um, and so that could hold true for some forms of implicit memory as well, some kinds of associations. Um, so that kind of, yeah. Discreet one-shot learning, um, episodes kind of. So it’s, it’s a way of saying that’s where the notion has its most proper use is in thinking about retaining information in that way. So episodes would certainly be a central feature of it, but maybe not in the rich sense that how we’ve thought about episodic memories.  

Paul    00:20:45    Hmm. But so you think that the, uh, that our taxonomy of memory is a hot mess, <laugh>  

Sarah    00:20:51    <laugh>? I do. I I do think that’s a technical term hot mess. Um, and yes, it is very much a <laugh>, it’s a hot mess. Um, yeah, it’s a, it’s, I’ve been, I was just thinking a few weeks ago about the fact that, you know, this kind of standard model that everyone uses, you know, they dista we distinguish between declarative and non-clear that we use these terms declarative and non declarative for that, which I went trying to track down like, where’s the first paper that calls it declarative? That’s such a strange mm-hmm. <affirmative>, you know, like, especially cuz now we study what we think are declarative memories in lots of animals where there’s no declaring happening. Right. Um, and so, and it does look like it’s maybe, uh, the kind of holdover of the early cognitive revolution days in psychology and a lot of list and word learning paradigms is where people first start talking about it.  

Sarah    00:21:40    I, I don’t know that I’ve, I haven’t found anything where I could say, oh, that’s the first use mm-hmm. <affirmative>, but that’s where you start seeing oh, man, and papers on Yeah. <laugh>. And it does have this kind of behavioral behaviorist tinge of, you know, like if you’re thinking about the memories, but in terms of like the expression that you see, um, where you see them. Um, and yeah, kind of how, I mean, episodic memory’s been thought about over time is complicated and the whole implicit side of the story is just, you know, everything that’s over there, those things are just sort of thrown over there and ignored, I think, um, by a lot of people doing memory work. And, uh, yeah.  

Paul    00:22:16    What do we need though? Do we need, um, more categories? Do we need a finer grained taxonomy, or do they, do we need to lessen the individual individuality of the categories? Because there’s, um, blurry lines between them. Do, do you know what we need? <laugh>  

Sarah    00:22:36    <laugh>. Yeah. It’s a, it’s a great question. I mean, I feel like half of the work, or a lot of the work about memory taxonomies in particular was really coming about at a time when there was great optimism about taking our psychological categories that, you know, the cognitive psychology had done a really good job of making precise and imaging was starting up, and we were like, we’re just trying to line everything up, right? And so the kind of memory systems view, and let’s find the cog, the neurocognitive systems that support that. I think of it as like really bound up in that optimism that as I see it has come sort of crashing down more broadly. All of the kind of worries about cognitive ontology and, and, you know, the ways that there might be, you know, massive neural reuse across systems. And how we think about those functions, I think is, is there’s that worry kind of in the background of that sort of project.  

Sarah    00:23:24    Um, but I, but more broadly, I think, you know, that we could have many tech taxonomies for different purposes. Um, so we could have one that says, let’s pay attention to the, the brain regions that are most centrally involved or that, you know, the kind of, and and that might give us one way of saying like, well, these are, you know, maybe all of these are a similar kind of operation, even if they’re, they’re not, but they’re on massively different kinds of content, right? Or, okay, maybe there’s ones that, you know, have content features in common, or maybe there’s ones that come from particular species or others. Um, but the current way of sorting things out does leave out a lot of things, right? It leaves out a lot of, there aren’t obvious places in which to put lots of kinds of memory that from everyday life seem like they’re important.  

Sarah    00:24:11    Like, um, I mean the, and I’m gonna say this and then I’m gonna find out there’s a whole literature in what these areas I just don’t know about. Yeah. Always, right? I mean, like <laugh> yeah, like memory for like, the sort of standard story about episodic memories, memory of episodes, and then semantic memory is understood as like memory for facts or kind of generalizations, but it’s not clear that that has a way of understanding, like the way that we have memories for like people and places and things that we know well, that I might have, you know, like a, a sort of memory of that person, which is over time just, you know, I’ve abstracted from all indivi. Maybe I rem of course I remember some experiences, but I also just remember, I remember that car. I remember my first car, right? Like, maybe I don’t remember any particular experiences about where I drove it. You know, I, I can think about that <laugh>, right? As a thing. Yeah. In particular separate from all that stuff. And like, well, where does that fit in the taxonomy? Um, or, and that’s just sort of one example of the moment. But yeah, lots of those kinds of things don’t have neat spaces to go, but you might think when you’re asking people about those, their memories, those are the kinds of things people go to rather centrally. Um, so  

Paul    00:25:28    Yeah, we don’t know where there  

Sarah    00:25:29    Could be other Yeah.  

Paul    00:25:30    It doesn’t fit cleanly in the <laugh> in the taxonomy. It’s like, it’s an abstraction. It’s kind of semantic, but, you know, and with, with the car example, because it’s, um, it is factual, but it’s also autobiographical. Yeah. I’d have to think about that more.  

Sarah    00:25:44    Yeah. And it’s got all this, it’s got imagery, it’s got, you know, it’s got all this stuff to it. Um, but yeah,  

Paul    00:25:52    Like zooming out though, like, um, <laugh> you were talking about the, the worry about cognitive ontologies. What’s your view on that or how have your, how have your thoughts been shaped over time, um, regarding the neuroscience to psychology divide or connection or lack thereof?  

Sarah    00:26:13    Yeah. Um, yeah, I guess, uh, like I, I mean, I’ve always been a little hesitant about some of the connections. I mean, this comes up in some of my work more directly, but I mean, I do think there are questions about levels of explanation, sort of how they fit onto and map to one another. So in the, I mean, when I was in graduate school and taking, uh, wash U had a standard kind of methods in F M R I course that all the grad students that were in the kind of psych neuro space took, and the students in my program took as well. And it was a team taught course by, you know, the whole slew of people at, at Wash U who did the different stages and phases of all the things that go into F M R I, which was absolutely fascinating to do.  

Sarah    00:26:59    And it, you know, there was, and that was a good place to be doing a lot of optimism about this kind of like, we’ve got these really well de defined kind of cognitive phenomena and effects and functions, and we’re getting better and better at the kind of task design to do in a scanner with the right kinds of analyses. And we’re lining these things up. And, and I remember even at the time being a little frustrated and thinking like, well, suppose my psycho psychological theory says that these are two distinct processes and you’ll come back to me and you tell me it happens in one brain area. That doesn’t obviously tell me that that’s wrong. Right? It might tell me that brain area is doing two functions, right? It might, um, that it’s doing two things or, you know, and the reverse could be true.  

Sarah    00:27:43    You could have one function and I could find out it’s in three or four locations that it’s happening. But like, of course there are ways that these things constrain each other, but in terms of like, which direction the inference goes, like, are we using what we know about these functions to figure out what the brain is doing? Are we using it? Is it like a referendum on psychology? Um, I was sort of worried about those questions from the beginning and people would, my colleagues who were, you know, my grad student friends who were in psychology and neuroscience would often roll their eyes. And like the philosopher’s always making this so complicated, <laugh>, um, <laugh>. Um, but I mean, you know, they took me seriously, but that wasn’t their concern. And I feel like the field more generally, from the scientist point of view, for, for their own reasons and motivations obviously is kind of in a similar space now. Those kind of, those worries were more peripheral, um, 10 or 15 years ago than they are, than they are now. Yeah. Um, and I don’t have, yeah, I don’t know that I have,  

Sarah    00:28:42    Uh, yeah, well worked out like views about how it’s gonna go. I worry about a kind of, I mean, I worry about what a bottom up ontology looks like. I think some people have been, you know, like let the brain tell its own story about what the categories are like. I think that might end up being one way to do it, but our ways of examining that have always been guided by what we think the functions are that we’re looking for and how we understand those phenomena. So I’m not, I’m not sure we’d even know how to interpret that, where we to get it. Um,  

Paul    00:29:14    Yeah. And we have to use words eventually anyway, so <laugh>.  

Sarah    00:29:18    Exactly. Yeah. Yeah.  

Paul    00:29:21    Uh, you were talking about, you know, the same brain area, maybe having two different functions and, um, you’ve written about imagination versus memory because, um, people have an account. So I wanna come back to that, but I, but I, I’m jumping the gun a little bit because, uh, I wanna talk about traces a little bit more because you, you focus so much That’s true. Uh, of that in your work, is optogenetics a like revolution in the idea of studying traces?  

Sarah    00:29:50    I, I think that it is. I mean, that’s the only, as someone, you know, I took the classes I had to take at in doing the kind of neurobiology cell biology stuff, but that was never my focus. I was much more interested in the psychology and the cognitive neuroscience, um, side of things. And, you know, so the mere fact that I’ve learned a lot about now write about optogenetics is in its own way. I kind of tested to me of like, okay, you know, I started reading these papers and thought like, uhoh, I have to go learn this stuff. I have to. Cuz it’s, yeah, it’s, yeah. It’s, it’s changing. I think, uh, lots of things about how we think about those things. So I’m, I’m very excited about it in general, um, uh, as a research program, but also as kind of a stage, um, of asking questions about memory.  

Paul    00:30:35    H how has it affected the, the philosophy of memory? I mean, has it settled questions? Is it opening up new cans of worms?  

Sarah    00:30:45    Yeah, I mean, so for the philosophy of memory, you know, as we were talking about earlier is a pretty new area. And unlike other parts of these sort of ways we think about the mind that philosophy’s been talking about for a while, those have been for other areas have been more hooked up to the science all for a while now. You know, there’s been a lot of work at the intersection of philosophy and psychology and neuroscience, so memory because it hadn’t been, had a lot of catching up to do, right? So philosophers, the traditional philosophical conception of, of memory is like, it is for a lot, you know, that memories are these things that are stored in your head and they’re, you pull them out when you need them and you, you know, hit play and there they go. Um, and, you know, philosophers have used appeals to that kind of thing to explain everything from knowledge to personal identity, moral responsibility, right?  

Sarah    00:31:29    I mean, memory’s an important thing. And so you have this kind of strong notion of it. So once you have people doing philosophy of memory, it becomes a real concern that like, uhoh, this is another place where the technical term hot mess applies. Well, I think like, turns out memory is a hot mess. Like, you know, you think you’re remembering and you’re not. So all of the false memory literature showing, you know, that we can be sort of sys memory can be systematically manipulated and distorted in ways you can’t detect, um, from the first person perspective that really like, so I guess like philosophy of memory. If, if it started by saying like, let’s pay attention to the science of memory, that’s where it had to go. That’s where everyone went first, myself included. Like uhoh, we gotta say something about, about that. Um, so maybe memory isn’t what people have always thought that it was, maybe we should think about it differently.  

Sarah    00:32:19    You ask about the optogenetics role. So I guess in that sense, because everything was so dominated by the false memory literature and the study of it in psychology and neuroscience, most of the science to which philosophers of memory have paid attention has been at that level. So optogenetics, part of its kind of basic significance is giving a really strong argument for paying attention to cellular, molecular level studies of memory, which, you know, would, of which there are of course decades, if not, you know, at this point more than a century of work. Uh, that’s a huge, huge place piece.  

Paul    00:32:54    But the has optogenetics settled whether there are memory traces, whether there is an Ingram. I mean, I know I had Tom Tomas Ryan on, and I, I think that he would reply that yes, it has <laugh>, but he’s a pro Ingram researcher,  

Sarah    00:33:08    <laugh>, um, yeah, I’m a huge fan of, of his work and the work in that area. Um, so I think it’s certainly has, it’s definitely, I mean, to my mind, yes, it’s settled that question. There are, there are traces I think where the people who don’t share my trace enthusiasm <laugh> would, would speak up is that it doesn’t, you know, maybe it’s only for these very low level memories. Maybe it’s, you know, like you’re showing them in these cases, you, you haven’t really, you know, obviously we’re a long ways from human episodic memories of someone’s seventh birthday or something like that. Um, and you can grant that, but still think like this is a major step, um, in that direction.  

Paul    00:33:51    One of the things that you write about, um, ar arguments against there being memory traces is, uh, people who are in favor of a neurodynamics account, uh, of memory. Could you just ex, um, explain what that means and then why you argue that, uh, neurodynamics, does that not actually pose a threat to the idea of traces mm-hmm. <affirmative>?  

Sarah    00:34:16    Yeah, so I guess, um, part of why I think it, I guess it depe it depends on exactly how the, the claim of neurodynamics gets made. Um, and I think on certain ways of making that claim, and for certain kinds of trace views, it could cause a problem. So especially if you have something like that sort of straw man conception of like memory traces just are these, um, you know, video recordings in the head, right? If you sort of prove like the brain can’t hold, the brain can’t do that, <laugh>, there’s no such thing that could play that role, then that would be, those kinds of views could be deeply challenged by, by neural dynamics. But more broadly from another side, like the sort of, I think it’s important to articulate exactly what the worry is about what neurodynamics is for memory, because finding out and paying attention to all of the dynamical features of neural activity doesn’t change the fact that this is what feels very similar to the, like two psychological processes, one brain region kind of worry, right? Like lots of things we do might be deeply dynamic and yet, um, right at the experience level, it doesn’t feel that way, right? So the brain is massively dynamic, you know, and maybe even perception too is a hot mess, but it feels to me like I’m looking at a constant world in front of me, <laugh>, right? <laugh>, um, I’m gonna see how many times I can use that <laugh>.  

Paul    00:35:37    I love it. Yeah.  

Sarah    00:35:38    Before we done <laugh>, um, ray it, so like, it doesn’t, you know, the dynamic, the dynamicity of the brain isn’t, isn’t like, uh, things can still be stable and feel stable and have a constancy to them in the midst of that at, at a separate level. Um, so it’s, you know, I think that is a kind of background thing. Also to keep in mind, another point that I always think about when people are worried about neurodynamics for traces is that one of the views that’s been, I mean, people make trouble for it nowadays, but a prominent thing that people have thought about memory in the brain for a long time involves systems consolidation and the idea that memory traces over time move from the hippocampus to prefrontal court. And they’re like that if you’re fine with a trace like picking up and moving and being housed by entirely different neurons.  

Sarah    00:36:28    Um, and that’s a view of people have, you know, obviously there are alternative views too, but lots of people have taken that idea seriously. Then you think that there is, there’s already built into the field some idea that like the same thing, right? That information can migrate to a new place and yet be a persistent version of this kind of thing. So if that idea is there, then there are ways of thinking about persistence or continu, you know, continuity of information that allow for those kinds of dynamics. It is a challenge. You can’t, you know, maybe certain kinds of mechanisms aren’t compatible with what we learn about certain features of neurodynamics. Um, but, you know, so whatever you say is, you know, are the mechanisms by which information persists, they better be, you know, resistant or, uh, the sort of resilient in the face of those kinds of things.  

Paul    00:37:20    But then you’re talking about information and that’s not a physical thing, right? So, okay, you have a memory in the hippocampus, it gets consolidated in the cortex, and those are kind of <laugh> two different traces I guess you would say, but two, two different implementations of the same trace because the traces information ha set me straight on this  

Sarah    00:37:41    <laugh>. Good. So I think that the trace of the, I mean, clearly it has a physical, we think of it as a physical mechanism. I think that’s a, it’s an inescapable feature, but part of how we identify and find that physical neural mechanism is by its informational sensitivity, right? I mean, that’s like over time you can tell the kind of story of getting to the end Graham, that, that Ryan and, and Jocelyn and others have, have found their way too as a story of getting more and more precise tools to say like, you know, the, this is, these are the cells involved because these are the ones that play the most information sensitive role, right? So the way that you tell mm-hmm <affirmative>, like, is this a, is this an Ingram relevant cell or not? Or is this an Ingram relevant process or not? Is the role it plays in encoding and retaining that information?  

Sarah    00:38:29    So it’s a mecca. We’re certainly looking for a mechanism, but it’s a mechanism that carries that information in the, the right way so that both features are, are there from the beginning. Um, and you might think like it’s a requirement that it have a physical mechanism, but the physical mechanism can change over time, right? It might, like maybe it migrates from one neuron to the other. Maybe it, maybe it moves or maybe it just, you know, updates the way it’s held over time, or parts of it degrade in ways that may in some ways relate to that information and not, I think these are actual puzzles we have to think about, about how we think carrying that information relates to what that mechanism is. Um, so yeah, they’re gonna, and I think this is the real puzzle and it’s part of what optogenetics and the suite of tools around it really let us do for the first time.  

Sarah    00:39:21    It’s like, not just watch these things get formed or watch these things when memories are being reactivated, but kind of track them over time, right? We can, we can now image them, we can mess with them when we want to. Um, I say we, I’m not doing this, they’re doing this. Um, but that lets you kind of keep an eye on these states over time, right? And from a philosopher’s perspective that lets you ask a lot of questions about like, well, how do you know you’re finding the same thing, right? Mm-hmm. <affirmative> one is by finding, re-identifying that very same mechanism, right? The neurons that you’ve stained, um, and seeing what happens with them. But another way is whether or not you’re eliciting the same behavior or representation that you think is involved mm-hmm. <affirmative>. Um, and I think it’s gonna be an open and very difficult question, what starts to happen as we notice that, well, the underlying physical mechanism we’re f if we, if we’re following it and it’s changes, it might take us here <laugh>, right? It might be, you know, but if we follow the information sensitivity, it could take us in a different direction and sort of how are we gonna think about which one really matters? Um, I don’t have a good answer to that question, but I’m a philosopher. My job is to pose really tricky questions, <laugh>, so I’m gonna try to pose that question very well and then let the smart people with the tools figure it out.  

Paul    00:40:34    Here’s a tricky question. Perhaps you can settle <laugh>. What, what is the boundary of a memory, I guess, that you could ask that of any representation in the brain, but, you know, thinking about a trace and the, you know, let’s say you have 50 cells, right? That are instantiate this memory trace or something like, is is, is that, is there a clear way to think about the boundaries of a memory?  

Sarah    00:40:58    Yeah, it’s a great, I mean, <laugh>, I’m not sure, um, what excites me about optogenetics and some of the around surrounding stuff is that I think maybe for the first time we could really pose that question and worry about it. Mm-hmm. Right? So, you know, previously lots of work about the kind of basic mechanisms of different kinds of memory, which, you know, involve a kind of implicit commitment to like, this particular memory is some of these particular cells. But for the first time, really we have like, now we can actually show you that we can literally count the cells and then we can reactivate it and we can recount them, and we can see as those studies find, like there’s very strong overlap, but it’s not precise mm-hmm. <affirmative>. Um, and I mean, I don’t know. So I don’t know what will happen as that gets, there’s gonna be a burden, I think, not just at the mechanism level, but at the content level for how we think about that.  

Sarah    00:41:50    Like, so what do you think the content of that memory is? Um, and you know, what might happen is that we start to explore like what would it mean for some of that content to go away or what that content to change mm-hmm. <affirmative> and how does that relate to the underlying changes that we’re seeing? And we might come up with like accounts that better and worse fit that data, um, you know, to make sense of that kind of story of what it is that we see. So yeah, I think, I mean, part of what I was so excited about optogenetics to begin with is it, it has that kind of precision that a philosopher wants for these ki you know, like, oh my gosh, you can literally <laugh> see exactly that, you know, you can count the neurons involved here. Um, but of course, once you can do that, you realize like, well, it’s not exactly the ones that we thought we were recruiting or which ones reactivate or how strongly they reactivate does vary over time. Um, so of course it’s more complicated than that. Um, so, but for the moment I’m just focused on the kind of, it’s <laugh>, it’s a hot mess. Um, you know, it’s exciting that we can do it. Um, like it’s great that we actually get to be worried about that problem, which is in a way a new problem.  

Paul    00:43:04    I’ve, I’ve, lately I’ve been enamored with process philosophy. I don’t know how familiar you are with, um, process philosophical, uh, philosophical accounts of, you know, various things. But, um, I, I, in preparation for, uh, talking with you, I looked up in, um, Dan Nicholson’s book, everything Flows to see if there was a chapter on the process philosophy of memory. And, uh, I, I didn’t find anything. I didn’t do an exhaustive search, but, hmm. This ki you know, approaching it from a process philosophy standpoint, um, I have not thought clearly about this, but, you know, just marrying that idea with the idea of a memory trace, which seems like a, a much more stable thing. And then thinking about the, the neurodynamics account of how everything is constantly, um mm-hmm. <affirmative> being, um, uh, you know, remade, rebuilt, changing, everything is constant, constantly in flux. Right? So I, I don’t know if you’ve thought about that, but, um, I thought I’d just ask you if you had  

Sarah    00:43:59    Yeah, so I don’t know the process stuff. Um, super, I, I don’t know well enough to say much of anything except that I think for a lot of these, um, kind of radical shifting views about cognition and, and the mind, um, it feels like memory is often a sticking point, <laugh>. So I know for like a lot of predictive processing views or a lot of kind of massive simulation style views, a lot of views that are, which, yeah, again, I don’t know the processing stuff as well, but a lot of them come from offloading things onto the environment or operating from very strong kind of general rules you can use, you know, kind of in, uh, across a ride wide range of things and memory. Yeah, it does, it is a kind of sticking point for a lot of these things because it’s, it’s offline, it’s, it’s more static or constant, um, sort of why and when it pops out what good it’s doing us sometimes all of that is, um, is less clear. And so it is, I mean, yeah, it feels like if I wanted to spend my time just picking on other people and other views I could spend, um, <laugh> I could spend time just making trouble for like putting memory into some of those frameworks.  

Paul    00:45:10    Hmm. Uh, maybe, you know, uh, maybe we can switch gears and talk about the relationship between memory and imagination. Um, I don’t know. I, I don’t know if it’s Daniel Shater who started this, this constructionist view of memory and kind of equating it with imagination. And I know that you’ve participated in some debates on this and you’ve written about it, so our memory and imagination the same thing. And, and what the hell is imagination? Anyway,  

Sarah    00:45:37    <laugh>, great question. Um, so no, and then great question. Um, and I’ll find a way to use hot mess again by saying that, like, I, I don’t <laugh>, I think, so the philosophy of imagination has been an area in philosophy for a lot longer than memory. Um, but I think our understanding of what imagination is, or the many different purposes to which it could be put has, is sort of all over the place. And I, so already, I think those ways they’re not well suited categories for one another. Um, I think there are big overlaps between memory and imagination, um, in the ways that like for a lot of imaginative activity, if you think of part of what it is doing is generating imagery or creating, you know, like in a kind of, what I think of as a default sense of what, I mean when I say imagination is like, I imagine like my six-year-old does, right?  

Sarah    00:46:33    I’m imagining this kind of superhero I wanna put together that has like, you know, different animal features and has different kinds of superpowers and it’s something I’ve never seen before, but it’s built out of elements that I’ve derived from experience in lots of different ways. Mm-hmm. <affirmative>, right? That’s part of what, and if on that kind of view, which I think is my sort of, you know, background view that I come in with, like, well, memory and imagination are related because memory is like the raw content that imagination uses. It’s kind of using the things you, you know, like I take what I remember about like what different animals can do and look like, and I, but I put it together in ways I’ve never seen to create a, you know, mythical superhero, um, of some  

Paul    00:47:16    Kind. Yeah. This is, so, um, I guess part of the debate here is that, uh, at least in F M R I, it seems that memory and imagination are, uh, constituted by highly overlapping brain areas, at least as measured by, um, oxygenation levels in the brain, which, you know, is not necessarily, like even if you had the exact same area, uh, the resolution with which you can measure activity in the brain using F M R I I, I don’t know that it would be able to tell you anything conclusive about how it was happening and whether it was exactly the same neurons and brain areas, you know?  

Sarah    00:47:54    Exactly. And I think for the, you might think that the same brand areas would be activated even if the story, uh, I was telling was true versus the stronger story that they’re the same. Right? Because if I’m making use of the content of my memory to do the imagining, then it would be, in fact, like in reading some of those studies, part of my sort of background, my knee-jerk reaction has always been like, like it would’ve been alarming if we hadn’t found that, if that result hadn’t been found. Right? <laugh>, like if you, especially given the <laugh>, especially given the methodology of those studies, which, you know, for reasons that are necessary for doing F M R I work, they’re highly similar. You know, they’re asking people to generate imaginative scenarios that they’ve well-matched to actual memories. Um, that’s gonna, you know, that’s going to, even if these two capacities are more separated, that’s gonna push them as close together as possible.  

Sarah    00:48:45    Hmm. Um, so, uh, so I have some skeptic, you know, sort of skepticism about how much you can infer on the basis of those studies, but also the kind of broader background motivation for those studies. So I think what’s led people to, um, articulate this view of, of this close relationship between memory and imagination is in part because of this perceived need to give a different account of what memory is because of all of the false memory work. So that has made people really concerned, like, well, the, the function of memory can’t be for remembering because, you know, I like if if it were, I wouldn’t be so bad at it, <laugh>. And so what is, what is memory doing? And so finding it’s overlap with imagination and thinking about it in those ways, I think is a kind of, is thought of as a way to save that project. And for me, I’ve kind of gotten off the boat sort of long before we get to the actual studies and thinking that that was necessary.  

Paul    00:49:44    Um, so the, the the constructivist view, which you adhere to, right?  

Sarah    00:49:49    In some ways Yeah.  

Paul    00:49:51    <laugh>. Oh, okay. Well tease that apart for me. So what maybe, uh, can you say what the constructivist, uh, view is and then, and then what you ag you’re a philosopher, you’re gonna be very nitpicky here, so let’s what a hot mess  

Sarah    00:50:03    <laugh>. Yeah, exactly. <laugh>, I’ll clean up the hot mess a little bit. I’ll make a move to clean up the hot mess move. Yes. Um, I say, um, I’ll like the, the constructivist part of the constructivist claim is just that, is this idea that memories are constructed at the time of retrieval, right? Or that there’s something, part of it is this rejection of an overly simplistic idea of retrieval is just the act of, you know, going and fetching the memory and, you know, re hitting play or turning the light on. Um, this idea that, you know, lots of things can have an effect on how it is you remember a past event. So information you’ve learned since then, but also, you know, the way I ask you questions, the mood that you’re in, probably your blood sugar, lots of things are going to affect, you know, how it is you actually represent or generate, um, and produce that memory.  

Sarah    00:50:56    So it is constructive in that sense. It’s not a just, you know, sort of discreet extraction from, from a memory store. And I, I’m fully in support of that. So it’s wrong to think that memories are just, you know, that retrieval is kind of like a neutral window. I think it’s actually shafter as someone who articulates the view that way. Um, that through which you can just directly probe memory. Um, but some people take the constructivist claim to be, um, that, to mean that there’s nothing that, that it’s entirely a process of constructing from, of course, from what steward information, but  

Paul    00:51:34    <laugh>, I mean, it has to come from somewhere.  

Sarah    00:51:36    Yeah. Yeah. So that’s why I think that that’s part of what is the basis for having this idea about traces is being a particular way of thinking about what information retention is, because of course, more radical constructivist don’t deny that you retain information from your past experiences, but they think that you do it in ways that are highly generalized or schematic or, you know, they’re based on sort of overall patterns of information, uh, and the kinds of things that you could learn in a more generalized way Hmm. Rather than keeping track of particular pieces of information. Um, so that’s the kind of, that’s the sort of difference. So they might think like you have a kind of broad network of, you know, like every experience I have in the world adds to my way of understanding the world, but it does so in this way that is, you know, maybe I’ve got a lot of <laugh>, I’ve got a lot of Bayesian things set up.  

Sarah    00:52:26    So I’m, I’m keeping track of what’s likely and what’s probable across the range of experiences that I have. And so if I have an experience that violates my expectations, that’s gonna change how I think about what the likelihoods are of the different features that were there co-occurring and, and different kinds of ways. Right? And so you are gonna store information, but just in ways that serve these general updating processes. So then having a memory is just gonna be a matter, or, you know, remembering something is just gonna be applying those sorts of predictive processes backwards. Mm-hmm. Um, and so I, I build something, I mean, not out of nothing, but not on the basis of some stored information about that event.  

Paul    00:53:07    I, is it, uh, accepted these days, um, that when you re retrieve a memory, um, by the act of retrieving it, whether you’re constructing it a new or from, you know, some, some register that, uh, it necessarily changes the reconstruction account of memory?  

Sarah    00:53:26    I think so I think that aspect of retrieval is largely accepted. I mean, there’s still some views and philosophy that try for some, at least, you know, they, they want some ideas of like, in really pure conditions, <laugh>, you could have, you know, kind of pure remembering. Um, so, so I think you might still get some people who want those kinds of things. Um, but I think as a broader, people are much more sensitive to that. The, the ways in which like subsequent information and experience and like what’s happening with you now changes the, the nature of that memory. Yeah.  

Paul    00:54:08    You don’t think that memory is for the future? Um, I, I kind of do. I I think that memory, you know, talking about the function of memory, what is it for mm-hmm. <affirmative>, how could it not be about your ongoing and future behavior?  

Sarah    00:54:24    Yeah. I mean, I’m probably gonna live to regret letting them punch up the title of that in <laugh>.  

Paul    00:54:30    Yeah. What is the title? Memory is Not For the Future, I believe is the actual title,  

Sarah    00:54:33    Something like that. I think it’s, yeah. Yeah. So it’s, you know, philosophers need at least seven people to read the things that they write down. So that’s one way to maybe, maybe I’ll get eight out of that. Um, no, so in some sense I agreed too. Um, but what I, what I haven’t agreed with is the way that that kind of claim is built into how people think about how that claim is often understood, um, for thinking about this sort of constructive approach, right? So I do think that memory is for the future and that, like the organism is only keeping stuff around in an individual case, but also, um, if we’re thinking about it evolutionary evolutionarily in terms of like things that are either good for it or at least not too costly <laugh> such that they die out, right? So at some point it has to pass that kind of threshold and that has a forward looking feature to it.  

Sarah    00:55:26    Um, so memory had to get by in that sort of, it had to get by in that framework in one way or another. The part of it that I don’t, the claim that episodic memory is for the future from some people like Shater and colleagues who have this sort of strong link between memory and imagination and constructivism view, is that they think that that means that if it’s for the future, then it’s not gonna retain information from particular past events. Right? It, that’s not what the function, it’s the system isn’t doing that. Instead the system is keeping track of lots of information in these more generalized and schematic ways, and then performing lots of simulations about what I, you know, about what’s happening now, what’s happening, what could happen tomorrow and when needed about what happened yesterday. Right? Um, and the sort of thought is like, because it’s for the future, it doesn’t keep track of particular events. Um, and I think, so I I’m perfectly happy with saying it’s for the future if that has an influence on how we keep track of information from past events. But that’s not, that doesn’t preclude saying that there are traces. Um, in fact, I think your simulations are only gonna work so well unless you actually do have something like that built into the view at some point. Hmm. Um,  

Paul    00:56:45    If you had, maybe this is not a coherent question, but, uh, if you had to give up one personally, would you rather give up your memory or your imagination, you know, since they’re not the same thing?  

Sarah    00:56:55    Yeah, it’s a good, um, <laugh>, I would definitely rather give up my imagination. I think mine is rather impoverished to begin with, probably. Um, I think <laugh>, I’m one of those, like, I was always the, like, I can’t abide a time travel movie. I’m one of those people that like, I’m noting the inconsistencies, science fiction sometimes. Like he can, you know, he’s not falling through the floor, but he can walk through a wall. Like someone explained it, like I’ve,  

Paul    00:57:18    Oh, sci-fi movies are worse. Yeah. <laugh>. They are, they are. I can’t, I can’t watch ’em either. Like, and I can’t enjoy them. Yeah,  

Sarah    00:57:26    Yeah, yeah. Exactly. I get so hung up on those things. I don’t get lost in the stories and the way I’m supposed to. So imagination in that sense. I think I’m already <laugh> already deficient in. So, um, yeah,  

Paul    00:57:39    <laugh>, that’s not gonna serve you well as a philosopher. Don’t you need a good imagination as a philosopher?  

Sarah    00:57:44    <laugh>? Oh. So I mean, it, I think philosophy really takes lots of kinds of people. You’ve gotta have the people with the imaginative big views and the other people that sit around and poke holes. And I think I’m, I’m, uh, more of the, that <laugh> that philosopher saying like, does it work? I don’t see how it works. Mm-hmm. <affirmative>. Yeah.  

Paul    00:58:02    Interesting. So here’s another, uh, kind of random question. So we’ve talked about errors in memory and there are different kinds of, you know, errors, but could misremembering things, uh, somehow be, um, an important feature of our intelligence or advantageous? Like what, is there anything advantageous about misremembering?  

Sarah    00:58:24    I think there’s a lot that’s advantageous about misremembering, um, in part because you see as a lot of the studies of it have shown, like with other kinds of memory deficits, um, things like, um, all size Alzheimer’s and other forms of dementia, you often lose those errors. Um, you like lose the sensitivity to some of the, like the D r dm, fx, um, the like list the cases of things where, you know, if I give you a list of related words, um, so like, you know, pillow, blanket, alarm, tired, melatonin, I don’t know, <laugh>, um, and then, you know, ask you later about things that were on that list, you’d be much more likely to think that sleep was on that list mm-hmm. <affirmative>, right? Those kinds of effects, those, um, diminish, uh, with me with like memory disorders. So it looks like there is some correlation between being susceptible to those kinds of things and having a memory that in some sense is working as it should.  

Sarah    00:59:19    Right? I mean, and, and it makes sense. I mean, it’s, if if your, if your goal in life is to perform well in psychology experiments of list learning <laugh>, then you’re not doing very well. Right? But if your goal is to, you know, remember, like, you know, the, of the list of things on your grocery list, like, were you trying to make soup or were you trying to bake a cake? Like that’s gonna help you in general. It’s better to come home with like two additional soup ingredients if you think you’re making soup, even if it’s not the exact ones you can probably make do, um, right. Like, that serves you well. Um, you know, it’s good if you’re talking to me about something that I figure out the topic of what you’re talking about and think about related things. That’s a, that’s a good thing, right?  

Sarah    01:00:02    I draw connections between my experiences. That’s a good thing. Um, so I do think those effects, um, I mean, I think it, so Roddy Rodger in a paper a long time ago, like talked about them in the same way that we talk about perceptual illusions and the kind of like funhouse sort of things. Like they’re, those experiments are really good for showing us, I think more how our memory works than how it doesn’t work, right? In the sense of like, you don’t go to the eye doctor after you’ve been through the house of mirrors because you think that something’s wrong with your vision. You’ve just been in a situation that shows you Right. That exploits those features. And if you think those experiments are well designed to, you know, test the kind of boundaries of your memory capacities, that’s really what they’re doing.  

Paul    01:00:45    Yeah. I wonder, one of the reasons why I ask is because, for instance, a computer, uh, has quote unquote perfect memory because it’s just, you know, a physical register, right? Yeah. That you can then retrieve and it doesn’t make mistakes. And, you know, often we, on this podcast, we talk about features of, you know, human intelligence and or brains, um, that may or may not be important for building artificial intelligence. And, and I don’t know, I was gonna ask you this earlier too. I don’t know if the concept of computer memory has sort of affected the way that we think of our own memory and, and, and we, we think of in terms of, of that permanent sort of computer storage and that’s how we should think about our own memories. Uh, do you think that that’s the case?  

Sarah    01:01:26    Y Yeah, I do. I mean, I think, I think that it’s gone in a certain way, and maybe this is not entirely, uh, sometimes I feel like I’m a bit of a Luddite about certain computer technology metaphors for, for memory and cognition. But I mean, I think a lot of the early computer designs are very much based on how we thought memory worked, right? And at the time, you know, it was a sort of sensible way of thinking about memory of having, you know, like taking everything, giving it an address, <laugh> storing it, and then calling it up by that address, and then here we go. Um, and so it might have been built by what we thought was a very sensible way of thinking about it. And that kind of view of memory in some respect or another is sort of in the background of a lot of, you know, philosophical and theoretical works that people wrote at those times.  

Sarah    01:02:14    And it’s only, and then of course, that as a technology and theory, the, the computer memory stuff has taken off on it’s right. Those things have taken off for their own, that’s turned out to be incredibly useful, I think useful in part because of how different it is from how we store information or like, we put all the things there, um, that it does so much better than that. Those things do so much better than we do. And since then we’ve learned a lot about human memory that is really hard to square with that model. Um, so, you know, I mean, for instance, the list effects that I was talking about a minute ago, the d the d r m deser or McDermott kind of paradigm where you think you remember the word sleep when you hear the list of related words, like that’s really hard to understand in the kind of computer model, right?  

Sarah    01:03:02    Sense, right? Like, if I manage to get to the address <laugh>, uh, if I manage to get to the memory of that list and extract it, which the, you know, like, it’d be strange that I get wrong what’s there in the ways that I do. Um, so those kinds of results are hard to fit into that model. Um, and so it seems like there, that’s one, I mean there’s I think other dis analogies too for, from that perspective, but, you know, parts of that still seem, I mean, there’s still parts of those models that seem quite right and that as a Tracy person, it’s hard not to give up on, like, there is kind of built into this, this idea of like, you keep track of things, you know, particular pieces of information and you put them in places and like, it’s hard for me not to think at, you know, that the Opto and Graham stuff is showing, like some forms of that are happening at some level for some pieces of, you know, some experiences and information. Um, but the overall organization I think has gotta be, has gotta be at least a little bit different.  

Paul    01:04:03    Did you just call yourself a Tracy person?  

Sarah    01:04:06    A <laugh>? I think I did, yes. Okay, okay. Someone said to me, not that long <laugh>, someone called it Team Trace not that long ago, and I was like, oh, I like that I could make shirts. Um, so, uh, yeah, so maybe that’s better than Tracy <laugh>,  

Paul    01:04:19    But, so, you know, it just kind of perseverating on on this idea a little bit. Would we want to build human-like memory in robots? I mean, don’t we want perfect memory artificial agents? Yeah.  

Sarah    01:04:35    So you, I think you would want them to be, you’re either gonna want, what is what we want out of a lot of our computing systems is that we want them to store a massive piece of information. When I ask for the thing that’s at address 1 0 4, I mean, the thing that’s at address 1 0 4, I don’t mean the thing that’s, you know, like, I’m gonna be very particular about what I want and it’s gonna allow me to store massive amounts of stuff. Um, so, so in that sense, yeah, we don’t want those capacities. Um, yeah, so I think it’s, it does raise interesting questions about like whether there is a way for something to have the system we do and do it in like parti. So could you have a good version? So for things we do have like episodic memory, like what would be the point or function of doing that?  

Sarah    01:05:20    Um, so I have spent a little bit of time collaborating with a couple of computer scientists who’ve thought about that a little, or like, there are plenty of people who’ve thought about it, um, even more. But it’s interesting to me to think about like, why would you, what’s the, what could the goal be of that? Um, it seems like the only thing I’ve heard that seems like it could be close to being the right kind of thing is that if you were in situations where you were trying to build kind of robot or drone models of things where you’re sending them to places that humans aren’t gonna go and you want them to report back to you, um, that you might want them to have something like those capacities. Um,  

Paul    01:05:57    Why, what, what, what  

Sarah    01:05:58    Even,  

Paul    01:05:59    Yeah. What would be the advantage of getting misinformation from your robot on Pluto?  

Sarah    01:06:04    Good. So you’d want it to have a working version of that. You’d want it to get that kind, you know, you’d want it to do episodic memory, but do like the ideal human version <laugh> that doesn’t change or distort over time. Um, and then that’s a question of like, is that really human memory you’re trying to build then? Or are you trying to build something like, you know, a one off that’s actually, you know, an upgrade?  

Paul    01:06:24    Is there any merit to the idea that, um, that is perhaps a way that we are creative by making mistakes and then analyzing the mistakes and somehow, you know, when you join two things that are not alike or whatever, that’s one form of creativity, but is is creativity related, do you think, at all? To misremembering and I do and memory errors,  

Sarah    01:06:48    I do think it’s related. And I think, um, and I think that’s one of the early directions that shater and some of the people that now think about membrane imagination is, is being related. We’re trying to go, so there are a couple of papers, I don’t know, there’s a reason that there’s not more, but there are a few papers that show that on certain me ways of measuring creativity, you get increased susceptibility to like the D R M effect and other of these false memory effects. Those things are very well correlated to one another. Um, and yes, I mean, it wouldn’t surprise me if there is some kind of relationship there that you’re kind of, you’re sort of putting together lots of different things, you know, if that’s kind of what you’re thinking about, that kind of does require you to sort of take apart <laugh> whatever memories you have of particular things or be willing to take them apart in certain ways.  

Paul    01:07:35    Is that what confabulation is also, that you’re, uh, my mom is the worst confabulator, by the way. And I don’t, I’m really worried about myself as I age, like, or how much I’m confabulating, you know, without knowledge of it. <laugh>  

Sarah    01:07:49    Yeah, it’s, it’s scary. I mean, it’s scary how those things work. I mean, I think that’s the, the way we sell ourselves on this is by saying like, actually it’s just side of how creative I am. Yeah. This is just a terrible side effect. Um, but yeah, I think there’s, I mean, there are, and I also think that there are probably individual differences across those kinds of things, and it would maybe make sense that there would be, um, in terms of, you know, how much those things happen. So for confabulation it’s so tricky. So I think it can happen in a couple of different ways. I mean, the kinds of ways that confabulation in the more clinical sense where you have this happening in, in broader cases of mental disorder or dysfunction, right? Like in schizophrenia or something like that mm-hmm. <affirmative>, those might be slightly different.  

Sarah    01:08:34    Kind of like, you know, there might be ways that you’re not doing a good job of Yeah. Thinking about how you generate <laugh> plausible representations of what you did the other day and, and weed them out. Um, but for other people, some of it is sort of could be source monitoring kinds of effects. I mean, that’s how you get some of the confab, like the Elizabeth Loftus style studies, right? I, you ask people lots of questions about some ask them to imagine something, and then over time, right? Mm-hmm. <affirmative>, you think about this event later and you’ve thought about it so much that it, it now feels, it has the feeling of a real event. And so you sort of misinterpret it as such. And I think, you know, sometimes people tell you stories. I have one of these confabulation things of my brother and I, when we were little, we were on a, like a vacation, like boogie boarding in California.  

Sarah    01:09:19    And I have this memory that I used to tell of like getting pulled out on a day when there were the sort of undercut, you know, the undercurrent and um, like getting sucked out and people having to come get me. And it being very, very scary. And it turns out it happened to my brother and took like years. I mean, we were, we were little, probably like six and eight at the time or something, maybe slightly younger. And so for years I thought it happened to me and it wasn’t until, you know, it’s not an event my family really talks about for, you know, for any reason. And I had mentioned it and everyone was like, that was not you <laugh>,  

Paul    01:09:49    But did you see it? Did you see it happen to your brother or  

Sarah    01:09:51    Yes, I saw it. I was, I saw it, I was right there. I was so worried, especially as the older sibling for not helping. And I think, I mean my mem my memory theorist justification for that is like, I took it so seriously and I probably freaked out and thought, what if that happened to me? Right? And over time, what stuck around was like a very first person image of what it was like for that to happen. Um, and it wasn’t until I got to the point of like, well, do you remember people pulling you out? I’m like, well, no, I don’t remember that part at all. Like, I have no memory of that. I just have this like, yeah. Um, parts of it seemed very vivid cuz they’re ones that I thought about  

Paul    01:10:27    And now you can’t, uh, not remember it that way probably. Right. You just can you kind of <laugh>, uh, I was gonna say imagine, but the maybe, you know, you simulate the, the thing that’s in your mind and then you realize that it’s not real. Is that the, is that the  

Sarah    01:10:41    Phenomenon? Exactly. Yeah. So it’s like, it still has all the same stuff to it, but I just, like, I don’t endorse it anymore in the same way. Oh yeah.  

Paul    01:10:49    I, my mom like will to my f like, I’ll be in the room. This happened so many times, I can’t, I can’t count the number and she’ll make up a story about something about me. And it’s, it is patently false and she just does it like, as if it’s very real and without a hitch in her voice, I, it drives me nuts. I love you mom, but it drives me nuts. You know, <laugh>,  

Sarah    01:11:12    That’s, is it you specific or is this, um, a more, it’s everybody to other, it’s  

Paul    01:11:17    Everybody’s. I’m not, I’m not special, but, you know, so <laugh>, I just, but I don’t know how to, there’s nothing I can do about it except I, I stopped calling her out on it because it’s like she needs it or something. So I don’t know.  

Sarah    01:11:29    Yeah, I think, I mean, there are, there are people in philosophy who think that a lot of these kinds of processes, they serve some other role, right? You know, that even if you don’t think, if your aim is, you know, saying true things about the world, this clearly doesn’t work, but maybe it is. Like they’re the stories you want to hear, they’re entertaining you, they’re making sense of things. Um, there could be lots of, lots of reasons for them that are perfectly, you know, psychologically healthy, at least for the person who’s engaged in them, even if they’re not, even if they’re frustrating. Um,  

Paul    01:12:01    We, we all have our own truth to tell or something is is that the Yeah, that’s a very modern,  

Sarah    01:12:05    Yeah. Something like that. Yeah. Yeah.  

Paul    01:12:07    <laugh> who, who has a better grasp on, on memory, uh, philosophers, neuroscientists, psychologists, who’s winning this, this race?  

Sarah    01:12:18    Oh, that’s a, um, I don’t think, I mean, I’d like to think that it should be some kind of joint enterprise of some sort. Um, who’s winning? I don’t, I mean, I, I’m not ready to call sides in some I like, I don’t know. And I don’t know that anyone could or should in those kinds of ways. I’m glad to have philosophers in the conversation now in ways that we weren’t for a while and that like some of the questions about what counts as a memory or doesn’t count, or where we wanna draw the boundaries. Um, I think philosophers can be helpful. I mean, maybe even just, especially now, not because philosophy is anything special, but because science, the science is so specialized, right? That you’re not just working on memory, but you’re working on episodic memory and not just episodic memory, but episodic memory in, you know, corvids and corvids on this set of, you know, with this set of measures.  

Sarah    01:13:13    And so you can read incredibly widely and still not at all come anywhere near the primate memory literature or the literature on memory in drosophila or the memory literature for trauma victims, you know, like, you know, and dementia patients, right? There’s, it’s massive. And so philosophers by not having to sustain a research program, get to kind of look around and talk about similarities and differences and try to put things into conversation. Um, and so I think, yeah, I don’t know that anyone’s gonna win that. I think kind of recognizing that <laugh> recognizing that it’s bigger than all of us is probably a good first step, especially for, I mean, philosophers need that too, right? And sort of thinking like, we don’t get to pronounce how things are without actually looking at what the data tell us memory is like. Yeah.  

Paul    01:14:03    Do you see the, um, <laugh>, okay, so my, actually, before we started recording, I was just asking you like whether I should, how, how I would proceed if I wanted to get into like a philosophy department or something. But, um, and so what I’m about to say may may rub you the wrong way, I don’t know. But the way that I see philosophy often is sort of, um, so you’re poking, you’re the type that pokes holes, right? And then you poke one hole and that opens up multiple other places where you can poke holes and mm-hmm. <affirmative>. So instead of solving anything, it seems like the questions become more detailed, more nitpicky and, and, and, uh, what, what you want in science and philosophy is to solve problems, right? But it seems like often philosophy opens up more problems and I’m wondering if, if how you view that trajectory specifically in the philosophy of memory.  

Sarah    01:14:49    Yeah, it’s a great question. And so I think, I think some, I think there are ways of curtailing parts of that question asking process, um, that, you know, like you can recognize it going in directions that sort of transcend what we’re able to measure just are limitations we have to take seriously. Mm-hmm. Um, and that’s a lesson I’ve always taken from doing work alongside scientists, um, at different points. It’s like, yeah, sometimes you just have to say like, look, this is, we want a good definition, you know, we wanna, well define our measures, we wanna, well define our criteria, but like once we’ve done that, we just, we’re gonna stick to it because that’s the only way to actually sustain an investigation. And maybe at the end of that, like, you know, it’d be nice if science weren’t structured such that like you end up kind of with a lot of sunk costs into like, well these are the tools we know <laugh>, right?  

Sarah    01:15:37    And so yeah, the retraining or re-exploring if that were more rewarded I think would be helpful for everyone. Um, but, you know, it makes sense to stick with the program for at least long enough to see it through, right. At least long enough to see like, what happens if this is the way we conceive of that thing. Hmm. Um, so I think Phil, I think philosophers, it’s good for philosophers who do this kind of work to spend enough time with scientists to recognize the different stages and ways in which the poking holes is helpful and in times at which that actually like is just gonna get in the way and you need to like, you know, maybe phrase that in a different way or wait until we’ve actually seen what happens from defining it in this way, um, as a way of going forward. So I, I do think that that’s, um, yeah, that can be philosophy needs to have that kind of, if it, if it, if the goal really is to answer the question. I mean, think for some philosophers the questions are their own right. Are their own reward, right? Um, right.  

Paul    01:16:32    There’s merit to that though too. I mean, I, I think that’s part of the fun right of philosophy is, is coming up with a good question.  

Sarah    01:16:40    Mm-hmm.  

Paul    01:16:40    <affirmative>, but, um, so, so, so then what’s your outlook on, uh, the philosophy of of, of memory and, you know, imagination? Um mm-hmm. <affirmative> are, are, you’re not gonna be out of a career anytime soon, for  

Sarah    01:16:54    Instance. No <laugh>. Yeah. No, I think we’re, I think this is um, is really like, it’s, it’s in some ways very early days. I think the number of topics that philosophers have picked up on in the kind of memory literature and of course there are debates and views. Um, it’s also, it is just very limited relative to the range of phenomena that are out there to talk about. So there’s gonna be plenty of that. Yeah. Um, and for the stuff that I’ve been interested in about traces, because especially because of all of the optogenetics work on the engram and such like that, I mean, it’s just, for me it’s an exciting time cuz you have a set of scientists who are excited about what these new tools are letting them do in a way that is opening them up not to quite, not quite to poking holes, but to asking questions in a way that like, at certain, you know, in, in certain places and stages of research projects, that’s not where you’re at. And so a philosopher is gonna be in your way mm-hmm. <affirmative> and right now somebody who wants to say like, but what is the memory trace? Like, that’s not getting me thrown out of conversations in the way it might have a decade ago,  

Paul    01:17:53    <laugh>. Oh, okay.  

Sarah    01:17:54    Good, good.  

Paul    01:17:56    Um, so, so like in inter interactions with neuroscientists and psychologists, I mean, is it a pretty friendly, uh, interaction or, or is there skepticism from their side or, you know, how do, how do you feel like you fit in there?  

Sarah    01:18:08    Yeah, I, it’s tricky. I mean, I think part of it is making the right kinds of space. So, you know, these are the kinds of things that, um, philosophers have a different kind of time in our hands and a different kind of like, the sort of space we have for these kinds of conversations are hard to find, especially because, you know, the kind of professional spaces where you would interact and have a conversation with someone about their work, there’s not a lot of those. Yeah. Um, there’ve been some, and they are usually really welcome in part cuz it takes a while to align terminology and align ways of asking questions and, and doing that sort of work. Uh, my experience has generally been that once you, if you build the kind of environment that allows people, kind of gives them the space to have those conversations, they can be incredibly productive. Mm-hmm. Um, but it’s, it takes a lot of doing to make that the case. Mm-hmm. Um, just cuz yeah, the, the the background environments are in kind of practices are so different. Yeah.  

Paul    01:19:08    Sarah, what, what, what is the, uh, title? What’s the working title of the book? Is it Memory Traces A Hot Mess?  

Sarah    01:19:15    <laugh>? No, it should be, although now like, I don’t think they’d let me get away with putting Hot Mess on the, on the cover. Um, it’s <laugh>, although now it now it sounds like a dare. And I’m, I’m interested, um, memories Trace that is the, um,  

Paul    01:19:28    Memories Trace. Oh, that’s nice. Um, yeah, well continued, uh, success and keep up the good work and let me know when, uh, when the book is ready to come out, send it to send it along and uh, we’ll have you back on if you’d to come  

Sarah    01:19:39    Back, back on. I would love that. Yeah, that would be fantastic. So thanks for this, this has been great. Um, and it’s a great resource in general, so this is fantastic.  

Paul    01:19:48    Oh, thank you. I alone produce Brain inspired. If you value this podcast, consider supporting it through Patreon to access full versions of all the episodes and to join our Discord community. Or if you wanna learn more about the intersection of neuroscience and ai, consider signing up for my online course, neuro Ai, the quest to explain intelligence. Go to Brain inspired.co. To learn more, to get in touch with me, email Paul Brennan inspired.co. You’re hearing music by the new year. Find them@thenewyear.net. Thank you. Thank you for your support. See you next time.