We cannot rethink the ethical and political dimensions of memory—centrally, its role in constituting personal and group identities—without at the same time rethinking the nature of memory. In what follows, I reconstruct from Sue Campbell's writings (especially Reference CampbellCampbell 2003; Reference Campbell, Koggel and Jacobsen2014b) a defense of the shift from a traditional view of memory (as a process for encoding, storing, and retrieving information) to a view of memory that is better suited for the business of constituting selves and identities in the contexts of our relationships with others—a shift from an epistemological to a political conception of memory.

The idea that philosophical attention to memory needs to turn away from a longstanding preoccupation with epistemology in order to examine the roles of memory in ethics and politics is predicated on a view of epistemology that has not already been sensitized to its own moral and political dimensions. But the epistemology of memory has been largely resistant to the changes (led by feminist and virtue epistemologies) that are conspicuous elsewhere—it is still very much a traditional epistemology that is, in Miranda Fricker's words, “impoverished by the lack of any theoretical framework conducive to revealing the ethical and political aspects of our epistemic conduct” (Reference FrickerFricker 2007, 2). Campbell's relational account of memory is designed to provide just such a framework.

In describing her account of memory in early sections, I will take the liberty of recasting matters in a way that I believe enables us to find the most palatable form for her dramatic anti‐individualism; in a later section I attempt to say what form that anti‐individualism should take. Finally, I will consider an objection to a relational theory of memory—indeed, to relational theories of cognition generally—and sketch what I think is the best strategy for responding to it.

I. Memory: A Traditional View

In the essay that now serves as the author's introduction to Our Faithfulness to the Past, Campbell describes the uncomfortable and largely mute response to a paper she presented at an epistemology workshop in 2006. At the end of her presentation one philosopher approached her and said, “the way you talk about memory makes epistemologists nervous” (Reference CampbellCampbell 2008/Reference Campbell, Koggel and Jacobsen2014b, 1). Footnote ¹ Campbell speculates about why that might be so—why what she has to say about memory should make anyone nervous. She suggests two hypotheses, both tied to the relational theory she first develops and defends in Relational Remembering. There, she describes the idea of a relational theory of memory:

In a relational account of it, remembering is a socially embedded and socially supported activity—both what and how we remember is continually readjusted in light of our changing needs and interests, which in turn are continually readjusted in light of the responses others have to our recollections, reminiscences, testimonies, confessions, and other narratives of the past. To this extent, relational memory is in line with the reconstructivist views to which cognitive and developmental psychologists were led in the 1980s and 90s, and which became the source of much skeptical anxiety about memory—or, as it turned out, much skeptical anxiety about women's memories.

An earlier archival model of how memory works supposed that the content of memory is determined solely by an original past experience, which is encoded, stored safe from corruption, and then retrieved as needed for cognitive consumption. On the archival model, successful remembering aims to maximize reproductive fidelity to the original experience. But the reconstructive view emphasizes that memory is highly selective, and that influences operate at every stage of the memory process. Prior experiences influence what is selected for encoding, and how it is encoded; continuing experiences alter the data in storage; context and interlocutors influence what is selected for retrieval and how it is remembered (Reference CampbellCampbell 2003, 14–15). In short, both what and how we remember are constructions or reconstructions of the past, made in light of our current needs and interests, and under the influence of those with whom we share our memories.

Where skeptics saw the reconstructive nature of memory—especially the influence of others on it—as evidence that memory is distorted and unreliable, Campbell sees its essentially relational character at work. The idea that the influence of others is invariably contaminating requires the extraordinary assumption that continued reflection on and consultation about the facts and their significance only distorts them. In Relational Remembering, Campbell explored in detail how that initially implausible assumption was made to seem obvious to so many—to memory scientists, to legal theorists, and to epistemologists—and how it was made to seem obvious in large part by pathologizing women, assigning them “overly‐suggestible” and unreliable natures (Reference CampbellCampbell 2003, 3). So although she agrees with the skeptical reconstructivists that its very nature makes memory highly selective, always perspectival, and malleable to influence, Campbell refuses to see these facts about memory as an occasion for alarm. This is surely one reason her views about memory might cause some epistemologists anxiety.

A second reason might be the impact of Campbell's relational perspective on the psychological individualism already entrenched in many epistemologists’ conception of knowers. She confesses that, in this respect, her work has often led her into territory that also makes her nervous. Campbell's “own disciplinary heritage,” she tells us, “[h]as been fiercely individualist about the mind”… and “the individualism of [her own philosophical] starting place causes [her] discomfort” (Reference CampbellCampbell 2008/Reference Campbell, Koggel and Jacobsen2014b, 7). Campbell's emerging anti‐individualism had always been evident in her work, beginning with her first book, Interpreting the Personal, in which she developed a deeply relational theory of the emotions (Reference CampbellCampbell 1997). But it is in her work on memory that some of the most unsettling implications become evident. In what follows I will expand on Campbell's suggestions for why her theory of memory might make epistemologists nervous. I emphasize the second source of philosophical anxiety—what she refers to as her anti‐individualism—since it lies at the very heart of her relational theory.

II. Memory's Complexity

According to a view we find throughout the epistemological literature in the final decades of the twentieth century, and in much work available today, remembering is a mental process or activity that, if successful, results in true beliefs about the past. Those beliefs are, when appropriately caused, memories. Further elaborations of this traditional view talk about the various stages of the process by which reliably true past‐tense beliefs are made available to us—including the initial encoding, reproduction, or representation of experiences, their storage, and their eventual retrieval. But such details do not change the basic format for traditional accounts, which are recognizable by three features. First, memory is characterized (often exclusively) in terms of its role in our epistemological enterprises—it is a faculty or resource upon which we draw in the conduct of inquiry. In effect, memory is portrayed as if it were a faculty or organ for knowledge of the past. Second, and as a corollary of the first feature, for memory “the primary cognitive value is truth”Footnote ²—to be faithful to the past is to have true beliefs about it. Third, remembering is represented as a mental process that leads to the production of a mental state, where the adjective “mental” is presumed to characterize processes and states located in the heads of inquirers.Footnote ³

Campbell's work asks us to turn away from this simple picture and begin our reflections on the nature and epistemology of memory by first acknowledging the tremendous complexity of memory, where that means acknowledging the considerable variety of phenomena we normally talk about under the heading of “memory.” Consider just a small sample of this variety.

There is, of course, the idea that memory is a faculty, the possession of which is essential to keeping track of and learning from the past. Then there are the many memories themselves—the particular representations or traces of the past that recall makes available to us for cognitive deployment. Among those memories we find these:

1. (i) Propositional (factual) memories: Memories of particular past facts, such as my knowledge that I was in Nevada last month. In the epistemological literature there has been a strong tendency to treat all memories as propositional.

2. (ii) Imagistic memories: Images from our past that serve as memories, as when I (vividly) recall the look on a friend's face when she learned of her mother's death.

3. (iii) Skill memories: We remember how to do certain things: to ride a bicycle, to factor polynomials, or to order drinks in Swedish.

4. (iv) Personal (autobiographical) memories: Memories whose expression takes the form of a sustained autobiographical narrative, such as remembering what it was like to go fishing on the Yellowstone River, to be raised by an alcoholic mother, or to grow up in rural Ontario.Footnote ⁴

A fact, a face, a number, a Swedish phrase, a poem, an argument, a skill, or how an entire childhood felt—these are all among the things that can be remembered.

Finally, the processes or activities by which we succeed in remembering facts, faces, poems, or occasions are equally various—from sudden recall, to studious reconstruction, automatically reciting, reminiscing together, testifying, confessing, and being reminded by others. We can be brought to remember something by looking through a photo album, by playing a piece of music that evokes a past occasion, by celebrating an anniversary, or by noticing the string tied to our finger. There are almost no limits to the activities that can be brought to bear in helping us to remember the past.

Campbell asks us to begin our reflections on the nature of memory by contrasting this rich array of phenomena, both of things that can be remembered and of the processes by which we succeed in remembering, with the traditional epistemologist's impoverished thought that remembering is a process in the head leading, when successful, to true beliefs about the past.

We can begin to understand Campbell's perspective on memory, and also impose some order on all the variety of memory phenomena, by deploying an old Aristotelian distinction between activities and achievements.Footnote ⁵ Just as “finding” indicates the successful completion of a search, or “arriving” the successful outcome of travel, so “remembering” typically indicates an achievement—remembering, like finding or arriving, is the successful outcome or completion of a process or activity.Footnote ⁶ Consider the case of searching and finding. There are no natural limits to the sorts of activities that might qualify as searching: a search can be brief or protracted, it can be done entirely in the head (as when I run over in thought the places I might have left the keys), it can consist of opening and closing drawers, or mounting a vast expedition to a remote continent. Furthermore, there are no natural limits to the kinds of things that can be found—they could be missing keys, the treasure, the solution to the puzzle, or time for contemplation. The crucial point is that there is no one activity or set of activities in which searching must consist and no one thing or kind of thing the finding of which constitutes the successful completion of a search.

Similarly, the achievement verb “to remember” labels the successful conclusion of a wide array of dissimilar activities. Following Campbell, let us call any of the diverse activities by which we manage the achievement of remembering simply “memory activities” (Reference Campbell, DesAutels and Urban WalkerCampbell 2004/Reference Campbell, Koggel and Jacobsen2014b, 29). Memory activity is the analogue for the activity of searching, and recalling something is the analogue for the achievement of finding. On this model the processes and activities that result in our remembering something do not form a natural kind. There is no one kind of activity or process for memory processes to be. Similarly, there is no one state—no one result—that constitutes the successful outcome or completion of our memory activities. On this way of reframing our issues, memory activities do the work done in the traditional theory by mental processes, and the achievement of remembering replaces the resulting mental products (memories conceived as traces or representations of the past). My proposal is simply that the variety of phenomena that come under the headings of memory and remembering are better accommodated by the act/achievement distinction than by the traditional distinction between mental processes and the mental products they leave as residue. Furthermore, the act/achievement distinction better captures the relationship between successfully remembering and the various activities or processes that are subordinated to that achievement.

If remembering something is analogous to finding something, and it indicates the successful completion or outcome of memory activities, then it makes sense for us to ask for the norm (or norms) in light of which the completion of memory activities will count as successful. In what does successfully remembering—that is, being faithful to the past—consist? The traditional epistemologist has a direct answer to our question: remembering consists in acquiring true beliefs about the past; memory processes and activities thus answer to the norm of truth. As we will see next, Campbell doubts that truth can serve as the principal norm for good remembering.

III. Truth and Accuracy

Campbell connects the norms in light of which memory is valued to the complexity of memory phenomenon.Footnote ⁷ She argues that if we keep sight of the immense variety among the things that can be remembered, and the variety of processes and activities involved in remembering, then we will see that truth cannot be the primary cognitive value in memory.

Before developing this idea, it is important to emphasize that at no point does Campbell doubt or downplay the importance of truth as a value for inquiry. In Relational Remembering, she argued that “we must close ranks around the idea that truth matters” (Reference CampbellCampbell 2003, 92). In Our Faithfulness to the Past, she argues that devaluing the truth about the past is inconsistent with integrity—which is itself “a central virtue of agency” (Reference Campbell, Koggel and JacobsenCampbell 2014b, 64). She thus follows Margaret Urban Walker, Michael Lynch, and others in the idea that “our shared moral understandings of integrity require our commitment to truth” (Reference Campbell, Haaken and ReaveyCampbell 2010/Reference Campbell, Koggel and Jacobsen2014b, 64; see Reference WalkerWalker 1998; Reference LynchLynch 2005).Footnote ⁸

Truth is nonetheless the wrong norm for explicating our faithfulness to the past; for this purpose we need the norm that Campbell calls accuracy. Truth and accuracy are distinct norms, since we can satisfy either without satisfying the other. The argument for this has two branches. First, recall that not all memories are even apt for being true or false. Remembering a person's surprised look might amount to having an image of her face, and though an image may succeed or fail at being faithful, images are neither true nor false. So the scope of accuracy will need to be considerably wider than that of truth. As Campbell writes: “Accuracy is… applicable both to kinds of representations that are truth apt and to those that are not. Expression of memory may fall into either category” (Reference CampbellCampbell 2006/Reference Campbell, Koggel and Jacobsen2014b, 36). Jonathan Dancy illustrates how this move is resisted when framing memory primarily as an epistemological resource. He recognizes an important category of memories (which he calls “perceptual memories”) consisting of images of past things, occasions, or events, but he explains the distinction between such memories and propositional memories by suggesting that to have a memory image is to have “a distinctive form of belief about one's past” (Reference DancyDancy 1985, 190). Dancy is led to the idea that memory images must be a special category of beliefs by the assumption that to remember is to have knowledge. If memory is to serve as a source of knowledge and justification, then remembering needs to deliver beliefs. By thus erasing the distinction between propositional and imagistic memories, as a way to preserve a view of memory as an organ for knowledge, Dancy masks the complexity of memory that Campbell insists should be our starting point. To restore that complexity we will need, as Campbell argues, a norm whose scope is much wider than the norm of truth.

The second branch of the argument for a distinct norm for assessing memory begins by noting that the extended narratives of autobiographical memory might consist only of truths while still distorting the past. In nostalgia, for example, we selectively remember just those details of the past that idealize it. The problem is not that nostalgia remembers only selectively—all memory is selective. The problem is that the details we elect to keep track of distort the past by idealizing it. “Nostalgia is a defect of memory accuracy: nostalgic memory is not faithful to the past…. When nostalgic, certain details of the past are remembered—those that contribute towards its idealization…. Concerns about nostalgic memory are not alleviated by showing that the details remembered are factual” (Reference Campbell, DesAutels and Urban WalkerCampbell 2004/Reference Campbell, Koggel and Jacobsen2014b, 25). So if having a mental image can be a case of successful remembering, and if nostalgia fails as successful remembering, then truth is neither necessary nor sufficient for successfully remembering. Truth is simply the wrong norm for evaluating memory.Footnote ⁹

Nostalgic memory also gives us a first glimpse of what accuracy will need to look like if it is to replace truth as the principal norm governing the many diverse activities and achievements of memory. In nostalgia, we may get the past right, but we do so in a way that gets its significance wrong. As Campbell says: “Neither reproductive fidelity nor the truth of declarative memory seems adequate to how successful remembering often tries to capture the significance of the past” (Reference CampbellCampbell 2006/Reference Campbell, Koggel and Jacobsen2014b, 35). Accurate memory, on her account of it, requires getting right the significance of the past for the present and the future. Of course, we can't get its significance right if we get the past wrong—so accuracy is not consistent with falsity. But, as nostalgic memory illustrates, we can get its significance wrong while getting the past right. Accuracy is therefore a more demanding standard to meet than truth.

Campbell draws together under the norm of accuracy both the diversity of things remembered and the variety of activities by which we come to remember: memory aims to get right the significance of the past for the present. In doing so she turns her back on the traditional epistemologist's idea of memory. If the primary cognitive value for memory is other than truth, and truth is the primary cognitive value for inquiry, then memory is not primarily an organ for knowledge.Footnote ¹⁰ I'll return to and expand on this point in a moment, but first we should get a better fix on the idea of significance that lies at the heart of the norm of accuracy.

IV. Significance and Relationality

Three observations about the idea of significance are in order. First, there are varieties and levels of significance that the past might have to the present. As Campbell says: “When we think of the significance of the past to the present expressed through human memory, we are dealing with a facet of remembering that can be personal, pragmatic, emotional, intellectual, social, political or ethical” (Reference Campbell, DesAutels and Urban WalkerCampbell 2004/Reference Campbell, Koggel and Jacobsen2014b, 26). Questions about the significance of the past to the present or to the future are often—at least in part—ethical questions in light of which we decide whether, and if so how, we can go on together from where we are now. So the idea of significance has the consequence that, within our conception of accuracy, both epistemic and ethical norms become entangled: epistemic norms because we are trying to get right the significance of past events or occasions; ethical norms because what we are trying to get right is the significance of those events or occasions.

Second, the significance the past now has for us is always a function of our present needs and interests and of our expectations for the future. So being faithful to the past requires redescribing or reinterpreting it in light of our changing and ongoing interests: “Our memories do not have to be fixed and singular in meaning… to bear witness to the reality of the past” (Reference Campbell, Haaken and ReaveyCampbell, 2010/Reference Campbell, Koggel and Jacobsen2014b, 60). This does not mean that in remembering the past we should tolerate its falsification—it simply means that in remembering it we aim to make the past useful for the present. Settling on the significance of the past is not a matter of replacing the facts about the past, as those are presented in memory, with interest‐laden interpretations of those facts. It's a matter of replacing the interpretation of the facts we had in the past, in light of the needs and interests we had then, with the interpretation we have reached after reflection, consultation with others, and in light of all we've learned since. Whether the new interpretation of those facts is better than our earlier interpretation is of course always open for further discussion.

Third, as we've already noted, on Campbell's account of it, memory is relational. Its relationality has several dimensions. First, like many other cognitive abilities, memory “develop[s] only in relations with other persons and only with the support of shared communal practices that foster these abilities” (Reference CampbellCampbell 2003, 17).Footnote ¹¹ Second, many of our memory activities are themselves social activities—we catch up on each other's lives since our last encounter, I testify before the jury, I confess to you, we reminisce together, and you remind me to do something. Thus, Campbell writes that “Significance has often to be understood in ways that reflect that we share memory; judgements that I have the significance of the past roughly right are rarely mine alone” (Reference CampbellCampbell 2006/Reference Campbell, Koggel and Jacobsen2014b, 31). But the idea of significance also introduces a third dimension of relationality. Others are constitutively implicated in the idea of significance—they don't merely assist us in discovering or understanding the significance of a past occasion; they can also play a role in bestowing on the past the significance that it has. Campbell expresses the idea this way: “if we share an occasion that is in some way significant to me, then both your responses to that occasion and your responses to my responses… may become a part of the significance of that occasion to me” (Reference CampbellCampbell 1997, 118). In other words, the responses of others to our autobiographical narratives become part of what gives to the past the particular significance that it now has for us.Footnote ¹²

Let's return now to the idea that accuracy, not truth, is the primary norm for evaluating memory and memory activities. There are two quite different ways to absorb this thought. A first reaction might be to think that accuracy should then replace truth as the goal of inquiry, or at least as the goal of our inquiries into the past.Footnote ¹³ This reaction only reasserts the idea that memory derives its character from its place in inquiry—that it is, first and foremost, an organ for knowledge. But Campbell does not call for any revision of our view of inquiry itself. Accuracy is not intended to replace truth as the goal of inquiry—not even those inquiries into the past for which memory is an essential resource. Her point is, rather, that when we see what being faithful to the past consists in, we also see that memory is not primarily in the business of inquiry. She wants us to stop reducing memory to a mere resource for knowledge and to learn to see its role in the work of building and shaping the relationships in which we are embedded, and in which we undertake our many projects of self‐constitution.

So I suggest that Campbell's idea is best seen as follows. Memory has already been thoroughly reconceived and reshaped by philosophers in order to serve—alongside Reason and Experience—as a faculty whose primary function is the acquisition of knowledge and the pursuit of truth.Footnote ¹⁴ From the perspective on memory that Campbell offers, this is about as helpful as saying that legs are designed by evolution for the pursuit of knowledge. After all, our two legs enabled our primate ancestors to peer over the tall grass of the savannah and to know what lay ahead of them, they carried us on our exploration of far continents, and they now take us to the library—all in the pursuit of knowledge. So we can, if we wish, see our legs primarily through the lens of inquiry and insist that legs, too, need an epistemology. But to do so would result in an impoverished and distorted understanding of legs—both their nature and their many roles in our lives.

To the best of my knowledge no one has yet proposed placing legs alongside reason and experience, but memory has been so positioned—philosophical work on memory has treated it primarily as an organ or faculty for knowledge about the past. Campbell, however, sees the place of memory differently: “Memory is one of the key cognitive abilities through which we develop personhood” (Reference CampbellCampbell 2003, 17), and “Remembering our pasts is one of the most important ways that we form, shape, and maintain relationships” (Reference CampbellCampbell 2014a/Reference Campbell, Koggel and Jacobsen2014b, 139). The work of constituting persons is not done by individual recollection but by the sharing of memories:

On the relational theory of persons, forming and maintaining relationships are central to projects of self‐constitution. Working out in detail what this means, and what role memory and the sharing of memories have in our projects of self‐constitution, in weaving and reweaving the relations with others on which we depend for personhood, is far beyond the scope of this essay.Footnote ¹⁵ Nonetheless, my suggestion is that the central concern of Campbell's work on memory has not been to provide an alternative epistemology for memory, but to reclaim memory from epistemology. And that, if it can be done, might well make some epistemologists nervous.

V. Anti‐Individualism

On the traditional epistemological account, remembering is highly individualistic—it is a mental process, and mental processes are processes in the head. But in order to give credit to the range and variety of activities by which we come to remember, and to the many forms that our resulting memories can take (images, propositions, narratives, and so on), we need to be pluralists about how memories are actualized or realized. And an obvious feature of our many memory activities is that they are not, except by occasional accident, processes in the head. As Campbell reminds us: “While sometimes experienced as a feature of our interiority, human remembering also takes place through action, narrative, and other modes of representation in public space and in the company of others” (Reference CampbellCampbell 2006/Reference Campbell, Koggel and Jacobsen2014b, 31), and “We remember… through visiting places or handling familiar objects, through public apology and political protest, through re‐learning a language or a heritage of skills” (Reference CampbellCampbell 2014a/Reference Campbell, Koggel and Jacobsen2014b, 139). The activities by which we come to remember—our memory processes—are then located in the world, along with the familiar objects we handle, the memorials we construct, and the reminders we assemble. So the complexity of memory that serves as Campbell's starting point, and that makes truth an unsuitable norm for memory, also provides reason to doubt that the activities and processes that constitute remembering are processes located in the head.

There are two different directions one might take this idea. Along one route we might say that Campbell is committed to the idea that not all mental processes occur in the head, and so we are led to the Extended Mind Thesis proposed by Andy Clark and David Chalmers (Reference Clark and ChalmersClark and Chalmers 1998). On their view, remembering is, at least on occasion, a mental process external to the head, and memories themselves (our mental traces of the past) can be encoded and stored on scraps of paper as easily as in the brain.Footnote ¹⁶ Along a second route Campbell need only deny that remembering is a mental process. This doesn't have to be a crazy view. Consider that we can add numbers “in our heads.” When we do so we qualify as thinking, and thinking is a mental process. But addition is not thereby made a mental process or activity—it is a mathematical operation by which any two numbers take us to a third number called their “sum.” Whether I perform the operation in my head, with my calculator, or on paper, I am adding numbers. Thinking is then one way to add numbers, but adding numbers is not therefore a mental process. Similarly, remembering can be one of those things that we sometimes do “in our heads” without thereby being a mental process.

So there are two options: We can count the many processes and activities of remembering—those activities whose successful culmination consists in our having memories—as playing out in the world external to our heads either (i) by extending the borders of the mind to include things and events outside the head, or (ii) by denying that some of the processes we once thought of as mental are really mental after all. Although in her last works Campbell was still exploring variations on Andy Clark's extended mind thesis, my own suspicion is that nothing in her account commits her to going one way rather than the other on this issue. And I also doubt that much of significance turns on which way we go. What matters for Campbell's account of memory is that we take seriously the idea that the practices of embodied agents, social practices in which we participate, and many familiar manipulations of our environment, are recognized as memory activities—that is, they are the very processes that the traditional theory of memory both classifies as “mental” and locates in the heads of rememberers. On Campbell's view of it, remembering can be done outside the head because remembering just is the business of trying to get right the significance of our pasts for our present. But nothing turns on whether that business is conducted in the head.

According to the Extended Mind Thesis, a shopping reminder—say, a piece of paper in my pocket with the word “milk” inscribed on it—qualifies as a memory state and, therefore, as a mental state located outside my head. But when our understanding of memory is reframed in terms of the act/achievement distinction, the use of the note, or the activity of using the note, is one of my memory activities. The pocketed piece of paper bearing the representation “milk” is not a part or component of my memory state, but the acts of writing the note, carrying it with me, and occasionally checking it, are all constituents of the process by which I successfully remember to bring home milk.Footnote ¹⁷ None of these are processes “in the head.” Thus, the anti‐individualism needed for Campbell's account of memory is already embodied in her account of the diversity of memory activities, which can include embodied and interpersonal activities, as well as private reflections. There is no reason to join a debate about the nature of the mental that is framed by a metaphorical contrast between inner and outer realms.

VI. Relationality and Animal Cognition: An Objection

I turn, briefly, to an objection to the idea that memory is relational. Recall the passage with which I began, where Campbell wrote that “[m]emory is one of the key cognitive abilities through which we develop personhood,” and that “the cognitive abilities necessary to being a person and hence to being a moral agent develop only in relations with other persons and only with the support of shared communal practices that foster these abilities” (Reference CampbellCampbell 2003, 17). There are two claims about memory here: (i) that it is a key cognitive ability in the development of personhood, and (ii) that it develops only in the context of other persons and communal practices. But it appears to follow from these claims that (iii) creatures lacking the appropriate relations with others, and lacking the support of the relevant communal practices—in short, a great many nonhuman species—will not have a faculty of memory. And that seems plainly false.

There are, in principle, three strategies for responding to the objection. The first—explored by John Dewey, who subscribed to deeply relational theories of both cognition and affectFootnote ¹⁸ —is to simply deny that nonhuman animals have memories.

But this strategy deprives us of an essential explanatory resource: it leaves us helpless to explain the dog's ability to find the bones she buried yesterday, or the elephant's ability to return to a seldom used waterhole.

A second strategy attributes a richer social life to nonhuman animals, in which they are embedded in the ways required for relational forms of cognition. Alva Noë, for example, writes of domestic animals that we can “close off the possibility of calling into question their status as bearers of minds” precisely because “they belong within the human sphere,” where “we pursue relations and projects with [them]” (Reference NoëNoë 2010, 36). If, however, we attempt to extend this strategy too far beyond domestic pets, we secure relational forms of cognition for other animals only by anthropomorphizing their worlds. Merely acknowledging the sometimes complex social worlds of certain other species will not meet the objection—a successful reply would need to say that all animals with memories have a social world that is sufficiently robust to make possible relational forms of cognition.Footnote ¹⁹ But many nonsocial animals also learn from experience—even paramecia in the petri dish have been thought to remember what they learn from past experience.Footnote ²⁰

There is a third and, I suggest, more sober strategy for reconciling a relational account of memory with the idea that even asocial nonpersons can remember. On the traditional view of memory, as a distinctive type of mental process leading to the formation of a distinctive type of belief state, it can easily seem that memory is an all‐or‐nothing affair. So a reply to the objection should once again emphasize the complexity of memory that is so central to Campbell's account: an endless variety of activities can be subordinated to a varied range of achievements, all under the rubric of remembering. But we can and should deny that this wide repertoire of activities and achievements is available, in its entirety, to nonsocial and nonlinguistic animals. Many nonhuman animals are cognitively and affectively equipped for only some of the forms that memory can take, and we can consistently attribute to them some forms of memory while denying them others. We can, for example, attribute to our pets learned abilities (skill memories) and imagistic memories while denying that they have the capacity for autobiographical memory, or while having principled reservations about assigning propositional memories to them. So there is no need to deny that other animals can remember—they learn from experience, they acquire and retain abilities, and they recognize previously encountered features of their world. But they nonetheless lack the full suite of abilities and capacities that we exploit in remembering, and much of what matters to us about memory is not within their reach. They operate with a more limited repertoire of memory activities, and their memory achievements are neither as various nor as thoroughly integrated as ours.

There is, then, nothing implausible about saying both that memory is relational, and that even the most asocial creatures can have memories. But the memories of nonsocial animals do not include the range and variety of phenomena that we find among our own memories, and that constitute us as persons—they lack the means to participate in core “activities involving self‐narratives” (Reference CampbellCampbell 2003, 17). For that, the right kinds of relations with others are needed. On Campbell's view, the forms of remembering most deeply implicated in the construction of our moral and political identities are the autobiographical memories to which we give narrative expression, and those are deeply relational forms of memory. When we reminisce with friends or confess to a priest, when we are subjects of therapy or testify as witnesses, we are then engaged in the practices of self‐narration “through which a human is configured into a person” (Reference CampbellCampbell, 2003, 41).

VII. Memory Reframed

The relational conception that has come to play a prominent role in feminist thought sees persons as constituted through their relationships with others. As a mundane object becomes transformed into a gift by being given a place in social practices, so the human animal becomes a person by being embedded in social relationships and institutional practices: “persons are socially embedded and… agents’ identities are formed within the context of social relationships and shaped by a complex of intersecting social determinants, such as race, class, gender, and ethnicity” (Reference Mackenzie, Stoljar, Mackenzie and StoljarMackenzie and Stoljar, 2000, 4). On this picture, the many “intersecting social determinants” that shape our relationships with others, as well as the forms of oppression and injustice that distort those relationships, are not merely features of the contexts in which we find ourselves—they also constitute the persons we are. But it is not only our personhood that we owe to relationality—it is within our many relationships that we develop and maintain what Annette Baier called “the essential arts of personhood” (Reference BaierBaier 1985, 84). These essential arts arguably include our capacities for reason, self‐consciousness, and autonomy.Footnote ²¹ Injustices of class, race, and gender then threaten both our status as persons and the cognitive and affective capacities that we develop and depend on for personhood. Importantly, these capacities include memory.

A person at any given moment is the result of ongoing interactions with others, but she is also the result of ongoing negotiations with her past, and with the person she once was. Memory plays an essential role in this complex relational process—it both links us to our past, and guides us toward our future. But on the anti‐individualistic view recommended here, this process is not a feature of our interior lives—it is embodied in our social world, in our interactions with and under the influence of others. By keeping records, sharing memories, celebrating anniversaries, and erecting public memorials, we negotiate the significance of our individual and shared pasts, and we make those pasts useful for the future. But these social practices are not merely outward expressions of memory—they are the very substance of remembering.

The psychological individualism Campbell turns away from treats memory as an aspect of our inner lives and—in part because of its interiority—as an aspect of persons that remains largely untouched by the social settings in which we remember. The outward expressions of our inner life are naturally subject to the turbulence of the social world, and so they can be affected by any injustices that structure that world; but psychological individualism encourages us to see gender, class, and race as features of our external circumstances that are in no way essential to what or how we remember. In the study of mind, the matters of greatest concern to feminist thinkers are pushed aside. But when we study memory in the context of a relational theory, matters of central concern to feminists are moved to center stage, and “we cannot talk about memory without discussing the social power that authority over the past secures” (Reference Campbell, Koggel and JacobsenCampbell 2014b, 1).

A relational account of memory is suited to do work for which traditional theories were poorly equipped. It enables us to recognize harms of oppression through the “collective reinterpretation of personal pasts in light of new social understandings” (Reference CampbellCampbell 2003, 20), and to do so without inviting skepticism about memory, antirealism about the past, or misplaced fear of memory distortion. Our ability to remember our own personal pasts—an ability that is deeply entangled with our ideas of personhood, selfhood, and autonomy—emerges from and acquires its character within social relationships and institutional practices. The success with which we grasp the significance of the past for the present and future can then be made to depend on the various social positions we occupy and the public practices in which we participate. To deny someone access to those positions and practices and to systematically discredit them as rememberers is also to challenge their status as persons. Campbell's own writings detail how women are often placed in the position of narrating a view of the past that challenges dominant representations, and how they have therefore been denied control over the understanding of their own personal pasts and, so, of the significance their past has for how they confront their futures.Footnote ²²

To accommodate the personal, ethical, and political dimensions of memory, we are compelled to rethink the nature of memory itself—to recover memory from the distortions that result from framing it exclusively as a faculty for knowing. By embellishing Campbell's relational account of memory with the distinction between activities and achievements, we make progress toward this end. The variety of achievements that qualify as successful remembering challenges the presumption that truth is the only norm for good remembering; a similar variety among our memory activities belies the thought that remembering is an inner (or mental) process; the same variety enables us to confront a potential objection to relational theorizing. When we acknowledge that the concept of person is relational (and, so social) and not metaphysical, the characteristics constitutive of personhood need to be reframed in such a way that their moral and political significance becomes evident. Campbell's work shows us what memory and remembering will have to look like in order for this to be accomplished and, so, what it means for persons to be fully relational.

None of the three original components of the traditional epistemologist's view of memory survive in Campbell's account: the activity of remembering is not a mental process located in the head, truth is not the primary cognitive value for memory, and memory is not a faculty for the pursuit of knowledge (though, like legs, memory can be co‐opted for that purpose). So perhaps there is enough in Campbell's account of memory to make epistemologists nervous.